Citation
Water and sanitation institutions and governance

Material Information

Title:
Water and sanitation institutions and governance
Alternate title:
Impact on service provisions in urban areas of low- and middle-income countries
Creator:
Manderino, Laurie Ann ( author )
Place of Publication:
Denver, Colo.
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
1 electronic file (318 pages) : ;

Thesis/Dissertation Information

Degree:
Doctorate ( Doctor of Philosophy)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
School of Public Affairs, CU Denver
Degree Disciplines:
Public Affairs
Committee Chair:
Heikkila, Tanya
Committee Members:
Martell, Christine
Weible, Christopher
Brett, John

Subjects

Subjects / Keywords:
Urbanization ( lcsh )
Community health services -- Finance ( lcsh )
Decentralization in government ( lcsh )
Public health -- Sanitation ( lcsh )
Poverty -- Research -- Developing countries ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Abstract:
Rapid global urbanization over the last few decades has intensified the challenge of providing adequate water and sanitation services to urban residents. Meeting this challenge has been the focus of domestic and international development efforts, including Millennium Development Goal 7.C. This research studies three institutional and governance attributes theorized to improve government service outcomes, testing hypotheses that the attributes are associated with greater country progress on providing urban water and sanitation access. The attributes are: a) decentralized services; b) sector-wide strategy and investment coordination; and c) civil society engagement. ( ,,,,,, )
Abstract:
Country-level experience is analyzed using a series of ordered logistic regression models for a sample of 75 low- and middle-income countries. UN GLAAS survey data is used to derive country-specific variables for the three attributes. These, along with control variables representing country background conditions, are analyzed relative to four country progress outcome variables, two each for water and sanitation. The outcome variables, (covering the 2000 to 2012 time period), are derived from the UN JMP dataset that tracks urban access rates by country. Based on results from these models, four country case studies look in-depth at implementation of the attributes and highlight aspects that can help or impede country progress. Overall, findings show that decentralization is helpful to sanitation progress, but not water progress, likely due to limitations of capacity and funding faced by subnational iv levels of government. Three explanations are proposed for why decentralization may impact water and sanitation differently. Results for sector planning were mostly inconclusive, except that it was shown helpful to water progress over the 12-year period. Study of this attribute would benefit from additional wide-scale data collection. Civil society engagement was consistently shown to help country progress in both water and sanitation, and several examples of engagement are profiled to demonstrate how it can improve service outcomes.
Abstract:
The last chapter relates findings to theories about government provision of public goods. The extent to which the three attributes help achieve efficiency, supply, equity, and social welfare goals is discussed. Finally, practical recommendations for strengthening sector institutions and governance are presented with application to governments and international aid donors.
Thesis:
Thesis (Ph.D.) - University of Colorado Denver.
Bibliography:
Includes bibliographic references
System Details:
System requirements: Adobe Reader.
General Note:
School of Public Affairs
Statement of Responsibility:
by Laurie Ann Manderino.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
946535401 ( OCLC )
ocn946535401
Classification:
LD1193.P86 2015d M35 ( lcc )

Downloads

This item has the following downloads:


Full Text
WATER AND SANITATION INSTITUTIONS AND GOVERNANCE:
IMPACT ON SERVICE PROVISION IN URBAN AREAS OF
LOW- AND MIDDLE-INCOME COUNTRIES
by
LAURIE ANN MANDERINO
B.A., California University of Pennsylvania, 1985
M.S., University of Colorado Denver, 1993
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
Public Affairs
2015


2015
LAURIE ANN MANDERINO
ALL RIGHTS RESERVED


This thesis for the Doctor of Philosophy degree by
Laurie Ann Manderino
has been approved for the
Public Affairs Program
by
Tanya Heikkila, Advisor
Christine Martell, Chair
Christopher Weible
John Brett
November 20, 2015
n


Manderino, Laurie Ann (Ph.D., Public Affairs)
Water and Sanitation Institutions and Governance: Impact on Service Provision in Urban
Areas of Low- and Middle-Income Countries
Thesis directed by Associate Professor Tanya Heikkila
ABSTRACT
Rapid global urbanization over the last few decades has intensified the challenge
of providing adequate water and sanitation services to urban residents. Meeting this
challenge has been the focus of domestic and international development efforts, including
Millennium Development Goal 7.C. This research studies three institutional and
governance attributes theorized to improve government service outcomes, testing
hypotheses that the attributes are associated with greater country progress on providing
urban water and sanitation access. The attributes are: a) decentralized services; b) sector-
wide strategy and investment coordination; and c) civil society engagement.
Country-level experience is analyzed using a series of ordered logistic regression
models for a sample of 75 low- and middle-income countries. UN GLAAS survey data is
used to derive country-specific variables for the three attributes. These, along with
control variables representing country background conditions, are analyzed relative to
four country progress outcome variables, two each for water and sanitation. The outcome
variables, (covering the 2000 to 2012 time period), are derived from the UN JMP dataset
that tracks urban access rates by country. Based on results from these models, four
country case studies look in-depth at implementation of the attributes and highlight
aspects that can help or impede country progress.
Overall, findings show that decentralization is helpful to sanitation progress, but
not water progress, likely due limitations of capacity and funding faced by sub-national
iii


levels of government. Three explanations are proposed for why decentralization may
impact water and sanitation differently. Results for sector planning were mostly
inconclusive, except that it was shown helpful to water progress over the 12-year period.
Study of this attribute would benefit from additional wide-scale data collection. Civil
society engagement was consistently shown to help country progress in both water and
sanitation, and several examples of engagement are profiled to demonstrate how it can
improve service outcomes.
The last chapter relates findings to theories about government provision of public
goods. The extent to which the three attributes help achieve efficiency, supply, equity,
and social welfare goals is discussed. Finally, practical recommendations for
strengthening sector institutions and governance are presented with application to
governments and international aid donors.
The form and content of this abstract are approved. I recommend its publication.
Approved: Tanya Heikkila
IV


ACKNOWLEDGEMENTS
I want to express my gratitude to Dr. Tanya Heikkila, my dissertation advisor, for
her ongoing guidance and support throughout the dissertation process, and also to Drs.
Christine Martell, Christopher Weible, and John Brett for their valuable input and
feedback on my research. Other faculty members at the School of Public Affairs,
particularly Dean Paul Teske and Dr. Peter deLeon, supported my academic development
and sustained participation in the PhD program. Staff at the School of Public Affairs,
especially Antoinette Sandoval and Rob Drouillard, helpfully guided me through
paperwork and computer issues. Finally, my fellow PhD students were great learning
partners and a source of beneficial advice throughout the program.
I am grateful to those individuals overseas contacted for the case studies who took
the time to respond to my research inquiry or to connect me with knowledgeable
colleagues in their country.
I want to also acknowledge the National Science Foundation who supported this
work through the Integrative Graduate Education and Research Traineeship (IGERT)
fellowship program (Grant #: DGE-0654378), and those faculty and staff at the
University of Colorado Denver who made this program possible.
v


TABLE OF CONTENTS
CHAPTER
I. INTRODUCTION...................................................1
II. BACKGROUND.....................................................7
III. LITERATURE REVIEW.............................................12
Water and Sanitation Service as Public Goods..................12
Water and Sanitation Institutions and Governance..............18
Literature on Attributes Selected for Study...................22
Country Background Conditions.................................49
Summary of Literature.........................................55
IV. QUANTITATIVE METHODS..........................................59
Sample Selection..............................................59
Outcome Variables.............................................63
Key Explanatory Variables.....................................67
Control Variables.............................................69
Data Validity and Limitations.................................75
V. QUANTITATIVE ANALYSIS.........................................82
Descriptive Statistics........................................83
Bivariate Analyses............................................88
Multivariate Analyses.........................................91
Questions for Further Exploration............................112
VI. SECTOR-WIDE STRATEGY AND
INVESTMENT COORDINATION......................................114
Sector Planning by Countries.................................114
vi


118
120
.122
.123
.124
.125
.127
.128
129
.148
.165
.177
.181
182
.196
.210
.221
.221
.223
224
.229
.236
.242
Relating Sector Planning to Achievement
Findings from Prior Research......
QUALITATIVE METHODS..............................
Case Selection...................................
Document Review..................................
Consultations with Country Experts...............
Scope of the Case Studies........................
DECENTRALIZED SERVICES...........................
Colombia Case Study..............................
Mali Case Study..................................
Summary and Analysis of Decentralization Cases...
Different Decentralization Impact on Water and Sanitation
CIVIL SOCIETY ENGAGEMENT.........................
Kenya Case Study.................................
South Africa Case Study..........................
Summary and Analysis of Civil Society Engagement Cases
SUMMARY AND CONCLUSIONS..........................
Introduction.....................................
Theoretical Background...........................
Research Methods.................................
Decentralized Services...........................
Sector-wide Strategy and Investment Coordination.
Civil Society Engagement.........................
Vll


Limitations of this Study..........................................248
Areas for Further Research.........................................249
Conclusion.........................................................250
BIBLIOGRAPHY.................................................................256
APPENDIX
A. Theoretical Concepts and Constructs for Research Questions.........279
B. List of Sample Countries with Data Notations.......................282
C. Variable Operationalization and Source Information.................285
D. Correlation Tables.................................................288
E. Frequency Tables for Key Explanatory
Variables and Outcome Groups.......................................290
F. Graphs Relating Sector Planning to Progress........................294
G. Persons Consulted for Case Studies.................................302
H. Case Study General Consultation Guides.............................304
viii


LIST OF TABLES
TABLE
4.1. Average Country Population: Comparison
of Sample Group and LMI Group...........................................60
4.2. GDP per Capita: Comparison of Sample Group and LMI Group................61
4.3. Urban Water and Sanitation Access: Comparison
of Sample Group and LMI Group...........................................62
4.4. Overview of All Variables...............................................63
4.5. Systems Considered as an Improved Source by JMP.........................64
5.1. Descriptive Statistics and Grouping Parameters for Outcome Variables....84
5.2. Descriptive Statistics for Key Explanatory
Variables and Control Variables.........................................86
5.3. Correlations between Explanatory Variables and Outcome Variables........89
5.4. Key Explanatory Variable Means by Outcome Group.........................91
5.5. Ordered Logistic Regression Results: Water A2012.........................94
5.6. Ordered Logistic Regression Results: Water Progress......................95
5.7. Ordered Logistic Regression Results: Sanitation A2012....................96
5.8. Ordered Logistic Regression Results: Sanitation Progress.................97
5.9. Summary of Ordered Logistic Regression
Results for Complete Models..............................................98
5.10. Odds Ratios for Key Explanatory Variables..............................101
5.11. Control Variable Significance..........................................109
8.1. Sources of Funds for Municipal Water and Sanitation
Investment in Colombia (1995 2010)....................................137
8.2. Water and Sanitation Investment by Different
Levels of Government in Colombia (2010).................................139
IX


8.3. Sources of Funds for Municipal Water and
Sanitation Investment in Mali (2006)..................................153
8.4. Decentralization Case Study Findings by Construct.......................165
9.1. Civil Society Engagement Case Study Findings by Construct...............211
10.1. Overview of All Variables...............................................225
x


LIST OF FIGURES
FIGURE
5.1. Predicted Probabilities for Decentralization...............................103
5.2. Predicted Probabilities for Sector Planning................................105
5.3. Predicted Probabilities for Civil Society Engagement.......................107
xi


LIST OF FORMULAS
FORMULA
4.1. Progress Outcome Variable Calculation..............................66
4.2. Urban Population Change Calculation................................70
5.1 Ordered Logistic Regression Model...................................92
xii


CHAPTERI
INTRODUCTION
At the United Nations (UN) Millennium Summit in September 2000, the largest-
ever gathering of world leaders adopted the Millennium Development Goals (MDG), an
integrated set of time-bound targets for extending the benefits of globalization to the
worlds poorest citizens. They committed to improving the lives of people living in
extreme poverty and demonstrating, by the year 2015, progress in addressing the most
pressing issues facing developing countries (Lenton et al. 2005; UN, 2014). A specific
target was set for increasing access to safe and sustainable water and basic sanitation
among unserved populations (Target 7.C). While the problem persists in both rural and
urban areas, the challenge is compounded in cities due to the accelerated pace of global
urbanization in the last few decades. Cities face numerous challenges in accommodating
new residents and providing them with basic services.
In 2012, an estimated 33 percent of the urban population in developing countries
lived in slum conditions, communities characterized by a lack of basic services such as
safe drinking water and adequate sanitation. This percentage has continually grown since
1990. The proportion of urban residents living in slums is particularly high in Sub-
Saharan Africa at 62 percent (UN, 2014). In total, there are an estimated 863 million
urban dwellers living in slums. Lack of clean water and sanitation in these communities
leads to a variety of health problems, including diarrheal disease for the affected
populations (WHO, 2014b).
As the global community strives to address this situation, efforts focus on
different aspects of the problem, such as water allocation and appropriate technologies
1


for low-resource settings. Other aspects being addressed relate to the institutional and
governance environments present in countries for water and sanitation service provision.
These environments are the focus of this dissertation. While institutions and governance
are further explored in the literature review chapter, basic definitions are presented here.
Institutions have been broadly defined as the rules of the game in a society...
[that] structure incentives in human exchange, whether political, social, or economic
(North 1990, p. 3). Institutional arrangements within governments determine, among
other things, who makes decisions and who allocates available resources. These
arrangements can play an important role in government performance in the delivery of
public goods1 (Cremer, Estache, & Seabright, 1994; Frank & Martinez-Vazquez, 2014).
Institutions operate within the realm of governance. Several different definitions
for governance have been developed by researchers and international organizations. A
widely accepted definition from the United Nations Development Program (UNDP) most
closely aligns with the intent of this research: the exercise of economic, political, and
administrative authority to manage a countrys affairs at all levels... [including]
mechanisms, processes, and institutions through which citizens and groups articulate their
interests, exercise their legal rights, meet their obligations, and mediate their differences
(UNDP, 1997, p.2).
So, while institutions determine decision and incentive structures within
governments, the broader concept of governance also encompasses the values and
processes by which government authority is exercised, including interactions with 1
1 Chapter III includes a detailed explanation of public goods.
2


citizens and organized groups outside of government. Accountability is a key theme
underpinning several aspects of governance (UNDP, 2015).
Understanding national institutional and governance environments is important
because they guide efforts of cities to provide new and upgraded water and sanitation
services that meet minimum international standards, something discussed more in a later
chapter. How cities approach decisions such as where to construct new infrastructure,
how to invest funds, and how to price services is to a large degree determined by the
institutional and governance context set by the national government. It determines,
among other things, the authority of cities to make investment decisions, how the public
should be informed and consulted, and what type of service quality monitoring may be
required.
The overall goal of this dissertation is to evaluate if certain national institutional
and governance attributes, generally considered indicative of good governance, have had
a positive impact in the water and sanitation sector, specifically for improving urban
water and sanitation service provision. This research focuses on three specific
institutional and governance attributes reflected in the following research questions.
Additional background information and hypotheses linked to these questions are
presented in the literature review chapter.
RQ1. What is the role of decentralization of water and sanitation services in a countrys
progress on providing urban water and sanitation access?
RQ2. What is the role of sector-wide strategy and investment coordination in a countrys
progress on providing urban water and sanitation access?
3


RQ3. What is the role of civil society engagement in a countrys progress on providing
urban water and sanitation access?
The three attributes were selected because: a) they have been identified in the
literature as having potential to contribute to effective government service provision and
donor aid effectiveness; b) they have not been studied extensively for the water and
sanitation sector, particularly on a large-scale basis; c) there was an observed gap of
research on potential interactions among the three attributes; d) they represent a departure
from previous approaches to water and sanitation service provision in many developing
countries; and e) relatively new UN data collection programs provide comparable data on
these attributes for multiple countries.
To study these attributes, the unit of analysis is a country and the population of
interest is low- and middle-income countries as classified by the World Bank (World
Bank website, 2015). The research is centered on the X (explanatory) and Y (outcome)
variable relationship, with spatial variation, intended to explore theoretical linkages
identified through literature review (Gerring, 2007). The overall time period being
studied is approximately 1990 to 2012.
This research adds to the existing literature on government service provision, (see
overview in Chapter III), by evaluating the impact of the three institutional and
governance attributes specifically in the context of urban water and sanitation service
provision, to see if their implementation in low- and middle-income countries really does
lead to better performance, as predicted by theory. It makes theoretical contributions by
uncovering the conditions under which the three attributes might be beneficial to
government provision of urban water and sanitation services in low- and middle-income
4


countries. The results of this research will also be of interest to domestic and
international development programs geared towards increasing urban water and
sanitation access and maintaining quality services.
The research is conducted in two phases. The first phase of research is a cross-
case analysis of 75 countries using ordered logistic regression models to assess the
impact of the three institutional and governance attributes (explanatory variables) on
country progress (outcome variables), while also controlling for several country
background conditions. The cross-case analysis uses country data from two UN datasets,
introduced in the next chapter. The phase one analysis focuses specifically on the 2000
to 2012 time period, for reasons of data availability further explained in Chapter IV.
The second phase of research looks more in-depth at the impact of individual
attributes. For two of the attributes, decentralized services and civil society engagement,
specific country case studies, (two for each attribute), chronicle country history with the
attribute-of-interest and further probe issues raised in the cross-case analysis. Cases
focus primarily on one attribute to better understand how it was implemented in a
country, and what can be learned about conditions that promote positive or negative
impacts. The case studies are developed from review of relevant documents and
consultations with knowledgeable persons in the selected countries.
The remaining attribute, sector-wide strategy and investment coordination, is
studied differently. Because there is not a long history of country implementation of this
attribute, instead of preparing case studies, the analysis looks at other studies that have
focused on this attribute for select groups of countries, and analyzes their findings along
with the phase one quantitative results to draw reasonable conclusions.
5


The focus of this research is on water and sanitation service provision, therefore
issues relating to water as a shared natural resource, including water rights, water quality,
and water supply considerations are not addressed by this study. However, reference is
made to some research on water resource management policy, as governance themes
often overlap with water and sanitation service provision. Also, while this research is
limited to water and sanitation service provision in urban areas, some studies of rural
areas are referenced when the findings have wider applicability.
The next chapter provides further background on global water and sanitation
challenges and the international efforts to address them. Subsequent chapters discuss
relevant literature for this study; the quantitative cross-case analysis and results;
qualitative findings for each of the three attributes, including country case studies; and
the overall analysis and conclusions.
6


CHAPTER II
BACKGROUND
Inadequate drinking water, sanitation and hygiene are estimated to cause more
than 842,000 diarrheal disease deaths per year and to generate disease caused by micro-
organisms and chemicals ingested through unsafe drinking water (WHO, 2014b). In
addition, adequate water, sanitation, and hygiene are critical for the prevention and care
of almost all the tropical diseases targeted for elimination and control by the World
Health Organization (WHO), such as soil-transmitted helminthiasis, guinea-worm
disease, trachoma and schistosomiasis, all of which affect mainly children (WHO, 2015).
In recognition of the importance of safe drinking water and sanitation to human
health, survival, and development, they were made the focus of MDG Target 7.C, to
"halve, by 2015, the proportion of people without sustainable access to safe drinking-
water and basic sanitation, based on comparison to the base year of 1990 (UN
Millennium Development Goals list). The drinking water part of MDG Target 7.C was
met in 2010, as the percentage of people lacking access to safe drinking water decreased
from 24% of the global population in 1990 to 11% in 2010 (UN, 2012). The sanitation
part of the target has not yet been met. Based on the most-recently available global data
(2015), 9% of the worlds population still lack access to safe drinking water and 32%
lack access to basic sanitation (WHO & UNICEF, 2000-2015).
Subsequent to the establishment of MDG Target 7.C in 2000, the UN Committee
on Economic, Cultural and Social Rights declared that access to water is a human right
and that water is a social and cultural good, not merely an economic commodity (UN
Committee on Economic, Social and Cultural Rights, 2002). In July 2010, the UN
7


General Assembly recognized the human right to water and sanitation and acknowledged
them as essential to the realization of all human rights (United Nations General
Assembly, 2010). Later, the UN Human Rights Council affirmed that the right to water
and sanitation is derived from the right to an adequate standard of living, and is therefore
covered by existing human rights treaties, placing an obligation on states to meet this
standard (UN Office of the High Commissioner for Human Rights, 2010, n.p.).
The importance of institutions and governance for achieving water and sanitation
access goals is broadly recognized in both the practitioner and academic communities.
Several prominent international water meetings in recent years, (for example, World
Water Forum (2000); Bonn International Conference on Freshwater (2001); World
Summit on Sustainable Development (2002); Commission on Sustainable Development
13th Session (2005); Budapest Water Summit (2013); and High-Level International
Conference on Water Cooperation (2013)), have highlighted water governance as a focal
issue with regard to both water resource management and service provision, as well as for
meeting the MDGs relating to water.
Among researchers, Saleth et al. (2007) emphasize the role of institutions in
influencing the magnitude and sustainability of MDG impacts in low-income countries.
They also explore linkages among past, ongoing, and planned policy interventions.
Koundouri et al. (2008, p.189-190) find that for meta-development goals such as the
MDGs, .. .the realization of the final goal is linked with the realization of several
intermediate, but related goals of a hierarchy of development interventions, all of which
require an effective institutional framework for their implementation and monitoring.
8


Researchers have also found effective governance to be a critical precondition for
efficient use of water and sanitation investments, as well as for operating and maintaining
infrastructure (Tropp, 2007; Rogers & Hall, 2003; Plummer & Slaymaker, 2007). These
views are discussed in more detail in the literature review chapter. Several international
organizations are working on water and sanitation governance, including the World
Banks Water and Sanitation Program (WSP) and the Global Water Partnership (GWP).
Their reports are referenced throughout this document.
Global Data Programs for the Sector
In this section, I introduce two sector-specific UN datasets that are used as
sources for country-level data for the cross-case analysis. One reports on country
institutional and governance environments and another measures country progress. I then
discuss two international groups that use data from these and other sources to plan
strategic interventions in conjunction with MDG Target 7.C and subsequent sector goals.
The first dataset is from the Global Analysis and Assessment of Sanitation and
Drinking-Water (GLAAS), a UN Water initiative implemented by the WHO, begun in
2008 as a pilot and expanded over time to include more countries. GLAAS surveys
countries biannually on water and sanitation topics, including the national institutional
and governance environment. The GLAAS process is intended to inform decision-
makers in developing countries and donor organizations; promote discussion and
identification of priorities and barriers to service provision; and promote a culture of
partnership and shared responsibility (WHO, 2011 & 2014a). I use recent GLAAS
survey data to derive explanatory variables representing country institutional and
governance attributes.
9


The second dataset comes from the WHO/UNICEF Joint Monitoring Programme
for Water Supply and Sanitation (JMP), the official UN mechanism for monitoring
country progress towards MDG Target 7.C, which has tracked individual country water
and sanitation access each year since 1990. JMP publishes global, regional and national
data for use by governments, donors, international organizations, and civil society (WHO
and UNICEF, 2012). The data are widely available through an on-line interactive
website that allows for the creation of individualized data tables. I used the JMP dataset
to derive the outcome variables representing country progress.
Recognizing an ongoing need for data collection relating to MDG Target 7.C and
subsequent international goals, a strategic advisory group (SAG) of GLAAS and JMP
representatives was formed in 2009 to bring independent, strategic thinking to further
development and implementation of JMP, and to provide guidance for rapid scale-up of
GLAAS reporting. SAG looks beyond the MDG deadline of 2015 towards development
of a comprehensive global monitoring system for drinking water and sanitation, and
harmonization of global and national monitoring needs, for tracking both development
goals and human rights criteria (WHO and UNICEF, 2011). The findings of this research
may help SAG identify items suited for ongoing monitoring.
Another strategic planning organization, Sanitation and Water for All (SWA), is a
global partnership of more than 90 developing country governments, donors, civil society
organizations, and other development partners, working to promote action and improve
accountability in programs designed to increase water and sanitation access. Annual
SWA meetings focus on planning, institutional strengthening, better resource utilization
and higher investment in the sector. Through the SWA process, countries make
10


commitments to address barriers to delivering sustainable water and sanitation services
(SWA website, 2014). Groups such as SWA can benefit from the findings of this
research as they plan coordinated partner actions to achieve water and sanitation goals.
The information highlighted in this chapter shows the magnitude of health
problems posed by inadequate access to water and sanitation and the high-level
recognition it has received as a development goal. It also highlights the prominent role of
institutions, governance, and data tracking programs in global efforts to accelerate the
pace of water and sanitation access, both currently and in the post-2015 period. In the
next chapter, literature relating to public service provision, institutions and governance,
and other relevant areas of public affairs is reviewed to form the basis for my research
questions and study design.
11


CHAPTER III
LITERATURE REVIEW
This chapter discusses relevant literature for this study, beginning with the
rationale for classifying water and sanitation service, the basis for the country progress
outcome variables, as a public good. It lays out a number of arguments for why water
and sanitation service qualifies as a public good and why it should therefore be the
responsibility of government. It then provides an overview of the literature on
institutions and governance, the subject of the explanatory variables, emphasizing the
important role they play in government service provision. These areas of literature
represent the general theoretical groundwork for this research.
Later sections review more specific literature relating to each of the three
attributes referred to in the research questions about country progress on providing urban
water and sanitation access. Some articles discuss these attributes in general, with
applicability to various types of government services, while others are specific to the
water and sanitation sector. The first topical area reviewed is decentralized government
services, the second is sector-wide strategy and investment coordination, (which also
relates to foreign aid), and the third is civil society engagement.
A final section provides a brief overview of literature about country
characteristics that are introduced as control variables in the statistical analysis, due to
their potential impact on country water and sanitation service provision.
Water and Sanitation Service as Public Goods
Water and sanitation service provision can be considered as a public good, one of
four types of basic goods consumed in the economy (Weimer & Vining, 2005).
12


Classifying it as public good has implications for how a country might establish policy
and organize institutions to provide the good. Social scientists classify goods into one of
four categories by two characteristics: subtractability and exclusion. Subtractability
refers to the extent to which one persons use of a good or service limits its consumption
by another person. Exclusion is the difficulty of restricting those in the community who
may benefit from the good or service. Goods that are characterized as having both low
subtractability, meaning that many people can jointly benefit from it, and low
excludability, meaning its access is difficult to restrict, are considered public goods
(Weimer & Vining, 2005).
The reasons why public goods are not considered suited to provision by the
private sector market include information asymmetry, monopoly power, externalities, and
allocative inefficiency that leaves some consumers underserved. Without adequate rules
governing who in the marketplace will provide a public good, and how, a public good
will often be undersupplied. Moreover, inequities in who benefits from the good may
exist with private provision, a problem that can be addressed through redistributive
government policies and implementation of targeted government programs (Weimer &
Vining, 2005).
Problems faced by the marketplace in providing public goods include difficulty
knowing consumer valuations for these goods, and problems of crowding and congestion
that may occur with low excludability goods (Weimer & Vining, 2005). One way to
address the latter problem is a system of usage rights for access to facilities that supply
water and provide sanitation. It has been shown that government and communities are
13


often in a better position to design and implement a system of joint rights and usage than
the private sector (Ostrom, 2003).
Justification for government provision of public goods stems from the overall
arguments for government intervention in the marketplace: (a) to correct market failures,
thus improving efficiency in the production and allocation of resources and goods; and
(b) to reallocate goods to achieve distributional and other societal values (Weimer &
Vining, 2005). For this sector, theory says that efficiency goals can be better served
through government provision if it addresses problems of undersupply and promotes
distributional and social welfare goals by extending services to marginalized groups,
thereby reducing health problems caused by poor water supply and sanitation.
The positive overall impact to society of water and sanitation access is the reason
it can also be considered as a merit good, something of value for the government to
promote, whether or not it can produce it efficiently (Musgrave & Musgrave, 1989). The
value to society comes via prevention of disease in children and adults, including
infectious diseases that can results from the ingestion of contaminants due to lack of
access to potable water and use of inadequate sanitation facilities. Given the negative
health externalities associated with a lack of clean water and sanitation facilities, there is
a societal rationale for the state and the global community to promote these goods. Doing
so minimizes communal health problems that potentially have global impact and extend
beyond national borders.
If social goals for water and sanitation are not addressed, there are detrimental
implications for the poorest members of society (Wijaya, 2005) and, as Rawls (1999)
proposed, equality in basic social goods is a fundamental right. Society should seek to
14


maximize benefits to those with the lowest level of prosperity, that is, persons with the
lowest utility on societys social welfare function.
Private Sector Involvement
The view of water and sanitation service as a public good best-suited for
government provision may seem at odds with country policies that allow for private
sector involvement in service provision. The key point here is that even when water and
sanitation services may be supplied by a private organization, if it is doing so under
contract with the government, or under some form of public-private partnership, the
government is still taking responsibility for provision of the public good to consumers,
and not leaving provision subject to the private market. All models of government
service provision, including those engaging the private sector, are impacted by the
national institutional and governance environments being studied for this research.
As noted by a UN expert, Governments can opt to have the private sector
provide water and sanitation. Likewise, they can choose public service provision. In
both cases, they must comply with their human rights obligation. Further, she stated
that governments must fully implement their obligations to create an enabling
environment and to regulate and monitor private providers (UN Office of the High
Commissioner for Human Rights, 2010, n.p.).
The introduction of private sector operators was intended to increase efficiencies
and overcome the high fixed costs associated with construction of water and sanitation
infrastructure (Kessides, 2004). However, the share of private sector involvement in the
sector globally has been proportionally small, and its impact can be difficult to evaluate.
Experience has often fallen short of expectations, both for investors and for recipient
15


countries, sometimes leading to highly contentious debates about the merits of this
approach. The initially anticipated flow of private investments to the sector did not
occur, often due to perceived investment risk resulting from inadequate investment
frameworks in some developing countries (OECD, 2009). Annual levels of private
investment in the sector through 2014 have been below the 2000 level. Projects are
concentrated in certain regions, particularly East Asia and Latin America
According to the World Bank Private Participation in Infrastructure Database
(2015), which measure private sector investment in different categories of infrastructure
investments, concessions are the most common types of arrangements. Concession
contracts are when a public entity grants the right and the obligation to provide a utility
service to a private company under terms and conditions specified in a contract or
license. The private sector partner takes over operational responsibility and at least part of
the commercial risk of service provision. The contract terms and conditions often specify
results to be achieved by the private sector partner (World Bank, 1995).
A 2004 overview report prepared for the World Bank noted the difficulty of
assessing the added value of private sector involvement in the sector because the
conditions built into individual service contracts between government and the private
sector can vary greatly and include specific conditions not made public. Also, the history
of private sector involvement in this sector is fairly recent, particularly as compared to
other infrastructure sectors such as transportation and communication, so there is not a
long history for analysis (Kessides, 2004).
In comparing public and private sector potential to contribute to developmental
goals such as MDG Target 7.C, Lobina and Hall (2008) found that the public sector has a
16


comparative advantage over the private sector for extending service delivery, with one
key factor being knowledge sharing among stakeholder networks. Their study, based on
a review of current practices in several large urban centers globally, found that in
transition and developing countries, reforms and good practices can create highly capable
public operators. In developing country urban areas, where systems of accountability,
transparency, and public participation are priorities, the informal stakeholder networks
that develop from these activities were found to be efficient ways of generating and
communicating knowledge. This exchange then translated into better public decision-
making around services. The authors see a crucial role for public providers and
recommend that donor aid be directed towards strengthening public sector utilities
towards MDG attainment (Lobina and Hall, 2008).
Summary
To synthesize the above literature, water and sanitation service access exhibits the
characteristics of a public good, and arguments can be made that the government has
advantages over the private marketplace for providing these goods. These arguments
center on improved service efficiency and greater equity of distribution. Also, because
provision in low-income settings may be complicated by problems of crowding and
congestion, some form of community collective action led by government may be
needed. Finally, given the global recognition of water as a basic human right, and the
negative health externalities associated with poor water supply and sanitation, their
placement in the special class of merit goods provides strong justification for government
provision, as well as for support from the international donor community.
17


Even in countries where the government may enter into service agreements with
private sector operators, the government still retains responsibility for service provision
and the operations are impacted by the institutional and governance environment.
The arguments discussed above justify government and donor participation in
water and sanitation service provision, leading to the next stage of literature review about
how institutions and governance impact the provision of public goods.
Water and Sanitation Institutions and Governance
How governments organize themselves to provide water and sanitation service
access, in terms of institutions and governance, is the focus of this section. Institutions
determine decision and incentive structures within governments, while the broader
concept of governance also encompasses the values and processes by which government
authority is exercised.
The literature reviewed here is mostly about water, as more scholarly articles are
available on water than sanitation. However, parallels can be drawn from one to the
other, as management of the two are often interlinked, and progress in both sub-sectors is
impacted by institutional and governance programs.
Institutions
Institutions have been broadly defined as the rules of the game in a society...
[that] structure incentives in human exchange, whether political, social, or economic
(North, 1990, p. 3). A major role of institutions in a society is to reduce uncertainty by
establishing a stable structure for human interaction (North, 1990). Institutional
arrangements within governments determine, among other things, who makes decisions
and who allocates available resources. These arrangements can play an important role in
18


government performance in the delivery of public goods (Cremer, Estache, & Seabright,
1994). Efficient service provision requires coordination within government and
sufficient institutional capacity (Frank & Martinez-Vazquez, 2014).
Institutions are key to designing, constructing, and managing water and sanitation
services. A definition of water institutions has been derived from the general definition
of institutions as: rules that taken together describe action situations, delineate action
sets, provide incentives and determine outcomes both in individual and collective
decisions related to water development, allocation, use and management (Saleth &
Dinar, 2005, p.2). The ability to effectively engage resources and capabilities toward
water and sanitation service provision depends largely on institutions, as these institutions
encompass the rules and mechanisms for: access to and allocation of resources;
placement of economic incentives; and systems of service design and delivery (Saleth &
Kadushkin, 2005).
There is a fundamental link between institutional reforms and changes in service
delivery, as demonstrated by a water sector framework developed by Saleth and Dinar
(2005). The framework shows that understanding the structure and environment of water
and related institutions, and the changes that take place within, can be useful for
approaching reform in these sectors. The framework considers the institutional structure
of water institutions, (further broken down into legal, policy, and organizational
components), and shows how important linkages exist both within water institutions and
also between these institutions and the broader political and economic environment in a
country. Understanding channels of institutional change are important for designing and
implementing reforms to realize desired changes in water sector performance.
19


Water and sanitation institutions are vital to the topic of this research because they
determine service delivery incentives and decision frameworks. The broader concept of
governance, discussed in the next section, incorporates the values and processes by which
government fulfills its service delivery function, including interactions with citizens and
organized groups outside of government.
Governance
Several different definitions for governance have been developed by researchers
and international organizations. A widely accepted definition from UNDP most closely
aligns with the intent of this research: the exercise of economic, political, and
administrative authority to manage a countrys affairs at all levels... [including]
mechanisms, processes, and institutions through which citizens and groups articulate their
interests, exercise their legal rights, meet their obligations, and mediate their differences
(UNDP, 1997, p.2).
The Global Water Partnership defines water governance as .. .the range of
political, social, economic and administrative systems that are in place to develop and
manage water resources, and the delivery of water services, at different levels of society
(Global Water Partnership, 2002, p. 1). Governance in the water sector addresses when
and how people get water and whose voices are heard in decision-making (Tropp, 2007;
Rogers & Hall, 2003; Plummer & Slaymaker, 2007). Effective water sector governance
includes approaches that are open and transparent, inclusive and communicative,
coherent and integrative, and equitable and ethical, as well as performance and operations
that are accountable, efficient, responsive and sustainable (Rogers & Hall 2003, p. 27-
29).
20


Governance in the water sector has not received the same level of attention as has
historically been focused on the technical and infrastructure aspects of water provision. It
started to be a focus of international development in the late 1980s, as it was realized that
much of the investment in technology and infrastructure by donors and the World Bank
did not realize its intended effect. There is now an increasing trend towards framing
water challenges in international development in terms of governance. Effective
governance has been linked to increased likelihood of sustainable water management by
Iribarnegaray and Seghezzo (2012) who developed a Sustainable Water Governance
Index showing value in participation/engagement between the public and water
authorities.
New governance approaches are broader than traditional inter-governmental
arrangements, and therefore more inclusive. Accountability is a key component of
governance, and new actors in the governance process are monitoring the actions of
government institutions at different levels. As a results, institutions have become more
transparent and more accountable, realizing a trend away from strict inter-governmental
relationships to relationships between government actors and nongovernmental actors,
including for global institutions and issues (Nye & Donahue, 2000). By forging strategic
coalitions with actors outside of government, governance practices are said to enhance
governments capacity to perform its functions (Peters & Pierre, 1998).
Tropp (2007) found that new forms of governance, which engage governments,
non-government organizations, the private sector, academic and research institutions, are
a way to stimulate innovative partnering for implementing global initiatives, including
the MDGs. However, these new forms of governance may require capacity building for
21


policy makers and managers so they may adapt to new forms of governance, including
coordinated and negotiated approaches. These new approaches apply to both large-scale
national issues and local-level service provision.
This overview of literature relating to institutions and governance shows the
important role that they can play in government service provision. The next section will
focus in on three specific attributes of institutions and governance and discuss how they
relate to urban water and sanitation service provision.
Literature on Attributes Selected for Study
This section overviews literature related to the three specific institutional and
governance attributes selected for study: a) decentralized services; b) sector-wide strategy
and investment coordination; and c) civil society engagement. At the end of each topical
discussion, a summary of the literature is provided which forms the basis for the
hypothesis linked to that attribute and the development of construct-level definitions.
Decentralized Services
National and regional water service providers in developing countries have often
historically failed to perform efficiently and in a sustainable manner, due to lack of
accountability to customers and communities dispersed across a large geographic area, as
well as constraints to operations from bureaucracy and politics (Pilgrim, Roche, Revels,
& Kalbermatten, 2004). With the onset of sector decentralization, implemented as part of
general institutional reforms in many developing countries, came an expectation of
improved services. The body of literature relating to decentralization, both in general and
specifically for the water and sanitation sector, reveals the rationale for decentralization
and highlights the conditions under which anticipated benefits are most likely realized.
22


Decentralization is defined as the transfer of planning, decision making, or
management functions from the central national government to organizations at the
regional or local levels (Handoussa, 2009, p. 217). While not explicitly stated in this
definition, many articles note that decentralized revenue-raising authority and
expenditure decisions, termed fiscal decentralization, is a key sub-component of
decentralization. This aspect is further discussed later in this section.
In a meta-analysis of US and international studies of government performance,
researchers found that the degree of centralization, size, and formalization of agencies
were among the key institutional variables associated with government service
performance (Forbes & Lynn, 2005, Hill & Lynn, 2005). In the water sector specifically,
decentralization was found to be one of the key policy elements having an impact on
sector performance in an analysis of water institutional reforms in 43 countries/regions
(Saleth & Dinar, 2004).
Driving decentralization in the water and sanitation sector is the belief that lower
levels of government can be more effective in responding to local conditions and meeting
the needs of local consumers, thereby making consumers more willing to maintain and
pay for services that they feel respond to their needs (Briscoe & Gam, 1995; Litvack &
Seddon, 1999). It is thought that local governments are in a better position to assess local
conditions, identify local needs, and make decisions about provision of collective goods
(Blair, 2000; Wunsch & Olowu, 1990).
Decentralization is supported by examining some of the inherent problems in
centralized approaches. Looking at institutional incentives in the provision of urban
sanitation, Evans (1995) found that when decisions about what level of service to provide
23


in different localities are left to the central government, there can be a problem of low
government accountability, particularly to low-income groups. The central government
can behave opportunistically and put political priorities ahead of social goals. Further,
there is likely no mechanism at the central government level for establishing user
preferences for the type of sanitation service to provide when overseeing multiple
communities, and central government organizational culture tends to work against the
development of innovative designs.
Decentralization has been promoted as an alternative institutional arrangement
that addresses the problems of centralized management. Literature on decentralization
explains how decentralization can promote political, efficiency and governance values
(Wolman, 1990; Pius Kulipossa, 2004), each of which are discussed here. Political
values can be served by spreading political power to a wider range of citizens as
compared to a centralized approach. Benefits include: political training and education for
local officials; political equality, rather than concentrated power, is enhanced; and
political security is better secured through increased participation in formal politics,
support for political parties and other practices. Decentralization may also improve inter-
organizational coordination (Pius Kulipossa, 2004).
Efficiency can also be enhanced by decentralization. Efficiency is influenced by
the levels of authority and jurisdictions present in an institutional environment, according
to polycentric governance theorists (V. Ostrom, Tiebout, & Warren, 1961). To be
effective, decentralization must be accompanied by incentives that motivate individual
and collective action (McGinnis, 1999; McGinnis, 2011). The appropriate degree of
24


decentralization depends upon which level of government will have the most incentive to
act to bring about desired outcomes (Cremer, et al., 1994).
Allocative efficiency can be improved because of the ability to capture diverse
preferences for public goods (Tiebout 1956; Oates, 1972; Musgrave 1997), and it also
encourages competition and efficiency, thereby offering citizens more choices and
options (Rondinelli, McCullough, & Johnson, 1989). Oates (1972) put forth a general
decentralization theorem that in the absence of cost savings from the centralized
provision of a public good, and as long as one communitys production does not interfere
with another community, welfare will be the same, if not better, when each community
designs a program to meet its specific needs, rather than being subjected to a standard,
centralized approach. More recently, Oates (2005) extended these ideas to consider
problems associated with soft budget constraints under decentralization, suggesting that
fiscal decentralization may require sufficient power at the central government level to
moderate excessive financial demands from local governments. Some level of
competition for funds among local governments can be helpful to avoid inefficient
spending, and central governments should provide a stable and enduring federal structure
to moderate local spending, without overly extending its role under a decentralized
system.
Decentralization can also be viewed as supportive of good governance, promoting
positive interactions between the government and the public. In particular, it can cause
government officials to be more responsive and accountable to citizens (Wolman, 1990).
According to Azfar, Kahkonen, Lanyi, Meagher, and Rutherford (2004), decentralization,
in its best sense, is a system where institutions and information flow produce effective
25


government and where accountability of government officials should increase, especially
if actions are transparent and civil society is included. Decentralization can address
credibility and information problems, especially if monitoring is shifted to a community,
which reduces the information needs of citizens to hold politicians accountable (Keefer &
Khemani, 2005).
Along with the perceived positive value outcomes from decentralization, the
literature has also explored potential problems with decentralized systems, the key
arguments of which are summarized below.
One such argument relates to power abuse within localities. Local provision of
public services may be subjected to inequalities due to the power of local elites who can
influence public funds expenditures more easily in a decentralized system (Bardhan &
Mookheijee, 2005; Mookheijee & Bardhan, 2005). Local government officials may be
more subject to capture by local interests that can result in overprovision to certain
powerful groups (Azfar et al., 2004; Bardhan & Mookherjee, 2000). Also, local officials
have the capability to withhold contributions needed for successful implementation, and
the central government may lack the authority to promote corrective action and will incur
monitoring costs for oversight of local authorities (Weimer & Vining, 2005).
When management systems are largely lacking within a country, some
infrastructure services may be more capably handled at a higher level of government
(Bahl & Wallace, 2005). There are potential economy of scale benefits with centrally-
planned projects, although the same might be realized under a decentralized system with
local communities partnering together for joint projects.
26


Another advantage of centralized authority is that it can equalize conditions
among rich and poor areas by transfer of tax revenues to areas most in need. Moreover,
in a developing country context, the intended efficiency benefits of fiscal decentralization
may not be fully realized due to: a lack of elections where the local population can reveal
preferences for public services; a lack of legal authority granted to local jurisdictions for
tax collection of expenditures; and limited administrative abilities to support increased
spending or revenue collection (Bahl & Linn, 1992). The impact of decentralized service
delivery depends on the method chosen for financing local governments, as restraints on
the revenue-raising capability of local governments can be limiting. If a central
government is generally struggling to provide basic services to the population, it may be
reluctant to give up power and resources to sub-national levels of government and
decentralization can be ineffective in practice (Jutting, 2005).
Turning now to country-specific studies of decentralization, this section highlights
what they have found about the conditions that promote success. A common theme is
that decentralization in developing countries may not realize all anticipated benefits if
certain enabling factors are not in place.
The success of decentralization may be contingent on local authorities being
granted both decision-making and revenue raising authority. In a study in Zimbabwe, it
was found that decentralization of responsibilities from the central government did not
actually result in decentralized authority. Local communities lacked autonomy in
decision making, which limited their ability to effectively manage local resources,
revenues, and personnel. The inability to raise revenues locally was a major impediment.
For effective urban management in Zimbabwe, and Africa in general, the author
27


recommended more efforts towards strengthening the role of democratically elected
officials and civil society (Davison, 2001). Another study of decentralization in Uganda
and the Philippines, looking at the health and education sectors, showed that local
governments have limited authority and are unable to adjust services to meet local
demands (Azfar, Kahkonen, & Meagher, 2001).
Studies focused on local water supply have shown that some success is possible
with local management, but support is needed to address institutional challenges and to
realize sustained improvements in service (Seppala, 2002; Pilgrim et al., 2004). Mugabi
and Njiru (2006) studied small communities in low-income countries and found that
while local authorities may be in a better position to understand local needs, they lack the
technical and management capacity to respond to the needs at hand, due to weak
institutional capacity. They recommend that the transition to decentralized water service
delivery be done in tandem with policies and programs that address expected challenges,
such as management capacity building and technical assistance programs, and granting
authority to raise revenue for financing of operations. Financial and managerial
autonomy, transparency and accountability, as well as professional support, are key
ingredients for realizing improved efficiency and effectiveness of service (Mugabi and
Njiru. 2006).
Similarly, Wilder and Romero (2006) found that continued state involvement was
required to ensure equity, accountability, transparency, and sustainability of decentralized
water institutions in Mexico. Marcus and Onjala (2008), in studying areas of Kenya and
Madagascar, found that national governments did not have in place the necessary legal,
juridical, and regulatory frameworks to promote successful decentralization. While
28


decentralization was expected to increase accountability and improve governance of
water systems, it actually resulted in central government pulling back from providing
leadership in water governance too soon, before problems of local funding gaps and lack
of local resources for management were addressed.
In rural India, Asthana (2003) found there is often a long-term strategy lacking
with decentralization and an authority closer to the community level will not always
provide better service. However, decentralization does offer the advantage to the local
level of learning by doing, which will likely eventually result in greater efficiencies and
less corruption than centralized service, especially if there is guidance from the state
government and oversight by the public (Asthana, 2003).
Role clarification between local and national government is also an important
element. Regulatory frameworks are a way to clarify roles, but putting these frameworks
into practice can be challenging. Marcus and Onjala (2008) suggest negotiations to
clarify state and local rights, responsibilities, and capacities with regard to
decentralizations, taking into consideration culturally embedded local norms.
As mentioned above, fiscal decentralization is important in addition to managerial
decentralization (Bardhan & Mookherjee, 2000; Mugabi & Njiru, 2006; Gillespie, 2005;
Rodinelli, 1990), because it allows sub-national governments to exercise more autonomy
in serving the local population. Gillespie (2005) pointed out that one of the factors
limiting finance to the water sector in many countries is failure of water sector
governance, including decentralization programs that do not clarify roles of central and
local governments, lack of financial autonomy of decentralized authorities, and lack of an
appropriate operations framework for utilities.
29


According to Rodinelli (1990), central restrictions on the financial authority of
local governments should be eased to facilitate decentralization, and there should be
capacity building to increase skills for raising local revenues, managing expenditures, and
delivering and maintaining efficient services. There is a risk of local governments over-
taxing and over-regulating that may hinder overall economic investment and growth.
This requires careful consideration of local government revenue-raising instruments and
expenditure plans (Bahl, Linn, & Wetzel, 2013). When shifting financing to sub-national
levels of government, it is important to create a governance environment that promotes
public services delivered in cost-effective ways. Even if funding to local authorities is
adequate, inappropriate management of funds can result in a perceived lack of financial
resources within government structures (Funke et al., 2007).
Taken together, the literature points to several anticipated benefits of
decentralized government services, including: local authorities being more responsive to
local needs and preferences than centralized systems; more efficient service provision
because programs can reflect diverse preference and be designed to meet its specific
needs, rather than being subjected to a standard, centralized approach; more accountable
and transparent systems, as the degree of information asymmetry is reduced; better
representation of good governance values, including equity in government services,
positive government interactions with citizens; and improved political equity and political
security with political training and education for local leaders.
There are also other studies that suggest that decentralization benefits may not be
fully realized if certain factors are not in place. For example, incomplete delegation of
decision-making, lack of authority for revenue raising and investments, unclear roles and
30


procedures for different levels of government, and lack of managerial and technical
capacity among local managers to adapt to new responsibilities. Other potential
problems with decentralization include power abuse by local authorities, lack of proper
incentives to motivate individual and collaboration action on the part of decentralized
authorities, unequal financial resources among different sub-national regions without the
benefit of resource equalization offered by a centralized approach. Decentralization
benefits may also be limited by circumstances that stem from economic, political, and
socio-cultural forces at work, such as historical tensions between central and local
authorities and a history of low citizen engagement with government.
In short, decentralization has the potential to address many of the water
infrastructure and planning challenges, but it requires an extensive range of institutional
pre-conditions to be successful; improving the institutional framework under which
local management initiatives operate (Mugabli & Njiru 2006, p. 191).
The literature highlights several conditions necessary for decentralization to
realize theorized benefits, and an interesting line of research is to explore to what extent
these conditions are present in countries where urban water and sanitation provision is
decentralized. So, while the hypothesis stated below assumes an overall positive impact
for decentralization, phase two of the two-phase study approach will provide a more
insightful analysis by focusing on the extent to which important aspects are present in
decentralized countries. Following are the research question and hypothesis related to
decentralization:
RQ1. What is the role of decentralization of water and sanitation services in a countrys
progress on providing urban water and sanitation access?
31


HI. A high level of decentralization in water and sanitation services is associated with
greater country progress on providing urban water and sanitation access.
The important aspects of decentralization discussed in the literature will be
studied through the constructs for decentralization as follows:
a) Decision-making authority for service provision is delegated to sub-national
levels of government.
b) Authority to manage expenditures and raise revenues is delegated to sub-national
levels of government.
c) There are clearly defined rights, responsibilities, and roles for sector management
actors at the national and sub-national levels.
d) Sub-national level authorities are monitored by communities and/or the central
government including, for example, regarding public funds expenditures and
equality in service provision.
e) Support is provided to build managerial and technical capacity of sub-national
levels of government.
Sector-wide Strategy and Investment Coordination
In this section, I review literature on international development aid effectiveness
and relate this to recent initiatives for coordinated national planning for water and
sanitation. I introduce key elements of theory and current recommendations for what
constitutes effective international development aid, and tie that discussion to relatively
new approaches that have emerged to address identified shortcomings of previous aid
modalities. These new approaches are grouped under the heading of sector-wide strategy
and investment coordination, defined for this research as: mechanisms for transparent,
country-led partnering with donors as part of coordinating, directing, and utilizing
international aid towards priority areas for the sector (adapted from Tropp, 2007).
32


Multi-lateral and bi-lateral donor aid programs assist countries with provision of
basic public services, such as schools, health facilities, and infrastructure. However,
donor aid programs over the last several decades have been found to lack a long-term
effectiveness in promoting economic growth (Gibson, Andersson et al., 2005; Riddell,
2007; Boone, 1996; Pedersen, 2001). Depending on how it is structured, international
development aid can actually create motivational problems and impede a countrys own
ability to promote development. A problem of cryptoimperisalism can emerge when
policymakers in developing countries are influenced by the ideas of foreign donors on
which they depend so heavily (V. Ostrom, 1999).
Financial dependency on foreign aid fosters a power structure between donor and
recipient countries that can create incentive problems. Buchanan (1977) used the term the
Samaritans Dilemma to characterize a motivational problem created when international
aid is viewed as a type of two-person game with ordinal payoffs. A Samaritan,
concerned about the well-being of others, enters into situations in which other people
might need help and decides to help or not to help. The recipient also decides how much
effort to put forth to receive the assistance. The dilemma is created when Samaritans
(international aid donors) achieve their goal of helping others no matter what level of
effort the recipient puts forth, and so this becomes their dominant strategy. An aid-
recipient country, once it understands this reality, will put forth a low level of effort as its
dominant strategy. This may lead to inefficiencies over time, with the distribution of
benefits skewed to the recipient. It also creates a motivational problem on the part of the
recipient. Under such a situation, the recipient loses skill and motivation, thus impeding
its own capability to manage and sustain programs over time (Buchanan, 1977).
33


Research on past donor aid programs has shown that the failure to fully achieve
goals of donor aid programs has not resulted so much from a lack of donor assistance
funds flowing to the less-developed countries, as it has from a lack of effective
institutions to manage and administer the funds (Auer, 2007; Gibson, Anderson et al.,
2005). If institutional arrangements do not promote ownership and coordination of
international aid on the part of the recipient country, international aid flows may fail to
achieve desired and sustainable results in improving the conditions in less developed
countries. In addition to funding projects that address primary public goods deficiencies,
development aid policy should also help design new institutional rules that guide
government decision-making and management over the long-term (Gibson, Andersson et
al., 2005)
Principal/agent theory can be applied to multiple relationships in the scheme of
foreign aid (Moe, 1984; Martens, Mummert, Murrell, & Seabright, 2002). The principal-
agent model is an economic model in which the principal enters into a contract or
agreement with the agent in the expectation that the agent will choose actions that
produce outcomes desired by the principal (Bendor: 1988, Miller: 1992).
Some applications of this model to foreign aid are: (a) donor aid implementing
agencies (agents), which are subject to multiple principals within their own countries,
including politicians, taxpayers and public officials; (b) complex incentive and control
relationships among donor agencies (principals) and implementing contractors (agents);
and (c) the direct relationships between foreign donors (principals) and aid-recipient
countries (agents). I explore relationship (c) in the following discussion, as my research
34


is focused on recipient countries. (See Martens et al., 2002 for further discussion of
relationships (a) and (b)).
Problems of asymmetric information can exist between principals and agents.
This leads to problems of adverse selection and moral hazard, which Rauchhaus (2009)
finds can obstruct humanitarian aid interventions. Looking first at adverse selection, this
stems from a situation where asymmetric information exists between the parties before
entering into a contract. With donor aid, there may be uncertainty concerning an agents
(recipient countrys) preferences prior to entering into a contract that are not revealed to
the principal (the donor) during the negotiation period.
Moral hazard, on the other hand, occurs when a party to an arrangement has an
opportunity to take action hidden from the other party once an agreement is in place, with
the principal unable to observe the agents behavior. This creates an opportunity for
agents to change their behavior in a way that increases their potential for gain. Both
adverse selection and moral hazard can apply to international aid agreements. The
problem is compounded by the fact that these agreements are often less binding and
enforceable than domestic political contracts, as they are essentially self-enforcing and
not bound by the law of any particular jurisdiction. Actions taken or not taken by the
agent (the aid-recipient country) may be difficult to assess and monitor by the principal
(the donor) or even by a neutral third party. These circumstances can foster a
commitment dilemma due to the weakly institutionalized nature of international politics.
Calmette and Kelkenny (2002) looked at asymmetric information problems in
donor aid and showed how they can lead to reduced country self-help. They present three
scenarios. First, when aid for public works is donated unconditionally, a recipient
35


government spends less on public works, levies the lowest taxes on the rich, and the
degree of self-help is at its lowest. Second, when charity is conditional, but information
sharing incomplete, countries are motivated to exaggerate their needs and downplay any
domestic self-help investment programs, creating an adverse selection problem. If
recipient countries perceive themselves as competing for aid with other countries,
recipient countries may adjust their needs to qualify for more aid (Pedersen, 2001).
Finally, when adequate information about internal development investment efforts is
known by the donor agency, it is in a better position to create conditions or incentives as
part of the aid terms, to mitigate potential negative effects of adverse selection, principal-
agent, and moral hazard problems (Calmette & Kelkenny, 2002).
In short, information asymmetry exists when a donor is unable to distinguish the
true level of need and effort of a country. Countries with more resources may hide their
capacity to address their own problems so that they receive higher transfers from donors,
thereby exhibiting shirking behavior. Corrupt, but potentially less needy, countries may
be able to capture informational rent. Calmette and Kelkenny (2002) suggest that having
incentives in place to encourage countries to correctly represent their domestic capacity
would help address this problem and increase the efficiency of aid, adding to overall
global welfare. These incentives can take the form of increased transparency
requirements and having mechanisms in place to monitor and penalize misrepresentation
of domestic efforts.
The recognition of potential incentive problems with traditional models of donor
aid has, among other things, spurned new thinking about how aid is approached at the
international level. Riddell (2007) found that donor aid is most effective when there is
36


predictability, efficiency in allocation, improved coordination, and meaningful
partnerships between donors and recipients. Gibson, Andersson et al. (2005) have added
that donors should focus on country ownership of development assistance programs to
encourage sustainability, and prioritize beneficiaries when designing development
programs.
An example of new-style governance is the sector-wide approach (SWAp), a
planning vehicle by which all significant domestic and donor sector investments in a
developing country are coordinated and directed towards the same overall objectives,
consistent with a sector strategy. The SWAp model has been used since the mid-1990s as
a method of international aid delivery in sectors such as health and education, based on a
set of operating principles rather than a specific package of policies or activities
(Development Assistance Committee, 2005). Donor support for a SWAp can take the
form of project aid, technical assistance or general budget support (OECD, 2006).
A SWAp allows development partners, (donors, governments, and other
stakeholders), to work towards a coordinated national program, rather than on project-
specific efforts. Characteristics of the SWAp include: broadening policy dialogue;
developing a single sector policy (that addresses both public and private sector issues),
and a common realistic expenditure program; common monitoring arrangements; and
more coordinated procedures for funding and procurement (WHO, 2000). With a SWAp,
aid partnership arrangements are shifted more towards recipient government leadership,
although there is joint cooperation in planning policy and projects. Key characteristics
are described by WHO:
37


Key characteristics of the SWAp should include: i) the partner government clearly
leads and owns the programme; and ii) a common effort by external partners to
support that programme, including provision of all or a major share of funding for
the sector, in support of the government's unified policy and expenditure
programme (WHO, 2013, p.l).
Both ongoing and new projects are screened to fit sector-wide priorities, aiming for more
comprehensive coverage of sector needs. Common procedures for project management
and funding are developed, increasingly led by the government. Additional benefits of
SWAps in recipient countries can include promotion of more local involvement in sector
development projects, more accountability of government officials, and increased
capacity building.
The SWAp approach is in contrast to the system of individual donor-driven
projects that sometimes led to fragmentation and duplication of efforts, and also placed
multiple demands on the recipient countrys economic and human resources for project
coordination. SWAps are intended to lead to more efficient use of financial resources
from national budgets, donors, and service users, maximizing overall investment in the
sector towards achieving national water and sanitation MDG targets (WHO, 2013; Harpe,
undated). Under a SWAp, there is less potential for hidden information, as details on all
planned sector projects and investments are widely shared among domestic and
international partners.
Similar to SWAps are country-specific poverty reduction strategy papers
(PRSPs), national planning documents stemming from the 1996 Enhanced Heavily
Indebted Poor Countries (HIPC) Initiative, by which G-7 governments offered low-
interest loans and other forms of debt relief to indebted nations. PRSPs are an approach
to reducing poverty and a method for structuring development assistance, intended to be
38


country-driven, partnership oriented, and developed with civil society participation. The
PRSP development process sought to promote increased transparency by governments
and donors (IMF, 2013). PRSPs have specific sections devoted to different development
challenges within a country and in some cases includes a section specifically on water
and sanitation infrastructure planning. PRSPs are a change from previous national
development strategies, which were often based on directives from the World Bank and
International Monetary Fund (IMF). While the IMF is currently transitioning from PRSPs
to economic development documents (EDDs) for some countries, these documents are
similar in nature in that they articulate and track an overall strategy for poverty reduction
and growth (IMF, 2015).
The degree to which the PRSP process has changed development assistance has
been assessed by several analysts. Cheru (2006) sees PRSPs as potentially important
instruments for streamlining and coordinating international aid flows, but that effort is
needed to ensure the PRSP approach remains participatory and results-oriented.
Abrahamsen (2004) sees the PRSP form of partnership as advantageous in that there is an
increased potential for inclusion, as well as an explicit commitment to the self-
governance of recipient states. Partnerships such as the PRSP help to produce self-
disciplined citizens and states by enlisting them as responsible agents in their own
development (Abrahamsen, 2004, p. 1464).
According to WHO (2013), countries that use both SWAps and PRSPs need to
work toward consistency between what is presented in both plans, including
arrangements for stakeholder buy-in, and plans for implementation, monitoring and
improved accountability. Countries with PRSPs may foster greater interest in developing
39


a SWAp, and having a SWAp can become a mechanism for work towards
implementation of the PRSP (WHO, 2013).
The review of literature on this topic highlights both the shortcomings of
traditional approaches to international development aid, stemming largely from
information, incentive and ownership problems, as well as the theorized potential of new
approaches designed to address these problems, such as SWAps and sector-specific
sections of PRSPs. These new approaches bring with them the expectation of greater
country ownership of sector projects, more predictability and efficiency in aid allocation,
and greater potential to monitor actions undertaken by donors and aid recipient countries.
As these approaches are relatively new to the sector, my second research
questions and hypothesis are designed to test if they are realizing the theorized benefits in
the realm of urban water and sanitation service provision:
RQ2. What is the role of sector-wide strategy and investment coordination in a countrys
progress on providing urban water and sanitation access?
H2. A high level of sector-wide strategy and investment coordination in water and
sanitation services is associated with greater country progress on providing urban water
and sanitation access.
To study the impact in greater detail, specific constructs were drawn from the
literature as follows:
a) The recipient country identifies priority projects and develops a realistic
investment plan for the sector
b) Development of partnering agreements between donors and recipient countries is
driven by the recipient country
c) A single agency is charged with coordinating all donor aid received in
conjunction with a sector strategy
40


d) Sector-wide strategies are prepared in collaboration with stakeholders and donors, who
are provided with information about the sector
e) A system of monitoring is in place for compliance with the terms of aid partnering
agreements
Civil Society Engagement
In this section, I review literature in the area of civil society engagement to see
how it informs a third research question related to water and sanitation governance. This
question looks specifically at the degree to which a countrys governance environment
supports civil society engagement in issues surrounding urban water and sanitation
service provision, and how this impacts progress in increasing water and sanitation
service coverage in urban areas.
The concept of civil society engagement is defined for this research as: the
existence of opportunities for the public and consumers to become involved in
government decision making, including program planning and implementation (adapted
from deLeon and deLeon, 2002). I consider civil society engagement to include
interactions between government and formal civil society organizations, as well as
between government and the public at large as consumers of water and sanitation
services. Therefore, the terms civil society engagement and public participation are both
used in this discussion.
The Organization for Economic Cooperation and Development (OECD) defines
civil society organizations (CSO) as, the multitude of associations around which society
voluntarily organizes itself and which represent a wide range of interests and ties. These
41


can include community-based organizations (CBO), indigenous peoples organizations
and non-government organizations (NGO), (OECD, 2006).2
Engagement with the beneficiaries of public services, often through organized
CSOs, is thought by many to enhance the effectiveness of government. The new public
administration movement in the US during the 1960s called for more public involvement
in public management, coinciding with the Great Society initiatives that called for greater
social equity in government programs, particularly for previously marginalized groups
(Thomas, 1995). DeLeon and deLeon (2002) advocated for increased opportunities for
citizen participation to promote democracy, not only in policy development, but also in
policy implementation through participation in public management decision making.
Similarly, political choice and citizen participation in the process of policy definition and
formation is emphasized by the field of discursive politics, so that analytical processes
may include a wide range of interests, arguments, and discourses. Discursive approaches
challenge the more traditional and dominant positivist, technocratic approaches to the
social sciences and place emphasis on social meaning and values, gained through
interactions between citizens and experts. This approach is touted as a way of
democratizing policy deliberations (Fischer, 2003).
Several authors studying water and sanitation service provision have advocated
for more engagement with civil society. Saleth and Dinar (2004) recommend that
decision processes for water accommodate the increasing role of user groups, non-
2NGOs are typically issue-based, legally registered, not-for profit, and have some
paid staff. CBOs are community-based, voluntary, membership organizations made up of
individuals who have joined together to further a common interest (WaterAid &
Freshwater Action Network, 2011).
42


governmental organizations, women, environmental and self-help groups. Others make
the case that water supply governance should include citizens participation to uphold
principles of social and environmental justice (Wijaya, 2005). Gillespie (2005) found
that limited opportunities for the public and consumers to become involved in decision
making about the sector can also impede sector finance.
In the context of water and sanitation services, an example of incorporating
citizen input would be consultations about technology choice that include input not only
from technological experts, but also from the community about their lifestyle and social
customs. Lynn (1994) said that what counts as a good solution depends a lot on the local
context. Mirumachi & Van Wyk, (2010) recommend that regulatory frameworks
incorporate mechanisms for consultation with users, so that they may express preferences
about service delivery. Local consultation mechanisms increase the likelihood that the
desires of water users are reflected in water management.
The central government has a role in generating the enabling conditions that make
this possible (Madrigal, Alpizar, & Schliiter, 2011). Two global water NGOs propose
that civil society engagement must go beyond basic consultation, to include equipping
citizens with knowledge, skills and confidence to undertake the advocacy for themselves
and eventually by themselves (WaterAid & Freshwater Action Network, 2011, p.9).
The effects of citizen participation depend on whether the approach is based on
understanding, not just on short-term incentives or pressure, and whether the public can
participate in articulating problem causes, which may then place them in a better position
to identify and adopt potential solutions (Oakley, 1991).
43


Civil society engagement also has the potential to improve redistributional equity
in the provision of public services. Greater participation can be a check against the often-
prevailing institutional culture, power relationships, and social constructions that favor
traditionally advantaged groups as the primary beneficiaries of government policy
(Schneider & Ingram, 1993, 1997), a situation perhaps even more pronounced in
countries where wide gaps of wealth exist among citizens. A governance environment
that favors greater participation could promote increased access to water and sanitation
services by all classes of citizens, not just traditionally advantaged groups.
In an urban setting, policies that promote broad-based participation and that
define strategies for actual involvement can be useful ingredients for sustainable service
delivery. The interests of poorer populations, particularly in urban areas, can be better
represented through open, inclusive, coherent, and equitable decision-making processes.
The poor can also gain more power and influence through governance systems that
facilitate direct participation in service provision, to include planning, installation,
management and monitoring of services (Sohail, Cavill, & Cotton, 2005; McGranahan &
Satterthwaite, 2006).
Oakley (1991) said that public participation should be appreciated not just for
idealistic and egalitarian purposes, but also as an important element to increase project
efficiency and effectiveness, and to encourage self-reliance. In a study of US
metropolitan governance, Oakerson and Parks (1999) suggest that interaction between
citizens and government officials can explain increased efficiencies in government
services, as it prompts officials to devise productive organizational arrangements to meet
the needs of local citizens.
44


Public participation leads to increased accountability in government. One of the
ways that institutional rational choice theory evaluates institutional outcomes is based on
the idea that in a democratic society, public leaders should be accountable to citizen
regarding the use of public funds and natural resources (E. Ostrom, 2005). This requires
that information about the preferences of citizens be made available to decision makers,
which can be achieved through various forms of civil society engagement. Management
transparency and accountability are critical to gain and maintain the trust of users and
investors in the water sector, including clearly defined roles and responsibilities,
independent monitoring, information disclosure, and consultations with consumers
(Mugabi & Njiru, 2006). In rural communities in Costa Rica, one of the most important
mechanisms linked to high performance in water provision was properly defined local
accountability for water services, as well as the ability of local leaders to generate
incentives for community involvement in sustainable solutions (Madrigal et al., 2011).
The 2004 World Development Report (World Bank, 2004) argues that for
government services to be effective, clients must be able to monitor and put pressure on
providers when needed. Hooper (2011) found that stakeholder engagement processes,
and accountability programs to monitor implementation actions, were associated with
effective river basin management, based on a global assessment of integrated water
resources management systems. Some of the decentralization studies discussed
previously have also found an inherent value of civil society engagement for improving
public services (Tropp, 2007; Iribamegaray & Seghezzo, 2012; Azfar et al., 2004;
Wolman, 1990; Mugabi & Njiru, 2006).
45


One avenue for public participation is having a feedback and complaint
mechanism in place for consumer input on water and sanitation services. The
importance of such mechanisms in low-income countries is emphasized by Li,
particularly when the government has privatized water service provision (Li, 2011). For
the new transfer-operate-transfer model of urban water supply projects in China, a form
of public-private partnership, researchers found that a critical success factor based on
four case studies was government supervision of the contractor and a mechanism by
which users could register complaints about the contractor with the local government.
For proven complaints, the local government then had the authority to order the
contractor to rectify and improve the situation (Meng, Zhao, & Shen, 2011).
A key success factor for the Phnom Penh Water Supply Authority in Cambodia, a
country challenged in governance capacity due to an unstable political history, is that it
forged social contracts with village associations for reporting thefts and leakages in the
system. Customer complaints and feedback were received through established channels,
responded to, and served to monitor problems within the system (Araral, 2008).
While the above studies found a positive association between public participation
and improved government services, promoting participation principles in countries that
have not previously adopted them poses unique challenges. Some general constraints
have been identified by Research Triangle Institute (2004):
poor public education systems;
lack of local, experienced non-governmental organizations; and
resistance from local government administrators and government officials to
opening up the decision making process
46


Even when public participation policies are in place, a gap may exist between
policy and how it is practiced in reality, often due to the political and cultural
environment. Traditional actors may play a predominant role in decision making and
new actors, such as civil society, may be present, but have limitations in their influence
on decision-making. In post-apartheid South Africa during the late 1990s, stakeholders
varied widely in their ability to understand and adopt government processes they were
not familiar with, and there was not a strong culture of participation in government
(Funke et al., 2007). While reformed water policies post-apartheid placed an emphasis
on inclusiveness and cooperative water governance, power disparities at the local level
often impeded the way that inclusiveness worked in practice, complicated by the
countrys history of non-inclusiveness (Mirumachi & Van Wyk, 2010).
Other considerations limit more inclusive water governance practices. In many
countries, inclusion of new actors in local-level discussions is viewed as needing
legitimization from the central government. While new participants may be present in
water discussions, they do not necessarily hold enough political power to influence
decision-making, and may perceive risks of low return on time and effort invested
(Mirumachi & Van Wyk, 2010). A study looking at health and education
decentralization in Uganda and the Philippines found that citizen influence at the local
level was constrained by limited access to information on the responsibilities and
performance of local government (Azfar et al., 2001).
Tevelow (2004) looked at the impact of NGOs and other CSOs in developing
countries, and found that while there are expanding networks among NGOs in different
countries, aided by Internet communication, there often remain institutional barriers to
47


their full participation in closed-door negotiations among governments and powerful
inter-governmental institutions, such as the World Bank and IMF. Empowerment may be
a social process that evolves over a period of time, and for true forms of cooperation to
develop and result in meaningful action, a long period of social capital formulation may
be required (Innes & Booher, 1999).
Even in the case of large-scale donor programs, the level of civil society influence
can vary. An independent study looking at two new forms of multi-stakeholder
partnerships in Africa (the UK-led Partners for Water and Sanitation and the European
Union Water Initiative), both of which have a strong focus on working towards MDG
achievement, found that governments remain the primary program partners and political
pressure remains the primary driver of activities (Stewart & Gray, 2006). For the
European Union program, governments were given a privileged position as partners,
relative to civil society. Further, the study found that the poorest citizens in developing
countries are often the most limited in terms of time available to commit to partnership
processes, due to activities required for their daily sustenance (Stewart & Gray, 2006).
For the most part, the literature on civil society engagement and public
participation points to several potential benefits of these activities, such as promoting
government transparency and accountability, improving redistributional equity, better
representing the interests of the poor, and improving the quality of government services
at various stages of the policy development and implementation process. To test these if
these findings hold true in the context of urban water and sanitation service provision, I
developed the third research question and hypothesis as follows:
48


RQ3. What is the role of civil society engagement in a countrys progress on providing
urban water and sanitation access?
H3. A high level of civil society engagement in water and sanitation services is
associated with greater country progress on providing urban water and sanitation access.
The following constructs for civil society engagement are designed to focus in on
how civil society engagement may be having a positive impact in different stages of
policy development and implementation:
a) Opportunities exist for consumers and civil society groups to become involved in
problem articulation.
b) Opportunities exist for consumers and civil society groups to become involved in
decision-making about policy development and planning.
c) Opportunities exist for consumers and civil society groups to become involved in
decision-making about implementation and management.
d) Sector operates with transparency and information disclosure, including for use
of funds and natural resources.
e) There is a feedback/complaint system in place.
In addition, the case studies will examine considerations specific to developing
countries that appeared in the literature, such as how well civil society engagement is
supported by political practices and cultural norms, and possible limitations posed by
education and time constraints among the population.
Country Background Conditions
Because country background conditions differ greatly from country to country,
several control variables are included in the statistical analyses to account for the
influence of these conditions on country progress. This section of the literature review
provides additional background information for inclusion of these variables.
49


Urban Migration
Over the last several decades, the rapid global pace of migration from rural to
urban areas has placed additional challenges on urban planners charged with provision of
water and sanitation services. Most of the infrastructure in established cities is dated, and
rapid population growth in cities is placing additional strain on system capacities for both
water and sanitation (Davison, 2001; OECD, 2003). Also, as city boundaries continue to
expand into previously unpopulated areas, infrastructure in these areas may be lacking
altogether.
Rapid population growth in cities often coincides with rising unemployment, as
well as strained or decreased municipal budgets (Davison, 2001), all of which create a
challenging environment for extension of water and sanitation services. In many
countries, growth in services has been inadequate to keep pace with the population
growth, and large groups of people have remained without access to the minimum of 25
liters per person per day of potable water that is recommended by the WHO (WHO,
1992). In the early 2000s, a shift in global attention occurred from focusing on
increasing supply in rural areas to the unique problems created by lack of access in poor
urban areas (UNCHS, 2003, 2006; UNDP, 2006; UNWWAP, 2003, 2006; WHO, 1992).
Recent migrants to urban and peri-urban areas in low-income countries typically
reside in poor, informal areas, which are often the lowest priority for water and sanitation
service provision (Evans, 2007). Ineffective governance systems can create disincentives
for utilities to connect poor households and for poor households to connect to services
(Bakker, Kooy, Shofiani, & Martijn, 2008).
50


Corruption
Corruption was among the six most important dimensions of governance included
in the Worldwide Governance Indicators, based on a study of more than 200 countries
during the period from 1996 to 2005 (Kaufmann, Kraay & Mastruzzi, 2006). Corruption
has been defined by Bardhan (1997) as the use of public office for private gains, and by
Shleifer and Vishny (1993) as the sale by government officials of government property
for personal gain. Corruption can lead to inefficient use of public funds for financing
public services, such as water and sanitation (Klitgaard, 1990).
The Water Integrity Network, a group of six international NGOs working to fight
corruption in water management, found that about 25-30% of state budgets on water
investments are lost due to corruption (Deen, 2006). The UN cited corruption as a
primary impediment to clean water access in the 2006 World Water Development Report,
and it has been estimated that corruption undermines efficiency in the water and
sanitation sector by 20-40% (Watkins, 2006). Corruption significantly reduces a
populations access to clean water and adequate sanitation, regardless of the level of
country income (Anbarci, Escaleras, & Register, 2009).
Corruption can further social inequality within a country, working against the
MDG Target 7.C objective of increasing access for unserved groups. Dudley (2000)
estimated that corruption diverts up to 30% of the billions of dollars spent each year for
international development loans. Misallocation of resources resulting from corruption
can exacerbate a situation of unequal wealth distribution where the poor are socially
excluded from goods and services (Riley, 2004; Gupta, Davoodi, & Alonso-Therme,
2002; Transparency International, 2008). Corruption can lead to expensive, capital-
51


intensive projects being favored by government officials, due to the possibility to receive
bribes or skim funds. These types of projects are more likely to serve high-income areas,
rather than address needs in low-income neighborhoods.
An environment of corruption can discourage investment in the water and
sanitation sector and result in underperforming and ineffective services (Mauro 1995;
Habib & Zurawicki, 2002). Estache and Kouassi (2002) studied 21 water companies in
Africa and found that nearly two-thirds of their operating costs were due to corruption. It
has been estimated that corruption can raise the price of a household water connection by
up to 30%, with opportunities for corruption at every point in the water delivery chain
(Transparency International, 2008).
Corruption in the water and sanitation sector exists beyond the national
government level, which is the focus of this research. Transparency International (2008)
summarized recent developments on the topic of water and corruption and points to
corruption during transactions among customers, contractors, and donors. This is
complicated by the nature of water and sanitation projects, which can involve multiple
levels of government, along with suppliers, contractors, and consultants. An overview
paper on this topic found approximately 50 different examples of how lack of
transparency, dishonesty or corruption can manifest itself in the water sector involving
not only government officials, but these other actors as well (Shordt, Stravato, Dietvorst,
& OLeary, 2006).
52


Political and Security Effectiveness
Fulfilling many of the operating principles identified for good water governance
transparency, accountability, efficiency, and inclusiveness requires effective political
systems. As stated by the Global Water Partnership:
The concept of governance of course encompasses laws, regulations and
institutions, but it also relates to government policies and actions, to domestic
activities, and to networks of influence, including international market forces, the
private sector and civil society. These in turn are affected by the political systems
within which they function (GWP Technical Committee, 2003, p. 4).
MDG achievement ultimately depends on not just water institutions within a country, but
also general institutions governing the economic, political, and social spheres, because
actual implementation of reforms can be curtailed due to political and resource
constraints (Saleth & Dinar, 2005).
Another country condition that can impede effective resource management and
public service provision is the general security environment, which can impede the ability
to mobilize financial and physical resources for development projects such as
infrastructure construction. As stated in the 2011 Global Report of the Center for
Systematic Peace, the qualities of conflict and governance must be included when
examining the potential for development (Marshall & Cole, 2011, p.3).
Literacy
The complexities inherent in providing water and sanitation services to urban
dwellers require both managerial and technical capacity at the responsible level of
government (Marcus & Onjala 2008; Funke et al, 2007; Mugabi & Njiru, 2006; Seppala,
2002). Skills needed include engineering and other technical specialties, system
management and maintenance, and financial management. The general literacy level of
53


the population can be an indicator of the pool of qualified individuals in a country who
can support water and sanitation service extension. It can also show the ability of the
general public, as water and sanitation consumers, to grasp information about technology
options, system operation, and usage costs, so they can provide input into government
and service provider decision making (Research Triangle Institute, 2004).
GDP per Capita
A countrys overall economic base, as measured by gross domestic product
(GDP), is one indication of its ability to apply funds towards water and sanitation
infrastructure investment. Studies of multiple countries have found positive associations
between overall country GDP, as well as GDP per capita, and water access among the
population. Fogden (2009) found a positive association between a countrys average
GDP per capita growth rate and the percent of the population with access to safe water
over a 35-year period. Onda, LoBuglio, and Bartram (2012) found GDP to be one of
several positive factors contributing to water access, even after adjusting published access
rates downward to account for issues of water safety.
Renewable Water Resources
When considering the potential for expanded urban water service, the availability
of water resources within a country is a relevant factor. Viewing it on a per capita basis
provides a better indication of the extent to which it can meet the needs of the domestic
population. The UN Food and Agriculture Organization collects water resources
information by country and publishes it through the AQUASTAT database (FAO, 2014).
Renewable water resources, inland waters renewed by the global water cycle, are
the main source of available water for human use. However, only part of these waters are
54


able to be accessed and used effectively because of limitations such as physical
conditions of accessibility, the potential for floods caused by excess flows, and
inconsistency in water flows which can make them difficult to capture and utilize (FAO,
2015). While these limitations can vary by country, measures of renewable water
resources per capita do provide some indication of the relative availability of water for
human use.
Foreign Aid to Water and Sanitation
Many countries in the sample group, particularly those in the low-income
category, are the recipients of foreign aid designated for the sector which, among other
things, goes towards the construction of infrastructure and other systems for delivery of
water and sanitation services. While several studies have looked at the impact of foreign
aid in general, (including those discussed above), a 2010 study (Botting et al, 2010)
looked specifically at the impact of foreign aid per capita for water and sanitation, by
relating it to changes in access to improved water and sanitation services from 2000 to
2006. The study found that countries in the highest tertile of aid per capita recipients had
odds ratios of 4 to 18 times more likely than those countries in the lowest tertile of aid per
capita recipients for achieving greater gains in access to improved water supply. The
increased odds for sanitation access were modest and largely insignificant.
Summary of Literature
This chapter presented key literature related to the overall objective of my
research. First was the theory that water and sanitation service exhibits the characteristics
of a public good, a type of good best suited for government provision due to supply
inefficiencies and distributional inequity likely to occur with private sector provision.
55


Further, access to water has been recognized by the UN as basic human right and it can
be considered as a merit good, one that society has an interest in promoting to avoid
negative externalities, such as health problems, associated with poor water supply and
sanitation. These arguments form the basis for government and international donor
participation in water and sanitation service provision.
Also discussed in this chapter was literature about the importance of effective
water institutions and governance practices for achieving water and sanitation access
goals. They are viewed by many authors as critical preconditions for achieving and
sustaining quality water and sanitation services. Institutions determine decision and
incentive structures within governments, encompassing the rules for: access to and
allocation of resources; placement of economic incentives; and systems of service design
and delivery. The broader concept of governance also encompasses the values and
processes by which government authority is exercised. New forms of governance, which
engage governments, non-government organizations, the private sector, academic and
research institutions, are viewed as a means of improving government service outcomes.
This research will explore the extent to which three specific institutional and governance
attributes are having a positive impact in the context of urban water and sanitation
services.
Literature on the first attribute of decentralization says it can be beneficial for
government service provision because lower levels of government can be more effective
in responding to local conditions and meeting the needs of local consumers. But, there
are also several articles noting factors that are important for realizing theorized
decentralization benefits, such as degree of decision-making and revenue raising
56


delegation, the political and socio-cultural environment, and the ability of local
governments to adapt to new responsibilities. The research question for this attribute is
designed to study the extent to which decentralization is having a positive impact on
urban water and sanitation provision, while also looking at whether necessary factors for
success are present in decentralized countries: RQ1. What is the role of decentralization
of water and sanitation services in a countrys progress on providing urban water and
sanitation access?
Other literature reviewed was about international development aid and
governance practices that engage donors and other stakeholders for sector-wide strategy
and investment coordination, the second attribute being studied. This highlighted the
shortcomings of traditional approaches to international development aid, stemming
largely from information, incentive and ownership problems between donors and aid-
recipient countries. The literature also explored the potential of new approaches designed
to address these problems, such as SWAps and sector-specific sections of PRSPs. These
new approaches bring with them the expectation of greater country ownership of sector
projects and more predictability and efficiency in aid allocation, ultimately improving
sector outcomes. The second research question is designed to study the extent to which
sector-wide strategy and investment coordination is having a positive impact on urban
water and sanitation service provision: RQ2. What is the role of sector-wide strategy and
investment coordination in a countrys progress on providing urban water and sanitation
access?
The third topical area of literature covered was for the third attribute of civil
society engagement. Overall, the literature on civil society engagement points to several
57


potential benefits of these activities, such as promoting government transparency and
accountability, improving redistributional equity, better representing the interests of the
poor, and improving the quality of government services at various stages of the policy
development and implementation process. To test these if these findings hold true in the
context of urban water and sanitation service provision, particularly in the socio-
economic environment of low- and middle-income countries, the following research
question was developed: RQ3. What is the role of civil society engagement in a countrys
progress on providing urban water and sanitation access?
The three research questions, hypotheses, and constructs developed from the
review of literature in this chapter will be examined starting with the next chapter on
quantitative methods, leading into the statistical analyses and the case studies.
58


CHAPTER IV
QUANTITATIVE METHODS
To test the three hypotheses, bivariate and multivariate statistical tests were
conducted using data from multiple countries. This chapter presents information about
sample selection, sample characteristics, variable measurement and operationalization,
and also discuss data validity and limitations.
Sample Selection
An initial sample of 75 countries was derived from those countries who responded
to the 2011 GLAAS survey3. This data source was identified as the most comprehensive
option for quantitative analysis, both in terms of number of country respondents and its
depth of information relevant to the operationalized constructs of the research questions.
The 75 respondent countries make up 54% of all 139 countries classified by the
World Bank as low- or middle-income (LMI) countries, a group that is most relevant for
MDG Target 7.C (World Bank website, 2015). A list of the 75 countries is included in
Appendix B, along with notations about missing and incomplete data, which are
introduced later in this chapter. Three countries (Haiti, India, and Equatorial Guinea) had
insufficient data for inclusion in any of the multivariate statistical tests, while other
countries had to be excluded only from certain tests. As a result, the actual percentage of
LMI countries analyzed for specific multivariate tests ranged from 47% to 49%.
To evaluate the representativeness of the sample group for generalizing results to
all LMI countries, various country characteristics were compared between the 72-country
3 The reader will recall that GLAAS is a UN Water data collection initiative
implemented by the World Health Organization (WHO)
59


sample used or multivariate tests (sample group) and all 139 LMI countries. Average
country population was compared, first against the entire 139-country LMI group, and
then again against the LMI group after removing the two most populous countries (China
and India), as their populations exceeds by five times or more the next most populous
country, Indonesia. Neither China nor India were part of the 72-country sample group.
As shown in Table 4.1, the sample group average country population, (33.2
million), is about 20% lower than the average country population for all 139 LMI
countries. However, when China and India are removed from the LMI group, the sample
group average country population is now about 50% higher than the average population
for the 137 remaining LMI countries. Looking at median values, the sample group
median is about 50% higher than the median for either LMI comparison group. Studying
this comparison, one can say that the sample group is generally skewed towards high
population countries within the LMI group.
Table 4.1
Average Country Population:
Comparison of Sample Group and LMI Group
N Average Country Population (millions) Median Country Population (millions)
Sample Group 724 33.2 13.7
All Low- and Middle-Income countries 139 41.3 9.5
All Low- and Middle-Income countries except for China and India 137 23.0 9.3
Source: World Bank, 2012b.
4 Three countries, (Haiti, India, and Equatorial Guinea), were not used in any
multivariate tests so are not included.
60


GDP per capita comparisons were also done.5 A comparisons was first done
between the sample group and all LMI countries with available data, and then also
against low-income and lower-middle income countries only6. (The World Bank further
divides the middle-income country classification into lower-middle-income and higher-
middle income groups.)
As show in Table 4.2, the sample group average GDP per capita falls just about
mid-way between these two comparison groups, whereas the sample group median is
closer to the median for the smaller comparison group. Looking particularly at the
median comparison, this shows that the sample group is more heavily weighted towards
low-income countries within the LMI group.
Table 4.2
GDP per Capita:
Comparison of Sample Group and LMI Group
N Average GDP per Capita (current US$) Median GDP per Capita (current US$)
Sample Group 722 2,867 1,251
All Low- and Middle-Income countries 135 4,051 3,256
All Low and only Lower-Middle-Income countries 80 1,710 1,354
Source: World Bank, 2012c.
5 Some country GDP per capita values may vary between those used for the
sample dataset and those summed for the total GDP per capita for low- and middle-
income countries, as these were downloaded at different points in time from
WorldBank.org. Any variations are thought to be nominal with little bearing on the
general comparison presented here. Income group classifications are from World Bank
website, 2015.
6 The 72-country sample consists of 28 Low Income Countries, 29 Lower-Middle
Income Countries, 14 Upper-Middle Income countries, and 1 High Income Country
(Oman), based on World Bank classifications.
61


A third set of comparisons was done for indicators of country progress on urban
water and sanitation access, specifically the urban population percent with access in the
years 2000 and 2012. This exercise was done to see whether or not the sample group,
those who responded to the GLAAS survey, was skewed towards low or high-access
countries. The results appear in Table 4.3.
Table 4.3
Urban Water and Sanitation Access:
Comparison of Sample Group and LMI Group Averages
Water Access (%) Sanitation Access (%)
Year 2000 2012 2000 2012
Sample Group used in multi- variate analyses (N = 66) 85.6 90.9 58.4 63.2
All Low- and Middle-Income countries (N ranges from 130 to 137) 89.6 93.1 67.6 71.5
Source: WHO & UNICEF, 2000-2015.
The comparison shows that for water, the sample group had marginally lower
access rates in 2000 and 2012 than did LMI countries for which data were available. For
sanitation, the sample group also had lower access rates in both years, with a wider
margin of difference. This reveals that the sample group is more heavily weighted
towards countries with low access rates.
In summary, the comparisons show that sample group is generally higher in
population, lower in GDP per capita, and lower in water and sanitation access rates, than
the overall LMI country population. Arguably, these characteristics make the sample
group that much more germane to the intention of MDG Target 7.C, as these countries
have greater needs and more limited domestic economic base.
62


In the following sections, I introduce and explain the outcome variables, key
explanatory variables, and control variables for the quantitative analyses. Table 4.4 is an
overview of all variables. Detailed information about variable sources and
operationalization is found in Appendix C.
Table 4.4
Overview of All Variables
Outcome Variables Key Explanatory Variables Control Variables
1. Water Access in 2012 2. Water 12-Year Progress 1. Water Decentralization 2. Water Sector Planning 3. Water Civil Society Engagement 1. Urban Population Change 2. Corruption Level 3. Political Fragility 4. Security Fragility 5. Literacy Rate 6. GDP per capita 7. Renewable Water Resources 8. Foreign Aid to Water and Sanitation1
3. Sanitation Access in 2012 4. Sanitation 12-Year Progress 4. Sanitation Decentralization 5. Sanitation Sector Planning 6. Sanitation Civil Society Engagement
foreign Aid to Water anc Sanitation was eventually dropped from the analysis for
reasons explained later in this chapter.
Outcome Variables
I conceived four outcome variables to represent country progress in extending
water and sanitation service to unserved residents in urban areas: two for water service,
and the same two for sanitation service. The two variables are defined as follows:
1) Access in 2012 (A2012): percent of the urban population with access to service
in 2012.
2) 12-year progress {Progress): the percent of the urban population that gained
access between 2000 and 2012, relative to the percent of the urban population that
lacked service in 2000. (An equation for this is presented later in this section).
Operational measures for these outcome variables were derived from the JMP
dataset introduced in Chapter II, the official UN mechanism tasked with monitoring
63


country progress towards MDG Target 7.C. Data were available from JMP for all
countries in the sample.
JMP produces a dataset annually that tracks the percentage of the population
lacking access to an improved source of water and an improved source of basic
sanitation, broken down by rural and urban areas (WHO & UNICEF, 2000-2015). Table
4.5 lists all system types that qualify as improved sources by the JMP program. These
criteria have remained consistent throughout the 12-year period from 2000 to 2012.
Table 4.5
Systems Considered as an Improved Source by JMP
Water Sanitation1
Piped water into dwelling Piped water into yard/plot Public tap/standpipe Tubewell/borehole Protected dug well Protected spring Rainwater collection Flush toilet Piped sewer system Septic tank Flush/pour flush to pit latrine Ventilated improved pit latrine Pit latrine with slab Composting toilet
'Only facilities which are not shared or are not public are considered improved.
Source: WHO & UNICEF, 2006, p.4.
I chose to equate service access, as referred to in my research questions and
measured in my outcome variables, with JMP definitions for improved sources. In other
words, I considered those who had service access in a country to be the same as those
who had improved sources as defined by the JMP criteria. In keeping with the research
parameters, I used only JMP data for urban areas.
JMP derives its estimates of the population lacking access to improved sources by
drawing on data collected by national statistics offices and international survey programs
that use nationally representative household surveys and national census data (UNICEF,
undated). Most of the data come from Demographic Health Surveys, Multiple Indicator
64


Cluster Surveys, Living Standards Measurement Studies, and World Health Surveys
(WHO & UNICEF, 2012). As possible, data are obtained directly from water and
sanitation service users, rather than service providers, and therefore are not impacted by
different data collection methods, definitions of access, or the political environment
surrounding data collection in a particular country (WHO & UNICEF, 2006). Population
data used by JMP when calculating access percentages are obtained from the UN
Population Division (WHO & UNICEF, 2006).
To derive the first outcome variable, Access in 2012 O42012), I started with JMP
country-specific data for the percentage of the urban population lacking access to an
improved source in 2012, and then transposed it to get the percentage of the urban
population with access to an improved source in 2012.
For the second outcome variable, 12-year progress (Progress), Formula 4.1 shows
the calculations used to derive the variable, along with two examples. I started with
i420i2 and subtracted i42000 (access percentage in the year 2000) to measure the 12-year
change in percent of the urban population with access to an improved source (labeled as
C for change).7 In most cases the 2012 value is greater than the 2000 value, resulting in a
positive figure, however there are also cases where the difference value is negative. I
then divided the 12-year change figure, positive or negative, by the percentage of the
population that lacked access in the year 2000 (labeled as G for goal), to derive the extent
of progress that was made over a 12-year period in reducing the percentage of the
7 The year 2000 was chosen, rather than the MDG base year of 1990, because
JMP data were incomplete for many of the sample countries for 1990 and subsequent
years.
65


population without access.8 The result constitutes the 12-year progress variable
(Progress), which is the proportion of those lacking access in 2000 (original goal for
increasing service) that gained access by 2012. Progress can range from a negative value
(for a decline) to a maximum of 1.0 if the entire population lacking access in 2000 gained
access over the 12-year period.
Formula 4.1
Progress Outcome Variable Calculation
^2000 = Percent of the urban population with access in 2000
^2012 = Percent of the urban population with access in 2012
G = 100 i420oo Percent of the urban population lacking access in 2000, considered
the initial goal for increasing access
C =A
2012
A
2000
C
Progress =
U
Change in the percent of the urban population with access over the
12-year study period
Example 1
^2000
G
Central Africa Republic Sanitation (improvement in overall access)
= 31.9%
= 68.1%
100%
i42oi2 =43.6%
C = 43.6% 31.9% = 11.7%
Progress = 11,7% = 0.17
68.1%
Example 2
^2000
G
Dominican Republic Water (decline in overall access)
= 90.7%
= 9.3%
= 100%
A2012 = 82.5%
C = 82.5% 90.7% = -8.2%
Progress = -8,2% = -0.88
9.3%
8 Other outcome variables were considered, such as the 12-year percent change
without comparing to the starting point, but no meaningful results were produced.
66


Individual country values calculated for A2012 and Progress were used to assign
countries to low, medium, and high achievement groups that were used for the statistical
analyses. The rationale for grouping countries into achievement groups, rather than
maintaining individual country values, is explained later in this chapter.
Key Explanatory Variables
The institutional and governance attributes represented by the key explanatory
variables, each linked to a research question, are: (a) degree of decentralization
(decentralization); (b) degree of sector-wide strategy and investment coordination (sector
planning); and (c) degree of civil society engagement (civil society). Each was measured
for both water and sanitation, and corresponding variables are written with the notations
W- and S-, respectively, before the shortened name, (for example, W-civil society). This
results in six key explanatory variables. Refer to Appendix A for operational definitions,
construct-level definitions, literature references, and source material for the key
explanatory variables.
To derive individual country observations for the six key explanatory variables, I
used select country-level data from the GLAAS 2011 survey, which addressed overall
policy, governance, and institutional environment for water and sanitation services
(WHO, 2011). A large portion of the GLAAS 2011 survey covered the nature and
impact of government policies and institutions. More information about GLAAS
questionnaire development, data collection, and data validation can be found in the 2012
GLAAS report (WHO & UN Water, 2012, Appendix A, p. 71-73).
Looking at temporal correspondence between the key explanatory variables and
the outcome variables, the use of 2011 data for the key explanatory variables corresponds
67


closely with the A2012 outcome variables measuring access in 2012. However, since the
Progress outcome variables reflect a 12-year period starting in 2000, corresponding
explanatory variables would ideally measure the same 12-year period. This was
explored, however, sufficient data from GLASS were not available for years prior to
20119, and no other comparable data source was identified. Therefore, the key
explanatory variables derived from 2011 data were used for all analyses.
The last column of Appendix A shows the specific 2011 GLAAS survey
questions/responses selected for inclusion due to their applicability to the construct
definitions listed in the middle column of Appendix A.10 Not all constructs have
corresponding GLAAS survey questions. Those constructs not measured by the GLAAS
survey responses were evaluated through document review and interviews during the
case studies. A coding system is used in the last column of Appendix A to show which
information came from the GLAAS survey, and which from documents and interviews.
In the next paragraph, I explain in detail how the GLAAS survey response data were
adapted for use.
For each GLAAS survey question, usually asked separately for water and
sanitation, country respondents provided one of three numbers representing the level of
development for particular programs or policies:
little indicated by 0
some indicated by 0.5
a great deal indicated by 1.0
9 Previous GLAAS surveys were conducted in 2008 and 2009, however, the
number of country respondents was about half that of 2012.
10 A few applicable questions were not included since almost all countries
responded with the same answer and it added nothing to the analysis.
68


To derive each countrys key explanatory variables, I averaged the numerical
responses (0.0, 0.5, or 1.0) for several applicable questions to arrive at composite value.
For example, to measure civil society engagement in the sanitation sub-sector, responses
to the two (G)-coded civil society probes in the third column of Appendix A, (when
asked specifically about sanitation), were averaged to arrive at a single composite value
for a countrys S-civil society level. Averaging multiple survey responses allows for
broader input and helped to mitigate against an inaccurate or misguided response to any
one particular survey question. Consistent with the response scale for individual
questions, composite values range from 0.0 to 1.0. Each countrys composite values were
then used as observations for the three key explanatory variables.
In a few cases where a respondent country did not answer one or more of the
questions used in calculating a composite value, an average was calculated from only
those question(s) with responses. This approach, which essentially applies the average
value from responses received to any unanswered question(s), was deemed reasonable, as
all questions are somewhat reflective of the construct of interest. Using a mean value is
acceptable for dealing with missing data in some cases when it would not markedly
impact the overall results (Tabachnick & Fidell, 2001, p.71). For a country that did not
respond to any of the questions relevant to a particular variable, the observation was
considered as missing. Refer to Appendix B for notes on partial and missing data by
country.
Control Variables
Eight control variables were selected for inclusion in the regression analyses to
reflect country background conditions that were thought to have potential influence on
69


the outcome variables. A variety of reputable data sources were considered as potential
sources for control variable data, with selection based on those that most closely
approximated the relevant background condition and were available for a large number of
sample countries. Data used are from the most recent year available, generally between
2011 and 2013. (Refer to Appendix C for detailed source information).
For countries lacking observations for a particular control variable, the variables
mean value was used as a substitute when deemed reasonable, to avoid exclusion of that
country from the overall analysis because of one missing variable. This approach is
considered acceptable in some situations when it would not markedly impact the overall
results (Tabachnick & Fidell, 2001, p.71). This maintained an overall larger sample size
and was judged to not have a major impact on the overall results. Mean value
substitutions are noted in Appendix B and also explained below.
Urban Population Change
The pace of urban migration within a country can be seen as an important factor
in its ability to increase the percentage of the urban population with access to water and
sanitation, as service providers are challenged to keep pace with this changing situation.
Reflecting this growth in a control variable required considering data over a 12-year
period, unlike most of the other control variables that were measured at one point in time.
The formula for the variable is as follows:
Formula 4.2
Urban Population Change Calculation
t/2012 = urban population % in year 2012
U2qoo = urban population % in year 2000
^2012 ~ ^2000
^2000
70


The variable represents the proportional growth over the 12-years, and not just the 12-
year absolute change. So, a country that increased from a 60% urban population in 2000
to a 75% urban population in 2012, would have a relative change in urban population of
0.25. Data were sourced from World Bank population data and were available for all
countries in the sample (World Bank, 2000-2012b).
Corruption Level
No specific indicator of corruption in the water or sanitation sector could be
identified for multiple countries, so instead, the Corruption Perceptions Index generated
by Transparency International for individual countries was used (Transparency
International, 2013).11 Corruption Perceptions Index values range from 0 (very corrupt)
to 100 (very clean). I transposed this scale so that a higher score represents a more
corrupt country. The transposed values for the 75 countries in the sample ranged from 37
(relatively clean) to 92 (very corrupt). Three of the countries were missing values (Fiji,
Maldives, and Samoa), so the sample mean value (70) was used.
Political Fragility and Security Fragility
Two separate indicators from the Integrated Network for Societal Conflict
Research (IINSCR), part of the Center for Systemic Peace, were adopted as variables to
represent national political and security environments. These measures were thought to
capture the relevant background conditions and were widely available for multiple
countries. 11
11 Commonly used measures of corruption tend to be very highly correlated
(Alesina & Weder, 2002; Treisman, 2000).
71


The INSCRpolitical effectiveness indicator reflects the political environment in a
country by assessing regime and governance stability. Scores range from 0 (effective) to
3 (highly ineffective). Two countries, Samoa and Timor-Leste, were without data and
the mean score of 1.5 was used.
The INSCR security effectiveness indicator measures a countrys general security
and vulnerability to political violence. Also ranging from 0 to 3.0, the mean security
effectiveness value of 0.7 was used for two countries that were missing observations,
again Samoa and Timor-Leste.
I modified the terminology from effectiveness indicators to fragility indicators
when adopting these measures as variables, to better reflect the meaning of the numerical
scale, where a higher number represents a worse situation.
Literacy Rate
The measure used to represent a countrys general education level in the analyses
is the percent of the population age 15 and over who can read and write.12 This provides
an indication of the populations capacity to learn technological and management skills
necessary to support service extension, and also the ability of the general population to
provide input into policy and programs. Country-specific data were sourced from the
Central Intelligence Agency (CIA) World Factbook and observations were available for
all 75 countries in the sample (CIA, 2011).
12 Other education indicators were considered, such as percent of the adult
population having completed secondary school, but data were missing for several
countries. Literacy rates correlate with UNDP 2010 secondary education data at 0.69,
and were available for more countries.
72


GDP per Capita
GDP per capita (in current US$) was used as a control variable to represent the
domestic economic base. This measure can vary widely, even among the LMI countries
in the sample group. Because of unacceptably high skewness and kurtosis, the data were
transformed to natural logarithm values, which then exhibited acceptable levels of
skewness and kurtosis for inclusion in the regression models. Values were available for
all 75 countries in the sample group.13 The data were sourced from the World Bank
Development Indicators14 (World Bank, 2012c).
Renewable Water Resources
While the focus of this study is on water supply and sanitation service delivery,
and not water availability, it was nonetheless felt that the availability of domestic water
resources would be a relevant factor for countries considering investment in service
extension, particularly for water supply. For this reason, data were sought to represent
water availability by country. The indicator total actual renewable water resources,
(cubic meters per year, per capita), as reported by UN Food and Agriculture Organization
(FAO) AQUASTAT, was selected to be used as a control variable (FAO, 2014).
AQUASTAT is FAO's global water information system, developed by the Land and
Water Division. The renewable water resources statistics provided in AQUASTAT are
long-term annual averages, typically 1961-1990.
13 Other indicators considered, such as the GINI index of income distribution, was
unavailable for several countries.
14 Mynamars 2012 GDP per capita was not available from the World Bank and
was instead sourced from UN data, https://data.un.org.
73


While this variable helps to account for vast differences in water availability
among countries, (ranging from more than 200,000 cubic meters per year, per capita, to
less than 1,000), it is recognized that this indicator does not get at the issue of ease of
water access or proximity to urban areas. Nor does it incorporate agreements for water
sharing across national boundaries. Attempts were made to locate such data, but it was
not available for a sufficient number of countries to include in the cross-case analysis.
Due to unacceptably high skewness and kurtosis for this indicator, it was
transformed to natural logarithm values for use as a control variable in the regression
analyses. Observations were available for all 75 countries in the sample group.15
Foreign Aid to Water and Sanitation
For all countries in the sample group, domestic financial resources for the sector
are supplemented by some level of donor aid, aid that can come from several different bi-
lateral and multi-lateral sources. The most comprehensive source of information
identified on these transfers was a dataset produced by the OECD Development
Assistance Committee (OECD, 2002-2012).
Using this dataset, I first totaled the official development assistance disbursed to
each country from 2002 to 2012 designated for the water and sanitation sector (in current
$US).16 (Complete disbursements data were unavailable for years prior to 2002, so these
years were not considered.) The 2002-2012 cumulative aid figure was then divided by a
countrys average population during the 10-year period, (derived, for simplicity, as an
15 Samoa, which had a zero value, was replaced with a nominal value of 10.
16 The OECD Development Assistance Committee includes in their figures all aid
provided by official agencies, including state and local governments, or by their
executive agencies (OECD, 2012).
74


average of the 2002 and 2012 country population figures), to calculate foreign aid per
capita from 2002 to 2012. The resulting country values exhibited unacceptably high
skewness and kurtosis and were transformed to the natural log values to achieve
acceptable levels for the regression analyses. Data were available for all 75 countries.
Some weaknesses of the OECD data are that it does not distinguish between aid to
water and aid to sanitation, and it incorporates funding for water programs outside the
scope of this study, such as water resources management, development, and protection.
Looking forward, this type of differentiated data may be available in coming years. New
efforts are underway by UN Water through its TrackFin initiative to develop
methodology to identify and uniquely track different types of financing to the water,
sanitation, and hygiene sectors in a coherent and consistent manner across several
countries.
Data Validity and Limitations
This section discusses how well the variables represent the constructs of interest,
the general validity of the data used, and potential limitations of the analyses.
Concept-Measure Validity
Concept-measure validity is an evaluation of whether the measures proposed are
accurate representations of the concepts of interest. The JMP, on which I based the
outcome variables, is the official UN mechanism for monitoring progress towards MDG
Target 7.C. Although JMP data differ in some cases from figures published by individual
national governments, JMP has the advantage of using a consistently applied
methodology for determining both water and sanitation access levels across multiple
countries over several years. Because of JMPs widely accepted status for tracking MDG
75


Target 7.C country progress, it was deemed the most appropriate measure to use for the
outcome variables, given the scope and intended global audience for this research.
The wording of MDG Target 7.C, and the way that progress is measured, is not
without its critics. Easterly (2009) questions measurement of individual country
development progress in numerical terms for initiatives like the MDGs, particularly when
comparing Sub-Saharan Africa to other world regions. Talking particularly about MDG
Target 7.C, he argues that the stated objective of reducing by half the percentage of
people without access to water and sanitation is structured and reported in a way that
undervalues the advancements of Sub-Saharan African countries:
Africa was relatively falling behind on reducing the percent WITHOUT access to
clean water, but it would have been relatively catching up if it had been measured
the conventional way of percent WITH access to clean water. The choice of
WITH and WITHOUT for percentage change measures depends on assumptions
on social welfare that have not been subjected to any serious scrutiny. If one just
uses the usual indicator of percent WITH access to clean water, Africa is catching
up to other regions (Easterly, 2009, p.33).
Easterlys concerns are somewhat addressed in this research by the choice to transpose
the JMP data, thereby defining the outcome variables in terms population percent with
access to water and sanitation.
For the three country institutional and governance attributes, the GLAAS survey
questions selected to develop the key explanatory variables, (see Appendix A, third
column), are fairly direct and pointed towards the theoretical concept of interest. For
example, the survey questions concerning decentralization ask about both about the
degree to which service delivery has been decentralized and the availability of funding at
the local level. Similarly, civil society engagement survey questions ask specifically
about citizen participation in program planning, budgeting, and implementation.
76


Data Validity
This section discusses potential data limitations and attempts to triangulate key
explanatory variable data with other sources of information.
As previously mentioned, GLAAS data do not capture all the constructs of
interest relating to the key explanatory variables, only those coded with (G) in column
three of Appendix A. Therefore, the content validity for the quantitative analyses is less
than complete and the remaining constructs, coded with (D/I) in Appendix A and
summarized below, will only be evaluated through document review and interviews as
part of the case studies.
Decentralization
There are clearly defined rights, responsibilities, and roles for sector management
actors at the national and sub-national levels
Sub-national level authorities are monitored by communities and/or the central
government for actions such as public funds expenditures and equality in service
provision
Support is provided to build managerial and technical capacity of sub-national
levels of government
Civil Society Engagement
Opportunities exist for consumers and civil society groups to become involved at
different stages of the policy and implementation process
The sector operates with transparency and information disclosure
As no case studies will be completed for sector-wide strategy and investment
coordination, further evaluation of constructs that were not represented in the quantitative
data is limited, however they are noted here so they might be addressed in future data
collection efforts:
77


Sector-wide Strategy and Investment Coordination:
Development of partnering agreements between donors and recipient countries is
driven by the recipient country
A single agency is charged with coordinating all donor aid received in
conjunction with a sector strategy
A system of monitoring is in place for compliance with the terms of aid partnering
agreements
Data Triangulation
For self-reported data such as GLAAS, there may be a tendency for countries to
report information that reflects positively on them, but which may not accurately reflect
the situation. The data undergoes some review by WHO after collection to guard against
such misrepresentation. Further, to check the validity of information reported through
GLAAS, I attempted to triangulate the key explanatory variables, (derived from GLAAS
data), with other independent data sources on the same topic.
For decentralization, an independent source of country ratings was identified and
correlated with country decentralization values. I used a country decentralization index
derived by Ivanyna and Shah (2014) that looked at the following: 1) the security and
existence of local government; 2) the relative importance of local government; 3)
political decentralization, (for example, elected rather than appointed local officials); 4)
fiscal decentralization, (for example, the ability to raise revenues); and 5) administrative
decentralization, (for example, local government authority for personnel hiring and
regulatory control). This index, particularly the last two items, corresponds well with the
construct-level definitions for decentralization outlined in Appendix A.
The correlation between the Ivanyna and Shah decentralization index and the W-
decentralization values for 72 overlapping countries is 0.33 (p < 0.01). The correlation
78


between the aggregate decentralization index and the S-decentralization values for 71
overlapping countries is 0.31 (p < 0.01). These correlation coefficients, while modest, do
indicate some positive correlation at a high level of significance, for both water and
sanitation. The difference could be explained by the fact that the comparison index
considers enabling factors for decentralization in all aspects of government and political
representation, and not just government services.
Turning to the other two explanatory variables, no comparable rating system
could be found for sector planning for a sufficient number of countries. However,
comparable measures were identified for civil society engagement. The World Alliance
for Citizen Participation (CIVICUS) rates countries using an Enabling Environment
Index (EEI) that measures the potential of citizens to participate in civil society, based on
the governance, socio-cultural and socio-economic environments in a country.17 The
2013 index includes 109 countries, 40 of which are also in the sample group (World
Alliance for Citizen Participation, 2013a). I measured the correlation between CIVICUS
EEI data and the civil society values and found a significant correlation (p <0.10) of 0.27
for the sanitation sector, but the correlation of 0.23 for the water sector was not
significant, showing no strong relationship in either case.
Another civil society measure was identified covering only countries in sub-
Saharan Africa. The 2011 CSO Sustainability Index developed by the US Agency for
International Development (USAID) includes 20 countries in the sample group. In
17 The EEI measures conditions that affect the capacity of citizens (whether
individually or collectively) to participate and engage in civil society. CIVICUS includes
non-organized forms of civil society as well as civil society organizations. (World
Alliance for Citizen Participation, 2013b, page 1).
79


addition to an overall index, there are sub-indices for different aspects of CSO
sustainability. The direction of the USAID indices are the inverse of that for civil society
values, so one would expect a negative correlation coefficient. For water, the correlation
coefficient for 20 countries was -0.62 (p < 0.01) for the overall index, and correlations for
the specific legal, organizational, and advocacy indices measured at -0.56, -0.73, and -
0.48, respectively, (all significant at p < 0.05). For sanitation, the correlation coefficient
was -0.51 for the overall index (p < 0.05), with correlations for the specific legal,
organizational, and advocacy indices measuring at -0.51, -0.60, and -0.38, respectively,
(the first two sub-indices had significance < 0.05; the third was not significant). Overall,
these results show a reasonably high correlation for both water and sanitation with the
USAID indices (USAID, 2011).
Internal and External Validity
A potential threat to internal validity in each of the tests for explanatory variable
impact is history, that is, events outside of the institutional and governance attribute being
studied that have impacted country progress on water and sanitation. This threat is
somewhat addressed by the use of control variables in the statistical analyses. It will also
be addressed in the case studies that go more in-depth into contextual evidence. The case
studies will also be a way of looking at potential issues of multi-collinearity among the
three independent variables.
Another limitation is that the collection time period for country institutional and
governance data, (2011 GLAAS Survey), does not directly correspond with the time
period represented in the outcome variables, (2000 through 2012). Responses provided
in 2011 are not necessarily indicative of the institutional and governance conditions
80


during the entire 2000 to 2012 time period represented by the Progress variables and
preceding the 2012 situation measured by the A2012 variables. While this limitation
cannot be remedied for the quantitative analyses, it is partially addressed by the case
studies that will look at conditions over a multiple-year time period.
External validity, the generalizability of results, means whether or not this
research yields information relevant to countries that were not part of the sample. In this
case, the ability to generalize the results to other countries water and sanitation
institutions would be supported by the fact that the sample represents about half of the
LMI population, and because the statistical models control for differing country
background characteristics. However, as previously noted, systematic differences exist
between the sample group and the overall LMI country population, and these should be
taken into consideration when generalizing results. The sample has, on average, higher
country populations, lower GDP per capita, and lower water and sanitation access rates.
81


CHAPTER V
QUANTITATIVE ANLAYSIS
This chapter explains the quantitative tests that were done to test the hypotheses
using country-specific data from the country sample group. It begins by reviewing the
research questions and presenting descriptive statistics for all variables. Next, bivariate
statistics are presented showing relationships between pairs of independent and outcome
variables. Finally the last part of the chapter presents a series of multivariate statistical
tests and analyzes the results to draw conclusions related to the research questions and
hypotheses.
RQ1. What is the role of decentralization of water and sanitation services in a countrys
progress on providing urban water and sanitation access?
HI. A high level of decentralization in water and sanitation services is associated with
greater country progress on providing urban water and sanitation access.
RQ2. What is the role of sector-wide strategy and investment coordination in a countrys
progress on providing urban water and sanitation access?
H2. A high level of sector-wide strategy and investment coordination in water and
sanitation services is associated with greater country progress on providing urban water
and sanitation access.
RQ3. What is the role of civil society engagement in a countrys progress on providing
urban water and sanitation access?
H3. A high level of civil society engagement in water and sanitation services is
associated with greater country progress on providing urban water and sanitation access.
82


Descriptive Statistics
This section presents descriptive statistics for the outcome variables, key
explanatory variables, and control variables. It also discusses how the outcome variables
were adapted to better suit the statistical analysis and what was revealed by analyzing
correlation values among variables.
Outcome Variables
Descriptive statistics for the four outcome variables revealed that individual
country values are not normally distributed and had unacceptable levels of skewness and
kurtosis. These characteristics make the outcome variables unsuitable for a standard
regression analysis. Therefore, instead of using individual country observations for the
statistical tests, I assigned countries to low, medium, and high achievement groups for
each outcome variable, thereby converting the variable to an ordered categorical variable
with three possible values.
The defining value range for each variables three groups was individually set at
logical breakpoints given the distribution of values.18 The three groups are labeled 0, 1
and 2 representing an ordered ranking of low, medium, and high achievement
respectively. Tables 5.1a through 5. Id show, for each outcome variable, descriptive
statistics for raw country data on the left side, and characteristics of the grouped variable
on the right side.
18 An attempt was made to create groups breaking at equal intervals on the
distribution scale, for example, between 0.10 and 0.20 then.0.20 and 0.30. This resulted
in excessive bunching of observations in some groups, with very few observations in
other groups.
83


Table 5.1
Descriptive Statistics and Grouping Parameters for Outcome Variables
Table 5.1a
Water A2012
Raw Country Data Grouped Variable
Obs 74
Minimum 52.3 Group Obs Value Range (%)
Maximum 100 0 -Low 16 0 > x < 85
Mean 90.71
Std. Dev. 9.82 1 -Medium 25 85 > x < 95
Variance 96.53
Skewness -1.72 2 -High 33 x > 95
Kurtosis 5.83
Table 5.1b
Water Progress
Raw Country Data Grouped Variable
Obs 72
Minimum -.99 Group Obs Value Range
Maximum 1.0 0 -Low 20 x< .05
Mean .24
Std. Dev. .49 1 -Medium 31 .05 > x < 0.50
Variance .24
Skewness -.61 2 -High 21 x > 0.50
Kurtosis 2.90
Table 5.1c
Sanitation A2012
Raw Country Data Grouped Variable
Obs 74
Minimum 15.7 Group Obs Value Range (%)
Maximum 100 0 -Low 20 x < 35
Mean 61.45
Std. Dev. 26.93 1 -Medium 27 35 > x < 80
Variance 725.08
Skewness -.12 2 -High 27 x> 80
Kurtosis 1.53
84


Table 5.1d
Sanitation Progress
Raw Country Data Grouped Variable
Obs 72
Minimum -.09 Group Obs Value Range
Maximum 1.0 0 -Low 24 x<0 .05
Mean .17
Std. Dev. .23 1 -Medium 28 0.05 >x< 0.20
Variance .05
Skewness 1.68 2 -High 20 x > 0.20
Kurtosis 5.45
All references to outcome variables from here forward will refer to the grouped
outcome variables, as will the notations A2012 and Progress.
Correlations among the outcome variables were checked and appear in Appendix
D. The correlation between the two water outcome variables is 0.38, and between the
two sanitation outcome variables is 0.53, suggesting moderate to strong positive
relationships between variables from the same sub-sector. When comparing water A20l2
and sanitation A2012, the correlation coefficient is 0.35, and when comparing Water
Progress and Sanitation Progress, the correlation coefficient is 0.33, both suggesting
moderate positive relationships.
Explanatory and Control Variables
Table 5.2 shows descriptive statistics for the six key explanatory variables,
(derived from the GLAAS response averaging process described above), as well as the
eight control variables. (Refer to Appendix C for variable units of measurement.) By
observing mean values for the key explanatory variables, with possible values from 0.0 to
1.0, one can see that the mean for W-sector planning, at 0.72, is towards the range
maximum. Mean values for the other five variables, ranging from 0.48 to 0.63, are close
85


to the mid-range value of 0.5. The relatively balanced distribution of responses for five
of the six key variables works against the idea that countries may have been inclined to
rate themselves as high or low to appear in a more positive light.
The 0.72 mean value for W-sectorplanning, as well as the second highest mean
value of 0.63 for S-sectorplanning, do indicate less variation in responses on that
attribute. This may be because GLAAS questions on the topic focus on sector plan
development, which many countries have done in some form, rather than sector plan
implementation.
Notice also that the mean values for sanitation are all about 0.10 lower than the
mean values for water on the same attribute, perhaps reflecting a pattern of less attention
paid to sanitation issues, a theme that is taken up again later in Chapter VIII.
Table 5.2
Descriptive Statistics for
Key Explanatory Variables and Control Variables
Variable Obs Mean Std. Dev. Min Max
Key Explanatory Variable
W-Decentralization 74 0.62 0.22 0.17 1.00
W-Sector Planning 73 0.72 0.26 0.13 1.00
W-Civil Society 71 0.58 0.29 0.00 1.00
S-Decentralization 73 0.53 0.32 0.00 1.00
S-Sector Planning 74 0.63 0.28 0.00 1.00
S-Civil Society 70 0.48 0.27 0.00 1.00
Control Variable
Urban Population Change 75 0.16 0.14 -0.10 0.61
Corruption Level 75 69.97 9.46 37.00 92.00
Political Fragility 75 1.52 1.01 0.00 3.00
Security Fragility 75 0.71 1.02 0.00 3.00
Literacy Rate 75 72.13 21.22 27.00 99.80
GDP per capita (log) 75 7.41 1.09 5.53 10.09
Renewable Water Resources (log) 75 8.40 1.79 2.30 12.21
Foreign Aid to Water & Sanitation (log) 75 2.74 1.31 -1.35 6.06
86


Correlations
Appendix D shows correlations among the six key explanatory variables. These
were generated to check for potential problems of multi-collinearity among explanatory
variables prior to running the multivariate models. The correlations to consider are those
among the three water variables, and those among the three sanitation variables, as none
of the statistical tests are run with the water and sanitation variables together. For the
three water variables, the highest correlation is 0.51, between W-sectorplanning and W-
civil society. Among the three sanitation variables, the highest correlation is 0.55, also
between S-sector planning and S-civil society. From the perspective of the statistical
model, these correlation levels were not considered so high as to oblige removing one or
the other variable. The fact that these two attributes are closely aligned is perhaps not
surprising, as both attributes depend, to some extent, on a climate of consultation and
openness within a society.
Correlations between the water and sanitation decentralization variables, sector
planning variables, and civil society variables are all relatively high (> 0.53). This is not
surprising, as we would expect there to be a level of similarity within a country on its
approach to water and to sanitation, give how inter-related the two sub-sectors are.
Appendix D also shows correlations among the eight control variables. These
were generated to check for potential problems of multi-collinearity among these
variables prior to running the multivariate models. The highest correlation is between
GDP per capita and literacy at 0.62. However this is still considered at an acceptable
level, and attempts at removing one of the other from the analyses did not improve
overall results.
87


Full Text

PAGE 1

WATER AND SANITATI ON INSTITUTIONS AND GOVERNANCE: IMPACT ON SERVICE PROVISION IN URBAN AREAS OF LOW AND MIDDLE INCOME COUNTRIES by LAURIE ANN MANDERINO B.A., California University of Pennsylvania, 1985 M.S., University of Colorado Denver, 1993 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Doctor of Philosophy Public Affairs 2015

PAGE 2

2015 LAURIE ANN MANDERINO ALL RIGHTS RESERVED

PAGE 3

ii This thesis for the Doctor of Philosophy degree by Laurie Ann Manderino has been approved for the Public Affairs Program by Tanya Heikkila, Advisor Christine Martell C h a i r Christopher Weible John Brett November 20, 2015

PAGE 4

iii Manderino, Laurie Ann (Ph.D., Public Affairs) Water and Sanitation Institutions and Governance: Impact on Service Provision in Urban Areas of Low and Middle Income Countries Thesis dire cted by Assoc iate Professor Tanya Heikkila ABSTRACT Rapid global urbanization over the last few decades has intensified the challenge of providing adequa te wa ter and sanitation service s to urban residents Meeting this challenge has been the focus of domestic and international development efforts, including M illennium Development Goal 7.C This re search studies three institu tional and governance attributes theorized to impr ove government service outcomes test ing hypotheses that the attributes are associated with great er country progress on providing urban water and sanitation access The attribut es are: a) de centralized services ; b) sector wide strategy and investment coordination; a nd c) civil society engagement. Country level experience is analyzed using a series of ordered logi stic regression models for a sample of 75 low and middle income countries. UN GLAAS survey data is used to derive country specific variables for the three attributes. These, along with control variables representing country background conditions, are analyzed relative to four country progress outcome variable s, two e ach for water and sanitation. The outcome variables, (covering the 2000 to 2012 time period), are derived from the UN JMP dataset that tracks urban access rates by country. Based on results from the se models, f our count ry case studies look in depth at im plementatio n of the attributes and highlight aspects that can help or impede country progr ess. Overall, findings show that decentralization is helpful to sanitation progress, but n ot water progress, likely due limitations of capacity and funding faced by sub national

PAGE 5

iv levels of government T hree explanations are propos ed for why decentralization may im pact water and sanitation differently Results for s ector planning were mostly inconclusive except that it was shown helpful to water progress over the 12 year period. Study of t his attribute would benefit from additional wide scale data collection. Civil society engagem ent was consistently shown to help country progress in b oth w ater and sanitation, and several examples of engagement are p rofiled to demonstrate how it can improve service outc omes The last chapter relates findings t o theories about government provision of public goods. The extent to which the three attributes help achieve efficiency, supply, equity, and social welfare goals is discussed. Finally, practical recommendations for strengthening sector institutions and governance are presented with application to governments and international aid donors. The form and content of this abstract are approved. I recommend its publication. Approved: Tanya Heikkila

PAGE 6

v ACKNOWLEDGEMENTS I want to express my gratitude to Dr. Tanya Heikkila, my dissertation advisor for her ongoing guidance and support throughout the dissertation process, and also to Drs. Christine Martell, Christopher Weible, and John Brett for their valuable input and feedback on my research Other faculty members at the School of Public Affairs, particularly Dean P aul Teske and Dr. Peter deLeon, supported my academic development and sustained participation in the PhD program. Staff at the School of Public Affairs, especially Antoinette Sandoval and Rob Drouillard helpfully guided me through paperwork and compute r issues. Finally, my fellow PhD students were great learning partners and a source of beneficial advice throughout the program. I am grateful to those individuals overseas contacted for the case studies who took the time to respond to my research inqui ry or to connect me with knowledgeable colleagues in their country. I want to also acknowledge the National Science Foundation who supported this work through the Integrative Graduate Education and Research Traineeship (IGERT) fellowship program (Grant #: DGE 0654378), and those faculty and staff at the University of Colorado Denver who made this program possible.

PAGE 7

vi TABLE OF CONTENTS CHAPTER I. INTRODUCTION 1 II. BACKGROUND 7 III. LITERATURE REVIEW .12 Water and Sanitation Service as Public Goods .12 Water and Sanitation Institutions and Governance .18 Literature on Attributes Selected for Study .22 Country Background Conditions .49 Summary of Literature 55 IV. QUANTITATIVE METHODS 59 Sample Selection 59 Outcome Variables 6 3 Key Explanatory Variables 67 Control Variables Data Validity and Limitations .75 V. QUANTITATIVE AN AL YSIS Descriptive Statistics .83 Bivariate Analyses ..88 Multivariate Analyses 91 Questions for Further Exploration ... 112 VI. SECTOR WIDE STRATEGY AND INVESTMENT COORDINATION .114 Sector Planning by Countries 114

PAGE 8

vii Relating Sector Planning to Achievement Findings from Prior 2 0 VII. QUALITATIVE METHODS C Document Review Consultations with Country Experts Scope of the Case Studies VIII. DECENTRALIZED SERVICES .. Colombia Case Study .. Mali Case Study Summary and Analysis of Decentralization Cases ..16 5 Different Decentralization Impa ct on Water and Sanitation ................ ......... 177 IX. CIVIL SOCIETY ENGAGEMENT Kenya Case Study South Africa Case Study ... 19 6 Summary and Analysis of Civil Society Engagement Case .2 1 0 X. SUMMARY AND CONCLUSIONS 1 Introduction .221 Theoretical Background Research Methods Decentralized Services Sector wide 236

PAGE 9

viii .. 248 .. 2 4 9 Conclusion 2 5 0 BIBLIOGRAPHY 256 APPENDIX A. Theoretical Concepts and Constructs for Research Questions 9 B. List of Sample Countries with Data Notations 2 C. Variable Operationalization and Source Information 5 D. Correlation Tables 8 E. Frequency Tables for Key Explanatory Variables and Outcome Groups ..29 0 F. Graphs Relating Sector Planning to Progress 4 G. Persons Consulted for Case Studies 0 2 H. Case Study General Consultation Guides 4

PAGE 10

ix LIST OF TABLES TABLE 4.1. Average Country Population: Comparison of Sample Group and LMI Group .. 60 4.2. GDP per Capita: Com parison of Sample Group and LMI Group 61 4.3 Urban Water and Sanitation Access: Comparison of Sample Group and LMI Group ... 62 4.4. Overview of All Variables ... 63 4.5 Systems Considered as an Improved Source by JMP .. .64 5.1 Descriptive Statistics and Grouping Parameters for Outcome Variables .. .. .84 5.2 Descriptive Statistics for Key Explanatory Variables and Control Variables .. .86 5.3 Correlations between Explanatory Variables and Outcome Variables ... 89 5.4 Key Explanatory Variable Means by Outcome Group ... .91 5.5 Ordered Logistic Regression Results: Water A 201 2 5.6 Ordered Logistic Regression Results: Water Progress ... ... 5.7 Ordered Logistic Regression Results: Sanitation A 2012 .. .96 5.8 Ordered Logistic Regression Results: Sanitation Progress ... .97 5.9 Summary of Ordered Logistic Regression Results for Complete Models 8 5.10 Odds Ratios for Key Explanatory Variables 5.11 Control Variable Significance 8.1 Sources of Funds for Municipal Water and Sanitation Investment in Colombia (1995 2010) ... ... 8.2 Water and Sanitation Investment by Different Levels of Government in Colombia (2010)

PAGE 11

x 8.3 Sources of Funds for Municipal Water and Sanitation Investment in Mali (2006) 153 8.4 Decentralization Case Study Findings by Construct 6 5 9.1 Civil Society Engagement Case Study Findings by Construct 10 .1.

PAGE 12

xi LIST OF FIGURES FIGURE 5.1. Predicted Probabilities for Decentralization .. 5.2. Predicted Probabilities for Sector Planning .................................. .. .. ..105 5.3. Predicted Probabilities for Civil Society Engag ement

PAGE 13

xii LIST OF FORMULAS FORMULA 4.1 Progress Outcome Variable Calculation 4.2 . 5.1

PAGE 14

1 CHAPTER I INTRODUCTION At the U nited N ations (UN) Millennium Summit in September 2000, the largest ever gathering of world leaders adopted the Millennium Development Goals (MDG) an integrated set of time bound targets for extending the benefits of globalization to the They committed to improving the lives of people living in extreme poverty and demonstrating by the year 2015 progress in addressing th e most pressing issues facing developing countries (Lenton et al. 2005 ; UN, 2014 ) A specific target was set for increasing access to safe and sustainable w ater and basic sanitation among unserved populations (Target 7.C) While the problem persists in both rural and urban areas, t h e ch allenge is compounded in cities due to the accelerated pace of global urbanization in the last few decades Cities face numerous challenges in accommodating new residents and providing them with basic services. In 2012, an estimated 33 percent of the urban population in developing countries lived in slum conditions, communities characterized by a lack of basic services such as safe drinking water and adequate sanitation This percentage h as continually grown since 1990 The proportion of urban residents living in slums is particularly high in S ub Sahara n Africa at 62 percent (UN, 2014). In total, there are an estimated 863 million urban dwellers living in slums. Lack of clean water and sanitation in these communities lead s to a variet y of health problems including diarrheal disease for the affected populations (WHO, 2014b). As the global community strives to address this situation, e f forts focu s on different aspect s of the problem, such as water allo cation and appropriate technolog ies

PAGE 15

2 for low resource settings Other aspect s being addressed relate to the institutional and governance en vironments present in countries for water and sanitation service provision These environment s are the focus of this dissertation. While i ns titutions and governance are further explored in the literature review chapter basic definitions are presented here. Institutions have been broadly defined as [that] structu re incentives in human exchange, whether political, social, or economic (North 1990, p. 3). Institutional arrangements within governments determine, among other things, who makes decisions and who allocates available resources. These arrangements can pl ay an important role in government performance in the delivery of public goods 1 (Cremer, Estache & Seabright 1994 ; Frank & Martinez Vazquez, 2014) I nstitutions operate within the realm of governance Several different definitions for governance have been developed by researchers and international organizations. A widely accepted definition from the United Nations Development Program ( UNDP ) most t he exercise of economic, political, and [including ] mechanisms, processes, and institutions through which citizens and groups articulate their interests, exercise their legal rights, meet t heir obligations, and mediate their differences (UNDP, 1997, p.2). So, while institutions determine decision and incentive structures within governments the broader concept of governan ce also encompasses the values and processes by which government autho rity is exercised, including interactions with 1 Chapter III includes a detailed explanation of public goods.

PAGE 16

3 citizens and organized groups outside of government Accountability is a key theme underpinning several aspects of governance (UNDP, 2015). Understanding national institutional and governance environments is important because they guide efforts of cities to provide new and upgraded water and sanitation services that meet minimum international standards, something discussed more in a later chapter. How cities approach decisions such as where to constr uct new infrastructure, how to invest funds, and how to price services is to a large degree determined by the institutional and governance context set b y the national government. It determine s among other things, the authority of cities to make investment decisi ons how the public should be informed and consulted, and what type of service quality monitoring may be required. The ov erall goal of this dissertation is to evaluate if certain national institutional and governance attributes, generally considered indi cative of good governance, have had a positive impact in the water and sanitation sector, specifically for improving urban w ater and sanitation service provision This research focuses on three specific instituti onal and governance attributes r eflected in the following research questions. Additional background information and h ypot heses linked to the se questions are presented i n the literature r eview chapter. RQ1. What is the role of decentralization of water and sanitation services in a coun progress on providing urban water and sanitation access? RQ2. What is the role of sector wide strategy and investment coordination progress on providing urban water and sanitation access?

PAGE 17

4 RQ3. What is the role of civil society en gagement urban water and sanitation access? The three attributes were selected because: a) they have been identified in the literature as having potential to contribute to effective government service provision and donor aid effectiveness; b) they have not been studied extensively for the water and sanitation sector, particularly on a large scal e basis; c) there was an observed gap of research on potential interactions among the three attributes; d) they represent a departure from previous approaches to water and sanitation service provision in many developing countries; and e ) relatively new UN data coll ection programs provide comparable data on these attrib utes for multiple countries. To study these attributes, the unit of analysis is a country and the population of interest is low and middle income countries as classified by the World Bank ( World Bank website, 2015) The research is centered on the X (explanatory) and Y (outcome ) variable relationship, with spatial variation intended to explore theoretical linkages identified through literature review (Gerring, 2007) The overall time period being studied is approximately 1990 to 2012. This research add s to the existing literature on government service provision (see overview in Chapter III), by evaluating the impact of the three institutional and governance attributes specifically in the context of urban water and sanitation service provision, to see if their implementation in low and middle income countries really does lead to better performance, as pred icted by theory. It make s theoretical contributions by uncovering the conditions unde r which the three attributes might be beneficial to government provision of urban water and sanitation services in low and middle income

PAGE 18

5 countries. The results of this research will also be of interest to domestic and international development programs geared towards increasing urban water and sanitation access and maintain ing quality services. The resear ch is conducted in two phases. The first phase of research is a cross case analysis o f 75 countries using ordered logi stic regression models to assess the impact of the three institutional and governance attributes (explanatory variables) on country progress (outcome variables), while also controlling for several country background conditions. T he cross case analysis uses country data from two UN datasets, introduced in the next chapter. The phase one analysis focus es specifically on the 2000 to 2012 time period, for reasons of data availability further explained in Chapter I V. The second phase of research looks more in depth at the impact of individual attributes. For two of the attributes, decentralized services and civil society engagement, specific country case studies, (two for each attribute), chronicle country history with the attri bute of interest and further probe i ssues raised in the cross case analysis. C ase s focus primarily on one attribute to better understand how it was implemented in a country and what can be learned about conditions that promote positive or negative impact s. T he case studies are developed from review of relevant documents and consultations with knowledgeable persons in the selected countries. The remaining attribute sector wide strategy and investm ent coordination, is studied differently. Because there is not a long history of countr y implementation of this attribute instead of preparing ca se studies the analysis looks at other studies that have focused on this attribute for select groups of countries, and analyzes their findings alon g with the p hase one quantitative results to draw reasonab le conclusions.

PAGE 19

6 Th e focus of this research is on water and sanitation service provision therefore i ssues relating to water as a shared nat ural resource, including water rights, water quality, and water supply considerations are not addressed by this study. However, reference is made to some research on water re source management policy, as governance themes often ov erlap with wat er and sanitation service provision. Also, while this research is lim ited to water and sanitation service provision in urban areas, some st udies of rural areas are referenced when the findings have wider applicability. The next chapter provides f urther background on global water and sanitation challenges and the international efforts to address them S ubsequent chapters discuss relevant literature for this study; the quantitative cross case analysis and results ; qualitative findings for each of the three attributes, including country case studies; and the overall analysis and conclusion s

PAGE 20

7 CHAPTER II BACKGROUND Inadequate drinking water, sanitation and hygiene are estimated to cause more than 842,000 diarrheal disease deaths per year and to generate disease caused by micro organisms and chemicals ingest ed through unsafe drinking water (WHO 2014b). In addition adequate water, sanitation, and hygiene are critical for the prevention and care of almost all the tropical diseases targeted for elimination and control by the W orld H ealth O rganization (WHO) such as soil transmitted helminthiasis, guinea worm disease, trachoma and schistosomiasis, all of which affect mainly children (WHO, 2015). In recognition of the importance of safe drinking water and sanitation to human health, survival, and d evelopmen t, they were mad e the focus of MDG Target 7.C, to "halve, by 2015, the proportion of people without sustainable access to safe drinking water and basic sanitation comparison to the base year of 1990 (UN Millennium Development Goals list). The d rinking water part of MDG Target 7.C was met in 2010, as the percentage of people lacking access to safe drinking water decreased from 24% of the global population in 1990 to 11% in 2010 (UN, 2012). The sanitation part of the t arget has not yet been met. B ased on the most recently available global data (2015), 9% of th access to safe drinking water and 32% lack access to basic sanitation (WHO & UNICEF, 2000 2015). S ubsequent to the establishment of MDG Target 7.C in 2000, the UN Committee on Economic, Cultural and Social Rights declared that access to water is a human right and that water is a social and cultural good, not merely an economic commodity ( UN Committee on Economic, Social and Cultural Rights, 2002 ). In July 20 10, the UN

PAGE 21

8 General Assembly recognized the human right to water and sanitation and acknowledged them as essential to the realization of all human rights (United Nations General Assembly, 2010). Later, the UN Human Rights Council affirmed that the right to water and sanitation is derived from the right to an adequate standard of living, and is therefore covered by existing human rights treaties, placing an obligation o n states to meet this standard (UN Office of the High Commissioner for Human Rights, 2010 n.p.). The importance of institutions and governance for achieving water and sanitation access goals is broadly recognized in both the practitioner and academic communities. Several prominent international water meetings in recent years, (for ex ample, Wo rld Water Forum (2000); Bonn International Conference on Freshwater (2001); World Summit on Sustainable Development (2002); Commission on Sustainable Development 13 th Session (2005); Budapest Water Summit (201 3); and High Level International Conference on Water Cooperation (2013)), have highlighted water governance as a focal issue with regard to both water resource manag ement and service provision, as well as for meeting the MDGs relating to water. Among researchers, Saleth et al. (2007) emphasize the role of institutions in influencing the magnitude and sustainability of MDG impacts in low income countries They also explore linkages among past, ongoing, and planned policy i nterventions. Koundouri et al. (2008, p.189 190) find that for meta development goals such as the intermediate, but related goals of a hierarchy of development interventions, all of which require an effective institutional framework fo

PAGE 22

9 Researcher s have also found e ffective governance t o be a critical precondition for efficient use of water and sanitation investments as well as for operating and maintain ing infrastructure (Tropp 2007 ; Rogers & Hall, 2003; Plummer & Slaymaker, 2007 ) These views are discussed in more detail in the literature review chapter Several i nternational organizations are work ing on water and sanitation governance including the World er and Sanitation Program (WSP) and the Global Water Partnership (GWP). T heir reports are referenced throughout this document. Global Data Programs for the Sector In this section, I introduce two sector specific UN datasets that are used as sources for country level data for the cross case analysis. One reports on country institutio nal and governance environments and another measure s country progress. I then discuss two international group s that use data from these and other sources to plan strategic interventions in conjunction with MDG Target 7.C and subsequent sector goals. The first dataset is from the Global Analysis and Assessment of Sanitation and Drinking Water (GLAAS) a UN Wate r initiative implemented by the WHO begun in 2008 as a pilot and expanded over time to include more countries. GLAAS surveys countries biannually on water and sanitation topics, including the national institutional and governance environment The GLAAS process is intended to inform decision makers in developing countries and donor organi zations ; promote discussion and identification of priorities an d barriers to service provision; and promote a culture of partnersh ip and shared responsibility (WHO 2011 & 2014 a ) I use r ecent GLAAS surv ey data to derive explanatory variables representing country institutional and governance attributes.

PAGE 23

10 The s econd data set comes from t he WHO/UNICEF Joint Monitoring Program me for Water Supply and Sani tation (JMP) the o fficial UN mechanism for monitoring country progress towards MDG Target 7.C which has tracked individual country wate r and sanitation access each year since 1990. JMP publishes global, regional and national data for use by governments, donors, international organizations, and ci vil society (WHO and UNICEF, 2012) The data are widely available through an on line interactive websit e that allows for the creation of individualized data tables. I used t he JMP dataset to d erive the outcome variables representing country progress. Recognizing an ongoing need for data collection relating to MDG Target 7.C and subsequent international goals, a strategic ad visory group (SAG) of GLAAS and JMP representatives was formed in 2009 to bring independent, strategic thinking to further development and implementation of JMP and to provide guidance for rapid scale up of GLAAS reporting. SAG looks beyond the MDG deadline of 2015 towards development of a comprehensive global monitoring system for drinking wate r and sanitation, and harmonization of global and national monitoring needs, for tracking both development goals and human rights criteria (WHO and UNICEF, 2011) The f indings of this research may help SAG identify item s suited for ongoing monitoring. Another strategic planning organization, Sanitation and Water for All (SWA), is a global partnership of more than 90 developing country governments, donors, civil society organizations, and other development partners, working to promote action and improve accountability in programs designed to increase water and sanitation access. Annual SWA meetings focus on planning, institutional strengthening, better resource utilization and higher investment in the sector. Through the SWA process, countries make

PAGE 24

11 comm itments to address barriers to delivering sustainable water and sanitation services (SWA website 2014 ) Groups such as SWA can benefit from the findings of this research as they plan coordinated partner actions to achieve water and sanitation goals. The information highlighted in this chapter shows the magnitude of health problems posed by inadequate access to water and sanitation and the high level recognition it has received as a development goal. It also highlights the prominent role of institutions, governance and data tracking programs in global efforts to accelerate the pace of water and sanitation access both currently and in the post 2015 period In the next chapter literature relating to public service provision, institu tions and governance, and other relevant areas o f public af fairs is reviewed to form the basis for my research questions and study design.

PAGE 25

12 CHAPTER III LITERATURE REVIEW This chapter discusses relevant literature for this study beginning with the rationale for classifying water and sanitation service the basis for the country progress outcome variables, as a public good It l ays out a number of arguments for why water and sanitation service qualifies as a public good and why it should therefore be the responsibility of government. It then provides an overview of the literature on institutions and governance the subject of the explanatory variables, emphasizing the important role they play in government service provi sion These areas of literature represent the general theoretical groundwork for this research. Later sections re view more specific literature relating to each of the three attributes referred to in the research questions about country progress on providing urban water and sanitation access Some articles discuss these attributes in general, with applicability to various types of governme nt services, while other s are specific to the water and sanitation sector. The first topical area reviewed is decentraliz ed government services t he second is sector wide strategy and investment coordination (which also relates to foreign aid), and the third is civil society engagement A final section provides a brief overview of literature about country characteristics that are introduced as control variables in the statistical analysi s due to their potential impact on countr y water an d sanitation service provision. Water and Sanitation Service as Public Goods W ater and sanitation service provision can be considered as a public good one of four types of basic goods consumed in the economy (Weimer & Vining, 2005 ).

PAGE 26

13 Classifying it as public good has implications f or how a country might establish policy and organize institutions to provide the good. S ocial scientists classify goods into one of four categories by two characteristics: subtractability and exclusion Subtractability by another person. Exclusion is the difficulty of restricting those in the community who may benefit from the good or service. G oods that are ch aracterized as having both low subtractability, meaning that many people can jointly benefit from it, and low excludability meaning its access is difficult to restrict, are considered public goods ( Weimer & Vining 2005 ). The reasons why public goods are not considered suited to provision by the private sector market include information asymmetry, mo nopoly power, externalities, and allocative inefficie ncy that leaves some consumers underserved. Without adequate rules governing who in the marketplace w ill provide a public good, and how, a public good wi ll often be undersupplied. Moreover, inequities in who benefits from the good may exist with private provision, a problem that can be addressed through redistributive government policies and implementati on of targeted government programs (Weimer & Vining, 2005) Problems faced by the marketplace in providing public g oods include difficulty knowing consumer valuations for these goods, and problems of crowding and congestion that may occur with low excludability goods ( Weimer & V ining 2005 ). One way to address the latter problem is a system of usage rights for access to facilities that supply water and provide sanitation. It has be en shown that government and communities are

PAGE 27

14 often in a better position to design and implement a system of joint rights and usage than the private sector ( O strom 2003). J ustificati on for government provision of public goods stems from the overall arguments for government intervention in the marketplace : (a) to correct market failures, thus improving efficiency in the production and allocation of r esources and goods; and (b) to reallocate g oods to achieve distributional and other societal values (Weimer & V ining, 2005). For this sector, theory says that e fficiency goals can be better served through government provision if it addresses problem s of undersupply and promotes d istribu tional and social welfare goals by extending s ervices to marginalized groups, thereby reducing health problems caused by poor water supply and sanitation. The positive overall impact to society of w ater and sanitation access is the reason it can also be considered as a merit good, something of value for the government to promote, whe ther or not it can produce it efficiently (Musgrave & Musgrave 1989 ). The value to society comes via prevention of disease in children and adults, including infe ctious diseases that can results from the ingestion of contaminants due to lac k of access to potable water and use of inadequate sanitation facilities Given the ne gative health externalities associated with a lack of clean water and sanitation facilities there is a societal rationale for the state and the global commu nity to promote these goods Doing so minimize s communa l health problems that potentially have global impact and extend beyond national borders. If social goals for water and sanitation are not addressed, there are detrimental implications for the poorest members of society (Wijaya, 2005) and, a s Rawls (1999) proposed equality in basic soci al goods is a fundamental right S ociety should seek to

PAGE 28

15 maximize benefits to those with the lowes t level of prosperity that is, persons with the lowest utility on so Private Sector Involvement The view of water and sanitation service as a public good best suited for government provision may seem at odds with country policies that allow for private sector involvement in service provision. The key point here is that even when water and sanitation services may be supplied by a private orga nization, if it is doing so under contract with the government, or under some form of public private partnership the government is still taking responsibility for provision of the public good to consu mers, and not leaving provision subject to the private market. All models of government service provision, including those engaging the private sector, are impacted by the national institutional and governance environments being studied for this research. As noted Governments can opt to have the private sector provide water and sanitation Likewise, they can choos e public service provision. In both cases, they must comply with their human rights o bligation Further, she stated enabling environment and to regulate and monitor private providers ( UN Office of the High Commissioner for Human Rights, 2010 n.p. ). The introduction of private sector operators was intended to increase efficiencies and overcome the high fixed costs associated with construction of water and sanitation infrastructure (Kessides, 2004). However, the share of private sector involvement in the sector globally has been proportionally small, and its impact can be difficult to evaluate. Experience has often fallen short of expectations, both for investors and for recipient

PAGE 29

16 countries, sometimes leading to highly contentious debates about the merits of this approach. The initially anticipated flow of private investments to the sector did not occur, often due to perceived investment risk resulting from inadequate investment frameworks in some developing countries (OECD, 2009). Annual levels of private investment in the sector through 2014 have been below the 2000 level. P rojects are concentrated in certain regions, particularly East Asia and Latin Ame rica According to the World Bank Private Participation in Infrastructure Databas e (2015), which measure private sector investment in different categories of infrastructure investments, conc essions are the most common type s of arrangements. Concession c ontracts are when a public entity grants the right and the obligation to provide a utility service to a private company under terms and conditions specified in a contract or license. The private sector partner takes over operational responsibility and at l east part of the commer cial risk of service provision. The contract terms and conditions often specify results to be achieved by the private sector partner (World Bank, 1995). A 2004 overview report prepared for the World Bank noted the difficulty of ass essing the added value of private sector involvement in the sector because the conditions built into individual service contracts between government and the private sector can vary greatly and include specific conditi ons not made public. Also, the history of private sector involvement in this sector is fairly recent, particularly as compared to other infrastructure sectors such as transportation and communication, so there is not a long history for analysis (Kessides, 2004). In comparing public and private sector potential to contribute to developmental goals such as MDG Target 7.C Lobina and Hall (2008) found that the public sector has a

PAGE 30

17 comparative advantage over the private sector for extending service delivery with one key factor being knowledg e sharing among stakeholder networks. The ir study based on a review of current practices in seve ral large urban centers globally, found that i n transition and developing countries, reforms and good practices can cr e ate h ighly capable public operators. I n developing country urban areas where systems of accountability, transparency, and public participation are priorities, t he informal stakeholder networks that develop from these activities were found to be efficient ways of generat ing and communicating k nowledge. This exchange then translated into better public decision making around services. The authors see a crucial role for public providers and recommend that donor aid be directed towards strengthening public sector ut ilities towards MDG attainment (Lobina and Hall, 2008) Summary To synthesize the above literature, wat er and sanitation service access exhibits the characteristics of a public good and a rguments can be made that the government has advantages over the private marketplace for prov iding these goods. These arguments center on improved service efficiency and greater equity of distribution. Also, b ecause provision in low income settings may be complicated by problems of crowding and congestion some f orm of community collec tive action led by government may be needed Finally, g iven the global recognition of water as a basic human right, and the negative health externalities associated with poor water supply and sanitation, their placement in the special class of merit goods provides strong justification for government provision, as well as for support from the international donor community.

PAGE 31

18 Even in countries where the government may enter into service agreements with private sector operators, the government still retains responsibility for service provision and the operations are impacted by the institutional and governance environment. The arguments discussed above justify government and donor participation in water and sanitation service provision, lead ing to the next stage of literature review about how institutions and governance impact the provision of public goods. Wate r and Sanitation Institutions and Governance How governments organize themselves to provide water and sanitation service access, in terms of institutions and governance is the focus of this se ction. Institutions determine decision and incentive structures within governments, while the broader concept of governance also encompasses the values and processes by which government authorit y is exercised. The literature reviewed here is mostly about water, as m ore scholarly articles are available on water than sanitation. However, parallels can be drawn f rom one to the other, as management of the two are often interlinked and progr ess in both sub sectors is impacted by institutional and governance programs Institutions (North 1990, p. 3). A m ajor role of institutions in a society is to reduce uncertainty by es tablishing a stable structure for human interaction (North, 1990) Institutional arrangements within governments determine, among other things, who makes decisions and who allocates available resources. These arrangements can play an important role in

PAGE 32

19 government performance in the delivery of public goods (Cremer, Estache, & Seabright 1994 ). E fficient service prov ision requires coordination within government and sufficient institutional capacity ( Frank & Martinez Vazquez, 2014 ). Institutions are key to designing, constructing, and managing wat er and sanitation services A definition of water institutions has been derived from the general definition sets, provide incentives and determine outcomes both in individual and collective decisions related to water development, a llocation, use and management Dinar, 2005, p.2). The ability to effectively engage resources and capabilities toward water and sanitation service provision depends largely on institutions, as these institutions encompass the rules and mechanism s for : access to and allocation of resources; placement of economic incentives; and systems of service design and delivery (Saleth & Kadushkin, 2005). There is a fundamental link between institutional reforms and changes in service delivery as demonstr ated by a water sector framework developed by Saleth and Dinar (2005). The framework shows that understanding the structure and environment of water and related institutions, and the changes that take place within, can be useful for approaching reform in these sectors. The framework considers the institutional structure of water institutions ( further broken down into legal, policy, and organizational components ) and shows how important linkages exist both within water institutions and also between these institutions and the broader political and economic environment in a country U nderstand ing channels of institutional change are important for designing and implementing reforms to realize desired change s in water sector performanc e.

PAGE 33

20 Water and sanitation institutions are vital to the topic of this research because they determine service delivery incentives and decision framewor ks. T he broader concept of governance discussed in the n ext section, incorporates the values and processes by which government fulfills its service delivery function including interactions with citizens and organized groups outside of government. Governance Several different definitions for governance have been developed by researchers and international organizations. A widely accepted definition from UNDP most closely the exercise of economic, political, and uding] mechanisms, processes, and institutions through which citizens and groups articulate their interests, exercise their legal rights, meet their obligations, and mediate their differences (UNDP, 1997, p.2). The Global Water Partnership defines water political, social, economic and administrative systems that are in place to develop and manage water resources, and the delivery of water services, at different levels of society (Global Water Partnership, 2002, p.1). Governa nce in the water sector addresses when and how people get water and whose voices are heard in decision making (Tropp, 2007; Rogers & Hall, 2003; Plummer & Slaymaker, 2007). E ffective water sector governance include s approaches that are open and transparen t, inclusive and communicative, coherent and integrative, and equitable and ethical, as well as performance and operations that are accountable, efficient, responsive and sustainabl e (Rogers & Hall 2003, p. 27 29).

PAGE 34

21 G overnance in the water sector has not received the same level of attention as has historically been focused on the technical and infrastructure aspects of water provision. It started to be a focus of international development in the late 1980s, as it was realized that mu ch of the investment in technology and infrastructure by donors and the World Bank did not realize its intended effect. There is now an increasing trend towards framing water challenges in international devel opment in terms of governance. Effective governance has been linke d to increased likelihood of sustainable water management by Iribarnegaray and Seghezzo (2012) who developed a Sustainable Water Governance Index showing value in participation/engagement between the public and water authorities N ew governance approache s are broader than traditional inter governmental arrangements, and therefore more inclusive. Accountability is a key component of governance, and new actors in the governance process are monitoring the actions of government institutions at different leve l s As a results, institutions have become more transparent and more accountable, realizing a trend away from strict inter governmental relationships to relationships between government actors and nongovernmental actors including for global institutions and issues (Nye & Donahue, 2000) B y forging strategic coalit ions with actors outside of government, governance practices are said to enhance Tropp (2007) found that new forms of go vernance, which engage governments, non government organizations, the private sector, academic and research institutions, are a way to stimulate innovative partnering for implementing global initiatives, including the MDGs. However, t hese new forms of governance may require c apacity building for

PAGE 35

22 policy makers and managers so they may adapt to new forms of governance including coordinated and negotiated approaches These new approaches apply to both la rge scale national issues and local level service p rovision This overview of literature relating to institutions and governance shows the important role that they can play in government service provision The next section will focus in on three specific attributes of institutions and governance and discuss how they relate to urban water and sanitation service provision. Literature on Attributes Selected for Study This sect ion overviews literature related t o the three specific institutional and governance attributes sele cted for study : a) decentralized services; b ) sector wide strategy an d investment coordination; and c ) civil society engagement. At the end of each topical discussion a summary of the literature is provided which forms the basis for the hypothesis linked to th at attribute and the development of con struct level definitions. Decentraliz ed Services National and regional water service providers in developing countries have often historicall y failed to perform efficiently and in a sustainable manner, due to lack of accountability to customers and communities dispersed across a large geographic area, as well as constraints to operations from bureaucracy and politics (Pil grim Roche, Revels, & Kalbermatten 2004). With the onset of sector d ecentralization implemented as part of general institutional reforms in many developing countries, came an exp ectation of improved service s The body of l iterature relating to decentralization, both in general and specifically for the water and sanitation sector reveal s the rationale for decentralization and highlights the conditions under which anticipated benefits are most likely realized.

PAGE 36

23 D ecentraliza management functions from the central national government to organizations at the regional or local levels ( Handoussa 2009, p. 217). While not explicitly stated in this definition, many articles note that decentralized revenue raising authority and expenditure decisions, termed fiscal decentralization, is a key sub component of decentralization. This aspect is further discussed later in this section. In a meta analysis of US and in ternational studies of government performance researchers found that the degree of c entralization, size, and formalization of agencies were among the key institutional variables associated with government service performance (Forbes & Lynn, 2005, Hill & Lynn, 2005). In the water sector specifically, decentralization was found to be one of the key policy elements having an impact on sector performance in an analysis of water institutional reforms in 43 countries/regions (Saleth & Dinar 2004 ). Driving decentralization in the water and sanitation sector is the belief that lower levels of government can be more effective in responding to local conditions and meeting the needs of local consumers, thereby making consumers more willing to maintain and pay for services that they feel respond to their needs (Briscoe & Garn, 1995; Litvack & Seddon, 1999). It is thought that local governments are in a better position to assess local conditions, identify local needs, and make decisions about pr ovision of collective goods (Blair, 2000; Wunsch & Olowu, 1990). Decentralization is supported by examining some of the inherent problems in centralized approaches. Looking at institutional incentives in the provision of urban sanitation, Evans (1995) found that when decisions about wh at level of service to provide

PAGE 37

2 4 in different localities are left to the central government t here can be a problem of low government acc ountability, pa rticularly to low income groups. The central government can behave opportunistically and put political priorities a head of social goals. Further, there is likely no mechanism at the central government level f or establishing user preference s for the type of sanitation service to provide when overseeing multiple communities and central government organizational culture tends to work against the developmen t of innovative designs. Decentralization has been promoted as an alternative institut ional arrangement that addresses the problems of centralized management. Literature on decentralization explains how decentralization can promote political, efficiency and governance values ( Wolman 1990 ; Pius Kulipossa 2004), each of which are discussed here Political values can be served by spreading political power to a wi der range of citizens as compared to a centralized approach Benefits include: political training and e ducation for local officials ; political equality, rather than conce ntrated power is enhanced; and political security is better secured through increased participa tion in formal politics, support for political parties and other practices. Decentralization may also improve inter organizational coordination (Pius Kulipossa 2004) Efficiency can also be enhanced by decentralization Efficiency is influenced by the levels of authority and jurisdictions present in an institutional environment, according to p olycentric governance theorists (V. Ostrom, Tiebout, & Warren, 1961). To be effective, decentralization must be accompanied by incentives that motivate individual and collective action (McGinnis, 1999; McGinnis, 2011) The appropriate degree of

PAGE 38

25 decentralization depends upon which leve l of government will have the most incentive to act to bring about de sired outcomes (Cremer, et al., 1994). Allocative efficiency can be improved because of the ability to capture diverse preferences for public goods (Tiebout 1956 ; Oates, 1972; Musgrave 1997 ), and it also encourages competition and efficiency, thereby offering citizens more choices and options (Rondinelli McCullough, & Johnson 1989). Oates (1972) put forth a general decentralization theorem that in the absence of cost savings from the centralized provision of a public good, and production does not interfere with another community, welfare will be the same, if not better, when each community designs a program to meet its specific needs, rather than being subjec ted to a standard, centralized approach. More recently, Oates (2005 ) extended these ideas to consider problems associated with soft budget constraints under decentralization, s uggest ing that fiscal decentralization may require sufficient power at the cent ral government level to moderate excessive financial demands from local governments. Some level of competition for funds among local governments can be helpful to avoid inefficient spending, and central governments should provide a stable and enduring federal structure to moderate local spending, without overly extending its role under a decentralized system. Decentralization can also be viewed as supportive of good governance, promoting positive interactions between the government and the public. In particular, it can cause government officials to be more responsiv e and accountable to citizens ( Wolman 1990 ). According to Azfar Kahkonen, Lanyi, Meagher, and Rutherford (2004), decentralization, in its best sense, is a system where institutions and information flow produce effective

PAGE 39

26 government and where accountability of government officials should increase, especially if actions are transparent and civil society is incl uded. Decentralization can address credibility and information problems, especially if monitoring is shifted to a community, which reduces the information need s of citizens to hold pol iticians accountable (Keefer & Khemani 2005). Along with the perceiv ed positive value outcome s from decentralization, the literature has also e xplored potential problems with decentralized systems, the key arguments of which are summarized below. One such argument relates to power abuse within localities L ocal provision of public services may be subjected to inequalities due to the power of local elites who can influence public funds expenditures more easily in a de centralized system (Bardhan & Mookherjee 2005 ; Mookherjee & Bardhan 2005 ) L ocal government officials may be more subject to capture by local interests that can result in overprovision to certain powerful groups (Azfar et al. 2004; Bardhan & Mookherjee 2000 ). A lso l ocal officials have the capability to withhold contributions needed for successful implementation, and the central government may lack the authority to promote corrective action and will incur monitoring costs for oversight of local authorities ( Weimer & V ining 2005 ). When management systems are largely lacking within a country, som e infrastructure services may be more capably handled at a high er level of government (Bahl & Wallace 2005). There are potential economy of scale benefits with centrally planned projects, although the same might be realized under a decentralized system with local communities partnering together for joint projects.

PAGE 40

27 An other advantage of centralized authority is that it can equalize conditions among rich and poor areas by transfer of tax re venues to areas most in need. Moreover in a developing country context, the intended efficiency benefits of fiscal decentralization may not be fully realized due to : a lack o f elections where the local population can reveal preferences for public services; a lack of legal authority granted to local jurisdictions for tax collection of expenditures ; and limited administrative abilities to support increased spending or revenue co llection (Bahl & Linn, 1992). The impact of decentralized service delivery depend s on the method chosen for financing local governments, as restraints on the revenue raising capability of local governments can be limit ing. If a c entral government is generally struggling to provide basic services to the population it may be reluctant to give up power and resources to sub na tional levels of government and decentralization can be ineffective in practice (Jutting, 2005) Turning now to c ountry specific studies of decentralization this section highlight s what they have found about the co nditions that promote success. A common theme is that decentralization in dev eloping countries may not realize all anticipated benefits if certain ena b ling factors are not in place. The success of decentralization may be contingent on local authorities be ing granted both decision making and revenue raising authority. In a study in Zimbabwe, it was found that decentralization of responsibilities from t he central government did not actually result in decentralized authority Local communities lacked autonomy in decision making, which limited their ability to effectively manage local reso urces, revenues, and personnel. The inability to raise revenues lo cally was a major impediment. For effective urban management in Zimbabwe and Africa in general, the author

PAGE 41

28 recommend ed m ore efforts towards strengthening the role of democratically elected officials and civil society (Davison, 2001). A nother study of decentralization in Uganda and the Philippines looking at the health and education sectors showed that local governments have limited authority and are unable to adjust services to meet local demands (Azfar, Khknen, & Meagher, 2001). Studie s focused on l ocal water supply have shown that some success is possible with local management, but support is needed to address institutional challenges and to realize sustained improvements in service (Seppala, 2002; Pilgrim et al., 2004). Mugabi and Nj iru (2006) studied small communities in low income countries and found that while local authorities may be in a better position to understand local needs, they lack the technical and management capacity to respond to the n eeds at hand, due to weak in stitutional capacity They recommend that the transition to decentralized water service delivery be done in tandem with policies and programs that address expected challenges, such as manage ment capacity building and technical assistance programs and gra nting authority to raise revenue for financing of operations. Financial and ma nagerial autonomy, transparency and accountability, as well as professional support, are key ingredients for realizing improved efficiency and effectiveness of service (Mugabi a nd Njiru. 2006). Similarly, Wilder and Romero (2006) found that continued state involvement was required to ensure equity, accountability, transparency, and sustainability of decentralized water institutions in Mexico. Marcus and Onjala (2008), in stu dying areas of Kenya and Madagascar, found that national governments did not have in place the necessary legal, juridical, and regulatory frameworks to promote successful decentralization. While

PAGE 42

29 decentralization was expected to increase accountability and improve governance of water systems, it actually resulted in central government pulling back from providing leadership in water governance too soon, before problems of local funding gaps and lack of local resources for management were addressed. In rural India, Asthana (2003) found there is often a long term strategy lacking with decentralization and an authority closer to the community level will not always provide better service. However, decentralization does offer the advantage to the local l evel of learning by doing, which will likely eventually result in greater efficiencies and less corruption than centralized service, especially if there is guidance from the state government and oversight by the public (Asthana, 2003). Role clarification between local and national gover nment is also an important element R egulatory frameworks are a way to clarify roles, but putting these frameworks into practice can be challenging. Marcus and Onjala (2008) suggest negotiations to clarify state and local rights, responsibilities, and capacities with regard to decentr alizations, taking into consideration culturally embedded local norms. As mentioned above, f iscal d e centralization is important in addition to managerial decentralization ( Bardhan & Mookherjee 2000; Mugabi & Njiru 2006; Gillespie, 2005; Rodinelli 1990 ) because it allows sub national government s to exercise more autonomy in se rving the local population. Gillespie (2005) pointed out that one of the factors limiting finance to the water sector in many countries is failure of water sector governance including decentralization programs that do not clarify roles of central and local governments, lack of financial autonomy of decentralized authorities, and lack of an appropriate operat ions framework for utilities

PAGE 43

30 According to Rodinelli (1990), central restrictions on the financial authority of local governments should be eased to facilitate decentralization, and there should be capacity building to increase skills for raising local revenues, managing expenditures, and delivering and m aintaining efficient services. T here is a risk of local governments over taxing and over regulating that may hinder overall economic investment and gr owth. This requires careful consideration of local government revenue raising instruments an d expenditure plans ( Bahl Linn & Wetzel 2013 ) When shifting financing to sub national levels of government, it is important to create a governance environment that promotes public services delivered in cost effective ways. Even if funding to local authorities is adequate, inappropriate management of funds can result in a perceived lack of financial resources within government structures (Funke et al., 2007). Taken together, the literature points to seve ral anticipated benefits of decentralized government services, including: local authorities being more responsive to local needs and preferences than centralized systems; more efficient service provision because program s can reflect diverse preference and be designed to meet its specific needs, rather than being subjected to a standard, centralized approac h; more accountable and transparent systems, as the degree of information asymmetry is reduced; better representation of good governance values, includin g equity in government services, positive government interactions with citizens; and improved political equity and political security with political training and education for local leaders. There are also o ther studies that suggest that decentralization b enefits may not be fully realized if certain factors are not in place. For example, incomplete delegation of decision making, lack of authority for revenue raising and investments unclear roles and

PAGE 44

31 procedures for different levels of government and lack of managerial and technical capacity among local managers to adapt to new responsibilities. Other potential problems with decentralization include power abuse by local authorities, lack of proper incentives to motivate individual and collaborat ion action on the part of decentralized authorities, unequal financial resources among different sub national regions without the benefit of resource equalization offered by a centralized approach. Decentralization benefits may also be limited by circumst ances that stem from economic, political, and socio cultural forces at work, such as historical tensions between central and local authorities and a history of low cit izen engagement with government. ecentralization has the potential to addre ss many of the water infrastructure and planning challenges, but it requires an extensive range of institutional pre conditions to be successful; -improving the institutional framework under which local management initiatives operate 06, p.191). The literature highlights several conditions necessary for decentralization to realize theorized benefits, and an interesting lin e of research is to explore to what extent these conditions are present in countries where urban water and sanita tion provision is decentralized. So, while the hypothesis stated below assumes an overall positive impact for decentralization, phase two of the two phase study a pp roach will provide a more insightful analysis by focusing on the exten t to which important aspects are present in decentralized countries. Following are the research question and hypothesis related to decentralization: progress on providing urban water an d sanitation access?

PAGE 45

32 H1. A high level of decentralization in water and sanitation services is associated with greater country progress on providing urban water and sanitation access. The important aspects of decentralization discussed in the literature will be studied through the constructs for decentralization as follows: a) Decision making authority for service provision is delegated to sub national levels of government. b) Authority to manage expenditures and raise revenues is delegated to sub national levels of government. c) There are clearly defined rights, responsibilities, and roles for sector management actors at the national and sub national levels. d) Sub national level authorities are monitored by communities and/or the central government including, for example, regarding public funds expenditures and equality in service provision. e) Support is provided to build managerial and technical capacity of sub national levels of government. Sector wide Strate gy and Investment Coordination In this section I review literature on international devel opment aid effectiveness and relate this to recent initiatives for coordinated national planning for water and sanitation I introduce key elements of theory and current recommendations for what constitutes effective international development aid, and tie that discussion to relatively new a pproaches that have emerged to address ide ntified shortcomings of previous aid modalities These new approaches are grouped under the heading of sector wide strategy and investment coordination defined for this research as: mechanisms for transparent, country led partnering with donors as part of coordinating, directing, and utilizing intern ational aid towards priority areas for the sector (adapted from Tropp, 2007).

PAGE 46

33 M ulti lateral and bi lateral donor aid programs assist c ountries with provision of basic public service s, such as schools, health facili ties, and infras tructure However, donor aid programs over the la st several decades have been found to lack a long term effectiveness in promoting economic growth (Gibson, Andersson et al. 2005 ; Riddell 2007 ; Boone 1996 ; Pedersen 2001 ). D epending on how it is structured, international development aid can actually create motivational problems and imped e a ability to promote development. A problem of c ryptoimperisalism can emerge whe n p olicymakers in developing countries are influenced by the ideas of foreign donors on which they depend so heavily (V Ostrom, 1999 ) F inancial dependency on foreign aid fosters a power structure between donor and recipient countries that can create incentive problems. Buchanan (1977) used the term the to characterize a motivational problem created when international a id is viewed as a type of two person game with ordinal payoffs. A Samaritan, concerned about the well being of others, enters into situations in which other people might need help and decides to help or not to help. The recipient also decides how much ef fort to put forth to receive the assistance. The dilemma is created when Samaritans (international aid donors) achieve their goal of helping others no matter what level of effort the recipient puts forth, and so this becom es their dominant strategy. An a id recipient country once it understand s this reality will put forth a low level of effort a s its dominant strategy. This may lead to inefficiencies over time, with the distribution of b enefits skewed to the recipient. It also creates a motivational pr oblem on the part of the recipient Under such a situation the recipient loses skill and motivation, thus impeding its own capability to manage and sustain programs over time (Buchanan 1977)

PAGE 47

34 Resear ch on past donor aid programs has shown that the failure to fully achieve goals of donor aid programs h as not resulted so much from a lack of donor assistance funds flowing to the le ss developed countries, as it has from a lack of effective institutions to mana ge and administer the funds (Auer 2007; Gibson, Anderson et al. 2005). I f institutional arrangements do not promote ownership and coordination of international aid on the part of the recipient country, international aid flows may fail to achieve desired and sustainable results in improving the conditions in le ss developed countries. In addition to funding projects t hat address primary public goods deficiencie s development aid policy should also help design new institutional rules that guide government decision making and management over the long ter m (Gibson, Andersson et al 2005) P rincipal/agent theory can be applied to multiple relationships in the scheme of foreign aid ( Moe, 1984; Martens Mummert, Murrell, & Seabright, 2002 ) The principal agent model is an economic model in which the principal enters into a contract or agreement with the agent in the expectation that the agent will choose actions that produce outcomes desired by the principal (Bendor: 1988, Miller: 1992) Some applications of this model to foreign aid are : (a ) donor aid implementing agencies (agents) which are subject to multiple principals within their own countries including politicians, taxpayers and publ ic officials ; (b ) complex incentive and control relationship s among donor agencie s (principals) and implementing contractors (agents) ; and (c ) the direct r elationships between foreign donor s (principals) and aid recipient countries (agents) I explore relationship (c) in th e following discussion a s my research

PAGE 48

35 is focused on recipient countries. ( S ee Martens et al., 2002 for further dis cussion of relationships (a) and (b) ) P roblems of asymmetric information can exist betw een principals and agents. This lead s to problems of adverse selection and moral hazard which Rauchhaus (2009) finds can obstruct humanitarian aid interventions Looking first at adverse selection, this stems from a situation where asymmetric information exists between the parties before entering into a contract With donor aid, t prior to entering into a contract that are not rev ealed to the principal (the donor) during the negotiation period Moral hazard on the other hand, occurs when a party to an arrangement has an opportunity to take action hidden from the other party once an agreement is in place with the principal unable t his creates a n opportunity for agents to change their behavior in a way t hat increases their potential for gain Both adverse selection and moral hazard can apply to in ternational aid agreements. The problem is compounde d by the fact that these agreements a re often less binding and enforceable tha n domestic political contracts, as they are essentially self enforcing and not bound by the law o f any particular jurisdiction. Actions taken or not taken by the agent (the aid recipient country) may be difficult to assess and monitor by the principal (the donor) or even by a neutral third party. These circumstances can foster a commitment dilemma d ue to the weakly institutionalized nature of international politics. Calmette and Kelkenny (2002) looked at asymmetric information problems in donor aid and showed how they can lead to reduced country self help They present three scenarios. First, w hen aid for public works is donated unconditionally a recipient

PAGE 49

36 government spends less on public works, levies the lowest taxes on the rich, and the degree of self help is at its lowest. Second, w hen charity is condition al but information sharing incomp lete, countries are motivated to exaggerate their needs and downplay any domestic self help investment program s creating an adve rse selection problem. I f recipient countries perceive themselves as competing for aid with other countries, recipient countri es may adjust their needs to qualify for more aid (Pedersen 2001). Finally, when adequate information about internal development investment efforts is known by the donor agency, it is in a better position to create conditions or incentives as part of the aid terms to mitigate potential negative effects of adverse selection, principal agent, and moral hazard problems (Calmette & Kelkenny 2002) In short, i nformation asymmetry exists when a donor is unable to distinguish the true lev el of need and effor t of a country. C ountries with more resources may hide their capacity to address thei r own problems so that they receive higher transfe rs from donors, thereby exhibiting shirking behavior. C orrupt, but potentially less needy countries may be able to capture informational rent. Calmette and Kelkenny (2002) suggest that having incentive s in place to encourage countries to correctly represent their domestic capacity would help address this problem and increase the efficiency of aid, adding to overall gl obal welfare. T he s e incentives can take the form of incr eased transparency requirements and having mechanism s in pl ace to monitor and penalize mis representation of domestic ef forts. The recognition of potential incentive problems with traditional models of donor aid has, among other things, spurned new thinking about how aid is approached at the international level. Riddell ( 2007) found that donor aid is most effective when there is

PAGE 50

37 predictability, efficiency in allocation, improved coordination, and me aningful partnerships between donors and recipients Gibson, Andersson et al. (2005 ) have added tha t d onors should focus on country ownership of development assistance programs to encourage sustainability, and prioritize beneficiaries whe n designing develo pment programs An example of new style governance is the sector wide approach ( SWAp ) a planning vehicle by which all significant domestic and donor sector investments in a developing country are coordinated and directed towards the same overall objectives, consistent with a sector strategy The SWAp model has been used since the mid 1990s as a method of international aid delivery in sectors such as health and education based on a set of operating principles rather than a specific packa ge of policies or activities (Development Assistance Committee, 2005) Donor support for a SWAp can take the form of project aid, technical assistance or general budget support (OECD, 2006). A SWA p allows development partners (donors, governments, and other stakeholders), to work towards a coordinated natio nal program, rather than on project specific efforts Characteristics of the SWAp include : broadening policy dialogue; developing a single sector policy (that addresses both public a nd private sector issue s) and a common realistic expenditure program; common monitoring arrangements; and more coordinated procedur es for funding and procurement ( WHO 2000). With a SWAp, aid partnership arrangement s are shifted more towar ds recipient government leadership, although there is joint cooperation in planning policy and projects. Key characteristics are described by WHO:

PAGE 51

38 Key characteristics of the SWAp should include: i) the partner government clearly leads and owns the programme; and ii) a common effort by external partners to support that programme, including provision of all or a major share of funding for the sector, in support of the government's unified p olicy and expenditure programme ( WHO, 2013 p .1 ) B oth ongoing and new p rojects are screened to fit sector wide priorities, aiming for more comprehensive coverage of sector needs C ommon procedures for project management and funding are developed increasing ly led by the government. Additional benefits of SWAps in recipient countries can include promotion of more local involvement in sector development projects, more accountability of government officials, and increased capacity building The SWAp approach is in contrast to the system of individual donor driven projects that sometimes led to fragmentation and duplication of efforts, and also placed coordination. SWAps are intended to lead to more efficient use of financial resources from nati onal budgets, donors, and service users, maximizing overall investment in t he sector towards achieving national water and sanitation MDG t arget s ( WHO, 2013; Harpe, undated ). Under a SWAp, t here is less potential for hidden information, as details on all planned sector projects and investments are widely shared among domestic and international partners. Similar to SWAps are c ountry specific poverty reduction strategy p apers (PRSPs), national planning documents stemming from the 1996 Enhanced Heavily Indebted Poor Countries (HIPC) Initiative, by which G 7 governments offer ed low interest loans and other forms of debt relief to indebted nations. PRSPs are an approach to reducing poverty and a method for structuring development assistance, intended to be

PAGE 52

39 country driven, partnership oriented, and developed with civil society participation The PRSP development process sought to promote increased transparency by governments and donors (IMF, 2013). PRSPs have specific sections devoted to different develo pm ent challenges within a country and in some cases includes a section specifically on water and sanitation infrastructure planning. PRSPs are a change from previous national development strategies, which were often based on directives from the World B ank and I nternational M onetary F und (IMF) While the IMF is currently transitioning from PRSPs to economic development d ocuments (EDD s ) for some countries, these documents are similar in nature in that they articulate and track an overall strategy for poverty reduction and growth (IMF, 2015). The degree to which the P R S P process has changed development assistance has been assessed by several analysts. Cheru (2006) sees PRSPs as potentially important instruments for streamlining and coordinating international aid flows, but that effort is needed to ensure the PRSP approach remains participatory and results oriented. Abrahamsen (2004) sees the PRSP form of partnersh ip as advantageous in that there is an increased potential for inclusion, as well as an explicit commitment to the self governance of recipient states. Partnerships such as the PRSP help to produce self disciplined citizens and s responsible agents in their own development ( Abrahamsen, 2004, p. 1464). According to WHO (2013), c ountries that use both SWAps and P R S Ps need to work toward consistency between what is presented in both plans, including arrangements for stakeholder buy in, and plans for implementation, monitoring and improved accountability. Countries with PRSPs may foster greater interest in developing

PAGE 53

40 a SWAp, and having a SWAp can become a mechanism for work towards implementation of the PRS P ( WHO, 2013) The r eview of literature on this topic highlights both the shortcomings of traditional approach es to international development aid, stemming largely from information, incent ive and ownership problems, as well as the theorized potential of new approaches designed to address these problems, such as SWAps and sector specific sections of PRSPs. These new approaches bring with them the expectation of great er country owner ship of sector projects, more predictability and efficiency in aid allocation, and greater potential to monitor actions undertaken by don ors and aid recipient countries. As these approaches are relatively new to the sector, m y second resear ch questions and hypothesis are designed to test if they are realizing the theorized benefits in the realm of urban water and sanitation service provision: RQ2. What is the role of sector progress on providing urban water and sanitation access? H2. A high level of sector wide strategy and investment coordination in water and sanitation services is associated with greater country progress on providing urban water and sanitation access. To study the impact in greater detail, specific constructs were drawn from the literature as follows: a) The recipient country identifies priority projects and develops a realistic investment plan for the sector b) Development of partnering agreements between donors and recipient countries is driven by the recipient country c) A single agency is charged with coordinating all donor aid received in conjunction with a sector strategy

PAGE 54

41 d) Sector wide strategies are prepared in collaboration with stakeholders and donors, who are provided with information about the sector e) A system of monitoring is in place for compliance with the terms of aid partnering agreements Civil Society Engagement In this section, I review l iterature in the area of civil society engagement to see how it informs a third research question related to water and sanita tion governance This question looks speci fically governance environment supports civil society engagement in issues surrounding urban water and sanitation service provision, and how this impacts progress in increasing water and sanitation service coverage in urban areas The concept of civil society engagement is defined for this research as: the existence of opportunities for the public and consumers to become involved in government decision making, including program planning and implementation (adapted from deLeon and deLeon, 2002). I consider civi l society engagement to include i nteractions between government and formal c ivil so ciety organizations as well as between government and the public at large as consumers of water and sanitation services. Therefore, the terms civil society engagement and public participation are both used in this discussion. The Organization for Economi c Cooperation and Development (OECD) defines voluntarily organizes itself and which represent a wide range of interests and ties. These

PAGE 55

42 can include community based o and non government organizations (NGO) (OECD, 2006). 2 Engagement with the beneficiaries of public services, often throu gh organized CSO s, is thought by many to enhance the effectiveness of government. T he new public a dministration movement in the US during the 1960s called for more public involvement in public management, coinciding with the Great Society initiatives that called for grea ter social equity in government programs, particularly for previously marginalized groups (Thomas 1 995) DeLeon and deL eon (2002) advocated for increased opportunities for citizen participation to promote democracy not only in policy development, but also in policy imple mentation through participation in public management decision making. Similarly, p olitical choice and citizen participation in the process of policy definition and formation is emphasized by t he field of discursive politics so that analytical process es may include a wide range of interests, arguments, and discourses. D iscu rsive approaches challenge the more traditional and do minant positivist, technocratic approaches to the social sciences and place emphasis on social meaning and values, gained through interactions between citizens and experts This approach is touted as a way of democratizing policy deliberations (Fischer 2003) Several authors studying water and sanitation service provision have advocated for more engagement wi th civil society. Saleth and Dinar (2004) recommend that decision processes for water accommodate the increasing role of user groups, non 2 NGOs are typically issue based, legally registered, not for profit, and have some paid staff. CBOs are community based, voluntary, membership organizations made up of individuals who have joined togethe r to further a common interest ( WaterAid & Freshwater Action Network, 2011 ).

PAGE 56

43 governmental organizations, women, environmental and self help groups. Others make the case th principles of social and environmental justice (Wijaya, 2005). Gillespie (20 05) found that limited opportunities for the public and consumers to become involved in decision making about the sector can also impede sector finance In the context of water and sanitation services an example of incorporating citizen input w ould be consultations about technology choice that include input not only from technological exper ts, but also from the community about their lifestyle and social customs. Lynn (1994) said that what counts as a good solution depends a lot on the local context. Mirumachi & Van Wyk, ( 2010 ) recommend that regulatory framework s incorporate mechanisms for consultatio n with users, so that th ey may express preferences about service delivery. Local consultation mechanisms increase the likelihood that the desires of water users are reflected in water management. T he central government has a role in generating the enabling conditions tha t make this possible (Madrigal, Alpzar, & Schlter 2011). Two global water NGO s propose that civil society engagement citizens with knowledge, skills and confidence to under take the advocacy for themselves and eventually by themselves (WaterAid & Freshwater Action Network, 2011, p.9 ). The effects of citizen p articipation depend on whether the approach is based on understanding, not just on short term incentives or pressure, and whether the public can participate in articulating problem causes, which may then place them in a better position to identify and adopt potential solutions (Oakley, 1991).

PAGE 57

44 Civil society engagement also has the pote ntial to improve redistributional equity in the provision of public services. Greater participation can be a check against the often prevailing institutional culture, power relationships, and social constructions that favor traditionally advantaged groups as the primary beneficiaries of government policy (Schneider & Ingram 1993, 1997), a situation perh aps even more pronounced in countries where wide gaps of wealth exist among citizens. A governance environment tha t favors greater participation c ould pro mote increased access to water and sanitation services by all classes of citizens, not just traditionally advantaged groups. In an urban setting, p ol icies that promote broad based participation and that define strategies for actual involvement can be useful ingredients for sustainable service delivery The interests of poorer populations, particularly in urban areas, c an be better represented through open, inclusive, c oherent, and equitable decision ma king processes. T he poor can also gain more power and influence through governan ce systems that facilitate direct participation in service provision, to include planning, installation, management and monitoring of service s ( Sohail Cavill, & Cotton 2005; McGranahan & Satterthwaite 2006). Oakley (1991) said that public participation should be appreciated not just for idealistic and egalitarian purposes, but also as an important element to increase project efficiency and effectiveness, and to encourage self reliance. In a study of US metropolitan governance, Oakerson and Parks (1999) suggest that interaction between citizens and government officials can explain increased efficiencies in government services, as it prompts officials to devise productive organizational arrangements to meet the needs of local citizens.

PAGE 58

45 Public participation leads to increased accountability in government. O ne of the ways that institutional rational choice theory evaluate s institutional outcomes is based on the idea that in a democratic society, public leaders should be accountable to citizen r egarding the use of public funds and natural resources (E. Ostrom, 2005) This requires that information abou t the preferences of citizens be made available to decision makers, which can be achieved through various forms of civil society engagement Mana gement transparency and accountability are critical to gain and maintain the trust of users and investors in the water sector including clear ly defined roles and responsibilities, indepe ndent monitoring, information disclosure and consultation s with cons umers (Mugabi & Njiru, 2006). In rural communities in Costa Rica, one of the most important mechanisms linked to high performance in water provision was properly defined local accountability for water s ervices, as well as the ability of local leaders to g enera te incentives for community involvement in sustaina ble solutions (Madrigal et al., 2011). The 2004 World Development Report (World Bank, 2004) argues that for government services to be effective, clients must be able to monitor and put pressure on p roviders when needed. Hooper (2011) found that stakeholder engagement processes, and accountability programs to monitor implementation actions, were associated with effectiv e river basin management based on a global assessment of integrated water resources management systems Some of the decentralization st udies discussed previously have also found an inherent value of civil society engagement for improving public services ( Tropp 2007; Iribarnega ray & Seghezzo 2012 ; Azfar et al. 2004; Wolman 1990; Mugabi & Njiru 2006)

PAGE 59

46 One avenue for public participation is h aving a feedback and complaint mechanism in place for consumer input on water and sanita tion services The importance of such mechanism s in low income countries is emphasized by Li particularly when the government has privatized water service provision (Li 2011). For the n ew transfer operate t ran s fer model of urban water supply projects in China, a form of public private partnership, researchers found that a critical success factor based on four case studies was government supervision of the contractor an d a mechanism by which users could register com plaints about the contract or with the local government. For proven complaint s, the local government then had the authority to order the contractor to rectify and improve the situation ( Meng Zhao, & Shen 2011) A key success factor for the Phnom Penh Water Supply Au thority in Cambodia, a country challenged in governance capacity due to an unstable political history, is that it forged social contracts with vi llage associations for reporting thefts and leakages in the system. Customer co mplaints and fe edback were received through established channels, respon ded to, and served to monitor problems within the system (Araral 2008) While the above studies found a positive association between public participation and impro ved government serv ices promoting participation principles in countries that ha ve not previously adopted them poses unique challenges. Some general constraints have been identified by Research Triangle Institute (2004): poor public education systems; lack of local, experienced non governmental organizations; and resistance from local government administrators and government officials to opening up the decision making process

PAGE 60

47 Even when public participation policies are in place, a gap may exist between policy and ho w it is practiced in reality, often due to the political and cultural environment. T raditional actors may play a predominant role in decision making and new actors, suc h as civil society, may be present but have limitations in t heir influence on decision making. In post apartheid South Africa during the late 1990s, stakeholders varied widely in their ability to understand and adopt gove rnment processes they were not familiar with, and there was not a strong culture of participation i n government (Funke et al., 2007). While refor med water policies post apartheid placed an emphasis on inclusiveness and cooperative water governance, po wer disparities at the local level often impede d the way that inclusiveness worked in practice comp lic ated by the history of non inclusiveness (M irumachi & Van Wyk 2010) Ot her considerations limit more inclusive water governance practices. In many countries, inclusion of new actors in local level discussions is viewed as needing legitimizati on from the central government. W hile new participants may be present in water discussions, they do not necessarily hold enough political power to influence decision making, and may perceive risks of low return on time and effort invested (M irumachi & Van Wyk 2010). A study looking at health and education decentralization in Uganda and the Philippines found that citizen influence at the local level was constrained by limited access to information on the responsibilities and performance of local governmen t (Azfar et al., 2001). Tevelow (2004) looked at the impact o f NGOs and other CSOs in developing countries, and found that while there are expanding networks amo ng NGO s in different countries, aided by Internet communication, there often remain institutional barriers to

PAGE 61

48 their full participation in closed door negotiations among governments and powerful inter governmental institutions, such as the World Bank and IMF E m powerment may be a social process that evolves over a period of time, and for true forms of cooperation to develop and result in m eaningful action, a long period of social capital formulation may be required (Innes & Booher 1999). Even in the case of larg e scale donor programs, t he level of civil society influence can vary. An independent study looking at two new forms of multi stak eholder partnerships in Africa (the UK led Pa rtners for Water and Sanitation and the European Union Water Initiative), both o f which have a strong focus on working towards MDG achievement, found that governments remain the primary program partners and political pressure remains the primary driver of activities (Stewart & Gray, 2006). For the European Union program, governments were given a privileged position as partners, relative to civil society. Further, the study found that the poorest citizens in developing countries are often the most limited in terms of time available to commit to partnership processes, due to activities required for their daily sustenance (Stewart & Gray, 2006). For the most part the l iterature on civil society engagement and public participation points to several potential benefits of these activities such as promoting government transparency and accountability improving redistributional equity better representing the interests of the poor, and improving the quality of government services at various stages of the policy development and implementation process To test these if these findings hold true in the context of urban water and sanitation service provision, I developed the third r esearch question and hypothesis as follows:

PAGE 62

49 urban water and sanitation acce ss? H3. A high level of civil society engagement in water and sanitation services is associated with greater country progress on providing urban water and sanitation access. The following constructs for civil society engagement are designed to focus in on how civil society engagement may be havin g a positive impact in different s tages of policy development and implementation : a) Opportunities exist for consumers and civil society groups to become involved in problem articulation. b) Opportunities exist for consumers and civil society groups to become involved in decision making about policy development and planning. c) Opportunities exist for consumers and civil society groups to become involved in decision making about implementation and management. d) Secto r operates with transparency and information disclosure including for use of funds and natural resources. e) There is a feedback/complaint system in place. In addition, the case studies will examine considerations specific to developing countries that appeared in the literature, such as how well civil society engagement is supported by political pract ices and cultural norms, and possible limitations posed by education and time constraints among the population. C ountry Background Conditio ns Because country background conditions differ greatly from country to country, several control variables are included in the statistic al analyses to account for the influence of these conditions on country progress T his section of the literature review provides additional backgro und information for inclusion of these variables.

PAGE 63

50 Urban Migration Over the last several decade s, the rapid global pace of migration from rural to urban areas has placed additional challenges on urban planners charged with provision of water and sanitation services. Most of the infrastructure in established cities is dated and rapid population growth in cities is pl acing additional strain on system capacitie s for both water and sanitation (Davison, 2001; O ECD 2003). Also, as city boundaries continue to expand into previously unpopulated areas, infrastructure in these areas may be lacking altogether. Rapid populati on growth in cities often coincides with risin g unemployment, as well as strained or decrease d municipal budgets (Davison, 2001), all of which create a challenging environment for extension of water and sanitation services. In many countries, growth in se rvices has been inadequate to keep pace with the population growth, and large groups of people have remained without access to the minimum of 25 liters per person per day of potable water that is recomm ended by the WHO (WHO, 1992). In the early 2000s, a s hift in global attention occurred from focusing on increasing supply in rural areas to the unique problems created by lack of access in poor urban areas (UNCHS, 2003, 2006; UNDP, 2006; UNWWAP, 2003, 2006; WHO, 1992). Recent migrants to urban and peri urban areas in low income countries typically reside in poor, informal areas, which are often the lowest priority for water and sanitation service provision (Evans, 2007). Ineffective governance systems can create disincentives for utilities to connect poor ho useholds and for poor households to connect to services (Bakker, Kooy, Shofiani, & Martijn, 2008).

PAGE 64

51 Corruption Corruption was among the six most important dimensions of g overnance included in the Worldw ide Governance Indicators, based on a study of more than 200 countries during the period from 1996 to 2005 (Kaufmann, Kraay & Mastruzzi, 2006). Corruption has been defined by Bardhan (1997) as the use of public office for private gains, and by Shleifer and Vishny (1993) as the sale by government officials of government property for personal gain. Corruption can lead to in efficient use of public funds for financing public services, such as water and sanitation (Klitgaar d, 1990). The Water Integrity Network, a group of six internationa l NGOs working to fight corruption in water management, found that about 25 30% of state budgets on water investments are lost due to corr uption (Deen, 2006). The UN cited corruption as a primary impediment to clean water access in the 2006 World Water Developm ent Report, and it has been estimated that corruption undermines efficiency in the water and sanitation sector by 20 40% (Watkins, 2006). Corruption significantly reduces a l of country income (Anbarci, Escaleras, & Register, 2009). Corruption can further social inequality within a country, working against the MDG Target 7.C objectiv e of increasing access for uns erved groups. Dudley (2000) estimated that corruption diverts up to 30% of the billions of dollars spent each year for international development loans. Misallocation of resources resulting from corruption can exacerbate a situation of unequal wealth distribution where the poor are socially excluded from goods and s ervices (Riley, 2004; Gupta, Davoodi, & Alonso Therme, 2002; Transparency International, 2008). Corruption can lead to expensive, capital

PAGE 65

52 intensive projects being favored by government officials, due to the possibility to receive bribes or skim funds. Th ese types of projects are more likely to serve high income areas, rather than a ddress needs in low income neighborhoods An environment of corruption can discourage investment in the water and sanitation sector and result in underperforming and ineffecti ve services (Mauro 1995; Habib & Zurawicki, 2002). Estache and Kouassi (2002) studied 21 water companies in Africa and found that nearly two thirds of their operating costs were due to corruption. It has been estimated that corruption can raise the price of a household water connection by up to 30%, with opportunities for corruption at every point in the water delivery chain (Transparency International, 2008). C orruption in the water and sanitation sector exists beyond the national government level, whi ch is the focus of this research. Transparency International (2008) summarized recent developments on the topic of water and corruption and points to corruption during transactions among cus tomers, contractors, and donors. This is complicated by t he natu re of water and sanitation projects which can involve multiple levels of government, along with suppliers, contractors, and consultants. An overview paper on this topic found approximately 50 different examples of how lack of transparency, dishonesty or corruption can manifest itself in the water sector involving not only government officials, but these other actors as well ( Shordt, Stravato, Dietvorst,

PAGE 66

53 Political and Security Effectiveness Fulfilling many of the ope rating principles iden tified for good water governance -transparency, accountability, efficiency, and inclusiveness -require s effective political systems. As stated by the G lobal W ater P artnership: T he concept of governance of course encompasses laws, regul ations and institutions but it also relates to government policies and actions, to domestic activities, and to networks of influence, including international market forces, the private sector and civil society. These in turn are affected by the political systems within which they function ( GWP Technical Committee 2003, p. 4). MDG achievement ultimately depends on not just water institutions within a country, but also general institutions governing the economic political, and social spheres, because a ctual implementation of reforms can be curtailed due to political and resource constraints (Saleth & Dinar, 2005). Another country condition that can impede effective resource m anage ment and public service provision is the general s ecurity environment wh ich can impede the ability to mobilize financial and physical resources for development projects such as infrastructure construction. As stated in the 2011 Global Report of the Center for included when examining the potential for development (Marshall & Cole, 2011 p.3). Literacy The co mplexities inherent in providing water and sanitation services to urba n dwellers require both managerial and technical capacity at the responsible level of government ( Marcus & Onjala 2008; Funke et al, 2007; Muga bi & Njiru, 2006; Seppala, 2002 ). Skills needed include engineering and other technical specialties, system management and mainten ance, and financial management. The general literacy level of

PAGE 67

54 the population can be an indicator of the pool of qualified individuals in a country who can support wat er and sanitation service extension. It can also show the ability of the general public a s water and sanitation consumers to grasp information about technology options, sys tem operation, and usage costs, so they can provide input into government and service provider decision making (Research Triangle Institute, 2004) GDP per Capita gross domestic product ( GDP ) is one indication of its ability to apply funds towards water and sanitation infrastructure investment. Studies of multiple countries have found positive associations between overall cou ntry GDP as well as GDP per capita and water access among the population. Fogden (2009) found a positive average GDP per capita growth rate and the percent of the population with access to safe water over a 35 year period. Onda LoBuglio, and Bartram (2012) found GDP to be one of several positive factors contributing to water access, even after adjusting published access rates downward to account for issues of water safety. Renewable Water Resources When considering the potential for expanded urban water service the availability of water resources within a country is a relevant factor. Viewing it on a per capita basis provides a better indication of the extent to which it can meet the needs of the d omestic population. T he UN Food and Agriculture Organization collects water resources information by country and publishes it through the AQUASTAT database (FAO, 2014) Renewable water resources, inland waters renewed by the global water cycle, are the main source of available water for human use However, only part of these waters are

PAGE 68

55 able to be accessed and used effectively because of limitations such as physical conditions of accessibility, the potential for floods caused by excess flows, and inconsi stenc y in water flows which can make them difficult to capture and utilize (FAO, 2015) While these limitations can vary by country, measures of renewable water resources per capita do provide some indication of the relative availability of water for huma n use. Foreign Aid to Water and Sanitation Many countries in the sample group, particularly those in the low income category, are the recipients of foreign aid designated for the sector which, among other things, goes towards the construction of infrastructure and other systems for delivery of wate r and sanitation services. While several studies have looked at the impact of foreign aid in general, ( including those discussed above ) a 2010 study ( Botting et al, 2010) looked specifically at the impact of foreign aid per capita for water and sanitati on, by relating it to change s in access to improved water and sanitation services from 2000 to 2006. The study found that countries in the highest tertile of aid per capita recipients had odds ratios of 4 to 18 times more likely than those countries in th e lowest tertile of aid per capita recipients for achieving greater gains in access to improved water supply. T he increased odds for sanitation access were modest and largely insignificant. Summary of Literature This chapter presented key literature related to the overall objective of my research. First was the theory th at water and sanitation service exhibits the characteristics of a public good, a type of good best suited for government provision due to supply inefficiencies and distributional ineq uity likely to occur with private sector provision.

PAGE 69

56 Further, access to water has been recognize d by the UN as basic human right and it can be considered as a merit good, one that society has an interest in promoting to avoid negative externalities, such a s health problems, associated with poor water supply and sanitation. These arguments form the basis for government and international donor participation in water a nd sanitation service provision. Also discussed in this chapter was literature about t he importance of effective water institutions and governance practices for achieving water and sanitation access goals. They are viewed by many authors as crit ical preconditions f or achieving and sustaining quality water and sanitation services. Institut ions determine decision and incentive structures within governments, encompassing the rules for : access to and allocation of resources; placement of economic incentives; and systems of service design and delivery The broader concept of governance also en compasses the values and processes by which government authority is exercised. N ew forms of governance, which engage governments, non government organizations, the private sector, academic and research institutions, are viewed as a means of improving government service outcomes. This research will explore the extent to which three specific institutional and governance attributes are having a positive impact in the context of urban water and sanitation services. Litera ture on the first attribute of decentralization says it can be beneficial for government service provision because lower levels of government can be more effective in responding to local conditions and meeting the needs of local consumers But, there are also several articles noting factors that are important f o r rea lizing theorized decentralization benefits, such as degree of decision making and revenue raising

PAGE 70

57 delegation, the political and socio cultural environment, and the ability of local governments to adapt to new responsibilities. The research question for this attribute is designed to study the extent to which decentralization is having a positive impact on urban water and sanitation provision, while also looking at whether necessary factors for success are pre sent in decentralized countries: RQ1. What is the role of decentralization providing ur ban wat er and sanitation access ? Other literature reviewed was about international development aid and g overnance p ractices that engage donors and other stakeholders for sector wide strategy a nd investment coordination, the second attribute being studied. This highlighted the shortcomings of traditional approaches to international development aid, stemming largely from information, incentive and ownership problems between donors and aid recipient countries. The literature also explored the potent ial of new approaches designed to address these problems, such as SWAps and sector specific sections of PRSPs. These new approaches bring with them the expectation of greater country ownership of sector projects and more predictability and efficiency in a id allocation, ultimately improving sector outcomes. The second research question is designed to study the extent to which sector wide strategy and investment coordination is having a positive impact on urban water and sanitation service provision: RQ2. W hat is the role of sector wide strategy and access? The third topical a rea of literature covered was for the third attribute of civil society engagement. Overall, t he literature on civil society engagement points to several

PAGE 71

58 potential benefits of these activities, such as promoting government transparency and accountability, improving redistributional equity, better representing the interests of the poor, and improvin g the quality of government services at various stages of the policy development and implementation process. To test these if these findings hold true in the context of urban water and sanitation service provision particularly in the socio economic envir onment of low and middle income countries the following res earch question was developed: RQ3. What is the role of civil society engagement progress on providing urban wat er and sanitation access ? T he three research questions, hypotheses and constructs developed from the review of literature in this chapter will be examined starting with the next chapter on quantitative methods, leading into the statistical analyses and the case studies.

PAGE 72

59 CHAPTER IV QUANTITATIVE METHODS To test the three hypotheses, bivariate and multivariate statistical tests were conducted using data from multiple countries This chapter present s information about sample selection sample characteristics variable measurement and operationalization, an d also discuss data validity and limitations Sample Selection An initial sample of 75 countries was derived from those countries who responded to the 2011 GLAAS survey 3 This data source was identifi ed as the most comprehensive option for quantitative analysis, both in terms of number of country respondents and its depth of information relevant to the operationalized constructs of the research questions The 75 respondent countries make up 54% of all 139 countries class ified by the World Bank as l ow or m iddle i nc ome (LMI) countries a group that is most relevant for MDG Target 7.C (World Bank website 2015 ) A list of the 75 cou ntries is included in Appendix B along with notations about missing and incomplete da ta, which are int roduced later in this chapter. Three countries (Haiti India, and Equatorial Guinea) had insufficient data for inclusion in any of the multivariate statistical tests, while other countries had to be excluded only from certain tests. As a result, the actual perce ntage of LMI countries analyzed for specific multivariate tests ranged from 47 % to 49% To evaluate the representativeness of the sample group for generalizing results t o all LMI countries, various country characteristics were compared between the 72 country 3 The reader will re call that GLAAS is a UN Water data collection initiative implemented by the World Health Organization (WHO)

PAGE 73

60 sample used or multivariate tests (sample group) and all 139 LMI countries A verage country population was compared, first against the entire 139 country LMI group, and then again against the LMI group after removing the two most populous countries (Chi na and India), as their population s exceeds by five times or more the next most populous country, Indonesia. Neither China nor India were part of the 72 country sample group. As shown in T able 4.1 the sample group average country population (33.2 million), is about 20% lower than the average country population for all 139 LMI countries. However, w hen China and India are removed from the LMI group the sample group average country population is now about 50% higher than the average population for the 137 remaining LMI countries. Looking at median values, the sample group median is about 50% higher than the median for either LMI comparison group. Studying this comparison, one can say that the sample group is generally skewed towa rds high population countries within the LMI group Table 4 .1 Average Country Population : Comparison of Sample Group and LMI Group N Average Country Population ( millions) Median Country Population ( millions) Sample Group 72 4 33.2 13.7 All Low and Middle Income countries 139 41.3 9.5 All Low and Middle Income countries except for China and India 137 23 .0 9.3 Source: World Bank, 2012b 4 Three countries, ( Haiti, India, and Equatorial Guinea ), were not used in any multivariate tests so are not included.

PAGE 74

61 G DP per capita comparisons were also done. 5 A comparisons was first done b etween the sample group and all LMI countries with available data and then also against low i ncome and l ower middle i ncome countries only 6 (T he World Bank further divides the middle i ncome country classification into l ower middle i ncome and h igher m id dle i ncome groups.) As show in Table 4 .2, the sample group average GDP per capita falls just about mid way between these two comparison groups, whereas the sample group median is closer to the median for the smaller comparison group. Looking particularly at the median comparison, this shows that the sample group is more heavily weighted towards low income coun tries within the LMI group. Table 4 .2 GDP per C apita : Comparison of Sample Group a nd LMI Group N Average GDP p er Capita ( current US$) Median GDP per Capita (current US$) Sample Group 72 2 2 867 1 251 All Low and Middle Income countries 135 4 051 3 256 Al l Low and only Lower Middle I ncome countries 80 1 710 1 354 Source: World Bank, 2012c 5 Some country GDP per capita values may vary between those used for the sample dataset and those summed for th e total GDP per capita for low and middle i ncome countries, as these were downloaded at different points in time from WorldBank.org. Any variations are thought to be nominal with little bearing on the general comparison presented here. Income group classifi cations are from World Bank website, 2015. 6 The 72 country sample consists of 28 Low Income Countries, 29 Lower Middle Income Countries, 14 Upper Middle Income countries, and 1 High Income Country (Oman), based on World Bank classifications.

PAGE 75

62 A third set of comparison s was done for indicators of country progress on urban water and sanitation access, specifically the urban population percent with access in the years 2000 and 2012. This exercise was done to see whether or not the sample group, those who responded to the GLAAS survey, was skewed towards low or high access countries. The results appear in Table 4 .3. Table 4 .3 Urban Water and Sanitation Access : Comparison of Sample Group and LMI Group Averages Water Access (%) Sanitation Access (%) Year 2000 2012 2000 2012 Sample Group used in multi variate analyses (N = 66) 85.6 90.9 58.4 63.2 All Low and M idd le I ncome countries (N ranges from 130 to 137) 89.6 93.1 67.6 71.5 Source: WHO & UNICEF, 2000 2015 The comparison shows that for water, the s ample group had marginally lower access rates in 2000 and 2012 than did LMI countries for which data were available. For sanitation the sample group also had lower access rates in both years, with a wider margin of difference. This reveals that th e sample group is more heavily weigh ted towards countries with low access rates. In summary, the comparisons show that sample group is generally higher in population, lower in GDP per capita, and lower in water and sanit ation access rates, than the overall LMI country population Arguably, t hese characteristics make the sam ple group that much more germane to the intention of MDG Target 7.C as these countries have greater needs and more limited domestic economic base.

PAGE 76

63 In the fol lowing sections, I introduce and explain the outcome variables, key explanatory var iables, and control variables for the quantitative analyses Table 4. 4 is an overview of all variables D etailed information about vari a ble sources and operational ization is found in Appendix C Table 4.4 Overview of All Variables Outcome Variables Key Explanatory Variables Control Variables 1. Water Access in 2012 2. Water 12 Year Progress 1. Water Decentralization 2. Water Sector Planning 3. Water Civil Society Engagement 1. Urban Population Change 2. Corruption Level 3. Political Fragility 4. Security Fragility 5. Literacy Rate 6. GDP per capita 7. Renewable Water Resources 8. Foreign Aid to Water and Sanitation 1 3. Sanitation Access in 2012 4. Sanitation 12 Year Progress 4. Sanitation Decentralization 5. Sanitation Sector Planning 6. Sani tation Civil Society Engagement 1 Foreign Aid to Water and Sanitation was eventually dropped from the analysis for reasons explained later in this chapter. Outcome Variables I conceived four outcome variables t o represent country progress in extending water and sanitation service to unserved residents in urban areas: two for water service and the same two for sanitation service. The two variables are defined as follows: 1) Access in 2012 ( A 2012 ): percent of t he urban population with access to service in 2012 2) 12 year progress ( Progress ) : the percent of the urban population that gained access between 2000 and 2012, relative to the percent of the urban population that lacked service in 2000. ( An equation for this is presented later in this section). O perational measures for these outcome variables were derived from the JMP dataset introduced in Chapter II the official UN mechanism tasked with monitoring

PAGE 77

64 country progress towards MDG Target 7.C. Data were available from JMP for all countries in the sample. JMP produces a dataset annually that tracks the percentage of the population lacking access to an improved source of water and an improved source of basic sanitation, broken down by rural and urban ar eas (WHO & UNICEF, 2000 2015 ). Table 4.5 lists all system types that qualify as improved sources by the JMP program. Th ese criteria have remained consistent throughout the 12 year period from 2000 to 2012. Table 4.5 Systems Considered as an Impr oved Source by JMP Water Sanitation 1 Piped water into dwelling Piped water into yard/plot Public tap/standpipe Tubewell/borehole Protected dug well Protected spring Rainwater collection Flush toilet Piped sewer system Septic tank Flush/pour flush to pit latrine Ventilated improved pit latrine Pit latrine with slab Composting toilet 1 O nly facilities which are not shared or are not public are considered improved Source: WHO & UNICEF, 2006 p.4 I chose to equate my research questions and measured in my outcome variables with JMP definitions for improved sources. In other words, I considered those who had service access in a country to be the same as those who had improved sources as defined by the JMP criteria. In keeping with the research parameters, I used only JMP data for urban areas. JMP derives its estim ates of the population lacking access to improved sources by drawing on data collected by national statistics offices and international survey programs t hat use nationally representative household surveys and national census data (UNICEF, undated) Most of the data come from Demographic Health Surveys, Multiple Indicator

PAGE 78

65 Cluster Surveys, Living Standards Measurement Studies, a nd World Health Surveys ( WHO & UNICEF, 2012). As possible data are obtained directly from water and sanitation service users, rather than service providers, and therefore are not impacted by different data collection methods, definitions of access, or th e political environment sur rounding data collection in a particular country (WHO & UNICEF, 2006). Po pulation data used by JMP when calculating access percentages are obtained from the UN Population Division (WHO & UNICEF, 2006). To derive the first out come variable, Access in 2012 ( ), I started with JMP count ry specific data for the percentage of the urban population lacking access to an improved source in 2012 and then transposed it to get the percentage of the urban population with access to an improved source in 2012. For the second o utcome variable, 12 year progress ( Progress ) Formula 4 .1 shows the calculations used to derive the variable, along with two example s. I started with and subtracted (access percentage in the year 2000) to measure the 12 year change in percent of the urban population with access to an improved source ( labeled as C for change ). 7 In most cases the 2012 value is greater than the 2000 value, resulting in a positive figure, however there are also cases where the difference value is negative. I then divide d the 12 year change figure, positive or negative, by the percentage of the po pulation that lacked access in the year 2000 ( labeled as G for goal ), to derive the extent of progress that was made over a 12 year period in reducing the percentage of the 7 The year 2000 was chosen, rather than the MDG base year of 1990, because JMP data were incomplete for many of the sample countries for 1990 and subsequent years.

PAGE 79

66 population without access. 8 The result constitutes the 12 year progress variable ( Progress ), which is the proportion of those lacking access in 2000 (original goal for increasing service) that gained access by 2012. Progress can range from a negative value (for a decline) to a maximum of 1.0 if the entire population lacking access in 2 000 gained access over the 12 year period. Formula 4 .1 Progress Outcome Variable Calculation = P ercent of the urban population with access in 2000 = P ercent of the urban population with access in 2012 Percent of the urban population lacking access in 2000, considered the initial goal for increasing access Change in the percent of the urban population with access over the 12 year study period Example 1: Central Africa Repu blic Sanitation (improvement in overall access) = 31.9% = 68.1% = 43.6% = 43.6% 31.9% = 11.7% = 11.7% = 0.17 68.1% Example 2: Dominican Republic Water (decline in overall access ) = 90.7% = 9.3% = 82.5% = 82.5% 90.7% = 8.2% = 8.2% = 0.88 9.3% 8 Other outcome variables were considered, su ch as the 12 year percent change without comparing to t he starting point but no meaningful results were produced. = 100% = 100%

PAGE 80

67 Individual country v alues calculated for and Progress were used to assign countries to low, medium, and high achievement groups that were used for the statistical analyses The rationale for grouping countries into achievement group s rather than maintaining individual country values is explained later in this chapter. Key Explanatory Variables The institutional and governance attributes represented by the key explanatory variables each linked to a research question, are: (a) degree of decentralization (decentralization) ; (b) degree of sector wide strategy and investment coordination (sector planning); and (c) degree of civil society engagement (civil society ). Each was measured for both water and sanitation, and corresponding variables are written with the notations W and S respectively, before the shortened name, (for example, W civil so ciety). This results in six key explanatory variables. Refer to App endix A for operational definitions, construct level definitio ns, literature references and source material for the key explanatory variables. To derive individual country observations for the six key explanatory variables, I used select country level data fr om the GLAAS 2011 survey, which addressed overall policy, governance, and institutional environment for water and sanitation services ( WHO, 2011). A large portion of the GLAAS 2011 survey covered the nature and impact of government policies and institutions More information about GLAAS questionnaire development, data collection, and data validation can be found i n the 2012 GLAAS report (WHO & UN Water, 2012 Appendix A, p. 71 73 ). Looking at temporal correspondence between the key explanatory variables and the outcome variables, the use of 2011 data for the key explanatory variables corresponds

PAGE 81

68 closely with the A 2012 outcome variables measuring access in 2012 However, s ince the Progress outcome variables reflect a 12 year period starting in 2000, corresponding explanatory variable s would ideally measure the same 12 year period. This was explored, however, sufficient data from GLASS were not available for years prior to 2011 9 an d no other comparable data source was identi fied. Therefore, the key explanatory variables derived from 2011 data were used for all analyses. The last column of Appendix A shows the specific 2011 GLAA S survey questions /responses selected for inclusion due to their applicability to the construct definitions listed in the middle column of Appendix A 10 Not all constructs have corresponding GLAAS survey questions. Those constructs not measured by t he GLAAS survey responses were evaluated t hrough document rev iew and interviews during the case studies. A coding system is used in the last column of Appendix A to show which information came from the GLAAS survey, and which from documents and interviews. In the next paragraph, I explain in detail how th e GLAAS survey response data were adapted for use. For each GLAAS survey question, usually asked separately for water and sanitation, country respondents provided one of three numbers representing the level of development for particular programs or policies: lit tle indicated by 0 some indicated by 0.5 a great deal indicated by 1.0 9 Previous GLAAS surveys were conducted in 2008 and 2009, h owever, the number of country respondents was about half that of 2012. 10 A few applicable questions were not included since almost a ll countries responded with the same answer and it added nothing to the analysis.

PAGE 82

69 umerical responses (0.0, 0.5, or 1.0) for several applicable questions to arrive at composite value For example, to measure civil society engagement in the sanitation sub sector, responses to the two ( G ) coded civil society probes in the third column of Appendix A, (when asked specifically about sanitation), were averaged to arrive at a single composite value for S civil society level. Averaging multiple survey responses allows for broader input and helped to mitigate against an inaccurate or misguided response to any one particular survey question. Consistent with the response scale for ind ividual questions, co mposite va lues range from 0.0 to 1.0 composite value s were then used as observations for the three key explanatory variables. In a few cases where a respondent country did not answer one or more of the question s used in calculating a composite value an average was calculated from only those question(s) with responses. This approach, which essentially applies the average value from responses received to any unanswered question(s), was deemed reasonable, as all questions are somewhat reflectiv e of the construct of interest. U sing a mean value is acceptable for dealing with missing data in some cases when it would not markedly impact the overall results (Tabachnick & Fidell, 2001, p.71). For a country that did n ot respond to any of the questions relevant to a p articular variable, the observation was considered as missin g Refer to Appendix B for notes on partial and missing data by country. Control Variables Eig ht control variables were select ed for inclusion in the regression analyses to reflect country background conditions that were thought to have potential influence on

PAGE 83

70 the outcome variables. A variety of reputable data sources were considered as potential sources for control variable data, with selection based on those that most closely approximated the relevant background condition and were available for a large number of sample countries. D ata used are from the most recent yea r available generally between 2011 and 2013. (R efer to Appendix C for detailed source information). For countries lacking observations for a particular control variable mean value was used as a substitute when deemed reasonable to avoid exclusion of that country from the overall analysis because of one missing variable This approach is considered acceptable in some s ituations when it would not markedly impact the overall results ( Tabachnick & Fidell, 2001, p.71 ) This maintained an overall larg er sample size and was judged to not have a major impact on the overall results Mean value subst itutions are noted in Appendix B and also explained below. Urban Population Change The pace of urban migration within a country can be seen as an important factor in its ability to increase the percentage of the ur ban population with access to water and sanitation, as service providers are challenged to keep pace with this changing situation. Reflecting this growth in a control variable required considering data over a 12 year period, unlike most of th e other contr ol variables that we re measur ed at one point in time. The formula for the variable is as follows: Formula 4 .2 Urban Population Change Calculation = urban population % in year 2012 = urban population % in year 2000

PAGE 84

71 The variable represent s the proportional growth over the 12 years, and not just the 12 year absolute change. So, a country that increased from a 60% urban population in 2000 to a 75% urban population in 2012, would have a relative change in urban po pulation of 0.25. Data were sourced from World Bank population data and were available for all countries in the sample (World Bank, 2000 2012 b ). Corruption Level No specific indicator of corruption in the water or sanitation sector could be identified for multiple countries so instead, the Corruption Perceptions Index generated by Transparency International for individual countries was used (Transparency International, 2013). 11 Corruption Perceptions Index values range from 0 (very corrupt) to 100 (ve ry clean). I transposed this scale so that a higher score represents a more corrupt country. The transposed values for the 75 countries in the sample ranged from 37 (relatively clean) to 92 (very corrupt). Three of th e countries were missing values (Fij i, Maldives, and Samoa), so the sample mean value (70) was used. Political Fragility and Security Fragility Two separate indicators from the Integrated Network for Societal Conflict Research (IINSCR), part of the Cente r for Systemic Peace, were adopted as variables to represent national political and security environments. These measu res were thought to capture the relevant background conditions and were widely available for multiple countries. 11 C ommonly used measures of corruption tend to be very highly correlated (Alesina & Weder, 2002; Treisman, 2000).

PAGE 85

72 The INSCR political effectiveness indicator reflects the politic al environment in a country by assessing regime and governance stability Scores range from 0 (effective) to 3 (highly ineffective). Two countries, Samoa and Timor Leste, were without data and t he mean score of 1.5 was used The INSCR security effectiveness indicator measures general security and vulnerability to political violence Also ranging from 0 to 3.0, the mean security effectiveness value of 0.7 was used for two co untries that were missing observations again Samoa and Timo r Leste. I modified the terminology from effectiveness indicators to fragility indicators when adopting these measures as variables to better reflect the meaning of the numerical scale, where a higher number represents a worse situation. Literacy Ra te in the analyses i s the percent of the population age 15 and over who can read and write. 12 This provides capacity to learn technological and management skills necessary to support service extension, and also the ability of the general population to provide input into policy and programs. Country specific data were sourced from the C entral Intelligence Agency (CIA) Wo rld Factbook and observations were available for all 75 countries in the sample (CIA, 2011). 12 Other education indicators were considered, such as percent of the adult population having completed seconda ry school, but dat a were missing for several countries. Literacy rates correlate with UNDP 2010 secondary education data at 0.69, and were available for more countries.

PAGE 86

73 GDP per C apita GDP per capita (in current US$) was used as a control variable to represent the domestic economic base This measure c an vary widely, even am ong the LMI countries in the sample group. Because of unacceptably high ske wness and kurtosis, the data were transformed to natural logarithm values, which then exhibited acceptable levels of skewness and kurtosis for inclusion in the regression models. Values were available for all 75 countries in the sample group. 13 The data were sourced from the World Bank Development Indicators 14 (World Bank, 2012c ). Renewable Water Resources While the focus of this study is on water supp ly and sanitation service delivery and not water availability, it was nonetheless felt that the availability of domestic water resources would be a relevant factor for countries considering invest ment in service extension, particularly for water supply. For this reason, data were sought to represent water availability by country. The indicator total actual renewable water resources (cubic meters per year, per capita), as reported by UN Food and Agriculture Organization (FAO) AQUASTAT was selected to be used as a control variab l e ( FAO, 2014). AQUASTAT is FAO's global water information system, developed by the Land and Water Division. T he renewable water r esources statistics provided in AQUASTA T are long term annual averages, typically 1961 1990 13 Other indicators considered, such as the GINI index of income distribution, was unavailable for several countries. 14 was instead sourced from UN data. https://data.un.org

PAGE 87

74 While this variable helps to account for vast differences in water availability among countries, ( ranging from more than 200,000 cubic meters per year, per capita to less than 1,000 ) it is recog nized that this indicator does not get at the issue of ease of water a ccess or proximity to urban areas. Nor does it incorporate agreements for water sharing across nat ional boundaries Attempts were made to locate such data, but it was not available for a sufficient number of countries to inclu d e in the cross ca se analysis. Due to unacceptably high skewness and kurtosis for this indicator, it was transformed to natural logarithm values for use as a control variable in the regression analyses. Observations were available for all 75 countries in the sample group 15 Foreign Aid to Water and Sanitation For all countries in the sample group, domestic finan cial resources for the sector are supplement ed by some level of donor aid, aid that can come from several different bi la teral and multi lateral sources. T he most comprehensive source of information identified on these transfers was a dataset produced by the OECD Development Assistance Committee (OECD, 2002 2012). Using this dataset, I first totaled the official development assistance disbursed to each country from 2002 to 2012 designated for the water and sanitation sector (in current $US). 16 ( Complete disbursements data were unavailable for years prior to 2002, so these years were not considered. ) The 2002 2012 cumulative aid figure was then divided by a country average population during the 10 year period (derived, for simplicity, as an 15 Samoa, which had a zero value, was replaced with a nominal value of 10. 16 The OECD Development Assistance Committee includes in their figures all aid provided by official agencies, including state and local governments, or by their executive agencies (OE CD, 2012).

PAGE 88

75 average of the 2002 and 2012 country population fig ures), to calculate foreign aid per capita from 2002 to 2012 The resulting country values exhibited unacceptably high skewness and kurtosis and were transformed to the natural log values to achieve acceptable levels for the regression analyses. Data were available for all 75 countries. Some weakness es of the OECD data are that it does not distinguish between aid to water and aid to sanitation and it incorporates funding for water programs outside the scope of this study such as water resources management, development, and protection. Looking forward, this type o f differentiated data may be available in coming years. New efforts are underway by UN Water through its TrackFin initiative to develop methodology to identify and uniquely track different types of financing to the water, sanitation, and hygiene sector s i n a coherent and consistent manner across several countries. Data Validity and Limitations This section discusses how well the variables represent the constructs of interest, the general vali dity of the data used, and potential limitations of the analyses. Concept Measure Validity Concept measure validity is an evaluatio n of whether the measures proposed are accurate representations o f the concepts of interest. The JMP, on which I based the outcome variables is the official UN mechanis m for monitoring progress towards MDG Target 7.C Although JMP data differ in some cases from figures published by individual national governments J MP has the advantage of using a consistently applied methodology for determining both water and sanitation access levels across multiple countries over several years. Because widely accepted status for tracking MDG

PAGE 89

76 Target 7.C country progress it was deemed the most appropriat e measure to use for the outcome variables given the scope and intended global audience for this research The wording of MDG Target 7.C, and the way that progress is measured is not without its critics. East erly (2009) questions measurement of indivi dual country development progress in numerical terms for initiatives like the MDGs, particularly when comparing Sub Sahara n Africa to other world regions. Talking particularly about MDG Target 7.C, he argues that the stated objective of reducing by half t he percentage of people without access to water and sanitation is structured and reported in a way that undervalues the advancements of Sub Saharan African countries: Africa was relatively falling behind on reducing the percent WITHOUT access to clean wa ter, but it would have been relatively catching up if it had been measured the conventional way of percent WITH access to clean water. The choice of WITH and WITHOUT for percentage change measures depends on assumptions on social welfare that have not been subjected to any serious scrutiny. If one just uses the usual indicator of percent WITH access to clean water, Africa is catching up to other regions (Easterly, 2009, p.33). somewhat addressed in this research by the choice to transpose the JMP data thereby defining the o utcome variables in terms population percent with access to water and sanitation. For the three country institutional and governance attributes, the GLAAS survey questions selected to develop the key expl anatory variables (see Appendix A, third column), are fairly direct and pointed towards the theoretical concept of interest. For exampl e, the survey questions concerning decentralization ask about both about the degree to which service delivery has been decentralized and the availability of funding at the local level. Similarly, civil society engagement survey questions ask specifically about citizen participation in program planning, budgeting, and implementation.

PAGE 90

77 Data Validity This section discusses potential data limitations and attempts to triangulate key explanatory variable data with other sources of information. A s previously mentioned, GLAAS data do not capture all the constructs of interest relating to the key explanatory variables, only thos e coded with (G) in column three of Appendix A. Therefore, the content validity for the quantitative analyses is less than complete and the remaining constructs, coded with (D/I) in Appendix A and summarized below, will only be evaluat ed throug h document review and interviews as part of the case studies Decentralization There are clearly defined rights, responsibilities, and roles for sector management actors at the n ational and sub national levels Sub national level authorities are monitored by communities and/or the central go vernment for actions such as public funds expenditures and equality in service provision Support is provided to build managerial and technical capacity of sub national l evels of government Civil Society Engagement Oppor tunities exist for consumers and civil socie ty groups to become involved at different stages of the p olicy and implementation process The sector operates with transparency and information disclosure As no case studies will be completed for sector wide strategy and investment c oordination furt her evaluation of constructs that were not represented in the quantitative data is limited, however they are noted here so they might be addressed in future data collection efforts:

PAGE 91

78 Sector wide Strategy and Inves tment Coordination: Development of partnering agreements between donors and recipient countries is driven by the recipient country A single agency is charged with coordinating all donor aid received in conjunction with a sector strategy A system of moni toring is in place for compliance with the terms of aid partnering agreements Data Triangulation For self reported data such as GLAAS, there may be a tendency for countries to report information that reflects positively on them but which may not accurately reflect the situation. The data undergoes some review by WHO after collection to guard against such misrepresentation. Further, to check the validity of information reported through GLAAS, I attempted to triangulate the key explanatory variabl es, (derived from GLAAS data), with other independent data sources on the same topic. For decentralization, an independent source of country ratings was identified and correlated with country decentralization values. I used a country decent ralization in dex derived by Ivanyna and Shah (2014) that looked at the following : 1) the security and existence of local government; 2) the relative importance of local government; 3) political decentralization, (for example. elected rather than appointed local officia ls); 4) fiscal decentralization, (for example, the ability to raise revenues); and 5) administrative decentralization, (for example, local government authority for personnel hiring and regul atory control). This index, particularly the last two item s, corr espond s well with the construct level definitions for decentraliz ation outlined in Appendix A The correlation between the Ivanyna and Shah decentralization index and the W decentralization values for 72 overlapping countries is 0.33 (p < 0.01 ). The correlation

PAGE 92

79 between the aggregate decentralization index and the S decentralization values for 71 overlappin g countries is 0.31 (p < 0.01 ). These correlation coefficients, while mo dest, do indicate some positive correlation at a high level of sign ificance, for both water and sanitation. T he difference could be explained by the fact that the comparison index considers enabling factors for decentralization in all aspects of government and p olitical representation, and not just government services Turning to the other two explanatory variables, no comparable rating system could be found for sector planning for a sufficient number of countries. However, comparable measures were identified for civil society engagement. The World Alliance for Citizen Participation (CIVICUS) rates countries using an Enabling Environment Index (EEI) that measures the potential of citizens to participate in civil society, based on the governance, socio cultural and socio economic environments in a country. 17 The 2013 ind ex includes 109 countries 40 of which are also in the sample group (World Alliance for Citizen Participation, 2013a). I measured the correlation between CIVICUS EEI data and the civil society values and found a significant correlation (p < 0.10) of 0.27 for the sanitation sector, but the correlation of 0.23 for the wa ter sector was not significant, showing no strong relationship in either case. Another civil society measure was identified covering only countries in sub Saharan Africa. The 2011 C SO Sust ainability Index developed by the US Agency for International Development ( USAID ) includes 20 countries in the sample group. In 17 The EEI measures conditions that affect the capacity of citizens (whether individually or collectively) to participate and engage in civil society. CIVICUS includes non organized forms of civil society as well as civil society organizations. (World Alliance for Citizen Participation, 2013b, page 1).

PAGE 93

80 addition to an overall index, there are sub indices for different aspects of CSO sustainability. The direction of the USAID i ndices are the inverse of that for civil society values, so one w ould expect a negative correlation coefficient. For water the correlation coefficient for 20 countries was 0.62 (p < 0.01 ) for the overall index, and correlations for the specific legal, o rganizational and advocacy indices measured at 0.56, 0.73, a nd 0.48, respectively, (all significant at p < 0.05 ) For sanitation the correlation coefficient was 0.51 fo r the overall index (p < 0.05 ), with correlations for the specific legal, organizational, and advocacy indices measuring at 0.51, 0.60, and 0.38, respectively, (t he first two sub indic es had significance < 0.05; the third was not significant ) Overall, the s e results sho w a reasona bly high correlation for both w ater and sanitation with the USAID indices ( USAID, 2011). Internal and External Validity A potential threat to internal validity in e ach of the tests for explanatory variable impact is history that is events outside of the institutional and governance attribute being studied that have impacted country progress on water and sanitati on This threat is somewhat addressed by the use of control variables in the statistical analyses. It will also be addres sed in the case studi es that go more in depth into contextual evidence. The case studies will also be a way of looking at potential issues of multi collinearity among the three in dependent variables. Another limitation is that the collection time period for country institutional and governance data (2011 GLAAS Survey), does not directly correspond with the time period represented in the outcome variables, (2000 through 2012) R esponses provided in 2011 are not necessarily indicative of the institutiona l an d governance conditions

PAGE 94

81 during the entire 2 000 to 2012 time period represented by the Progress variable s and preceding the 2012 situation measured by the A 2012 variable s While this limitation cannot be re medied for the quantitative analyses, it is pa rtially addressed by the case studie s that will look at conditions over a multiple year time period. External validity, the ge neralizability of results, means whether or not th is research yiel ds information relevant to countries that were not part of the sample. In this institutions would be supported by the fact that the sample represents about half of the LMI population, and because t he statistical models control for differing country background characteristics. However, a s previously noted, systematic difference s exist between the sample group and the overall LMI country popu lation, and these should be taken into consideration when g eneralizing results. The sample has, on average, higher country populations, lower GDP per capita, and lower water and sanitation access rates.

PAGE 95

82 CHAPTER V QUANTITATIVE ANLAYSIS Th is chapter explains the quantita tive tests that were d one to t est the hyp otheses using country specific data from the country sample group It begins by reviewing the resea rch questions and presenti ng descriptive statistics for all variables. Next, bivariate stati stics are presented showing relationships between pairs of inde pendent and outcome variables. Finally the last part of the chapter presents a series of multivariate stat istical tests and analyzes the results to draw conclusions related to the research questions and hypotheses. RQ1. progress on providing urban wat er and sanitation access ? H1 A high level of decentralization in water and sanitation services is associated with greater country progre ss on providing urban wat er and sanitation access. RQ2. What is the role of sector progress on providing urban wat er and sanitation access ? H2 A high level of sector wide strategy and investmen t coordination in water and sanitation services is associated with greater country progress on providing urban wat er and sanitation access. RQ3. What is the role of civil society engagement providing urban wat er and sanitation ac cess ? H3 A high level of civil society engagement in water and sanitatio n services is associated with greater country progress on providing urban wat er and sanitation access.

PAGE 96

83 Descriptive Statistics This section presents descriptive statistics for the outcome variables, key explanatory variables, and control v ariables. It also discusses how the outcome variables were adapted to better suit the statistical analysis and what was revealed by analyzing correlation values among variables. Outcome Variab les Descriptive statistics for the four outcome variables revealed that individual country values are not normally distributed and had unacceptable levels of skewness and kurtosis. These characteristics make the outcome variables unsuitable for a standard regression analysis. Therefore, instead of using individual country observations for the statistical tests, I assigned countries to low, medium, and high achievement groups for each outcome variable, thereby converting the variable to an ordered categori cal variable with three possible values. The defining value range for three groups was individually set at logical breakpoints given the distribution of values. 18 The three groups are labeled 0, 1 and 2 represent ing an ordered ranking of low, medium, and high achievement respectively. Tables 5.1a through 5.1d show, for each outcome variable, descriptive statistics for raw country data on the left side, and characteristics of the grouped variable on the right side. 18 An attempt was made to create groups breaking at equal intervals on the distribution scale, for example, between 0.10 and 0.20 then.0.20 and 0.3 0. This resulted in excessive bunching of observations in some groups, with very few observations in other groups.

PAGE 97

84 Table 5.1 Descripti ve Statistics and Grouping Parameters for Outcome Variables Table 5.1 a Water A 2012 Raw Country Data Grouped Variable Obs 74 Minimum 52.3 Group Obs Value Range (%) Maximum 100 0 Low 16 0 > x < 85 Mean 90.71 Std. Dev. 9.82 1 Medium 25 85 x < 95 Variance 96.53 Skewness 1.72 2 High 33 x 95 Kurtosis 5.83 Table 5.1 b Water Progress Raw Country Data Grouped Variable Obs 72 Minimum .99 Group Obs Value Range Maximum 1.0 0 Low 20 x < .05 Mean .24 Std. Dev. .49 1 Medium 31 .05 x < 0.50 Variance .24 Skewness .61 2 High 21 x 0.50 Kurtosis 2.90 Table 5.1 c Sanitation A 2012 Raw Country Data Grouped Variable Obs 74 Minimum 15.7 Group Obs Value Range (%) Maximum 100 0 Low 20 x < 35 Mean 61.45 Std. Dev. 26.93 1 Medium 27 35 x < 80 Variance 725.08 Skewness .12 2 High 27 80 Kurtosis 1.53

PAGE 98

85 Table 5.1 d Sanitation Progress Raw Country Data Grouped Variable Obs 72 Minimum .09 Group Obs Value Range Maximum 1.0 0 Low 24 x < 0 .05 Mean .17 Std. Dev. .23 1 Medium 28 x < 0 .20 Variance .05 Skewness 1.68 2 High 20 0 .20 Kurtosis 5.45 All references to outcome variables from here forward will refer to the grouped outcome variables as will the notations A 2012 and Progress Correlations among the outcome variables were checked and appear in Appendix D. The correlation between the two water outcome variables is 0.38, and between the two sanitation outcome variables is 0.53, suggesting moderate to strong positive relationships between variables from the same sub sector. When comparing water and sanitation the correlation coefficient is 0.35, and when comparing W ater Progress and S ani tation Progress the correlation coefficient is 0.33, both suggesting moderate positive relationships. Explanatory and Control Variables Table 5 .2 shows descri ptive statistics for the six key explanatory varia bles ( derived from the GLAAS response averaging process described above ) as well as the eight cont rol variables. (Refer to Appendix C for vari able units of measurement. ) By observing mean value s for the key explanatory variables, with possible values from 0.0 to 1.0, one can see that the m ean for W sector planning at 0.72, is towards the range maximum. M ean values for the other five variables, ranging from 0.48 to 0.63, are close

PAGE 99

86 to the mid range value of 0.5. The relatively balanced distribution of responses for five of the six key vari ables works against the idea that countries may have been inclined to rate themselves as high or low to appear in a more positive light. The 0.72 mean value for W sector planning as well as the second highest mean value of 0.63 for S sector planning do indicate less variation in responses on that attribute. This may be because GLAAS question s on the topic focus on sector plan development which many countries have done in some form, rather than sector plan implementation Notice also that the mean va lues for sanitation are all about 0.10 lower than the mean values fo r water on the same attribute, perhaps reflecting a pattern of less attention paid to sanitation issues a theme that is t aken up again later in Chapter VIII Table 5 .2 Descriptive Statistics for Key Explanatory Variables and Control Variables Variable Obs Mean Std. Dev. Min Max Key Explanatory Variable W Decentralization 74 0.62 0.22 0.17 1.00 W Sector Planning 73 0.72 0.26 0.13 1.00 W Civil Society 71 0.58 0.29 0.00 1.00 S Decentralization 73 0.53 0.32 0.00 1.00 S Sector Planning 74 0.63 0.28 0.00 1.00 S Civil Society 70 0.48 0.27 0.00 1.00 Control Variable Urban Population Change 75 0.16 0.14 0.10 0.61 Corruption Level 75 69.97 9.46 37.00 92.00 Political Fragility 75 1.52 1.01 0.00 3.00 Security Fragility 75 0.71 1.02 0.00 3.00 Literacy Rate 75 72.13 21.22 27.00 99.80 GDP per capita (log) 75 7.41 1.09 5.53 10.09 Renewable Water Resources (log) 75 8.40 1.79 2.30 12.21 Foreign Aid to Water & Sanitation (log) 75 2.74 1.31 1.35 6.06

PAGE 100

87 Correlations Appendix D shows correlations among th e six key explanatory variables These were generated to check for potential problems of multi collinearity among explanatory variables prior to running the multivariate models. The correlations to consider are those among the three water variables and those among the three sanitation variables, as none of the statistical tests are run with the water and sanitation variables together. F or the three water variables, the highest correlation is 0.51, between W sector planning and W civil society Among the three sanitation variables, the h ighes t correlation is 0.55, also between S sector planning and S civil society From the perspective o f the statistical model, these correlation levels were not considered so high as to oblige removing one or the other variable. The fact that these tw o attributes are closely aligned is perhaps not surprising, as both attributes depend, to some extent, on a climate of consultation and openness within a society. Correlations between the water and sanitation decentralization variables sector planning variables, and civil society va riables are all relatively high ( ). This is not surprising, as we wou ld expect there to be a level of similarity within a country on its approach to water and to sanitation, give how i nter related the two sub sectors are. Appendix D also shows c orrelations a mong the eight control variables These were generated to check for potential problems of multi collinearity among these variables prior to running the multivariate model s. The highest correlation is between GDP per capita and literacy at 0.62. H owever this is still considered at an a cceptable level, and attempts at removing one of the other from the analyses did not improve overall results.

PAGE 101

8 8 Finally, correlations among explanatory variables and control variables were checked, separately for water and sanitation, as no tests are run w ith water and sanitation variables together. The results appear in Appendix D and show that no variable pairings had a correlation level that was problematic. The highest correlation pairings by sub sector are between W civil society and corruption ( 0. 36), and between S civil society and GDP per capita (0.37). Bivariate Analyse s In this section, I discuss what was learned by looking at different types of relationships between pairs of in dependent and outcome variables, including frequency of Group Frequencies App endix E shows frequency tables that relate country key explanatory variable composite values with low, medium, and high o utcome group s for each of the four outcome variable s. The presentation is grouped according to the three attribu tes : decentralization, secto r planning, and civil society. T he sector planning frequency tables show country values clustered at the high end of the value range in contrast to results for decentralization and civil societ y that show the highest frequencies at mid range values. Correlations The correlations between explanatory and outcome variables were analyzed to check if the directions of bivariate relationships are as expected, based on the key explanatory variable hypotheses and the discussio n of control variables presented in earlier chapters. The results are shown in Table 5.3.

PAGE 102

89 Table 5 .3 Correlations between Explanatory Variables and Outcome Variables 1 Explanatory Variable Water A 2012 Water Progress Sanitation A 2012 Sanitation Progress W Decentralization 0.14 0.23 --W Sector Planning 0.06 0.14 --W Civil Society 0.24 0.14 --S Decentralization --0.26 0.29 S Sector Planning --0.12 0.13 S Civil Society --0.45 0.28 Urban Population Change 0.20 0.10 0.15 0.05 Corruption Level 0.26 0.03 0.01 0.01 Political Fragility 0.05 0.17 0.26 0.12 Security Fragility 0.28 0.19 0.11 0.23 Literacy Rate 0.29 0.21 0.62 0.08 GDP per capita (log) 0.32 0.05 0.64 0.36 Renewable Water Resources (log) 0.02 0.17 0.16 0.18 Foreign Aid to Water & Sanitation (log) 0.03 0.03 0.10 0.10 1 N = 66 for both wat er and sanitation; a few countries vary between the two groups. Looking first at the water sector, the W decentralization variable moves in a negative direction relative to the outcome variables, which contradicts my hypothesis. The other two key water variables are positively correlated with the outcome variables, consistent with my hypotheses. Th e first four control vari ables mostly correlate negatively with the two outcome variables as expected except for corruption level with P rogress (slight positive), and political fragility with Pr ogress T he last four control variables all positively correlate with as expected, but mostly negatively correlate with Progress The negative correlations with Progress are unexpected. The exception is renewable water r esources, which is positively correlated with Progress It is unclear why the unexpected control variables relationships occur, but the topic of co ntrol variable influence is taken up again later in the discussion of multivariate results.

PAGE 103

90 For the sanitation sector, the three key explanatory variables are positively correlated with both outcome variables, consistent with my hypotheses. Also, the first four control variables, all c orrelate negatively with both outcome variables, as would be expected The last four control variables mostly correlate positively with the two outcome variables, again as expected. The exceptions are renewable water resources, which correlates negatively with and foreign aid, which correlates negatively with both outcome variables. After observing the unexp ected negative correlation between foreign a id and thr ee of the four outcome variables and considering that the donor aid figures on which the variabl e was based cover aspects o f the water sector outside the scope of this study I decided to drop this as a control va riable. It did not perform as expected which was to register the added benefit of dono r financial resources to a country Also, prelimin ary multivariate models that included the variable found it to be in significant in explanatory value. Dropping this variable left seven remaining control variables. Differences of Group Means Another way of analyzing bivariate relationships is to look at differences of explanatory variable means among outcome variable groups. Means by group, are presented in Table 5.4 for the six key explanatory variables. The rows highlighted in bold indica te statistically significant differences among means for that explanatory variable and outcome variable pairing although not a lways with a linear relationship. The two sanitation variable pairings with significance demonstrate increasing linear relations hips, but the two significant water variable pairings do not.

PAGE 104

91 Although not statistically significant, W decentralization shows a Low group mean that is greater than the High group mean for both outcome variables. While this is in conflict with my hypoth esis, it is similar to the findings from the correl ations exercise, where decentralization correlated negatively with outcomes. For sanitation, we see a relationship between S decentralization and the outcome variables c onsistent with my hypothesis, meani ng that greater decentralization is associated with better outcomes. Table 5 .4 Key Explanatory Variable Means by Outcome Group A 2012 Progress Low Medium High Low Medium High W Decentralization 0.72 0.54 0.63 0.72 0.60 0.58 W Sector Planning 0.70 0.66 0.74 0.66 0.71 0.76 W Civil Society 0.53 0.45 0.68 0.54 0.54 0.64 S Decentralization 0.45 0.53 0.63 0.49 0.48 0.74 S Sector Planning 0.64 0.54 0.71 0.55 0.70 0.64 S Civil Society 0.34 0.38 0.63 0.38 0.50 0.56 Bold figures show statistically significant diff erences among the group means (p < 0.10) O nly countries used for multivariate tests were included here; those countries not tested due to missing data and outlier pairings were removed. Multivariate Analyses This section details multivariate statistical tests that est imate the effect of the key exp lanatory variables on explaining each of the four outcome variables. S tandard regression analysis was attempted using individual country outcome data but as previously noted, data distributions were not normall y distributed and there were persiste nt problems with outlier s. Deletion of all outliers would have reduced the sample size to an unacceptable level. Therefore each outcome variable was converted to three ordinal groups, (low, medium, and high) I then used ordered logistic regression (ologit)

PAGE 105

92 for the statistical test s as this test is appropriate for a categorical outcome variable when the categories are ordered. Stata statistical software was used to r u n the ologit models, also referred to a s a proportional odds models An ologit model predicts probabilities that a case result will fall in one of the cat egories designated for outcomes, based on values of the explanatory variables for a particular case. The predicted ordinal outcome variable, (low, medium, and high in this case), is a function of another continuous latent variable Y* that is not measured. The Y* variable has various threshold points. The predicted outcome group for a particular case Y i depends on the whether or not the Y* has crossed a particular threshold for the designated categories, (three in this case), as follows (Hamilton 2012): Formula 5.1 Ordered Logistic Regression Model Y i = 0 if Y* i 1 Y i = 1 if 1 i 2 Y i = 2 if Y* i 2 Y i = observed (predicted) ordinal outcome variable for case i Y* i = unobserved continuous latent variable for case i 1 = minimum value for the medium outcome category 2 = minimum value for the high outcome category Ologit results can be evaluated using the Pseudo R 2 value, (representing the McFadden R 2 ), which can theoretically range from 0.0 to 1.0 and represents the goodness of fit of the model. T he Pseudo R 2 is meant to approximate the R 2 of OLS regression models A hi gher Pseudo R 2 value ind icates a better fitting model. Logis tic regression results can also be evaluated by the significance of chi 2 which shows the statistical significance of the overall model as compared to the null model with no predictors (Stata Dat a Analysis Examples, 2014).

PAGE 106

93 When using ologit one must be careful to have a sufficiently large sample size relative to the number of explanatory variables, in order for the results to be meaningful. For this research, t he number of country observations used in each of the four ologit test s, (one for each outcome variable), range d from 66 to 68 since not all countries coul d be used for each test due to missing data no ted in Appendix B. With ten explanatory variables inclu ded for each test three key variables and seven control variables -t he se sample sizes translate to a ratio of about seven observat ions per explanatory variable This ratio of observations per explanatory variable was judged sufficient based on guidelines from Vittinghoff & McCulloch (2007), who showed that prediction problems are uncommon with five to nine observations per explanatory variable. 19 They further stated that a model showing statistical significance with five to nine observa tions per explanatory variable is meaningful ; only in some situations is extra caution warranted, such as when highly significant associations are hypothesized prior to testing. When running the ologit test for each outcome variable I first did a test u sing only the seven control variables, and then a complete test where the three key exp lanatory variables were also included for a total of ten variables. This allows one to observe the added value to the model results of the three key explanatory variables. Tables 5.5 through 5.8 show the coefficients, standard errors, and level of significance of the explanatory variables for each of the four ologit tests Table 5.9 is a summary of results from all four ologit tests. This summary format allows one to easily observe patterns of statistical significance and direction for the key explanatory variables. 19 Vittinghoff & McCulloch focused primarily on confidence interval coverage for 1 and the related type I error rate of the test of H 0 1 = 0), secondarily on bias in the estim 1 and indirectly on variability and power.

PAGE 107

94 Table 5 .5 Ordered Logistic Regression Results : Water A 2012 Explanatory Variable Controls Only Model Coefficient (Standard Error) Complete Model Coefficient (Standard Error) Simplified Model Coefficient (Standard Error) W Decentralization -3.36** (1.56) 3.16** (1.49) W Sector Planning -0.28 (1.25) -W Civil Society -2.72** (1.30) 2.29** (1.15) Urban Population Change 3.44* (1.82) 3.53* (2.08) 3.65* (2.03) Corruption Level 0.06** (0.03) 0.05 (0.03) 0.05 (.03) Political Fragility 0.41 (0.29) 0.60* (0.33) 0.58* (0.33) Security Fragility 0.33 (0.24) 0.41 (0.26) 0.42 (0.26) Literacy Rate 0.02 (0.01) 0.02 (0.02) 0.02 (0.01) GDP per capita (log) 0.53 (0.32) 0.70* (0.39) 0.69* (0.37) Renewable Water Resources (log) 0.05 (0.14) 0.15 (0.17) -Observations 74 1 68 1,2 68 1,2 Prob > chi 2 0.00 0.00 0.00 Pseudo R 2 0.15 0.21 0.21 *** p < 0.01; ** p < 0.05; p < 0.10 1 One country from the sample of 75 was missing 2012 data (Eq. Guinea). 2 Two countries had missing data on a key explanatory variable (Haiti and India). Four countries were removed due to outlier pairings (Azerbaijan, Brazil, Iran, and Senegal).

PAGE 108

95 Table 5 .6 Ordered Logistic Regression Results: Water Progress Explanatory Variable Controls Only Model Coefficient (Standard Error) Complete Model Coefficient (Standard Error) Simplified Model Coefficient (Standard Error) W Decentralization -4.10*** (1.49) 4.50*** (1.42) W Sector Planning -2.43* (1.30) 2.69** (1.28) W Civil Society -2.49** (1.22) 2.57** (1.20) Urban Population Change 3.20* (1.76) 4.05* (2.23) 4.25* (2.21) Corruption Level 0.01 (0.03) 0.03 (0.03) 0.03 (.03) Political Fragility 0.05 (0.25) 0.13 (0.30) -Security Fragility 0.25 (0.26) 0.26 (0.29) -Literacy Rate 0.03** (0.01) 0.05*** (0.02) 0.05*** (0.02) GDP per capita (log) 0.19 (0.29) 0.47 (0.34) 0.42 (.33) Renewable Water Resources (log) 0.19 (0.13) 0.35** (0.16) 0.37** (0.16) Observations 72 1 66 1,2 66 1,2 Prob > chi 2 0.18 0.00 0.00 Pseudo R 2 0.07 0.18 0.18 *** p < 0.01; ** p < 0.05; p < 0.10. 1 One country was missing 2012 data (Eq. Guinea). One country was missing 2000 data (South Sudan). One country st arted and ended at 100% access and was excluded (Lebanon). 2 Two countries had missing data on a key explanatory variable (Haiti and India). Four countries were removed due to outlier pairings (Azerbaijan, Brazil, Iran and Senegal).

PAGE 109

96 Table 5. 7 Ordered Logistic Regression Results: Sanita tion A 2012 Explanatory Variable Controls Only Model Coefficient (Standard Error) Complete Model Coefficient (Standard Error) Simplified Model Coefficient (Standard Error) S Decentralization -2.35** (1.06) 2.28** (1.03) S Sector Planning -0.95 (1.22) -S Civil Society -2.93** (1.34) 3.08** (1.25) Urban Population Change 0.92 (2.02) 1.80 (1.48) -Corruption Level 0.04 (0.03) 0.06 (0.04) 0.06* (0.03) Political Fragility 0.00 (0.29) 0.10 (0.35) -Security Fragility 0.19 (0.24) 0.74** (0.32) 0.67** (0.31) Literacy Rate 0.04*** (0.02) 0.05** (0.02) 0.04** (0.02) GDP per capita (log) 1.22*** (0.38) 1.50*** (0.45) 1.38*** (0.41) Renewable Water Resources (log) 0.27* (0.16) 0.25 (0.20) -Observations 74 1 68 1,2 68 1,2 Prob > chi 2 0.00 0.00 0.00 Pseudo R 2 0.31 0.43 0.41 *** p < 0.01; ** p < 0.05; p < 0.10. 1 One country from the sample of 75 was missing 2012 data (Eq. Guinea). 2 Two countries had missing data on a key explanatory variable (Haiti and India). Four countries were removed due to outlier pairings (Afghanistan, Burundi, Central African Republic and Mali).

PAGE 110

97 Table 5 .8 Ordered Logistic Regression Results: Sanitation Progress Explanatory Variable Controls Only Model Coefficient (Standard Error) Complete Model Coefficient (Standard Error) Simplified Model (a) Coefficient (Standard Error) Simplified Model (b) Coefficient (Standard Error) S Decentralization -3.00*** (0.95) 2.94*** (0.93) 3.04*** (0.95) S Sector Planning -0.45 (1.19) -1.95** (1.03) S Civil Society -2.35* (1.36) 2.60** (1.20) -Urban Population Change 0.93 (1.78) 0.66 (2.13) --Corruption Level 0.04 (0.03) 0.07** (0.04) 0.08** (.03) 0.08** (.04) Political Fragility 0.16 (0.26) 0.09 (0.30) --Security Fragility 0.23 (0.23) 0.93*** (0.33) 0.92*** (0.33) 0.97*** (0.32) Literacy Rate 0.03* (0.01) 0.05** (0.02) 0.05*** (0.02) -GDP per capita (log) 0.98*** (0.33) 1.37*** (0.41) 1.35*** (0.40) 0.72*** (0.27) Renewable Water Resources (log) 0.21 (0.13) 0.44** (0.18) 0.42** (0.17) 0.37** (0.16) Observations 72 1 66 1,2 66 1,2 65 1,2 Prob > chi 2 0.03 0.00 0.00 0.00 Pseudo R 2 0.10 0.26 0.26 .22 *** p < 0.01; ** p < 0.05; p < 0.10 Note: Simplified model (a) included civil society without sector planning; (b) included sector planning without civil society. Significance was lost when both were included. 1 One country was missing 2012 data (Eq. Guinea); one was missi ng 2000 data (South Sudan); one st arted and ended at 100% access and was excluded (Lebanon). 2 Two countries had missing key explana tory variable data (Haiti and India); four were removed as out lier s (Afghanistan, Burundi, Central African Republic and Mali.). Zimbabwe was removed only when running Sector Planning without Civil Society

PAGE 111

98 Table 5 .9 Summary of Ordered Logistic Regression Results for Complete Models Key Explanatory Variable Water A 2012 Coefficient (Standard Error) Water Progress Coefficient (Standard Error) Sanitation A 2012 Coefficient (Standard Error) Sanitation Progress Coefficient (Standard Error) Decentralization (by sector) 3.36** (1.56) 4.10*** (1.49) 2.35** (1.06) 3.00*** (0.95) Sector Planning (by sector) 0.28 (1.25) 2.43* (1.30) 0.95 (1.22) 0.45 (1.19) Civil Society (by sector) 2.72** (1.30) 2.49** (1.22) 2.93** (1.34) 2.35* (1.36) Observations 68 66 68 66 Prob > chi 2 0.00 0.00 0.00 0.00 Pseudo R 2 0.21 0.18 0.43 0.26 *** p < 0.01; ** p < 0.05; p < 0.10 In Tables 5.5 through 5.8, one column shows test results using only the control variables, another shows the complete model with the controls plus the key explanatory variables, and a last column introduces a simplified model, where explanatory variables that were not statistically significant are removed, as long as their removal did not markedly reduce the Pseudo R 2 value. The simplified models, with a reduced number of explanatory variables, serve to maximize the num ber of observations per explanatory variable adding predictive strength to the model. Also, with few explanatory variables, data collection efforts can be simplified should the model subsequently be applied to a new sample of countries or to an alternate time period for these same countries. The coefficients reported in Tables 5.5 through 5.8 are expressed in log odds units and should not be read as ordinary least squares (OLS) regression coefficients. They are most easily interpreted by considering the predicted probabilities and odds ratios based on the coef ficients, something that is presented later in this chapter. The positive or negative sign of the coefficient indicates the direction of the relationship.

PAGE 112

99 As mentioned above, Pseudo R 2 values reported in the tables are the McFadden R 2 and meant to approximate the R 2 of OLS regression models The Prob > chi 2 value reported in each table is the probability of getting a result 20 as extreme as, or more so, than the observed under the null hypothes is, which is that all the coefficients in the model are zero and there is no effect of the predictor variables (Stata Annotated Output) Summary of Logistic Regression Results The following discussion includes general observations that can be drawn by looking at the results of all four ologit models, as well as specific findings relating to the three key explanatory variables that were studied : decentralization, sector planning, and civil society. First some findings across all four mode ls: The four complete models had Pseudo R 2 val ues ranging from 0.18 to 0.43, and t he Pseudo R 2 always increased between the controls only models and the complete models, demonstrating an added predictive value of the three key variables. Chi 2 was statistically significant at p < 0.01 for all four complete models, but only for two if the controls only models. Aside from the predictive value of the model as a whole, the significance of the three key explanatory variables is most important for testing my hypotheses. The important findings for these variables are: 20 The log likelihood values are not reported in the tables, but do appear in the Stata output.

PAGE 113

100 Decentralization : W decentralization i s statistically significant for (p < 0 .05) and for Progress (p < .01) However because both display negative coefficients, this means that an lowering of its outcome group for water. S decentralization is statistically significant for (p < 0.05), and for P rogress (p < .01 ) With both outcome variables displaying positive coefficients, this means that a higher level of S decentralization results in higher country achievement for sanitation. Sector Planning : The association between W sector planning and achievement is significant only for the W ater Progress outcome variable, where there is a positive coefficient and slight statistical significance (p < 0.10 ) T his modest statistical significance coupled with the lack of any statistical significance for the variable, doe s not point towards a large impact for sector planning with water. For S sector planning there is no statistical significance for either outcome variable, and there is no consistency in the direction of the coefficient, with one boing positive and one negative. Therefore, no meaningful conclusions can be drawn about the impact of sector planning on the sanitation sector. Civil Society : W civil society is statistically significant for both water outcome variables (p < 0.05 ) and with positive coefficients This reveals that an increased level of civil society engagement in the water secto r support s higher country achievement.

PAGE 114

101 S civil society i s statistically significant wi th a positive coefficient for (p < 0.05), an d also statistically significant with a positive coefficient for Progress (p < 0.10). These results demonstrates that an increased level of civil society engagement result s in higher country achievement for sanitation. Odds Ratios Another way of interpreting the ologit results is throu gh observation of odds ratios, presented in Table 5.10 for each of the four logit tests A one unit change in an explanatory variable will increase the odds of a country being in the high versus the medium and low outcome groups by the odds ratio value, given that all other variables are held constant (Stata Annotated Output Torres Reyna 2014 ) The ratios show that an increased explanatory variable value increases the odds of being the high group particularly for civil society and for decentralization in the sanitation sector, as these odds ratios are all greater than 10. For the other pairings, an increase in the explanatory variable value barely help s the odds of being in the high group, with the exception o f sector planning and Water Progress Table 5.10 Odds Ratios for Key Explanatory Variables Explanatory Variable Water A 2012 Odds Ratio (Standard Error) Water Progress Odds Ratio (Standard Error) Sanitation A 2012 Odds Ratio (Standard Error) Sanitation Progress Odds Ratio (Standard Error) Decentralization by sector .03 (.05) .02 (.02) 10.51 (11.12) 20.10 (19.09) Sector Planning by sector .75 (.94) 11.40 (14.86) .39 (.47) 1.57 (1.87) Civil Society by sector 15.24 (19.88) 12.09 (14.72) 18.63 (27.65) 10.53 (14.35) Note: Odds ratios > 10 appear is bold.

PAGE 115

102 Predicted Probabilities Another way of presenting the ologit results is to graph, for any given value on a key explanatory variable, the probabilities of being in the low, medium, or high outcome gro up with all other explanatory variables being taken at their means. These re sults are presented in Figures 5.1 5.2, and 5.3 for decentralization, sector planning, and civil society respectively. The explanatory variable is represented on the X axis, and the predicted probabilities for the three groups on the Y axis, summing to 100 for any given X value T he slopes provide i nsight into how the predicted outcome variable group changes as the key explanatory variable increases In most cases, the proba bility of being in the high outcome group increases as the key explanatory variable increases, and conversely, the probability of being i n the low outcome group declines. The line for medium group probability is generally bell shaped, peaking around mid r ange values of the explanatory variable. Graphs displaying these patterns are consistent with my hypotheses. Two notable exceptions to this patte rn appear in the first two figures, Figures 5.1a and 5.1b where opposite sloped lines are observed for W decentralization as related to both outcome variables for water Also, for some of the sector planning graphs, there is are no obvious pattern s to the slopes, corresponding with those variable pairings that did not show statistical significanc e in the ologit results

PAGE 116

103 Figure 5.1: Predicted Probabilites for Decentralization 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 0 0.2 0.4 0.6 0.8 1 Predicted Probability Decentralization Decentralization and Water A 2012 Group Low Medium High 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 0.2 0.4 0.6 0.8 1 Predicted Probability Decentralization Decentralization and Water Progress Group Low Medium High

PAGE 117

104 Figure 5.1: Predicted Probabilites for Decentralization (cont.) 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 0.2 0.4 0.6 0.8 1 Predicted Probability Decentralization Decentralization and Sanitation A 2012 Group Low Medium High 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 0.2 0.4 0.6 0.8 1 Predicted Probability Decentralization Decentralization and Sanitation Progress Group Low Medium High

PAGE 118

105 Figure 5.2: Predicted Probabilities for Sector Planning 0 0.1 0.2 0.3 0.4 0.5 0.6 0 0.2 0.4 0.6 0.8 1 Predicted Probability Sector Planning Sector Planning and Water A 2012 Group Low Medium High 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 0.2 0.4 0.6 0.8 1 Predicted Probability Sector Planning Sector Planning and Water Progress Group Low Medium High

PAGE 119

106 Figure 5.2: Predicted Probabilities for Sector Planning (cont.) 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 0.2 0.4 0.6 0.8 1 Predicted Probability Sector Planning Sector Planning and Sanitation A 2012 Group Low Medium HIgh 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 0.2 0.4 0.6 0.8 1 Predicted Probability Sector Planning Sector Planning and Sanitation Progress Group Low Medium High

PAGE 120

107 Figure 5.3 : Predicted Probabilites for Civil Society Engagement 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.2 0.4 0.6 0.8 1 Predicted Probability Civil Society Engagement Low Medium High 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 0.2 0.4 0.6 0.8 1 Predicted Probability Civil Society Engagement Civil Society and Water Progress Group Low Medium High

PAGE 121

108 Figure 5.3 : Predicted Probabilites for Civil Society Engagement (cont.) 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 0.2 0.4 0.6 0.8 1 Predicted Probability Civil Society Engagement Civil Society and Sanitation A 2012 Group Low Medium High 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 0.2 0.4 0.6 0.8 1 Predicted Probability Civil Society Engagement Civil Society and Sanitation Progress Group Low Medium High

PAGE 122

109 Control Variables Looking now a t the control variables, Table 5.11 summari zes where each control v ariable showed significance ( p < 0.10 ) in a test and the direction of the coefficient. Four control variables exhibited patterns of influence for at least two outcome variabl es in the expected direction and appear in bold: 1) urban population change; 2) security fragility; 3) GDP per capita; and 4) renewable water resources. Table 5.11 Control Variable Significance Water Water Progress Sanitation Sanitation Progress Urban Population Change Corruption Level + Political Fragility + Security Fragility Literacy Rate + GDP per capita (log) + + + Renewable Water Resources (log) + + Notes: Significance indicated at p < 0.10 + = positive coefficient = negative coefficient. B old ed variable name = expected direction of coefficient Interaction Variables To explore the potential significance of interactions among the three key control variables, I generated additional variables for both water and sanitation calculated by multiplying two variables together as follows: decentralization and sector planning decentra lization and civil society sector planning and civil society

PAGE 123

110 The olgoit models were then re run with the a ddition of the in teraction variables With two exceptions, none of the interaction variables showed statistical significance in their contribution to the models. The two that were significant were both sector planning and civil society interaction variables, one for Water Progress and the other for Sanitation A 2012 This significant interaction is consistent with the relatively high correlation observed between sector plann ing and civil society ( 0 .51 for water, and 0 .55 for sanitation) Inclusion of interaction variables in the models often modified the significance level of other ex planatory variables and, with the addition of new variables the number of explanatory va riables became too large for the sample size, according to statist ical guidelines. Therefore, it did not seem beneficial to include the interaction va riables as they are mostly insignificant and their inclusion risks expanding the number of explanatory variables beyond what is supported by the sample size. Conclusions Several c onclusions can be drawn a bout the research questions and hypotheses, based on t he entirety of quantitative analysi s results Each hypothesis is recapped below along with relevant findings and conclusions. H1 A high level of decentralization in water and sanitation services is associated with greater country progress on providing urban wat er and sanitation access. For the water sub sector, there was strong evidence against H1, as demonstrated by the correlation direction, the ologit results, the odds ratios, and the predicted probabilities. The reasons for this result may relate to the literature that found decentralization benefits may not be fully realized due to incomplete implementation of all aspects of decentralization, as well

PAGE 124

111 as challenges posed by the economic, political, and socio cultural environment in developing countries For the sanitation sub sector, as demonstrated by the correlation direction, the ologit resu lts, the odds ratios, and the predicted probabilities, a conclusion can be drawn that higher decentralization is helpful to sanitatio n achievement, consistent with H1. H2 A high level of sector wide strategy and investment coordination in water and sanitation services is associated with greater country progress on providing urban wat er and sanitation access. Findings related to sector planning were inconclusive. While th e correlations show a positive relationship with all outcome variables, the ologit results and predicted probability graphs show varied directional relat ionships. Only in the case of W ater Progress was the variable shown to be significa nt in the ologit te sts (p < 0.10) and to have a high odds ratio. This lack of a clear pattern may stem from t he relatively high means for both sector planning variables reflecting minimal differentiation amo ng countries. The ologit results do not provide supporting evidence for H2. Nonetheless, the fact that sector planning may have helped W ater P rogress over t he 12 year period is noteworthy. H3 A high level of civil society engagement in water and sanitation services is associated with greater country progress on providing urban wat er and sanitation access. The results consistently support H3, as civil society was positively correlated with all four outcome variables, wa s statistically significant in all ologit test s, and demonstrated the expected relationships in the odds ratios and predicted probability graphs. The reader is reminded that not all constructs of key explanatory variables were captured in the quantitative data for the ologit tests. For example, GLAAS questions

PAGE 125

112 averaged to develop the composite decentralization values focused on the first three constructs of decentralization as listed in Appendix A, namely decision making authority, funding, and investment decisions. The remaining constructs, (clear r oles for different levels of government, monitoring of sub national levels of government, and the extent of capacity building for sub national levels of government ), were not measure d They will instead be studi ed through the case studies, along with a de eper exploration of the already measured constructs. Questions for Further Exploration In addition to the above conclusions that correspond to hypotheses some broader issues have emerged from the quantitative results. These issues are de tailed below along with questions for further analysis: 1) Decentralization was shown to be highly significant (p < 0.01) for predicting Water Progress with a negative coefficient, and also highly significant for predicting Sanitation Progress with a positi ve coefficient. The same coefficients and direction were observed for the A 2012 variables, but with a lower level of significance (p < 0.05). Considering these findings, what are some possible explanations for why decentralization impacts the two sub sec tors differently? 2) Since the sector planning variables did not show a large degree of differentiation among countries, what can be ascertained from other studies about the extent to which water and sanitation sector planning is being implemented in LMI cou ntries? Can any conclusions be drawn by relating sector planning data from other studies to my outcome variables?

PAGE 126

113 3) As findings on the impact of sector planning were largely inconclusive, and the variable was significant in predicting only one of the outco me variables ( Water Progress ), how do these results compare to other studies looking at the impact of sector planning on country water and sanitation access? 4) Given that civil society engagement was positively associated with country achievement across al l four outcome va riables, and mean composite values of the variable, (0.48 to 0.58), were modest relative to the maximum possible value of 1.0, what can be learned from countries employing new forms of civil society engagement that might be beneficial to other countries seek ing to heighten their level of engagement? Decentralization and civil society engagement are explored through country case studies, (two per attribute), presented in later chapters. For sector planning, since there is not a long hist ory of countr y implementation the detail available from a ny one country would be minimal. Therefore, instead of preparing ca se studies other topical studies were identified that looked at the impact of sector planning for select groups of countries, pro viding data and inform ation towards further evaluation of sector planning constructs and questions. This is the topic of the next chapter.

PAGE 127

114 CHAPTER VI SECTOR WIDE STRATEGY AND INVESTMENT COORDINATION Th is chapter responds to the two questions related to sector wide strategy and investment coordination recapped below. 1) Since the sector planning variables did not show a large degree of differentiation among countries, what can be ascertained from other studies about the extent to wh ich water and sanitation sector planning is being implemented in LMI countries? Is it possible to relate this alternate data to my outcome variables, even if for a smaller sample of countries? 2) As findings on the impact of sector planning were largely in conclusive, and the variable was significant in predicting only one of the outcome variables ( Water Progress ), how do these results compare to other studies looking at the impact of sector planning on country water and sanitation access? The discussion first look s at what other studies found about the extent to which countries are engaged in sector planning, and then analyze how country data from these other studies relate to my outcome variables. Finally, a section is included on general findings about sector planning impact on country water and sanitation achievement. S ector Planning by Countries Here, I review findings from other studies about the extent to w hich countries have included water and sanitation in P overty Reduction Strategy Papers (P RSP ) and the challenges faced in preparing country led sect or specific plans Several of the studies identified focus on sub Sahar an Africa countries They found i n general that the inclusion of water and sanitation issues in PRSPs has been weak, (although improving

PAGE 128

115 over time), and the development of sector wide approach (SWAp) documents for the sector has been slow. In the following discussion, PRSPs are discussed first, f ollowed by sector specific planning. Poverty Reduct ion Strategy Plans In comparison to health and education issues, water supply and sanitation frequently received less coverage in PRSPs, based on a 2004 study of 10 developing countries (Newborne, 2004) Even though water and sanitation consistently ranked among the to p three priorities during participatory poverty assessments (exercises conducted to identify country needs and plan action ) coverage in PRSPs did not adequately reflect the priority in detailed targets or action s. The sector was often insufficiently addres sed and planned for in spending budgets financing strategies, and monitoring and e valuation plans W ater and sanitation sector budgets that do appear in PRSPs tend to focus on inputs and out puts rather than strategically projecting outcomes and impacts (WSP, 2006b), and these planned budgets do not always fully transfer to actual budget preparation and execution at the national level (Slaymaker & Newborne, 2004). Achieving a realistic inves tment plan is complicated by the fact that there are multiple public and private stakeholders in the water and sanitation sector Again c omparing to the health and education sectors, the water and sanitation sector has major donor funded projects that are not part of national budget s and therefore not incorporated into the PRSPs (Slaymaker & Newborne, 2004). Substantial donor funding can make it difficult for sector authorities to advocate for alloc ations from the national budget when other sectors are competing for the same government funds (Newborne, 2004).

PAGE 129

116 Another reason the sector may be underrepresented in PRSPs is that the secondary benefits of improved water and sanitation on health and education programs often go unrecognized It has been sugg ested that authorities should more aggressively pursue water and sanitation PRSP promotion in coordination with those responsible for the related sectors that receive secondary benefits (Overseas Development Institute, 2004) Sanitation wa s usually give n little or no attention during PRSP development (WSP, 2003; Mehta, 2002; Newborne, 2004). For example, in a group of ten sub Saharan PRSPs prepared prior to 2002, nine of ten included coverage of water in the poverty diagnostic and assessment exercise, b ut only two out of ten included sanitation (Mehta, 2002, p. 23). Even though sanitation consistently had lower coverage rates among the population, it was viewed as having a lower priority for poverty alleviation. A s initial plans have undergone updates and revisions t he attention paid to water and sanitation in the PRSP process has improved During the transition f rom interim to full PRSPs between 20 03 and 2006, there was increased coverage of the sector, as well in follow on documents, such as progress reports and PRSP updates (WSP, 2003). P RSPs have helped to improve coordination with donor programs and have, in many cases, laid the groundwork for development of sector specific plans (WSP, 2006b). Challenges of Country Led Sector Planning S everal s tudies highlight the complexities of having true country ownership of planning exercises like PRSPs and SWAps. A 2004 study found that donors and NGOs may develop project p lans at the local and regional level s which are not adequately included in national plans due to weak links among national, regional and local authorities (Slaymaker & Newborne, 2004). More recently, a 2011 study found that

PAGE 130

117 water sector SWAps are still mostly driven by donors, with national authorities adopting the approach as a necessary step for attracting greater financial resources (EuropeAid Development and Cooperation Directorate, 2011). With diminished country ownership, national government follo w through on implementation of plans may be relatively weaker than if they were initiating and dr iving the process more directly. In many countries, responsibilities for water and sanitation policy and management extend acr oss several different ministries. St udies show that sector planning can be hindered by this dispersion of responsibility and possible competition among different departments. A 2004 study of four countries found c oordination between institutions responsible for water and sanitation serv ices and those responsible for wate r resource management to be weak (Overseas Development Institute, 2004). A lso, when different governmental institutions and ministries compete for the same pool of financial resources, the incentive to cooperate under one lead agency for PRSP or SWAp development may be compromised (Slaymaker & Newborne, 2004). M otivational problems can arise when participating agencies perceive that the lead agency will receive credit for ach ieving plan goals, with limited recog nition of the role of other government entities in its accomplishment (EuropeAid Development and Cooperation Directorate, 2011) Information p resented in later chapters for the Colombia and Mali case studies illustrate how responsibilities for water and san itation at the national level can be fragmented a nd complex. Such conditions c o mplicate co hesive plan ning and project prioritization. One example is in consistent costing and budgeting approaches during national investment plan development (Newborne, 2004 ) This can result in different sub

PAGE 131

118 sector plans, for example, rural and urban, that originate from different areas of the government and are not mutually supportive (Slaymaker & Newborne, 2004). Observations about national pla n monitoring from other stu dies are summarized here While sector progress is sometimes tracked thr ough the PRSP process, (to the extent that it is covered in the document), detailed sector tracking was found to be minimal and underfunded in several studies, hurting efforts to prioritize scarce resources and accurately measure results (WSP, 2006b; Newborne, 2004; WSP, 2003). A study from the early 20 00s recommended strengthening sector wide monitoring and evaluation, in addition to project specific evaluation and focusing more on goal related outcome monitoring, rather than just output monitoring (Mehta, 2002) Focusing on outcomes helps evaluate if donor projects support goals articulated in sector plans, rather than just realizing an output The same study found expenditure tracking to be inadequate and suggested broad based surveys to evaluate services from a customer perspective. The next section undertakes the analysis proposed in question 1) at the start of this chapter: relating alternate data to my outcome variables. Relating Sector Planning to Achievement I tried relating sector planning measures from other studies to my cou ntry outcome variables similar to what was done in t he cross case analysis, but for a smaller number of countries due to limited data. T wo outside studies were identifie d that rank countries according to the degree to which a country included water an d sanitation in national planning exercises First, I used data from a 2003 study that rated sub Saharan African countries on the extent to which water supply and sanitation were incorporated full PRSP ( Water and Sanitation Program, 2003, p.10) I created both

PAGE 132

119 scatter plots and linear pre dictor plots showing these scores against my four outcom e variables for ten countries that are also in my sample group The results appear graphically in Appendix F Series 1. Three of the four linear prediction graphs show an upward slope, consistent with my hypotheses. The exception is Water A 2012 Secondly, I did the same type of analysis with outside data measuring the extent to which countries had developed national water strategies, using scores for sub Saharan African countries estimated based on 2006 data (Water Partnership Program, 2010 p.11 12 ). I compiled three applicable sub indicators for each country which the study rated as good, fair or poor, and assigned a value of f ive, three, or one respectively I then totaled the three indicator scores to arrive at a value between 3 and 15 representing sector plannin g for each of 13 countries that are also in my sample group The three sub indicators measured the existence of: 1) a national strategy; 2) a sector financing plan; and 3) a monitoring and evaluation program for the water sector. 21 Appendix F Series 2 sh ows the country total scores plotted against my four outcome variables in both scatter graphs and linear prediction graphs. The two predicted water lines have upwards slopes, consistent with my hypotheses, and the two predicted sanitation lines have downward slopes, in opposition to my hypotheses. So, this would imply that sector planning helped water achievement, but not sanitation achievement. The next section discusses findings from another recent study that looked specifically at the impact of SWAps for country water and sanitation achievement. 21 These scores were based on information in Water and Sanitation Program ( 2006a ) and the evaluation undertaken of M&E Systems in the Draft Synthesis Report for the Pan African Water Sector M&E Assessment.

PAGE 133

120 Findings from Prior Research Another way to assess the implications of the findi ngs from this dissertation is to compare them to other studies that ha ve examined the relationship between sector planning and country water and sanitation access achievement Only one study was identified that had done this type of analysis. A 2011 study of seven developing countries used e mpirical evidence from the seven countries, ( both existing data and information obtained through interviews and workshops ), and compared this t o existing measures of country progress, including JMP data (EuropeAid Development and Cooper ation Directorate, 2011) T he study found that the impact of a country having a SWAp is unclear. It noted the difficulty of attributing country advances in water and sanitation coverage solely to the preparation of a SWAp due, among other things, to the difficulty of isolating it fro m other reforms in the sector. For most countries, the addition of sector w ide planning into the mix of secto r initiatives was likely at a stage when access progress was already underway, but more coordination and planning were needed to reach the remaining populations that lacked access. This is supported by data from the previ ously mentioned study showing the seven study countries had achieved the most rapid growth in coverage prior to 2000 before SWAps were adopted, and more modest growth rates between 2000 and 2008 (EuropeAid Development and Cooper ation Directorate, 2011) T hese temporal differences in achievement may explain why the sector planning variable was not significant in my cross case analyses that used 2000 to 2012 service access data. The report goes on to say that SWAps may be important for longer term sustainability of the sector if attention is paid to maintaining infrastructure and securing

PAGE 134

121 access rates over time (EuropeAid Development and Cooper ation Directorate, 2011) In that se nse, benefits of a SWAp may not be evident for 20 years or so, and continued study of their imp act overtime would be of value. Conclusion The above analysis has provided insight into the two questions posed at the start of the chapter related to sector wi de strategy and investment coordination. Findings from other studies indicate that inclusion of water and sanitation issues in PRSPs has been weak, (although improving over time), and the development of SWAp documents for the sector has been slow Sanita tion often receives little or no coverage in these documents. Challenges involved in conducting comprehensive sector planning include: fostering country ownership so that SWAps are not just done to comply with donor requirements; donors agreeing to includ e their projects and comply with plan modalities; different government agencies contributing to plan development and receiving due recognition for sector accomplishments; having consistent budgeting system to streamline investment plan development; and reg ularly monitoring sector progress relative to plan goals. While there is no consistent conclusion about the contribution of sector planning to country water and sanitation achievement either from my analysis or prior studies, there is some evidence that it is helpful to long term water access achievement and may be important in the long term to help sustain gains made in the sector. T he activity warrant s additional data collection and analysis to cover plan development, plan updates, and systems for mon itoring aid partnering agreements The next chapter begins the discussion of the country case studies by explaining the qualit ative analysis methods used to prepare the cases.

PAGE 135

122 CHAPTER VII QUALITATIVE METHODS The qualitative p rovide s greater insight into key explanatory variables through closer examination of individual constructs over time particularly those that were not well represented in the GLASS survey data. In addition, th e qualitative research looked at the broader question s stemming from the quantitative results that were listed at the end of Chapter V I did t wo countr y case studies for d ecentralization and two for civil society engagement, for a total of four. I chose to study extreme cases for both variables. extreme value on an independent variable. As per Gerring (2007 p. 104), extreme cases exploratory ended fashion (Gerring, 2007 p.105). For this research, the rela tionships revealed between these two explanatory variables and the outcome variables in the cross case analysis are significant and consistent and thus well understood. Decentralization has a positive and significant relationship with sanitation achievement, but a negative and significant relationship with water achievement. Civil society engagement has a positive and significant relationship with achievement in both sub sectors. In choosing extreme case countries, I selected based on a high value of the explanatory variable, indicative of a depth of experience with implementing the explanatory variable of interest.

PAGE 136

123 The following secti ons discuss the countries selected and the methods used for collecting and evaluating country specific i nfor mation through document review, interviews, and questionnaires. Case S election T o further study decentralization, I chose Columbia and Mali, two countries that had high decentralization value s 22 experience with sector decentralization dating back to the early 1990 s, and a reasonable amount of written material available about the country experience. These characteristics made both countries s uitable for a detailed analysis, yet the two countries differ in some key background condi tions, such as income lev el and urban access rates for water and sanitation in the year 2000. Also, in Mali, high decentralization was accompanied by high achievement in the water sector, although not in the sanitation sector, which is contrary to the cross case findings. Among o ther things, the Mali case explore s reasons why this opposing result was realized. To further explore civil society engagement, South Africa and Kenya were selected for case studies because of their high value s on this variable, 23 and because they introd uced during the 2000 to 2012 timeframe, some inn ovative civil society init iatives at the national level with country wide influence Studying the experience of these countries will hel p to provide additional insight into the relationship between civil so ciety engagement and country ac hievement Because both are located in sub Saharan Africa, where many of the countries facing the greatest challenges for increased access are 22 Colombia 1.0 for both water and sanitation; Mali 0.83 for water and 0.75 for sanitation. 23 Kenya 1.0 for water and 0.75 for sanitation; South Afric a 0.75 for both water and sanitation.

PAGE 137

124 located, their experiences may be particularly relevant to the large group of LMI countries located in this region, due cultural and income level similarities. Document Review To begin the case research, a n on line search was conducted for relevant documents, searching by country name and search terms related to the variable of inter est, such a were then evaluated for relevancy based on the topical coverage, level of detail, date written, and the objectivity of the source. Most documents originated from international organizations working in water and s anitation, academic journals national governments, and non profit organizations. D ocuments were then reviewed and coded for evidence of the attribute of interest using the construct level definitions from Appen dix A as a guide, a s well as the more general topical questions stemmi ng from the cross case analysis, such as why decentralization may have had a different impact on water and sanitation, and what civil society engagement techniques have been particularly impactful. During t he review, I also looked for information on the impact of the attribute of interest and other relevant events that would influence country progress. Approximately 25 documents were reviewed per country, although not all contained relevant information. I n evaluating the information and conclusions presented in documents, consideration was given to the perspective of the source and, t o the extent possible, key findings and conclusions were supported by multiple documents originating from different sources. In coding the material, I searched for mention of specific terms and activities that fit within the identified constructs for that attribute, and then studied the

PAGE 138

125 material thoroughly to understand the context and timeframe in which the activity was imple mented, so that it could be appropriately reflected in the case narrative. Information from the coded documents was then synthesized to develop country narratives detailing implementation timing and modalities. During this process, contradictions and ga ps in information were ident ified and noted to be addressed as possible, through consultations with country experts. Consultations with Country Experts During the document review and other on line research the names of certain individuals emerged as having experience with a particular country and topic of interest. T en or more persons were identified per coun try from diverse professional affiliations and backgrounds further screened based on their expertise with the topic, the length of time the y have been involved with the issue, and/or the ir potential to fill in information gaps in a count ry narrative. Several individuals per country were then c ontacted by email but not all responded to the initial request or to follow up correspondence Among those who responded, a total of 14 persons ultimately provided information for the case studies. Their names and affiliations are listed in Appendix G I ndiv iduals agreeing to consult were provided with background information on the study and asked to provide consent for participation as per Colorado Multiple Institutional Review Board guidelines governing all University of Colorado Denver research activities Some consultations were i nterviews conducted via Skype, (for about 1 hour) after which responses were recorded and documented in interview summaries. In other cases, input was received in response to qu estions sent via email. The questionnaire approach was adopted due to time z one considerations, l imited internet

PAGE 139

126 connection, or t he general preference of the respondent For several individuals in Mali, emails and questions were translated and provided in French. A few of the persons sent questionnaires declined to answer due to time considerations but instead sent written materials they felt were releva nt to the questions posed and the general area of inquiry. Respondents were able to request that all or part o f the i nformation they provided be kept anonymous, although no one made such a request. As recommended by Rubin and Rubin (2005), the interview questions, (sometimes conveyed through questionnaires), were guided by the purpose of the research and the typ es of conclusions desired. Append ix H includes the general guides used for collecting information from consultants, grouped by case t opic. Depending on the background of the consultant, more specif ic questions were also asked about particular construct s for the attribute of interest, such as the details of fiscal funding for decentralization, or the response rate of complaint programs for civil society engagement. The consultations f ocused on how a country implements the key attribute being studied, and what has been the p erceived impact of the attribute on country progress in extending water and sanitation service access Q uestions were also as ked relating to the other two attributes being studied that were not the primary focus of the case study so that a cursory evaluation of variable i nteractions could be conducted. For example, in cases focused on civil society engagement, questions about decentralization and sector planning were also asked. The interview summaries and questionnaire responses, along with any written material forwarde d were then coded and incorporated into the c ase narratives. C oding and review of the material was similar to that discussed above for the document review. I

PAGE 140

127 looked for mention of specific terms and act ivities that fit within the identified constructs for that attribute as well as the context and timeframe in which th e activity was implemented. Scope of the Case Studies While the quantitative analysis focused on the 12 year period from 2000 to 2012, th e decentralization case studies explore events starting in the late 1980s that laid the groundwor k for decentralization In keeping with the focus of this research, the emphasis is on reforms and progress in urban areas, leaving out many of the specific circumstances and challenges faced by rural areas. The level of detail differs by country, depending on the available information, (both documents and consultation s), and the relevancy of the country experience to the attributes and constructs being studied. In the scope of the case studies, it is not possible to provide in depth coverage of all relevant topics, such as legal reforms or specific financial and budgeting details, as these are often intertwined with wider reforms at the country level and delve into areas of specialized knowledge. What is featured are select elements that most correspond with the key explanatory variable constructs as defined for this research. Also, w hile a diligent attempt was made to accurately report on programs and events within countries, some inaccuracies may exist, p articularly about specific dates, agency/department na mes and donor project details. The author regrets any errors or omissions and exp ects that they would not markedly alter the overall case conclusions.

PAGE 141

128 CHAPTER VIII DECENTRALIZED SERVICES This chapter examines decentralization implementation in the sector by drawing on the experience of two countries: Colombia and Mali. These countries are similar in that they had high composite decentralization value s 24 but are different in overall country circumstances and in how decentralization was implemented, thereby providing varied experiences from which to draw useful insights. The Colombia case, w hich is considerably longer due to more extensive experience and inform ation, is presented first. I chroni cle the evolution of the decentralization program in each count ry and highlight key shifts in program implementation In addition some discussion is devoted to the role of sector planning and civil society engagement in each country, so that the reader may apprec iate to what extent these attribute s coincided or inte racted with decentralization. P rivate sector engagement in service provision, something that was promoted in both countrie s, is discussed in brief At the end of the chapter a compilation analysis is done using findings from both countries to evaluate t he decentralization constructs listed in Appendix A and recapped below : a) Decision making authority for service provision is delegated to sub national levels of government b) Authority to manage expenditures and raise revenues is delegated to sub national le vels of government c) There are clearly defined rights, responsibilities, and roles for sector management actors at the national and sub national levels 24 Colombia 1.0 for both water and sanitation; Mali 0.83 for water and 0.75 for sanitation.

PAGE 142

129 d) Sub national level authorities are monitored by communities and/or the central government including, for example, regarding public funds expenditures and equality in service provision e) Support is provided to build managerial and technical capacity of sub national levels of government Finally, a last section address es the finding from the quantitative analysis regarding the opposing influence of decentralization on water and san itation the subject of a questi on posed at the end of Chapter V Colombia Case Study Colombia has a significant history with decentralization, dating ba ck to the early 1990s, with legislative reforms continu ing into the 2000s. There are 32 departmental level governments and about 1,100 municipalities, with great variation in size. It has a very high urban population, comprising 72% of the total populati on in 2000 and increas ing to 76% by 2012. GDP per capita in Colombia increased three fold from 2000 to 2012 (World Bank, 2012c) The GINI index, which measures the extent to which the distribution of income among individuals or households within the econ omy deviates from a perfectly equal distribution, decreased from 58.7 in the year 20 00 to 53.5 in 2012, trending towards a more even distribution of wealth (World Bank, 2000 2012 a ). I searched for compari son data to validate the Colombia composite decentralization value derived from GLAAS survey responses. Relevant indices w ere available in a 2014 article ( Ivanyna & S hah, 2014) that provided scores for all countries on different aspects of decentralization, (fiscal, political, and administrative), as well as an overall country index. (This same i ndex was referenced in Chapter IV under Data Validity and Limitations .) For the overall index, Colombia is the second highest rated decentralized country among 73 countries in my sample with ratings, surpas sed only by

PAGE 143

130 Brazil. It is the 4 th highest for fiscal decentralization, 13 th highest for political, and 7 th highest for administrative. Given its relative ranking among the sample countries, the composite water and sanitation decentralization value s, 25 both at 1.0, seem reasonable (Ivanyna & Shah, 2014). Decentralization Reforms Colombia is considered a forerunner of decentralization among developing countries (Rojas & Frank, 2002; Samad Lozano Gracia, & Panman 2012). The decentralization shift in Colombia was not unique to this country, but part of a wave of political and economic liberalization in the region (Chacon, 2013). Strikes during the 1980s reflected discontent on the part of the population, and decentralization responded to increasing pr essure among citizens to democratize the local electoral process and was also seen as a way to address social inequalities and the increasing threat of guerrilla violence (Falleti, 2010). A significant political step for Colombia was the passage of Law 11 in 1986, which for the first time in the twentieth century established that mayors should be popularly elected for two year terms, rather than appointed, with no possibility of immediate re election (Snchez Torres & Pachon, 2013; Granados & Snchez, 2014) Related legislation in 1986 introduced a platform for local referendums and citizen participation in municipal affairs (Chang Memon, & Imura 2004). At this time, Colombia had about 1,800 water utilities, most of which were publicly managed. Only abo ut 45 of the companies operated in large cities (population of 25 The reader wi ll recall that a composite value was developed for each of the key variables, drawing on country responses to relevant questions in the GLAAS survey.

PAGE 144

131 100,000 or more), and many of the others operated under INSFOPAL, a national centralized water company for small and medium sized cities. INSFOPAL was administratively and financially troubled and was abolished in 1987. Regional entities that operated under INSFOAPL were transferred to provincial government management and water service provision responsibility moved universally to municipalities (Frangano et al., 2001; World Bank, 2012 a ). In 1991, another major government reform took place with enactment of a new constitution that integrated the political, fiscal, and administrative aspects of decentralization, seeking to address the severe problems of inequality and poverty of the country (C hacon, 2013). It introduced to the water and sanitation sector principles of competition, regulation of monopolies, and the promotion of private sector investors and operators (Rojas & Frank, 2002; World Bank, 2012 a ). The legal framework establish ed b y the new constitution clearly separated the responsibilities of policy making and service provision for the sector The central government was to set policies, regulate and oversee services, while municipalities were responsible to ensure universal provi sion and service quality. In support of this change, the new constitution increased the tax base and rate of fiscal resources transferred from the central government directly to municipalities. At the same time, it gave greater fiscal authority to munic ipal and departmental authorities to administer and modify taxes to meet their obligation to perform public functions. Historically, most funding for the sector originated from government budget allocations, such as legal transfers to municipalities and c entral government grants (C hacon, 2013; Echavarra, Rentera, & Steiner 2002; Frangano et al., 2001).

PAGE 145

132 Notwithstanding this change, there were huge differences among mun icipalities in their ability raise own source revenues Large urban centers had much greater ability and means, due to taxes paid by large size firms located in these areas ( Ramirez Interview Summary, 2015) Refining Decentralization The new decentralization framework was further defined in 1993 by Law 60 that required municipalities receiving transfers from the central government to spend a specified percent of transfers on different sector programs. Water and basic sanit ation were designated at 20%. While the central government continued to be the main source of fund s to finance sector specific investments, municipalities had autonomy for decision making, including deciding expenditure amounts, structuring investments, and execu ting investments ( Echavarra et al., 2002; Frangano et al., 2001; Granados & Snchez, 2014) Law 60 was notable because in addition to mandating a fixed percentage be spent on different sectors, it increased over time the level of resources allocated to sub national levels of government (for all services). The law allowed for a gradual incre ase in the amount transferred directly to municipalities from 14% of total national fiscal revenues in 1993 to a maximum of 22% in 2001 (Chacon, 2013; Echavarra et al., 2002). This guaranteed funding stream provided municipalities with a level of fundi ng security for the sector, thereby helping decentralization in this sector move at a faster pace than in other sectors that were not part of the guaranteed funds such as transportation ( Ramirez Interview Summary, 2015) While the National Planning Department was responsible for monitoring expenditures by sub national levels of government, it was limited by the constitutional

PAGE 146

133 autonomy granted to departments and municipalities, and some incentive issues developed After Law 60 providing guaranteed fu nding, there were decreased incentives for municipalities to increase their own revenue generating capacity and their main source of financing for sector investments was central government transfers (Snchez Torres & Pachon, 2013). With this duality, a negative side effect of decentralization was to weaken the fiscal discipline of sub nationa l levels of government and between 1991 and 2002, some departmental level governments defaulted on bank loans. The fiscal imbalance that Colombia experienced in th e 1990s when expenditures exceeded revenues, was largely attributed to increased central government transfers to regional governments. The overall situation reflected a failed institutional framework for decentralization that allowed reduced fiscal disci pline on the part of sub national governments (Echavarra et al., 2002 ; Giugale Lafourcade, & Luff 2003). M unicipalities that wanted to increase their expenditure in the s ector needed to increase their own fiscal capacity, from sources such as tariffs. Additional fiscal reforms from 2000 to 2002 brought even more fiscal responsibility. As incentives were modified, the trend changed to increasing the a verage share of a mun own resources utilized relative to central government transfers although transfers remained the dominant source (Ramrez et al., 2014; Snchez Torres & Pachon, 2013). Regarding the efficiency of municipal investments in the sector Silva an d Andia (2006) found evidence of sharp spending inefficiencies when they related the level of

PAGE 147

134 central government allocated resources with changes in water and sanitation coverage 26 from 1998 to 2002, showing that central government funds were not being spen t efficiently at the local level. However a study by the Inter Ame rican Development Bank (IDB) of the period 1994 to 2005 found that local fiscal effort positively and significantly impacts water coverage progress, and that differences in water coverage across municipalities in Colombia are linked to the fiscal effort 27 put forth by the municipality (Snchez Torres & Pachon, 2013). Spending efficiency was also a focus of by Law 142 in 1994, which encouraged competition and privat e sector participation. It mandated that all existing public services entities convert to legally autonomous stock companies, whether municipal, private, or mixed ownership. Municipalities were also given power to grant concessions to private sector entities. According to one survey of 550 municipalities, about half made some type of reform between 1994 and 2004, changing from direct municipal service provision to services provided by a publically managed utility or private company (Granados & Snchez, 2014). (More discussion on private sector involvement in a later section.) Roles and Responsibilities Following the constitutional reforms, the sector was characterized by overlapping responsibilities, a lack of clarity of roles and responsibilities for different levels of gov ernment, and weak accountability provisions (Echavarra et al., 2002). Although 26 Coverage is considered to be synonymous with the term access 27 Fiscal ef fort is defined as policy actions that the local government undertakes to augment local revenu e, cal culated as the proportion of total local spending on water financed with local taxes, as opposed to other sources such as national government transfer and royalties with specified uses.

PAGE 148

135 municipalities were given responsibility for service provision and given new fiscal authority, policies were set by the central government and it remained the main source of f unds to finance investments. A USAID funded study in 2001 found the early years following the shift in responsibility to be characterized by conflicting roles, overstaffed operations, inefficiencies, and insufficient resources for system maintenance and improvement (Frangano et al., 2001). Smaller municipalities (2,500 to 100,000 residents) faced particular challenges, as they had previously been served by INSFOPAL (Granados & Snchez, 2014). The USAID report commented on difficulties in the first years of the decentralization program, whic h they attributed to political interference, institutional fragmentation, and overlapping responsibilities: The transition to a decentralized scheme has been erratic and costly, while the capacity of central government entities to guide the process has l agged behind expectations. Only a limited group of cities has been able to take advantage of these difficulties and successfully face the challenge of decentralization by increasing autonomy, fostering community participation, and designing and implementi n g their own suitable solutions (Frangano et al., 2001, p.21). The report suggested further clarification of roles for municipal and departmental governments with regard to sector decentralization. At the national level, responsibilities were spread among several different institutions, which some saw as negatively impacting sector efficiency (Giugale et al., 2003). In 2006, efforts to set overall policy were consolidated within the newly created Vice Ministry of Water and Sanitation within the Ministry o f Environment, Housing and Territorial Development. However, t he sector policy was not independent but took its lead from the overall policy framework and allocation of public resources established by the National Planning Department.

PAGE 149

136 In comparison to the national level, the role s of department al and municipal governments were initially more focused; they were given responsibility to develop sanitation and wastewater treatment infrastructure, and to this end, were vested with authority to pass local decrees, impose taxes, and conduct planning in coordination with the national system of planning (World Bank, 2012 a ). Their scope of authority included deciding who provides services, who monitors providers, and whether coverage should be increased to new areas of a municipality (Snchez Torres & Pachon, 2013). But i n recent years, developments have somewhat constricted independent municipal decision making, thro ugh promotion of regiona l investment programs and by requirements for central government project reviews These developments further discussed below have created some ambiguity around the roles of different government levels under decentralization. G eneral Revenue Sharing System T o further refine the decentralization funding scheme Law 715 in 2001 modified Law 60 and consolidated three national transfer systems into a Genera l Revenue Sharing system (SGP) for central government transfers to municipalities. It included transfers specifically targeted to poorer a nd less populated areas to be used either for infrastructure building or to subsidize fees for poor households (Granados & Snchez, 2014; World Bank, 2012a). It also required that loc al user fee systems takes a cross subsidization approach, with residential users from low socio economic levels receiving fee discounts covered by higher fees charged to high socio economic levels and commercial and industrial users (Granados & Snchez, 20 14). Finally, it mandated that transfers to

PAGE 150

137 municipalities for water and sanitation have a real annual increase of 2% from 2002 to 2005, and 2.5% from 2006 to 2008 ( World Bank, 2005 ). This funding mandate was designed to stabilize sector funding, follow ing a severe economic crisis in the late 1990s, during which transfers to municipalities fell d ue to a drop in national income (Ramirez Interview Summary, 2015). However, some saw the mandated percent increases as having been set below the expected growth of the national economy, thereby actua lly diminishing the level of funds designated for the sector (Lopez Murcia, 2015a). From the late 1990s to 2010 municipal per capita spending in the water and san itation sector increased steadily A breakdown of so urces of funds for municipal investments over a 15 year period appears in Table 8 .1 Table 8 .1 Sources of Funds for Municipal W ater and Sanitation Investment in Colombia ( 1995 2010 ) Source: Samad et al. 2012 p. 173 28 R oyalties from natural reso urces and extractive industries can be assigned to territories where the nonrenewable resources are produced or assigned by the national government to finance priority investment pr ojects in non producer areas ( Samad et al., 2012 ) Source of Finance Share of Municipal Investment Central Government Transfers (SGP) 40% Own Resources 24% Royalties 28 12% Other 10% National Co financing 6% Credit 5% Departmental Co financing 3%

PAGE 151

138 The funding breakdown shows a relative ly small share (24%) of own sources meaning locally generated funds. As per th e previously cited Silva and Andia 2006 study, the predominant reliance on SGP central government transfers would have been associated with spending inefficiencies. A similar observation was made in a 20 12 Colombia Urbanization Review repor t prepared for the World Bank: The fragmented nature of funding in this sector has led to a regional distribution of resources that is not necessarily in line with national investment priorities, and indeed, may generate incentives for some municipalities to avoid making adequate tariff efforts (Samad et al., 2012, p.174). The need for continued reform of the funding structure to increase efficiency of spending was reflected in specific recommendations from the U rbanization Review report, which included: introducing output and performance based incentive systems; studying the impact of spending conditions, such as earmarking, on investment efficiency; and providing technical assistance to smaller cities and rural areas to help improve spending efficiency (Samad et al., 2012). In 2012, a significant policy change allowed for a wider distribution of royalties that was particularly helpful to smaller and lower income municipalities. Previously, 80% of royalties wen t to about 50% of municipalities (mainly based on where the royalty based resources were produced), but subsequently, the central government distributed these to a wider range of municipalities based on poverty level, to meet basic needs and help address u nemployment (Ramirez Interview Summary, 2015). Investment Decision Autonomy Table 8.2 shows a breakdown of 2010 sector investment by government level:

PAGE 152

139 Table 8 .2 Water and Sanitation Investment by Different Levels of Government in Colombia (2010) (Billions of 2010 Colombian pesos) Central Government 0.3 Counties 1.0 Municipalities 3.5 Total 4.8 Source: Ramrez et al., 2014. p.33 Table 1, as taken from National Planning Department, 2010 administrative fiscal regis ters While the majority of investment decisions are made at the municipal level, it is subject to some constraint s since new fiscal reforms were from 2007 onward If any central government transfers are being used towards a project, (which we see from Table 8 1 is often the case), then the project needs to go through a review process at the Ministry of Housing. This is the case for the vast majority of cities that have some reliance on central government funds, and can be viewed as a trend towards recentralization of the sector (Lopez Murcia Interview Summary, 2015b). M unicipalities are also required to meet certain public service outcome targets, many stemming from the national deve lopment plan, in order to stay certified. Losing certification mea nt either that: the relevant Ministry designs a special improvement plan that is monitored and requires regular reporting; an order to reallocate the use of central government transfers is issued or; in extreme cases, the national government takes ov er se rvice provision for up to five years (Lopez Murcia, 2015b).. Also b eginning in 2007 the central governm ent encouraged d epartmental level planning and investment to take advantage of economies of scale and shared resources, and possibly creating better con ditions for private sector involvement. A special account was created to fund departmental plans, diverting some funds away from municipalities,

PAGE 153

140 which created some tension between mayors and governors. Municipalities were encouraged, but not required, to also apply their financial resources towards this approach. While this appeared to be a good approach in theory, in practice it has only been successful in some areas (Ramirez Interview Summary, 2015; Rojas Interview Summary, 2015) Some possible reasons include: 1) local councils and local mayors would prefer to retain local control of funds; 2) if local officials are from a different political party than the departmental level officials, this may result in a lack of trust; 3) there is a perceptio n of a lack of capacity at the departmental level to successfully carry out regional plans; and 4) there is confusion about who maintains legal rights over the funds put into regional schemes the departmental level or the local level of government (Lopez Murcia Interview Summary, 2015b). In terms of providing additional benefits to poorer municipalities, the d epartmental approa ch does not offer any particular advantage, as services received are in proportio n to what a municipality contributes. Also, the central government maintains some a pproval over departmental plans, which can be viewed as another trend towards recentralization of the sec tor (Lopez Murcia, 2015a). Monitoring and Accountability An argument frequently made in support of decentralizati on is that it will increase the acc ountability of local government and improve service quality large ly due to the power of local information exchange and officials tailoring services to local needs and conditions The literature specific to the Colombian experience is somewhat mix ed in this regard. Krause (2007) analyzed water and sanitation services in developing countries

PAGE 154

141 and found particularly in Colombia, examples of poor governance by sub national governments negatively impact ing the internal efficiency of service delivery and the extent of service coverage. While service problems might be addressed with greater central government monitoring and oversight (Giugale et al., 2003), at present, there is not suffi cient capacity within the government to monitor the large number of municipalities, and subsequent ly potential problems often go unchecked (Rojas Interview Summary, 2015). Another study also found a negative impact of decentralization for some Colombian c ommunities, s uggesting that the amount of public resources available for capture at the local level has been a key determinant of local political violence (Chacon, 2013). Conversely Faguet and Snchez (2014) analyzed empirical evidence f rom the education and health sectors in Colombia and found that, in general, decentralization is generating accountability in local government and improving the quality of public service. Impact on Equity Recall that a general principle of th e 1991 Constitution was that new system s were to address problems of inequ ality and poverty in Colombia. T he expectation was that a decentralized system would better represent the interests of the poor and improve public services at the local level. To that end, one can examine the impact that reforms have had on general living conditions, a major component of which is water and sanitation service. A 2014 study found that the percentage of households living in s ub standard housing conditions was redu ced by more than 50% from 1993 to 2005 ( Ramirez, Bedoya, & Diaz, 2014 ) The improved conditions in urban areas are largely attributed to higher per capita incomes, cost efficiencies of serving a more clustered population, and to

PAGE 155

142 a lesser degree additiona l resource s and competencies provided to urban areas through decentralization measures ( Ramirez Interview Summary, 2015). Specifically for water and sanitation the cross subsidies mandated by Law 715 have been effective for transferring resourc es within some cities, particularly those that have high income neighborhoods and a relatively high overall tax base. But, as low income urban popul ations grow and form the primar y population in some cities, the system of cross subsidies becomes less effective, as there are not enough users in the high income categories to pay the higher fees that support subsidies (Rojas Interview Summary, 2015; Lopez Murcia Interview Summary, 2015b.) Another type of approach that has been successful in increasing service in poor neighborhoods has been the inclusion of performance targets in operator contracts (Giugale et al., 2003). Capacity Building At the start of decentralization in Colombia, municipalities had little support or skills to manage the new responsibilities, which was particularly challenging for the smaller municipal ities and those in rural areas Even with the allocation of funds for the sector from the central government, there was still a dependency on the central government and donors for program managem ent skills among small and rural municipalities Many smaller municipalities lack the institutional capacity to operate auto nomous public utility companies (Lopez Murcia Interview Summary, 2015 b ), and some faced challenges in decidin g among different tech nology options. T here were some cases of corruption among local officials and situations of consultants trying to sell inappropriate water and sanitation technologie s ( Rojas Interview Summary, 2015)

PAGE 156

143 Central g overnment programs tried to build the necessa ry skills at the local government level S oftware was developed that would help municipalities model and choose among different water and s anitation technologies, as well as accounting and billing software. In addition, educational materials were distrib uted on the legalities of creating water enterprises and calculating tariffs These tools were thought to be helpful, although the extent of their use is uncertain. (Rojas Interview Summary, 2015). Even when local officials are trained on such management skills, they may be replaced during political changes with those less knowledgeable about the sector. M unicipalities have also had the opportunity to learn from successful project e xamples in othe r Colomb ian municipalities that were studied with lessons distributed. The cities of Cartagena and Barranquillam were provided with World Bank support in the 1990s for capacity building of local authorities, including skills for fina ncing and executing tr ansactions, such as joint ventures. Success in these cities encouraged scaling up of reforms in other cities, as mayors became awar e of the approach and progress. It also partly motivated a much wider scale World Bank project (World Bank, 2012a; World Ba nk, undated/b). The World Bank Water Sector Reform Assistance Project (from 2001 to 2010) conducted institutional strengthening to help build the capacity of the Ministry of Environment, Housing and Territorial Development to promote, structure, and superv ise the introduction of private operators throughout the country. The Project conducted a pilot project that helped influence legislative reforms and executive decrees, improving the enabling framework for private sector participation (World Bank, 2012a).

PAGE 157

144 Private Sector Participation While i t is difficult within the scope of this research to isolate the impact of private sector reforms from other reforms in the sect or during the study period, some relevant findings are presented here. The legal reforms gave municipalities the option to remain as service providers by converting to public companies, to contract out to specialized providers, or to form some type of joint venture. With this framework in place, private firms began to operate in Colom bia. In the late 1990s, several major international utility companies came to Colombia, but they have not develop ed a large presence By 2013, only about 35% of municipalities had reformed their water services to involve private or semi private companies in some way (Snchez Torres & Pachon, 2013) wi th the most significant impact being in large and mid size urban areas ( Rojas Interview Summary, Ramirez Interview Summary, 2015 Lopez Murcia Interview Summary 2015b ). At present, there are only 14 wholly p rivate utilities in the country (Rojas Interview Summary, 2015). Looking at the efficiency impact of private sector involvement, t he previously mentioned IDB study of Colombia from 1994 to 2005 considered factors influencing water coverage by municipal ity and found that the public or private nature of service providers was not statistically significant in explaining increases in water coverage across municipalities, nor was it significant for affordability of service (Snchez Torres & Pachon, 2013). An other study found that municipalities with more than 2500 subscribers that had changed from direct municipal provision to a specialized company (either private, joint, or municipal), between 1993 and 2005, had a four percent higher increase in sew age coverage but a 6% lower increase in water coverage (Granados & Snchez,

PAGE 158

145 2014). Similar to these findings, Snchez Torres & Pachon (2013) reviewed several studies of private sector participation in Colombia and found most had inconclusive or mixed results about the overall benefits of private sector participation for service coverage. However, Barrera Osorio and Olive ra (2007) observed positive effects on coverage as a result of the involvement of the private sector in water and sanitation in Colombia, particularly in urban areas. Others note the general benefits of private sector involvement in Colombia: the stability afforded by long term operator contracts and the steadily increasing capacity and asset base of private operators (Samad et al., 2012; W orld Bank, 2012 a ). The limited evidence to date of private sector participation having had a positive impact on service coverage may grow over time as both municipalities and operators gain more experience. Sector Planning Colombia developed a SWAp for the environmental sector in cooperation with the Dutch government in 2007. This document was aimed more at overall environmental planning, river basin management, and biodiversity than specifically at water and sanitation service provision (Newborne Zuleta, & Lleras 2010). Col ombia has not prepared a PRSP 29 None of the information reviewed indicated a ny type of comprehensive national level plan for water and sanitation ser vice provision. The plans mentioned most were the departmental level plans p reviously discussed. 29 As per IMF website : http://www.imf.org/external/np/prsp/prsp.aspx#HeadingC updated as of May 7, 2015.

PAGE 159

146 Civil Society Engagement Particularly in comparison to other countries in the region, civil society is not very engaged on water and sanitation issues in Colombia, or on social issues in general. Only in recent years has a national consciousness developed on social issues, which could develop into a larger role for civil society There had been a movement a few years ago for universal basic free water, but it was unsuccessful ( Rojas Interview Summary, 2015) There are no prominent national civil society groups focusing on the sec tor, nor is there a network of civil society organizations (C SOs ) for water and sanitation. Big cities, more so than small cities and rural communities, have lobbying capacity at the central government lev el and are therefore more consulted on national level sector planning. There remains a feeling that the central government does not fully understand the realities of small cities and rural areas, yet the central government is often in a position to approv e projects and designs for these areas (Rojas Interview Summary, 2015). There is a fairly new CSO focused on rural areas called AQUACOL that is becoming known to the central government and influential on laws and policies (Rojas Interview Summary, 2015). Colombia Case Conclusion Following are some ge neral conclusions from the Colombia case study The composite decentralization values derived for the quantitative analysis, ( 1.0 for both water and sanitation ), are generally validated by findings o f the case study, considering th e specific element s that were measured by the GLAAS data : service provision decentralization, operational decision making autonomy and funding made available to local level governments for carrying out this respon sibility. However, what

PAGE 160

147 the GLAAS data fails to capture is specifics such as how funding mechanisms are executed, if any spending controls are in place, and what type of approval local investment decisions are subject to by the central government. Looki ng back at the more than 20 year history of decentralization in Colombia, w hat can be said about its impact i s that the particularly harmful aspects were the dissolution of INSFOPAL and the transition of decision making responsibility to sub national gover nments before they had benefited from any type of learning through case examples or capacity building programs. This would have been particularly challenging for smaller municipalities, which is where most of the areas lacking coverage were located. Ther e was also not yet an experienced group of private sector companies with whom municipalities could contract for services, nor were they experienced in managing such contracts prior to the demonstration projects. Further, the lack of adequate fiscal guidel ines may have hurt investment efficiency and failed to encourage partnerships among nearby municipalities to benefit from economies of scale, which can be particularly useful for water infrastructure. According to the JMP data, urban water access in Colombia showed a very slight decline from 1990 to about 2002, then remained virtually sta gnant through 2012. Urban sanitation access showed a moderate increase between 1990 and 2012 (from 82% to 85%). 30 Given that Colombia had already achieved 95% access for water by 1990, and that attaining the last 5% of access may face particular challenges, there still is a wide disparity between the progress for water and for sanitation during the time period studied 30 The vast majority of sanitation in urban areas of Colombia is to a piped sewer system.

PAGE 161

148 The chal lenge of building additional sanitation infrastructure in urban areas already covered by water service c ould have become easier once additional financial resources and a guaranteed revenue stream were available, which could explain why we see the small imp rovement in sanitation coverage. This presumably wo uld have been an easier step tha n extending water service t o the last 5% of areas without service, which may include small cities lacking sufficient capacity. This case shows that when implemented abru ptly, without fiscal constraints to encourage efficient inve stment, and prior to wide scale capacit y building, decentralization can slow progress, even in a high GDP growth environment like Colombia. A higher level of sector planning at the national level to better coordinate various initiatives could have been helpful, and greater civil society engagement would most certainly have been helpful based on the findings of the cross case analysis. Mali Case Study M ali is one of the least de veloped countries in the world. E ven though GDP per capita increased almost three fold from 2000 to 2012, i t ranked 176 out of 187 countries in the UNDP 2012 Human Development Index, which considers achievement in key dimensions of human development, such as having a long and healthy life, being knowledgeable and hav ing a decent standard of living ( UNDP, 1980 2013 ) In 2002, around the beginning of the study period, 64% of the population were classified as living in poverty, including 21% living in extreme poverty ( IMF, 2002 ). In addition to the ongoing challenge of being a low income country, Mali also experience d a 27% growth rate in urban population during t he study period, resulting in 38 % of the total country population l iving in urban areas in 2012 (World Bank 2000 2012 b ). Mali is organized

PAGE 162

149 into 8 regions and 49 districts, plus the capital district of Bamako The districts are further divided into municipalities, also known as communes, of which there were about 700 during the study period ( IMF, 2002 : Coulibaly, 2010 ) Mali experienced major political upheaval in 2012, including an attack on government forces by Is lamic militant supported rebels in the north of the country, and a coup that deposed the central government. While significant, t h ese events were towards the end o f the study period and did not directly coincide with the decentralization progr am events discussed in this chapter The following sections experience with sector decentra lization Among o ther things, the discussion examine s hips (high decentralization associated with high water achievement and low sanitation achievement) can be explained in terms of measur ement error or some other element that was not evident from the statistical analysis. composite decentralization value s (derived by averaging relevant questions from the GLAAS dataset), I looked at the Ivanyna & Shah (2014) decentralization indices. For the ir overall index, Mali ranked in the lower third of countries in my sample. Similarly, Mali anking s for the fiscal and adminis trative decentralization sub indices were both in the lower quartile of sample countries. The p olitical decentralization sub index ranking was higher, with Mal i ranking above 60% of sample countries. If we consid er t he overall, fiscal and administrative indices to be more relevant to the water and sanitation sector than political decentralization, then t he Mali

PAGE 163

150 decentralization values used in the quantitative model, (0.83 and 0.75), seem to be too high. Looking at another country comparison a 2008 Water Aid study compared Mali to 15 other African and Asian countries on several aspects of decentralization. Mali ranked just below the mid point for the overall decentralization index, but in the lower quartile for l ocal government influence on expenditures in the water sector This second measure influence of local government on sector projects funded from other sources, such as NGOs implemented in th eir locality (WaterAid, 2008). Unlike the Colombia case, the relatively high composite decentralization value s for Mali used in the quantitative analysis do not seem to correspond with the se more modest outside rankings. The supposition that water and sanitation in Mali is actually only moderately decentralized is explored through the case study and revisited when discussing conclusions of the case study. Decentralization Reforms The current political structure has its roots in pol itical reforms from the early 1990s, which modified a highly centralized political system in place from 1968. In 1991, a popular revolt and rebellion in the north part of the country triggered a military coup that brought in the Third Republic com mitted to giving more autonomy to localities A new constitution in 1992 provided the principles of decentralization and a 1993 law set out the framework, but it left many of the specifics undefi ned ( USAID, 1998; The Hunger Project, undated : Coulibaly, 2010 ).

PAGE 164

151 A 1996 law created the commune, or municipal level of government in Mali, increasing the number of local government entities from 19 to mo re than 700. Water was among three areas that were prioritized for statutory transfer of responsibilities to sub national authori ties. 31 ( Coulibaly, 2010 ). The needs of urban areas were specifically highlighted in a 1996 Urban Development Strategy, which stressed the need to increase living standards and fight urban poverty as part of decentralization. Various legislation provided the guiding principles for decentralization including local control of development projects (Farvacque Vitkovic Casalis, Diop, & Eghoff 2007) Despite the statutory transfer of responsibilities to sub national levels of government, m any of the operational aspects of decentralization were undefined and the transfer of resources did fully no t correspond (USAID, 1998; Coulibaly, 2010 ) There was a lack of clarity in redefined roles for long existing institutions and the roles of newly created institutions. I n 1998 USAID noted a lack of resources at the local level to implement programs related to their new responsibilities, a symptom o f their very low tax base and saw no clear plan on how to distribute needed resources to poorer regions of the country (USAID, 1998). In 2002, a Water Law was passed and i n 2004, the decentralization framework was further defined in a National Decentr alization Policy document covering the period of 2005 to 2014. Notable policy objectives include d increasing lo cal fiscal revenues to enhance their ability to finance pro grams locally, and harmonizing procedures for assistance from external donors. Objec tives specifically targeted to local government development include d : enhancement of their authority in managing development 31 The other two sectors were health and education.

PAGE 165

152 projects; a decreased role for central government staff; development of local citizenship; and development of the private sector to provide servic es for local governments (The Hunger Project, undated; Farvacque Vitkovic et al., 2007). Whereas previously, village management committees oversaw water operations, under the Water Law of 2002, local governments were discouraged from direct ly operating water services. They were instead encouraged to delegate services under contract to a private operator or formal associations of users (WSP, 2014). Fiscal Decentralization In Mali, local governme nts budgets are composed of : (i) local tax revenues collected with the help of the tax authorities; (ii) government budget transfers (solidarity subsidies to make up for regional disparities); and (iii) investment grants through the National Local Governme nt Investment Agency (World B ank, 2008 ). Because most l ocal governments have a very low tax base relative to needs, this leaves a wide gap in funding to be filled by central governments transfers, investment grants, and donor finan cial support (USAID, 1998). Central government transfers are set on an ad hoc basis and do not provide a predictable source of funding, nor is the distribution among local governments transparent. Further, local taxation parameters are determined by the national assembly and collect ed by deconcentrated central government offices (The Cities Alliance & UCLG Africa, 2013). A share of Heavily Indebted Poor Countries ( HIPC ) resources is allocated to local governments (about 15% in 2003), also through deconcentrat ed central government o ffices ( IMF 2002 ). Looking at the overall funding situation, a World Bank study found the transfer of resources to local

PAGE 166

153 governments to support the decentralized transfer of sector responsibilities insufficient (World Bank, undated/c). The 2008 Water Ai d report profiled local level financing for water and sanitation in 15 low income countries in African and Asia making it pos sible to compare Mali with s imilar income countries Mali ranked relatively high ting a large share of donor and non governmental organization projects, and low funding from the central government, highlighting the unpredictability of central government transfers to local government Considering all internal and external sources, Mali was in the lower half of the 15 country ranking on per capita spending on water (WaterAid, 2008). To illustrate the predominance of off budget financing, (those sources that do not appear in the national budget), a breakdown of sector financing sou rces f or the year 2006 appears below in Table 8.3: Table 8 .3 Sources of Funds for Municipal Water and Sanitation Investment in Mali (2006) Source of Funds Percent Off Budget (bypass ing national and local budgets, such as direct donor projects) 55.9 Local Government (from local gov ernment revenues and inter governmental transfers) 24.9 Sect or Projects ( national government projects or donor projects included in the national sector budget) 13 Community Water Service Provider 6.1 S ource : WaterAid, 2 008 p.47

PAGE 167

154 The poorest municipalities are particularly dependent on outside sources, including off budget financing and central government transfers. However, b udget allocations do not reflect differences in need, so p oorer municipalities are given little in the way of supplemental financial support (World Bank, 2008 ). While t here is a national transfer system called th e equalization fund in place to provide varying amounts to different locales based on local g overnment characteristics, these transfers are relatively small and often late, thereby not prov iding a reliable source for poorer municipalities to plan investment projects (The Cities Alliance & UCLG Africa, 2013). In 2011, Mali passed a new law to adjust the fiscal resources availabl e to sub national levels of government, acknowledging that the current system provided insufficient financial resources for them to carry out the responsibilities tran sferred under decentralization. The new law was designed to address we ak revenue raising ability at the local level, attributed to several circumstance s, including : lack of skills is this area among locally elected officials ; insufficient human and financi al resources for tax collection; and lack of penalty application for unpaid taxes (Direc tion Nationale des Collectivits Territoriales & Lux Dev, 2013). The legal changes clarify the roles of different government actors in tax collection at the local level, introduce a real estate tax to benefit the local lev el of government, and reinforce procedures for tax collection and enforceme nt of the tax code. They also attempt to address what was recognized as insufficient communication and technical support between national level agencies and local government (Direction Nationale des Collectivit s Territoriales & Lux Dev, 2013). These legal changes help to counter what had been identified as a major impediment to decentralization in Mali. A researcher at the University of Bamako

PAGE 168

155 evaluated at the decentralization experience in Mali for educati on, water and health services for the period prior to 2010. Among the conclusi ons drawn were that the complex tax system set up under decentralization did not conform to the realities of the country and was not well understood by either tax payers or tax col lecting authorities, hurting local autho rities in their ability to fulfill responsibilities under decentralization (Sidibe, undated). Major Donor Projects As highlighted above, d onor financial support has been important for addressing the water and sanit ation access challenge in Mali. The World Bank conducted three major projects from 1979 to 2005 aimed at improving urban services and supporting decentralization (Farvacque Vitkovic et al., 2007). One project from 2001 to 2005, with an investment of $80 million, yielded substantial results in targeted cities, especi ally the major city of Bamako In reviewing these projects, the World Bank considered some key challenges remaining to be lack of financia l resources at the local level and lack of management capa city, especially in medium and small sized cities Similarly, a 2007 assessment by the African Development Fund found a lack of participation and ownership of donor projects by the local population, largely attributed to limitations in financial and h uman resources for proje ct oversight. Roles and Responsibilities Following the transfer of author ity to local governments for several funct ions, including water there was resistance on the part of some central government ministries to implement the decent ralization reforms (USAID, 1998), and the division of responsibilities among different levels of government was unclear at times (The Cities

PAGE 169

156 Alliance & UCLG Africa, 2013). Overlapping roles and responsibilities between various institutions was a key const raint to sector development notably for sanitation where responsibilities were shared among the Ministry of Mines, Energy and Water, the Ministry of the Environment and Sanitation, and the Ministry of Health (African Development Fund, 2007; USAID, 2010a ). The limits of super vision for the national level were not clearly defined and, therefore, the level of autonomy for local institutions was unclear (Coulibaly, 2010). Monitoring and Accountability No programs could be identified directed towards monitoring sub national performance in provision of pu blic services, nor any laws requiring such assessment by the national government. The local government code requires independent audits of local government on a regular basis but, in practice, these ar e done infrequently (The Cities Alliance & UCLG Africa, 2013). Water Aid saw a need for ongoing evaluation systems for newly constructed water and sanitation projects, and UNDP found monitoring of coverage in the water sector to be lacking in a country ass essment report, although it made note of plans for an improved national monitoring system (UNDP, 2009). Capacity Building From the start of decentralization in the mid 1990s, local communities were limited in their ability to take on new responsibilitie s by a lack of capacity and training at the commune level (World Bank, 2008 ; UNDP, 2009 ). The national government has offered some technical assistance progra ms and advisory su pport (Sebetao, 2015), with additional support coming from donor pro grams in re cent years. The W orld B ank Country Assistance Strategy for 2004 to 2006 emphasized the building and maintaining

PAGE 170

157 of infrastructure with the participation of decentralized authorities and building the capacity of such authorities (Farvacque Vitkovic et al., 2007). In 2009, USAID began the Shared Governance Program 2, with the objective of improving local government capacity to deliver basic public services s uch as water and sanitation Most recently, a $64 million project supported by the Danish and Swedish Governments was launched in 2010 for sanitation offering technical assistance to sub national governments and promoting greater public involvement in national policy form ation (Hanjahanj a, 2011) The national government has committed to develop and implement a human resources strategy for the water and sanitation sector by the end of 2015 ( SWA, undated ). Private Sector Participation The transfer of water and sanitation responsibilities to local governments was accompanied by a push for private sector involvement. Since the Water Law in 2002, the central government has promoted delegation to formal users associations or private operators under contracted to municipalities ( WSP 2014). T he 2002 PRSP also emphasized private sector participation (African Development Fund, 2007; Farvacque Vitkovic et al., 2007 ; IMF 2008). However, as recently as 2007, the African Development Fund characterized the capacity of the private sector to participate in service provision as weak. A major service provider has been Ele ctric and Water Company of Mali (EDM) a semi private water company owned by the Mali Central government and Aga Khan Foundation, which has operated in about 16 urban areas und er a long term contra ct with the central government The progress of EDM in making new connections has been closely linked to the availability of subsidies offered to consumers There was a

PAGE 171

158 substantial drop in the number of new connections between 2002 an d 2004, as EDM cut the level of subsidies offered to customers getting new connections in low income areas (Farvacque Vitkovic et al., 2007 ; USAID, 2010a ), suggesting a willingness to pay or ability to pay component. EDM was returned to strict public mana gement in 2005 UNDP later observed that EDM ability to expand had been restricted by the scope of its limiting it from operating on a commercial basis (UNDP, 2009). As a results, the private sector is not a major participant in financing the sector (Sebetao, 2015). Sector Planning There are several examples of Mali conducting sector planning exercises that involve multiple partners and long rang e planning. The country developed a national drinking water supply and sanitatio n development strategy in 2000, followed by a National Plan for Access to Drinking Water cover ing the period of 2004 to 2015. This drew on consultations with technical and financial partners that were h eld in 2004. The p lan included specific projects to build and renovate water infrastructure at the regional and sub regional levels, towards a goal of achieving 75% national coverage by 2015 (Farvacque Vitkovic et al., 2007). In 2006, the p lan was used as the basis for the national Water and Sanitation Sector Prog ramme (PROSEA) The development of PROSEA was supported by The World Bank, aimed at improving access fo r the most disadvantaged groups ( World Bank, 2008 ; African Development Fund, 200 7). PROSEA put in place a unified comprehensive national program and budget framework for the sector. P ROSEA was acknowledged as having improved coordination among donor s in the sector (Sebet a o, 2015; WaterAid,

PAGE 172

159 2008) whereas prior to 2006, donor projects employed different approaches and execution methods, without much inter connectivity or coordination at the country level ( World Bank, 2008 ; African Development Fund, 2007). A further step towards coordinating donor aid was taken in 2008 when a s pecific donor coordi nation strategy was developed. The first comprehens ive SWAp was published in 2 012, subsequently updated with five new pledge s in 2014, including a commitment of 5% of the national budget to the sector, strengthening of the national mon itoring system, and creating a human re sources strategy for the sector ( SWA, undated ) The updated SWAp represents a major step forward in coordinated sector planning and seeks to address some of the identified weaknesses in the national program. The 20 02 Mali PRSP ha d decentralization and equitable development as goals, including initiatives to help underdeveloped regions of the country advance by building infrastructure and facilities (Farvacque Vitkovic et al., 2007). It was used as a framework for a ll medium term policies and as a basis for negotiating with technical and financial development partners ( IMF, 2002 ). Updates to the PRSP have been more specific, including detailed objectives for the water sector for the period 2007 2011 (Farvacque Vitkovic et al., 2007). Annual g oals were set to increase water access nationwide with a specific reference to central government disengagement in favor of local government authority (I MF, 2008 ). A related document, the Budget Program by Objective for the period of 2008 10, specifically estimated what means would be required to meet the MDG objectives by 2015 (World Bank, 2008 ).

PAGE 173

160 Another recent planning initiative specifically addressed the sanitation sub sector. In 2008, the Ministry of Environment and Sanitation developed a National Strategy to Transfer Capacities for Sanitation which, among other things, clarified responsibilities in the sub sector (UNDP, 2009). An action plan was p repared in 2010 that prioritized needs for sanitation and made plans for progress evaluation inclu ding regular data tracking (Hanjahanja, 2011). Civil Society Engagement During the 1990s, Mali experienced a large growth in the number of NGOs. The numbe r registered with the government increased from about 50 in 1991 to 600 in 1998 (USAID, 1998). By 2011, there were more than 1200 registere d CSOs, and numerous informal and religious groups not included in that number. Many are engaged in providing socia l services in the areas of health, education, water, and sanitation. Generally, these CSOs rely heavily on donor funds and many suffer from internal financial, organizational and structural weaknesses that impede their growth (USAID, 2011). Decentralizat ion has, however, provided CSOs with access to some local resources allocated through the Economic and Social Development Program of the PRSP. From 2000 on, the national government had in place frameworks for cooperation with CSOs and held an annual consultation with CSOs in most years. The general climate for civil society in Mali was assessed in a 2011 USAID report that evaluated 23 sub Sahara n African countries. Mali ranked favo rably in three areas that most closely coincide with my constructs for civil society engagement:

PAGE 174

161 Legal Environment for CSOs to freely particulate in policy activities: 5th out of 23 CSOs role in Advocacy including l ines of communication and opportunities to collaborate with policymakers: 9th out of 23 Infrastructure for CSOs around the country to share information and network: 4th out of 23 CSOs are represented on several national advisory boards and monitoring gro ups, including those for economic and social development There are no legal restrictions on the ability of CSOs to particulate in civic education, governance, and anticorruption and to participate in the development and implementation of public policy (U SAID, 2011). For the water and sanitation sector, joint sector revi ews have been held since 2007 that include multipl e stakeholders. CSO advocacy has had a major impact on nati onal policy (Sebetao, 2015), as input from organizations such as the NGO Coalition for A ccess to Water, Sanitatio n, and Hygiene have increasingly been taken int o consideration. Several of the most active CSOs in the sector have received international donor funding. Advocacy activities by civil society resulted in creation of a civil societ y watchdog group that has studied sector impediments and made recommendations on topics such as financing and sanitation that have been adopted by the government (WaterAid & Freshwater Action Network, 2011) At the local level, while there are legal provisions which require municipal councils to consult with community leaders such as neighborhood chiefs or village water chiefs before taking action of certain subjects of community concern (USAID, 2010b; The Cities Allianc e & UCLG Africa, 201 3), Mali lacks any legislative requirement for local governments to consult more broadly with citizens before undertaking spending

PAGE 175

162 decisions on public services, such as water and sanitation (The Cities Alliance & UCLG Africa, 2013). One outreach program th at began nationally has since been duplicated in some local communities. Dialogue Day held at the national level since 1992, brin gs citizens and government together to discuss water and sanitation issues. Q uestions and comments are solici ted through o ut reach to the community screened though an independent group, and then provided to local authorities. A public hearing is held a short time later, chaired by a neutral party, during which the local council responds to the questions and comments. Not all local communities participate in the program, nor are they required to do so. As of 2011, hearings were held in about 25 communities. A secondary benefit of the hearings is that citizens feel more comfortable approaching local authorities on water and sa nitation issues during other times of the year, and it can also increase the local funds are spent (WaterAid & Freshwater Action Network, 2011) The Government of Mali h as recognized the contributions of CSOs to community development and social services and the role some play in water and sanitation service provision. In some locales they are provided with government funds to implement pr ograms, such as resources allocated through Economic and Social Development Programs, part of the PRSP process (The Cities Alliance & UCLG Africa, 2013). Wh en building infrastructure, CSOs can recover up to 20% of their costs by allowing beneficiaries to participate in the constru ction and the monitoring of the project (USAID, 2011).

PAGE 176

163 Mali Case Conclusion Here I present some general conclusions from the Mali case study In the next section findings related to the decentralization constructs for both cases are synthesized and f urther a nalyzed From 1990 onward, Mali saw a large increase in urban water access, from 53% in 1990 to 91% in 2012. For urban sanitation, the increase was only 33% to 35%. Both trends are contrary to the cross case analysis results, which found high d ecentralization associated with low water achievement and high sa nitation achievement. H owever, the case study reveals other circumstances and background conditions the help to explain the unexpected variable relationship are as follows: 1) Mali does not exhibit the criteria for decentralization to a large degree, as per my constructs and the comparative indices; 2) the large achievement in water coverage is mostly due to large donor projects in urban areas that are independent of decentralization; and 3) country achievement benefited f rom sector planning exercises, donor aid coordination, and civil society input to policymaking, independent o f any decentralization impact. Here, I further explore point 1) from t he last paragraph and attempt to validate M composite decentralization values derived for the quantitative analysis, ( 0.83 for water and 0.75 for sanitation ) The values seem high given the findings of the case study relating to the specific elements that were captured by the GLAAS data: serv ice provision decentralization, operational decision making autonomy, and funding made available to local level governments for carrying out this responsibility. Particularly the last element, relating to funding availability to local governments, is not p resent to a large degree in

PAGE 177

164 Mali. Also, o ther rankings place Mali at or below the mid range on certain decentralization measure specific to the water sector. Overall, the fact that the composite decentralization values for Mali derived from the GLASS survey are not well supported by other information illustrates the difficulty of capturing nuanced information through a general survey instrument such as GLAAS, and also the fact that the survey, an d therefore the composite values, did not incorporate the constructs of clarity of institutional roles, monitoring, and capacity building. However, r ather than bring into question the general validity of all country composite values derived from GLAAS, th e fact that this case demonstrated a relationship contrary to that of the overall model results, (by having high decentralization associated with high achievement), is, in reality, l ower. A situation of low decentralization and high achievement is consistent with the overall model results. In many ways, the lack of full realization of decentralization in the water sector mirrors that of the country as a whole. A 2013 local governm ent assessment report found that many of the responsibility transfers planned under the 1996 decentralization law had not been fully realized (The Cities Alliance & UCLG Africa, 2013). Also, w ith what we now see as the limitations that accompanied decentralization in Mali, particularly the lack of financial resources available to local com munities, it is not surprising that sanitation was not widely advanced during this period, as it is often a given lower priority than water S anitation planning o nly became a focus in 2008 with preparation of a national sanitation strategy

PAGE 178

165 Summary and Analysis of Decentralization Cases This section provides a summary and analysis of information presented in both case studies, beginning with Table 8 .4 that lists key devel opments for each construct by country along with general observations. Afterwards, a written analysis draw s conclusions about how these develop ments impacted progress in the two countries. Table 8 .4 Decentralization Case Study Findings by Construct Colombia Mali General Observations Decision Authority INSFOPAL (a national company) is dissolved shifting responsibility for provision to medium and small cities ; d epartmental governments take over some entities previously with INSFOPAL Central government sets policies; specific authorities are spread among several Ministries and agencies Decision autonomy and authority to conduct plan ning granted to municipalities ; m unicipal governments can define levels of cross subsidies Recent national government actions su ch as requiring proj ects to undergo national review and encouraging departmental level planning are seen by some as a trend towards rec entralization New constitutions provides principles of d ecentralization Political reforms transfer water responsibilities to sub national governments Local governments are encouraged to contract services to private operators or formal users associations While respon sibility for services is legally placed with local governments in bo th countries, programs were subject to oversight by the central government, often unde r a confusing array of agencies In the case of Colombia, municipalities have been encouraged to participate in regionally planned programs, but this approach has not s hown wide spread success

PAGE 179

166 Colombia Mali General Observations Sources of Revenue More authority given to municipalities to administer and modify taxes and perform public functions (Constitution, 1991), however, large urban centers have greates t ability to raise tax revenues Central government transfers to municipalities set to increase a fixed percentage from 1993 to 2005 SGP revenue sharing system is created to consolidate transfers; it includes specific transfers targeted to poor and less populated mu nicipalities Local governments rely predominantly on central government transfers and donor aid (including HIPC resourc es), due to low local tax base Central government tra nsfer amounts are unpredictable Funds deemed insufficient for carrying out water and sanitation responsibilities Complex system of taxes setup under decentralization was not well understood and ineffective for raising revenues at the local level; new tax system introduced in 2011 attempts to clarify roles and communicat ion and improve tax collection Comparing the two countries, we see the advantage for municipalities in Colombia that had a guaranteed amount of funding for the sector from the central government, allowing for long range investment planning In both countries, municipalit ies that lack a sufficient local tax base remain overly dependent on central government transfers Poor and less populated municipalities in Colombia be nefited from targeted transfers Particularly in Mali, many small municipalities lack sufficient finan cial resources to fulfill the respo nsibility for service provision Investment Decisions Law requires that 20% of transfers received from central government be spent on the water and sanitation sector (Law 60, 1993) Municipal level given primary responsibility for deciding spending amounts, structuring investments, and Mali ranks relatively low on per capita spending on water Donor funding dominates as a source of funds for investment projects Lack of participatio n and ownership of donor projects by local population is noted, implying that investment The Colombia experience shows that fiscal design s that mitigate moral hazard problems are needed to promote efficient investment spending at the l ocal government leve l In Mali, due to the dominance of donor funding, municipalities only impact investment decisions to the extent that

PAGE 180

167 Colombia Mali General Observations executing investments (Law 60, 1993) Issues develop due to incentive problems and weak fiscal discipline at sub national levels of government Fiscal reforms modify incentives; increases in investment must now come from local sources Cross subsidies mandated, giving reduced fees to lower i ncome areas, subsidized by higher fees for other users Funds generated locally for the sector are better spent that central government transfers, as per study Municipalities encouraged to apply investment funds towards regionally planned initiatives M ore than 70% of sector investment decisions nation wide managed by municipalities (2010) More recently, projects using any amount of central government funds must be reviewed and approved at the national level, perhaps a trend towards recentralization decisions are made outside of the local community they are involved in planning and implementing these projects

PAGE 181

168 Colombia Mali General Observations Roles and Responsibilities Decentralization creates a clear separation of responsibilities: Central government to set policies for the sector Municipal governments responsible to ensure universal provision and quality Observers see overlapping responsibilities, lack of clarity in roles (as recently as 2015), and institutional conflict at centr al government level Division of responsibilities among central government agencies is un clear and sometimes overlapping In both countries, a situation of confusing roles and responsibilities ensued for a decade or more after the i ntroduction of decentrali zation Monitoring of Sub National Performance Observers see weak accountability at all government levels Lack of fiscal monitoring leads to expenditure and credit problems at sub national government levels World Bank report sees need for more monitoring and evaluation of the effi ciency of municipal investments Central government lacks the capacity to monitor the large number of municipalities There is no national monitoring program for local government s ector performance Monitoring of cover age and service quality of sub national governments found to be lacking Neither country has a sufficient monitoring system for overseein g local government performance

PAGE 182

169 Colombia Mali General Observations Capacity Building At the introduction of decentralization, municipalities had little support or skills to manage the responsibility Learning facilitated by case study examples and lessons disseminated, such as the Cartagena and Barranquillam joint venture contracts with private sector Small cities, especially, need supporting for selection of technologies, and some recent initiatives, such as software, have been targeted to those needs In early stages of decentralization, there was a lack of training and capacity buildi ng for local governments, b ut this has increased over time Beginning in 2004, donors support programs for capacity building, including a major one focused on sanitation In both countries, responsibilities are handed to local authorities prior to a ny capacity building of technical and managerial skills. It is often a decade or so later before any large scale capacity programs are in place Lack of knowledge of technologies is a big challenge fo r smaller cities in particular Private Sector Partic ipation Efficiency of service provision emphasized, competition encouraged, and private sector participation promoted All existing public services entities are to convert to legally autonomous stock companies, whether private, municipal, or mixed ownership Municipalities given power to grant concessions to the private sector Private sector participation is promoted since the 2002 Water Law and the PRSP There is weak capacity and presence of private sector operators, with the exception of EDM, which is co owned and subsidized by the national government, operati ng predominantly in urban areas In both countries, the push for private sector participation coincides with or closely follows decentralization reforms. However, both a lack of knowl edge about technologies and financial aspects makes contracting to the private sector challenging particularly for small cities There is no clear evidence that private sector engagement in itself leads to increased coverage, except in the case of governm ent sub sidized new connections in Mali

PAGE 183

170 Colombia Mali General Observations World Bank Project (2001 2010): builds national capacity to promote and supervise private operators finds private sector models have been more effective in medium size cities than in small cities Empirical studies of Colombia show no clear relationship between private sector eng agement and increased coverage Sector Planning A SWAp is developed in 2007, but mostly for water resources management There does not appear to be a comprehensive national plan for water and sanitation services, however there are some plans at the d epartmental level of government Planning exercises and documents occur beginning in 2000, incorporat ing stakeholder and donor input: PRSP (2002) and updates include g oals for increased water access National PROSEA plan is developed (2006) and donor coordination among donors Plan to coordinate donor aid is developed (2008) S WAp for water and sanitation is devel oped (2012) and updated (2014) A specific sanitation sector strategy is developed (2008) A series of planning initiatives in Mali were avenues both for stakeholder input and b etter coordination of donor aid

PAGE 184

171 Colombia Mali General Observations Civil Society Engagement Civil society is not very engaged on water and sanitation issues in Colombia, or on social issues in general Big cities, more so than small cities and rural communities provide input on sector issues No prominent national civil society group for water and sanitation issues No consistent national government procedures for consulting with civil society Country rates relatively high among Sub Saharan countries on havi ng a favorable climate for CSOs Annual sector specific consultations are held with CSO groups an d they serve on advisory boards Civil society is consulted in PRSP, SWAp, and PROSEA preparation CSO advocacy h as had a major impact on policy Mali has regular forums for consulting with civil society and has done so for multiple planning exercises Colombia is much less developed in the area of civil society engagement for the sector Drawing on this comparative analysis, the following sections discuss conclusions that can be generalized to a larger population of countries. Decision Authority In both countries, decentralization reforms clearly placed responsibility for water and sanitation service at sub national levels of government and provided them with decision authority, however their ability to exercise this authority effectively depended on a number of other factors that were not always in place. When Colombia decentralize d in 1993, many of the larger cities had already maintained their own service companies, but it posed a significant challenge for small and medium sized municipalities. Subseque ntly, in organizing services, they suffered from a lack of capacity, particular ly in the early years. Their authority in making managerial decisions for the sector remained consistent

PAGE 185

172 and extended to decisions about private sector contracting, investing in new service areas, and designing cross subsidy programs. While not required t o, they were encouraged in 2007 to participate in regionally p lanned projects and investments, and pr ojects using central government funds are now subject to review by a national level board. Both of these developments, while seeking to improve investment efficiency, also somewhat diminish the decision authority of local government. In Mali, the delegation of authority and responsibility for service provision was in place, but its execution was challenged by f inancial limitations In effect, most dec isions about services in their area were being made by donor s, with varying roles in decisions by local communities. Sources of Revenue The cases show that the origin and consistency of funding greatly influence how municipalities are able to plan for an d execute investments. From the early stages of decentralization in Colombia, authority was given to municipalities to raise revenue in the form of taxes and tariffs. Revenue policies underwent some early refinements which served to strengthen the positi on of municipalities. Dating back to 1993, they had a legally guaranteed allocation coming from the central government for the sector, with specified increases by year. Both of these elements would have provided a degree of financial security to municipa lities and encouraged long term investment planning. For poorer municipalities, the 1993 modification that provided them, (on average), with a high per capita allotment of funding, would have provided increased resources to extend access to unserved areas The addition of other specially targeted transfers to poor and

PAGE 186

173 low population municipalities under the SGP transfer system (2001) would have gone further in this regard. In Mali, the overall lack of financial resources for this extremely impoverished country impact municipalities in two ways: lack of a sufficient local tax base to raise local revenues and lack of reliable transfers from the central government. Consequently, municipalities are limited in their ability to implement their authority for s ervice provision and expanded services to unserved areas, and this often is realized through donor projects outside of government budgets. F indings can be summarized as follows. For central government funds, knowing that they are subject to payment fo rmulas and guaranteed over time allows for more eff ective long range planning. When the re is a reliance on donor funding, municipalities lack a guarantee of long term fun ding and often feel less ownership of projects The greatest ability to control revenue for investments comes from local funding sources, but for small or low income municipalities the ability to raise funds through a local tax base may be limited. Expenditures from local resources may also be more efficient than those using other fu nd sources Investment Decisions Turning to authority over spending and investments, the situation regarding revenue sources often dictates how investment decisions are made as well. The reality in Mali is that investments in the sector are funded by, a nd therefore largely determined by, donors. To the extent that municipalities are consulted and involved in investment planning, as well as represented at the national level in donor coordination and planning initiatives, they will play an increasing role directing investment s decisions, even if they

PAGE 187

174 are not the primary source of investment funds. This will strengthen their ability to assume this role in the long term. In Colombia, as the 2010 breakdown of investment by government level shows, the central government makes less than 10% of investments in the sector, demonstrating strong realization of decentralized spending. T he national policy mandated that a minimum of 20% of transfers from the central govern ment go to water and sanitation. In one way, t his limited the freedom of local governments to set their own al location, but ultimately it helped the sector not to be crowded out by spending on other key sectors, such as health and education. Similarly, the mandate for cross subsidies removed that dec ision from the municipal government, although it did allow them the freedom to design program specifics. Again, this stipulation was likely an effective strategy towards the goal of increasing access to poor areas The overspending and excessive debt by some departmental and municipal level governments in Colombia underscore the importance of monitoring and oversight This was subsequently addressed with more restrictive fiscal guidelines, however additional approval requirements related to spending can also have an effect of constraining local authority. Roles and Responsibilities More so than a confusion of roles at sub n ational levels, both countries experienced confusion and overlap of roles at the central government level, which meant municipa lities were accountable to multiple parties and may not have known where to turn for need ed information and resources. In Mali, ini tial hesitation on the part of so me

PAGE 188

175 central government ministries to adapt to decentralization and new rules created obstacl es to development. In Colombia, the early stages of decentralization were characterized by a variety of rules and directives from diff erent central government agencie s, which may have been contradictory at times. This shows the importance of countries l imiting the number of regulatory agencies with sector authority establishing clear roles and authorities from the beginning, and making sure municipalities understand these roles. Monitoring of Sub national Performance There was a lack of monitoring of sub national government performance in both countries, particularly in the early years of decentralization. Oversight of the fiscal aspects of the decentralized sector in Colombia had been weak and was subsequently addressed with tighter guidelines. Ot her types of monitoring may be still be lacking. I t is unclear that sufficient programs are in place, either at the municipal level o r through the national level to monit or service quality and equitable access by different socio economic groups. In Mal i, there is also a lack of capacity for monitoring of performance of sub national levels of government, limited by the lack of financial resources that significantly constrain the ability to implement programs. Having monitoring programs in place from th e start of decentralization, to better track progress and needs for support among municipalities would have avoided some of the difficulties faced by municipalities Disparity in technical resources and financial capacity among municipalities can be identified and better targeted, (by government and donors), to advance service coverage at a faster rate.

PAGE 189

176 Capacity Building Both countries entered into decentralization with sub national authorities largely unprepared for the managerial, technical, and financial, tasks that would be required to assume responsibility for service provision. Perhaps it is because both countries envisioned these expertise coming from private sector entities, the contracting of which was encouraged in both cases, but has not really taken hold on a large scale. Donor programs have been established in Mali to address c apacity building needs, but mostly in recent year s. In Colombia, capacity building programs have been primarily in the form of case studies and pilot projects rather than wide spread training programs The fiscal management problems encountered in the early stages of decentralization show that more general financial capacity building early on would have been valuable. The experience of these countries shows that skills development for local government agencies is crucial. Even if the private sector is en gaged in some way, local government needs a range of skills to be effe ctive in developing and monitoring programs. Private Sector Involvement Contracting of the private sector, promoted by national government policy in both countries, has so far proven most ef f ective for medium sized cities and less so for smaller cities, perhaps due to a smaller pool of technical and managerial resources from which to draw (many larger cities had already had their own companies). Going forward, private sector provide rs who have gained experience could benefit those municipalities that lag behind in achievement, b ut more training is needed for municipal governments to succe ssfully enter into these arrangements. The impact of private sector involvement on

PAGE 190

17 7 achieving higher access should continue to be studied as there may be a stronger effect over time Private sector experience in Mali is more limited, dominated by t he exper ience of EDM. There have not yet been sufficient programs or pilot projects to build capacity of domestic private providers and there is no significant presence of international companies. User associations have largely stepped in to fill the void for ma ny small and rural communities. Conclusion of Case Studies These cases provide insight the practical implementation aspects of water and sanitation decentralization. We see what approaches have worked well and which have yet to be fully realized due to financial or capacity limitations, thereby limiting the potential for decentralization to fully support country achievement i n water and sanitation coverage. Lessons learned from the more than twenty year sector decentralization programs in Colombia and Mali can be of value to other countries that are beginning or exten ding decentralization reforms. Different Decentralization Impact on Water and Sanitation Drawing on the case study findings, I now address the broad question posed at t he end of Chapter V concerning the impact of decentralization: Decentralization was shown to be highly significant (p < 0.01) for predicting Water Progress with a negative coefficient, and also highly significant for predicting Sanitation Progress with a positive coefficien t. The same coefficients and direction were observed for the A 2012 variables, but with a lower level of

PAGE 191

178 significance (p < 0.05). Considering these findings, what are some possible explanations for why decentralization impacts the two sub sectors differen tly? I offer below three proposition s for the difference the examination of which could form the basis for additional research: Proposition 1: Incomplete Realization of Decentralization: In many countries, decentralization has not been f ully realized, to encompass all aspects as contained in my constructs, such as sufficient revenue streams and technical and managerial capacity of the part of local governments. Therefore, even those countries with high composite decentralization value s derived for thi s research, such as Colomb ia and Mali, are not fully realizing sector decentralization as defined in the literature and adapted for this research So, rather than concluding that decentralization is associated with a lack of water access achievement, it can be said that incomplete decentralization is associated with a lack of achievement Had one been able to comprehensive ly examine decentralization constructs for multiple countries as was done in the case studies, more countries might have measured lower on decentralization and we might have seen a different statistical relationship emerge between dec entralization and achievement. Proposition 2: Sanitation Only Recently Became a Development Priority S anitation development has historically received less attention and been viewed a s less of a priority than water, both globally an d in a domestic contex t, and was given less coverage than water in PRSP s ( WSP 2003; Mehta, 2002; Newborne 2004) In addition, t he three key explanator y variables each had lower mean composite values for sanitation than for water, by about 0 .10, (refer back to Table 5.2 ) and average sanitation access

PAGE 192

179 rates in both 2000 and 2012 were about 25% lower than water access rates in the sample group (refer back to Table 4.3). If we take the case of Colombia, some explanation for the difference is that more people within national government understood different water technologies, whereas t hey may hav e only thought about sewa ge systems (a relatively high cost options) when it came to options for sanitation ( Rojas Interview Summary, 2015) The lack of political will to focus on sanitation may have also stemmed from a perceived lack of import ance on the part of the population as compared to water Interviewees in Colombia generally felt that demand from the population is greater for water, and therefore more reso urces are allocated to increasing water access ( Rojas Interview Summary Ramirez Interview Summary, 2015) Even in urban areas, the need for sanitation had not been as well understood by the public and they would question why they should pay for sanitation ( Rojas Interview Summary, 2015) reduction has been increasingly recognized and promoted in the global development community particularly since the mid 2000s, at a point when many gains in water had already been realized (Newborne, 2004 ) Chapter IX will discuss how sanitation programs became a priority in Kenya around 200 7 and in South Africa in 2011, and we saw that Mali developed their sanitation strategy in 2008. Even i n Colombia, sanitation did not really become a governmen t priority until the early 2000 s largely as a result of do nor emphasis, and with increased awareness of different sanitation options ( Rojas Interview Summary, 2015) So, country gains in sanita tion f rom 2000 to 2012 may have had more to do with evolving development priorities than with decentralization programs.

PAGE 193

180 Proposition 3: Sani tation is Easier to Address at the Local Level of Government Sanitation most often appears in national budgets on the same budget line as water, and it is often dealt with a s an add on to water planning. While a few countries, such as Mali, have set up separate working groups and planning exercises for sanitation, this was rare in the period of 2000 to 2012 (Newborne, 2004 ) Decentralization may have provided an opportunity for municipalities to fund small scale sanitation service providers well suited to small cities and low income areas of large cities, something that would have been too minor in scope for a program managed by the centra l government Kenya has several examples of micro service providers successfully supporting a range of sanitation options in low income urban areas (Eales, 2008; Tukahirwa, Mol, & Oosterveer 2013). Moreover, t he nature of sanitation service may be such that local g overnments, once given the authority /resources to manage it, could develop appropriate infrastructure options (including alternative s to piped sewa ge systems), without a great deal of central government technical and managerial suppor t Depending on the technology sanitation also can require less regional cooperation than managing a shared resource such as water, and there may also be less advantage from economies of scale for large cooperative infra structure projects. The next chapter present s the tw o country case studies focused on civil society engagement and end s with a summary of key points, similar in format to this chapter.

PAGE 194

181 CHAPTER IX CIVIL SOCIETY ENGAGEMENT This chapter aims to provide deeper insight into different ways that civil society engagement can be implemented for the water and sanitation sector, and in what ways these approaches achieve the objectives for civil society engagement detailed in the constructs as follows : e) Opportunities exist for consumers and civil society gro ups to become involved in problem articulation f) Opportunities exist for consumers and civil society groups to become involved in decision making about policy development and planning g) Opportunities exist for consumers and civil society groups to become involved in decision making about implementation and management h) Sector operates with transparency and information disclosure including for use of funds and natural resources e) There is a feedback/complaint system in place The two countries to be studie d are Kenya and South Africa which both exhibit re latively high composite value s for civil society 32 and have introduced specific civil society engagement initiatives to the sector since the year 2000. Their experience s with differe nt engagement techniques are chronicled along with feedback about the e ffectiveness of the initiatives, so that other country programs may benefit from the experience. In addition to looking at civil socie ty engagement, some coverage is given to decentralizat ion and sector planning, so the reader may appreciate to what extent these 32 Kenya 1.0 for water and 0.75 for sanitation; South Africa 0.75 for both water and sanitation

PAGE 195

182 attribute s interacted with civil society engagement. At the end of the chapter a compilation analysis is done using findings from both countries to evaluat e the constructs and approac hes, drawing insights as possible. Kenya Case Study Kenya is classified as a low i ncome country by the World Bank. Even though annual GDP per capita almost tripled between 2000 and 2012, the 2012 figure is still less than $1,200. Kenya historically w as divided into eight provinces headed by a provincial commissioners, but a new c onstitution in 2010 divided it into 47 semi autonomous counties headed by elected governors T oday, t he county level of government oversees many services, including water and sanitation, as further explained later in this section. the total country population in 2012 (World Bank, 2000 2012b) It experienced significant urban migration from 2000 to 2012, a 23% growth rate that places it in the top 20% of the sample countries for growth Much of the growth occurred in informal settl ements, such as those adjacent to the capital city of Nairobi. These settlements often developed without any permanent infrastructure in place composite civil society value s used for the quantitative analysis, I looked at comparative country rankings. Kenya ranked near both the mean and median values for the previously mentioned CIVICUS EEI among the 40 countries in my sample with CIVICUS EEI scores. However, for the 2011 USAID CSO Sustainability Index, which overlapped with 20 countries in my sample and had an overall higher correlation, Kenya ranked the second highest, surpassed only by South Africa. This second comparison, more so than the first, seems to validate the relatively

PAGE 196

183 high composite civil society values derived for Kenya from the GLAAS dataset (1.0 for water and 0.75 for sanitation). General Climate for Civil Society A 2012 report from USAID found approximately 6,500 active non governmental organizations ( NGOs ) registered in Kenya. This is in a ddition to more than 300,000 community based organizations ( CBOs ) which operate as informal groups or without official status. In general, civil society organizations (C SOs ), which include b oth of the above group types are able to opera te without state interference They increasing ly are engaged in l obbying activities and provide input into the passage of legislation and the development of policy T he right to participate in formulating laws and policie s is stated in the constitution However, there a re no specific laws guiding public participation requirements, and the extent to which civil society is invited to engage fo r specific sectors is subjective and depends on the issue (USAID, 2012). For the water and sanitation sector, civil society is see n as a key partner in development of policies (Opiyo Response, 2015). Today, there are several CSOs active at the national level on water and sani tation issues. Water and Sanitation Civil Society Engagement sanitation sector was the passage of the Water Law in 2002, coinciding with a national civil awareness campaign that led up to the 2002 general election. Through the public awareness campaign, citizens became more aware of their rights and demanded greate r participation in governmental affairs. Civil society groups increasingly focus ed on rights advocacy and ensuring government fulfillment of their responsibilities for supplying water and sanitation extending CSOs

PAGE 197

184 beyond their customary role in community development work (Nyamu Musembi & Musyoki, 2004 ). The Water Act of 2002 included provisions to promote civi l society engagement, such as requirements for consulting with the public on m ajor decisions and for placing members of the public on governing boards (ORGUT Consulting, 2010) It als o established roles for agencies at different levels of government, generated around the concepts of decen tralization and participation It required that the minister responsible for water f ormulate and publish in the Kenya Gazette a national water management strategy based on the outcome of public consultations (WSP, 2011; Nyangena, 2008). The Water Act covered both water and sanitation, and included a push for commerc ialization of public u tilities. The Act created the national W ater Services Regulatory Board (WASREB), which holds a wide range of responsibilities, several of which relate to consumers and civil society. For example, it is responsible for advising water providers on procedur es for handling consumer complaints, seeing that consumers are protected, and developing guidelines for tariff setting (iHub Research, 2012; WSP, 2011 ). WASREB shares some regulatory authority with the Ministry of Water and Irrigation, including oversight and reporting on service providers (WSP & AMCOW, 2011a). In association with WASREB, eight independent water service boards were established in different parts of the country, monitored by WASREB. The eight boards are responsible for establishing service provision agreements with local water companies, many of which a re publically owned ( WSP, 2011 ).

PAGE 198

185 A Water Services Trust Fund (WSTF) was established in 2002 to mobilize financial resources and provide assistance for water and sanitation investments in are as that lacked service coverage. The Fund has resource allocation procedures that prioritize low income areas, thus reducing political interferences in the allocation of funds (ORGUT Consulting, 2010). The WSTF is capitalized by various sources, includi ng national budget allocations and donor contributions, and is linked with the Water Service Boards in its effort to reach underserved populations. It is considered to have some of the best practices for citizen participation among the different sector sp ecific funds in Kenya (Kenya Human Rights Commission & Social and Publi c Accountability Network, 2010), and the project cycle system used by the Fund is considered to be transparent and consistent (ORGUT Consulting, 2010). Even w ith the establishment of s everal new entities following the Water Act of 2002, communication among different stakeholders was insufficient (Richards & Doering, et al., 2008) A few years later, new outr each tools were introduced that began to address that shortcoming. The ard (CRC) introduced in 2007, i s an innovative step towards engaging consumers in water and sanitation service management. Based on a feedback program for public agencies in Bangalore, India the CRC was suggested by the ter and Sanitation Program ( WSP ) to be piloted in Kenya based on a positive assessment of overall freedom of expression and st rength of civil society in the country ( WSP, 2011 ).

PAGE 199

186 The program was piloted in the three most populous cities in Kenya 33 with NGO s as coordinators, but involving a wide group of stakeholders i n the exercise. The objective wa s to collect consumer fe edback and report back to policy and d ecision makers in the sector, consequently providing an avenue for civil society to have input in to management and operations ( WSP, 2011 ; Eales, 2008). The CRC process entail surveys distributed to the public for rating the quality of water services, the results of which are used to identify gaps in servi ce (World Bank, undated/a ). T he exercise was particularly useful for highlighting weaknesses in the policy and legislation framework for urban sanitation. The lack of a coherent management plan for the sub sector was found to be detrimental to informal settlement areas in all three pilot cities (WSP 2011). A positive outcome of the CRC program was to strengthen overall citizen eng agement, perceiving their role as not just providing feedback on existing service, but also advocating at the national level for inc reased sector resources The Kenya Alliance of Residents Associations, the group that led the CRC in Nairobi, eventually formed KEWASNET, a network of water and sanitation CSOs KEWASNET membership grew rapidly, (from 10 CSOs in 2008 to 50 by 2011) with financial support from internationa l donors. I t continually increased its pu blic profile and the extent to which it was engaged in sector discussions with government ag encies and donors ( WaterAid & Freshwater Action Network, 2011 ). KEW A SNET sponsors a website and other resources for CSOs such as national level Learning Forums on a range of sector topics, some of which are open to non KEWASNET stakeholders, including government officials (WaterAid & Freshwater Action Network, 2011) 33 Nairobi, Mombasa and Kisum u.

PAGE 200

187 In reviewing the CRC experience in Kenya, the WSP recom mended that implementation of this approach involve policymakers, service providers, and citizen representatives from the start, and to have early support from a senior government agency that will consider the resulting policy recommendations. In addition they recommended establishing institutionalized avenues for citizen participation at different stages of the policy development and implementation process (WSP, 2011) In 2008, the central government, with support from the World Bank and the WSP, establ ished the Water and Sanitation Services Improvement Project. A key initiative under this program was the establishment of local institutions known as Water Action Groups (WAGs ), designed to engage the public in decision processes for identifying and solving water problems on a regular basis (iHub Research, 2012). WAGs are citizen committees with a range of responsibilities to facilitate interaction between water regulatory institutions and the pub lic ( WSP, 2011 ; Water Integrity Network, 2012; World Bank, und ated/a ). The WAG approach was piloted in 2010 in four urban areas 34 (WASREB, 2010). WAG m embers are appointed from urban resident associations ( WSP, 2011 ), and provided with two days of techni cal training by WASREB (World Bank, 2013). WAG m embers who work on a voluntary basis are required to introduce themselves to the community so that the public is aware of their role and activities (Water Integrity Network, 2012). Some of the key functio ns of WAGs relating to civil society engagement are: 34 Nairobi, Mombasa, Kakamega and Kisumu

PAGE 201

188 Dissemination of information, including to receive and disseminate information that is relevant to consumers from both WASREB and Water Services Boards Engaging with service providers to actions, with a view to improving customer orientation and accountability Facilitating continuous stakeholder involvement in water supply and sanitation services Monitoring and providing feedback to service prov iders on consumer experiences, and resolution of consumer complaints Provide information on policy development through annual reports and other mechanisms (WASREB, 2010, p.3 4) WAG outreach mechanisms include : a) focus group discussions about consumer ex periences with service; b) meetings that bring together service providers with community leaders to address concerns in a neutral, non antagonistic, environment; c) public hearings with participation by government institutions; and d) structured meetings a mong WAGs, utilities and regional water boards geared towards better understanding and issue resolution. A 2012 Water Integrity Network c ase information sheet stated that WAGs helped to deliver on their objective of consumer protection and soliciting con sumer input into sector planning and operations, strengthening the recognition of wate r as a human right In keeping with this consultative approach, the central government in 2012 invited CSOs to participate in Sector Working Groups convened in conjunction with development of the National Medium Term Plan covering the period of 2013 2017 These G roups included representatives from government, civil society, and other stakeholders, who work ed w ith specific Ministries in po licy development. The G roup for

PAGE 202

189 Environment, Water, and Sanitation had among their activities to devise a system for ongoing monitoring of water service delivery (USAID, 2012 ). Information Sharing and Reporting The national government shares information on pending policies and current activates in a variety of ways. The Ministry of Environment, Water and Natural and gives a notice for comments via a popular daily newspaper (Opiyo Response, 2015). D ifferent government entities involved with the sector have developed websites and some issue regular newsletters about operations. A 2011 report noted that the compre hensiveness of the websites varies by agency, and not all issue annual reports (WSP & AMCOW, 2011a ). WASREB annual reports detailing agency activities and finances are available, back to 2006, on the WASREB website and also at regional board offices In addition, Impact report s, dating back to 2008, provide information on the performance of water se rvice providers throughout the country, including coverage rates for both water and sanitation. The 2014 report covers 100 out of 101 formalized water service providers and ranks them on performance (WASERB, 2010; WASREB, 2014). Other monitoring and reporting activities in Kenya include quarterly and annual public expenditure reviews done by sector ministries, and a list of Sector Undertakings prepared following meetings. Generally speaking, inclusion of CSOs in service monitoring and o versight acti vities is very low. ( Opiyo Response, 2015). And, a ccording to a 2011 independent report, although the Ministry of Water and Irrigation conducts annual performance reviews evaluating key sector indicators, ongoing monitoring an d evaluation for specific

PAGE 203

190 plans and strategies is lacking such as for t he Pro Poor Implementation Plan. This Plan mandates that service providers create expansion plans to bring access to areas of greatest need (WSP & AMCOW, 2011a) One system is marking progress in track ing da ta related to the needs of low income areas. The MajiData website tracks water and sanitation service indicators, specifically for urban low income areas and provides opportunities for information exchange (WSP & AMCOW, 2011a). Its creation was an outco me of recommendations from the 2007 annual sector r eview meeting (MajiData website). Other communication tools have also been set up using information technology. WASREB has an informa tion system named WARIS that reports information and a set of key indicators for water service boards and water and sanitation service providers. Feedback and Complaint Mechanisms One of the activities of the WAGs discussed above, was to fi eld customer complaints. W ithin the first year of the WAG pilot, they reportedly handled more than 400 complaints, ( many of which had been pending for some time), on issues ranging from billing and metering to dilapidated infrastructure. About two thirds of these comp laints were considered to be resolved to the satisfaction of the consumer, according to WASREB. WAGs also worked with utilities to improve their overall complaint management system, sometimes facing resistance and even threats from the utilities after rep orting cases of corruption. WAG complaint resolution progress is in contrast to the broader national experience, where there is a large degree of dissatisfaction in 2012 among consumers on how complaints are handled by water authorities, due to slow or in adequate response ( Water Integrity Network, 2012; iHub Research, 2012).

PAGE 204

191 Attempting to build on this positive experience of WAG complaint resolution, WASREB established a n interactive program in 2013 MajiVoice, designed to collect feedback from consumers using computers or mobile phones Here, they can report corruption, submit a service complaint, ask billing questions, or provide comments and also track the progress of their inquiry (World Bank, undated/a). MajiVoice currently operates in four urban areas, (encompassing five services providers), and has processed more than 85,000 consumer complaints to date. More than 90% of the complaints recei ved have reached closure (Anzagi Response, 2015). S ervice providers are directly responsible to respond to complaints, however, an internal website dashboard allows WASREB to monitor progress in complaint resolution by various service providers. service provider rose almost tenfold after MajiVoice, and res olution rates climbed from 46% to 94%, with a shorter resolution time. The five top areas of complaints nationwide, ordered from most to least frequent, are: Billing, Meter Reading, Sewer Blockage, No Water, and Water Leak (Anzagi Response, 2015). While M ajiVoice is targeted to those with water/sanitation services, and not designed to address the population without access, it strengthens citizen engagement and service provider accountability, ultimately benefiting service quality, consistency and sustainab ility. A recent survey of service provider staff found that more than 90% found it easier to address and follow up on consumer complaints with the MajiVoice system (Anzagi Response, 2015). MajiVoice also enables service providers to communicate quickly with consumers via SMS broadcast for priority issues.

PAGE 205

192 P rior to the initiation of MajiVoice, KEWASNET found that the main reason cited in focus groups as to why people do no t complain over poor service delivery is because they do not have anyone with w hom to register their complaint (Shivaii, undated). The MajiVoice system seems to be making great strides in addressing this def iciency. Project Implementation at the Local Level At the municipal level, several NGOs and CBOs are engaged in servic e delivery to disadvantaged communities The national government views them as key partners for service delivery and thus has created an environment for them to participate effectively (Opiyo Response, 2015). For example, in the Laikipia East District of Kenya, co mmunity participation contributed to the sustainability of donor funded water and sanitation projects in t hat area, by fostering more project ownership, more projects completed on time, and timelier repair and maintenance of projects. Community members se rved on water management committees along with project staff (Gathuru, 2014). Nairobi is unique in having the Environmental Sanitation and Hygiene Policy 2007, a policy that specifically targets sanitation needs and calls for a communi ty based particip atory approach to addressing sanitation issues. KEWASNET, the sector NGO/CBO umbrella organization, cooperates with the city at the policy level on sanitation issue s ( Tukahirwa, et al., 2013 ). Sector Decentralization In 2010, Kenya adopted a new constitu tion that divided the country into 47 counties, to which both political power and government function s were devolved, making

PAGE 206

193 only two government levels The division of sector responsibilities is such that the national level is responsible for international waters, water resources, and water management, and the county level is r esponsible for service delivery The counties now oversee local service providers and work in cooper ation with the eight regional water service boards (Ndii, 2010). Sector Planning Kenya prepared a SWAp for th e sector in 2006, in conjunction with th e first annual sector r eview meeting t o discuss sector wide progress and targets. At the conference, Partn ership Principles were agreed among various ministries and sector development partners (Uyterwaal & Dietvorst, 2007 ; Richards & Doering, et al., 2008). This was a step in the transition from donor support of individual projects to donor overall sector budget support, however, it did not contain all elements of a fully developed sector strategy, such as fully developed strategies and monitoring and evalu u, 2009). In 2009, a Sector Investm ent Plan (SIP) was developed. Challenges for fully integrating these sector plans into practice were noted in a 2011 report: Key issues include limited donor awareness and engagement with the SIP as a viable sector plan, low levels of donor financing to the WSTF, an increasing number of agencies (10 in 2005, 16 in 2008) and projects (35 in 2005, 80 in relative to other countries in the region is a real chal lenge (WSP & AMCOW, 2011a, p.17). So, while these planning documents represent positive steps in the realm of sector planning, there are challenges that remain in coordinat ing multiple sector agencies and proje cts, including some resistance on the part of donors in aligning their priorities with a

PAGE 207

194 broader national plan (ORGUT Consulting, 2010). The full potential of these national p lanning exercises may only be realized several years into the future. On a smaller scale, the c ity of Nairobi, by far the l argest urban area in Kenya, is currently cooperating with the World Bank to implement the Nairobi Metropolitan Services Improvement Project ( NaMSIP ) This project is intended to strengthen urban services and infrastructure in the Nairobi metropolitan region and water and sewa ge services are key components of the program. O ne of the key objectives is institutional reform and planning with support extending to each of the 15 local authorities located within Nairobi. There are also provisions for cons ultations with and engagement of stakeholders in project implementation (World Bank, 2011; Opiyo Response, 2015). Kenya Case Conclusion Following are some ge neral conclusions from the Kenya case study The composite civil society engagement va lues derived for the quantitative analysis, ( 1.0 for water and 0.75 for sanitation ), seem appropriate given findings of the case study, keeping in mind those specific elements that were expected to be captured by the GLAAS data: opportunities exist for the public to be involved during sector planning and implementation, and effective fe edback/complaint mechanisms are in place. E xamples of innovative programs introduced and expanded by Kenya to engage civil society and facility consumer feedback seem to support the maximum composite value of 1.0 for water and a relatively high value for sanitation These composite values are also consistent with the comparative USAID CSO Sustainability Index introduced at the start of this case study. What is less clear is the reason for the discrepancy between these high

PAGE 208

195 composite values and the mid ra nge Kenya score in the CIVICUS EEI, the other comparative index introduced above. T he civil soc iety engagement programs introduced in Kenya mostly since 2007 effectively engage civil society at various stages: from problem arti culation and policy development to program implementation and managem ent. These broad access tools are a means of involving a wide range of urban dwellers, eve n those in informal settlements, and provide some protection against neglecting the needs of low income residents The establishment of KEWASNET, as an outcome of the pilot projects now supports a CSO network that can advocate for needed policy and service improvements in the sector. The sector has historically demonstrated a certain degree of transparency, in its requirement to consult with the p ublic on policy and to conduct a nnual s ector r eviews. The availability of information on agency websites, continually expanded over time, now includes detailed reports on service provider performance, providin g citizens with information and resources to put pressure on counties where service is sub standard. However, the role of civil society in conducing monitoring and oversight could be greater. The contribution of the WAGs to complaint resolution and the establishme nt of the MajiVoice website where consumers can issue complaints and even report alleged s. This initiative can be credited to sector planning practices, since it was an out come of an annual s ector r eview meeting. The WSTF has had a positive impact with projects in low income urban areas in recent years, working in partnership with civil society. The 2010 political changes

PAGE 209

196 solidify responsibility at the county level for wa ter and sanitation service provision. It will be interesting to observe how this change impacts the sector over time. South Africa Case Study A major political shift occurred in South Africa in 1994, when the apartheid system was abolished ushering in a new era of equality for all citizens. At this point in time, there were extreme disparities between the living conditions for different segments of society (Lane, 2004). New systems of governance were adopted in an attempt to eliminate these disparities and a new constitution took effect in 1997. The Bill of Rights stated that everyone has the right to have access to sufficient water and sanitation. South Africa experienced a 10% growth rate in urban population between 2000 and 2012, by which 63% of th e population lived in urban areas. As an upper m iddle i ncome country, South Africa uses mostly domestic financial resources for water and sanitation service extension and relies relatively little on donor funds (Lane, 2004; Fox & Liebenthal, 2006). It had also achieved almost universal urban water access by the sta rt of the study period in 2000, although the urban sanitation rate was only 78%. So, in this country environment, lack of funds does not seem to be obstacle to achievement at least in the w ater sector, but as the case show s weak local government performance was an impediment to greater progress in sanitation. composite civil society values used for the quantitative analysis I looked at comparative country rankings. South Africa was the highest rated country for the CIVICUS EEI among the 40 countries in my sample with CIVICUS EEI scores, and also highest rated for the 2011 USAID CSO Sustainability Index among 20 countries i n my sample that were rated, both indicating a superior civil

PAGE 210

197 society environment in the country. Based on these comparative ratings, t he composite civil society value s for South Africa derived from the GLAAS dataset are somewhat mod est at 0.75 for both w ater and sanitation. T his may reflect limitations specific to the water and sanitation sector as compared to the country as a w hole. These limitations are further discussed throughout the case below. General Climate for Civil Society I n the post aparth eid era, the general climate for civil society in South Africa is considered to be good. According to the USAID 2011 CSO Sustainability Report for sub variety of laws, a nd operate s freely and openly array or organizations, including voluntary associations, interest based associations, advocacy organizations, and professional not for profit organizations. In 2011, there were more tha n 76,000 registered not for profit organizations, with more than half working on issues of social services, development, and housing (USAID, 2011). In addition, there are organizations dedicated specifically to promoting civil society activity. For examp le, the Socio Economic Rights Institute of South Africa (SERI) is a non profit organization that assists social movements and supports civil society coordination and mobilization, with one area of focus being basic services such as water and sanitation (Ti ssington, 2011). The courts have guaranteed the right for public participation and civil society engagement as relates to pending legislation and public policy development. There is a requirement that adequate forums relating to new national policies or laws be held, so t hat civil society feedback can be given, and the views expressed taken into account

PAGE 211

198 There are similar protections for participation at the local and provincial government levels. However, the USAID study found tha t, in reality, the tendency is for gover nment to work more with paid consultants rather than directly with CSOs on issues, even though CSOs are generally viewed positively by both the government and the public (USAID, 2011) Larger, more established CSOs have an advantage in terms of interactio n with the government and new organizations trying to establish themselves have a more difficult time, including problems of delayed registrations ( Munnik Wilson & Pereira 2014; USAID, 2011). Moreover, t here is sometimes a lack of connection between grass roots interests of the population and the organized work of CSOs. At the grass roots level, there are frequent protests over poor living conditions in particular areas, including water and sanitation se rvices (USAID, 2011; Tissington, 2011 ; Munnik et al., 2014 ). Lo cal level concerns including lack of service and service disruptions, may not be adequately tackled by large CSOs that receive the most attention from the national government (USAID, 2011). Regulatory Development In 1997, a new constitution came into effect in South Africa and there was an emphasis throughout the government on programs to meet basic needs (Lane, 2004). The Bill of Rights states that everyone has the right to have access to sufficient water and sanitation and that the government should take steps to achieve the realization of this right. Inclusion of this premise had been the focus of lobbying by several human rights organizations (Fox & Liebenthal, 2006). The Constitution mandates that local

PAGE 212

199 government be responsible for potable water supply systems and domestic waste water and sewage disposal (Tissington, 2011). Also in 1997, the Water Services Act was passed, defining the role of the Department of Water Affairs (DWA) as regulator of the sector, and municipalities as responsible for service provision. DWA is mandated to monitor the performance of water service provide rs, including municipalities tha t perform this function directly (Tissington, 2011; ISF UTS, 2011). In 2003, a Strategic Framework for water services was developed. It again placed responsibility for water and sanitation service provision with local governments, requiring the national and provincial governments to support and strengthen the capacity of local governments to assume this role (Tissington, 2011; Department of Water and Sanitation, 2004). It also formalized a Water Services Sector Leadership Group drawn from sector stakeho lder groups including CSOs and water companies, to pro vide policy and strategic direction to the regulatory process (Jones and Williamson, 2005 Mtolo Response, 2015 ). dominated by large co mmercial water users who have elite access to policy makers (Munnik et al, 2014). Civil Society Engagement in the Sector The new constitution incorporated the principle of public participation in all parts of government, and encouraged the involvement of community org anizations in local government. Likewise, the Strategic Framework for the water sector had as objectives to: a) support capacity development of CSOs; b) promote the active involvement of civil

PAGE 213

200 society in the provision of sustainable and affor dable water; and c) encourage CSO participation in monitoring sector performance (Mjoli, 2010). In 2004, a s part of a national sector strategy, the Masibambane 35 Water Services Sector Support Programme, a civil society engagement strategy was prepared wit h an action plan following in 2005. The strategy not only sought to fulfill the constitutional mandate and the Strategic Framework objectives, but also to meet requirements under agreements for sector support from the European Union (Department of Water a nd Sanitation, 2004). Specific elements of meaningful civil society engagement were detailed in the strategy, including: Strong awareness of and participation in water and sanitation planning and delivery processes by communities and CBOs; Increased num bers of NGOs, CBOs and community members actively engaged in local level planning processes for water and sanitation services delivery; Greater proportions of water and sanitation programme financing flowing through and being managed by NGOs and CBOs; Increased numbers of NGOs and CBOs, constructively involved in policy review and reformulation processes a t provincial and n ational level (Department of Water and Sanitation, 2004, p. ii). However, an independent evalu ation report of the Masibambane progr am completed in 2007 found that the ob jectives of the civil society engagement strategy were not being met, and that there had been limited progress in this regard since Masibambane was initiated in 2000 (Everatt, et al., 2007). Since that time, CSOs have been increasingly consulted in the regulatory process at the national level, including providing input on dr aft documents and meeting with 35

PAGE 214

201 national regulatory staff ( Moremong Response, 2015 ; Mtolo Response, 2015 ) However, there is a perception among some that government remains ambivalent about the role of civil society. A ccording to for government officials, consultation is too often seen as something to be ticked off on a list, and usually as a means of telling the publ The situation at the local level was also found to be lacking M irumachi & Van Wyk (2010) found that power disparities at the local level can impede the way that consultation works in practice, perhaps co inclusiveness during the apartheid era. Even though newer civil society actors may be present in water discussions, they do not necessarily hold the political power needed to influence decision making. Similarly, a 2005 study of collaboration in the water sector found that although CSOs were pr esent in discussions, they were a relatively subdued presence and not well organized (Jones and Williamson, 2005). Mirumachi & Van Wyk, (2010) recommend ed a regulatory framework that details specific mechanisms for communication and consultation with user s, so that they may express preferences about service delivery and related issues. Such mechanism would help to overcome a history of low citizen engagement with government, as noted by both Mirumachi and Van Wyk (2010) and Funke et al. (2007) In recen t years, more has been done to foster engagement at the local community level. The government collects input using existing community structures, such as ward committees and CSOs, helping to foster a sense of community ownership. Community water and sani tation forums are being established in all municipalities to improve two way communication and as a means for CSOs and consumers to pro vide feedback on

PAGE 215

202 service issues, with input sent to relevant government offices for intervention (Moremong Response, 2015 ; Mtolo Response, 2015 ). There are also provincial sector for ums, to which officials of CSOs water companies, and government departments belong, where national policy positions and pending programs are presented and discussed (Mtolo Response, 2015). The Mvula Trust, established to assist the government in implementing the water service policy and has helped to pilot new policies and project implementation approaches. It works nationwide on advocacy issues, and also in specific communities where it s upports community based projects. It has helped to network CSOs working in the sector and strengthen their capacity for advocacy, including a two year Turning Advocacy into Action program from 2007 to 2009 that highlighted examples of municipalities engag ing with the public. The Trust implemented a pilot project in two municipalities near Durban called which educates government officials and civil society on water rights and responsibilities and created a platform for interaction among these groups (Morrison, undated). Mvula Trust is also one of several non profit and community associations organizing to provide water and sanitation services in South African communities (Fox & Liebenthal, 2006; Mjoli, 2010 ). Social justice has been a focus of two key civil society groups focused on the water sector in South Africa. T he Durban based Centre for Civil Society has over several years has taken a human rights based approach to water and sanitation access. The Cent re has brought attention to how commercialization and neo liberal pricing policies have led to water price increases that negatively impacted the urban poor, exacerbating a situation of extreme income inequality that exist s in South Africa The Centre Director

PAGE 216

203 has said that protests are necessary to campaign against such policies, in the absence of adequate social spending and attention to the needs of the poor (Bond, 2013 & 2014). The South African Water Caucus (SAWC), a network of NGOs, CBOs, a trade union and individual activists in the South Africa n water sector, has been an active network since 2001. It works on national policy level issues and conducts research related to the sector, to include information sharing, networking, education, lobbying, and public awareness raising. Because of its wide ran ging membership, it often serv es as an intermediary between government and local c ommunities when conflict develops SAWC uses a dece ntralized leadership appr oach to engage local government s to addres s service pr oblems such as disruptions in water service. On e activity originated by an SAWC member is the Water Dialogues project, which brings together multiple stakeholders within a municipal area. This program has done much to increase understanding of the nature and problems in local government water services delivery. In its lobbying role, SAWC has participate d in all important national policy processes since its inception focusing on networking and educating CSOs Local Government Effectiveness B eginning with the Constitution, and extending through the Water Services Act, the obligation for providing water and sanitation services is given to local government s, who have the option to engage outside entitie s as service providers, through contracts o r joint ventures. T he Municipal Systems Act addresses community participation, with details on mechanisms, process, and procedures to be followed when preparing, implementing, and reviewing development plans, budget, and decisions related to municipal ser vices (Tissington, 2011).

PAGE 217

204 However, what is actually implemented at the local government level ma y not fulfill legal requirements in several respects. Studies cite the weakness of local government to manage water and services, more so than lack of funds, as an obstacle to extension of water and sanitation services to unserved populations (Schreiner & Hassan, 2011; Fox & Liebenthal, 2006; Tissington, 2011; WSP, 2002). Generally speaking, there is a lack of responsiveness by electe d officials to community needs Frequent protests throughout the country have resulted from dissatisfaction over servi ce provision and corruption is a problem both at the local and provincial levels of government (USAID, 2011 ; Munnik et al, 2014 ). A vast majority of municipalities failed a general government audit in 2011 (Tissington, 2011). Challenges facing local government include a shortage of man agement skills and capacity, especially in smaller and rural municipalities (Tissington, 2011). Municipalities try different technical approaches that may not be appropriate for the community, particularly in the area of household sanitation, where there is less knowledge of different options The Department of Provincial and Local Government developed a Five Year Local Government Strategic Agenda for 2007 to 2012 that recognized the need for technical assistan ce and knowledge acquisition by new staff at the municipal level (Department of Water and Environmental Affairs, 2008). A 2011 concept paper by the D epartment of W ater A ffairs (DWA) further identified i ssues to be addressed included policy inconsistency, instability among different areas of local g overnment, weak support mechanisms, poor political management and ethics, and weak service delivery capacity.

PAGE 218

205 Accordingly, t raining and su pport to local government were a key foci of the most recent Masibambane sector plan. DWA introduc ed an incentive b ased regulatory program for municipalitie s in 2008 aimed a service improvement The Blue Drop program rewards municipalities through a certification program based on performance in both water service and water quality The 2013 program report covered more than 150 municipal water authorities and more than 1,000 differen t water supply systems with broad input going into the evaluations. manageme nt and operational decisions to support and improve on long term sustainable water services (Department of Wat er and Sanitation, 2013, p. 1). As explained in the report, the findings are intended to be cautionary rather than punitive in nature. The ave rage score has continually improved from the first reporting in 2009, his incentive based regulatory approach seems to have raised awareness and created a stimulus for gradual and sustainable improvement (Centre f or Environme ntal Rights, 2012 p. 9). Sanitation Sanitation is given par ticular coverage in this case study to illustrate how protests were key to motivating additional government action on the issue and to contribute to the general question of sanitation versus wate r priority. South Africa produced a White Paper on Bas ic Household Sanitation in 2001. Principles of the document included community participation through meaningful involvement in projects and decisions relating to community facilities, such as schools an d clinics (Tissington, 2011). However, sanitation progress continued to lag in the early

PAGE 219

206 2000s. There was confusion of roles among national level agencies with responsibility for sanitation, particularly when sanitation was transferred to the Department of Human Settlements from DWA in 2009. According to a 2011 report by the SERI NGO national policies inadequately address ed sanitation, with particular gaps in policy for peri urban areas and informal settlements ( Tissington, 2011 ) Problems i mpacting local management include a lack of clarity on standards and appropriate technological options, and a lack of strategic planning and consultation with communities w hen deve loping plans In Cape Town, the Social Justice Coalition (SJC) initiated a sanitation campaign to bring attention to lack of access in informal settlements. It acquired a broad membership, which hel ped to both le gitimize it in the eyes of the c ity and validate its demands SJC adopted a nonpartisan approach to its campaigns, to distance itself from political affiliation and campaigning, and ultimately compelled city authorities to improve access to basic sanitation in Cape Town (Overy, 2013). Following several community based campaigns and protests ab out lack of a dequate sani tation throughout the country, a revised policy framework for sanitation was issued by the government in 2011 building on the 2001 white paper. It includes sections on community participation, monitoring and evaluation, and institutional arrangements for sanitation, and stresses protection of the poor (Tissington, 2011). Some saw this as a milestone in the recognition of sanitation as a mainstream issue (Overy, 2013 ) and that the issue is now receiving much needed attention (Moremong Response, 2015).

PAGE 220

207 Reporting and Monitoring This section evaluates the extent to which the national governmen t in South Africa reports information about the sector to the public For the sector as a whole, t he government has monitoring and evaluation systems that track : a) performance against national targets an d spending policies; b) e quity of subsidy allocations; and c) percentage of persons with access to water and sanitation (WSP & AMCOW, 2011b). The previously mentioned Masibambane sector planning program uses a single water services reporting format to track service performance throughout the country. However, several findings point to the need for continued refinement of data. An evaluation done of the Masibambane program found that the reliability and usefulness of the data was questionable, citing issues of quality for the primary data on which the national figures were based, and the process by which these data were analyzed for national planning (Department of Water and Sanitation, 2011). Many municipalities need more support for data collection and analysis, particularly those facing rapid population growth And, with numerical meas ures, there is not enough emphasis on the quality and functionality of infrastructure, or the accessibility and sustainability of sy stems (Tissington, 2011). A recent study from the Centre for Environmental Rights, based on consultation with civil society representatives, found that sector compliance mo nitoring and enforcement data are not given sufficient political and institutional priority, and was not readily available to the public. They recommended strengthening this capacity, co llecting and publishing more monitoring and evaluation information and reporting more about enforcement activities (Centre f or Environmental Rights, 2012 ).

PAGE 221

208 Decentralization The South African government consists of three levels: national, provincial and local. A decentralized approach to water and sanitation was introduced with the new constitution in 1996 and reaffirmed with the 2003 Strategic Frame work. Experience shows that local governments do, in fact, independently make decisions for the sector, however, the effectiveness of local programs vary due to differences in capacity and political will Revenue collection and spending are at th e discre tion of the local water service authority, as well as decisions about engaging the private sector (WSP & AMCOW, 2011b : Moremong Response, 2015 ). Sector Planning The Masibambane program discussed above was essentially a sector wide approach (SWAp) developed for South Africa in the early 20 00s, facilitated by national level budget reform s 36 It initially focused on just three provinces with the greatest need for additional services, but has been expanded to encompass the entire country. The development approach was typical of a SWAp, incorporati ng input from government, civil society, service providers, and donors (WSP & AMCOW, 2011b ; Everatt, et al. 2007) It was intended to support the shift from nationally driven water and sanitation del ivery to a more decentralized system Plans addressed governance issues, donor coordination, a consolidated sector budget, and plans towar ds achievement of sector goals, encompassing a range of issues, including appropriate technology, gender mainstreamin g, and greater civil society engagement (ISF UTS, 2011). The program has helped to form the Water Information 36 See https://www.dwa.gov.za/masibambane/about.aspx for more information.

PAGE 222

209 Network for South Africa (WIN SA) a platform for sharing lessons and knowledge in the sector, not only within South Africa, but globally as well WIN SA produced a Learning Manual and Guide, distributed to more than 150 CSOs. The Masibambane plan was regularly updated until 2011 Around this time, donor funds that supported the program were shifted from South Africa to other countries with greater financial needs (Mtolo Response, 2015). South Africa Case Conclusion Following are some ge neral conclusions from the South Africa case study The co mposite civil society engagement values derived for the quantitative analysis, ( 0.75 for both water and sanitation ), seem to be supported by findings of the case study, keeping in mind those specific elements that were expected to be captured by the GLAAS data: opportunities exist for the public to be involved during sector planning and implementation, and effective feedback/complaint mechanisms are in place The fact that the composite values are below the maximum values of 1.0 seems warranted by the fact that, until recently, there were not effective feedback and complaint mechanisms. This also provides some explanation as to why the composite values for this sector are not as superior as the top ranking given to South Africa in both the general USAID CSO Sustainability Index and the CIVICUS EEI. While composite values are relative high, and the civil society climate in South A frica is highly rated on other country comparison indices, it seems that more can be done in the water and sanitation sector to: a) coordinate and network CSOs; b) make it easier for new CSOs to establish themselves; and c) have more consistent and institu tionalized engagement of CSOs at the national and municipal level s. These steps

PAGE 223

210 would promote an environment where problems are identified and addressed early, before developing into protest situations, and where concerns about equity in service can be hi ghlighted and addressed The stated goals for civil society engagement that are part of the Masibambane program would be increasingly effective if also incorporated at the local government level where service decisions are made and where failures relate d to responsiveness and ethics have been identified. Groups working towards social justice, such as the Centre for Civil Society and SAWC seem to be putting the most pressure on municipalities where service is lacking, and the poor are especially impacted Programs like Blue Drop help keep municipalities accountable by highlighting municipal performance in water service and providing informati on by which the public can push for improvements What seems to be lacking is a similar performance measuring tool for sanitation which lags far behind water in coverage and has been the subject of protests and widespread dissatisfaction. Also of value would be involving civil society more in monitoring programs, improving the quality of data collecte d from municipalities and expanding the measures to include the functionality, accessibility and sustainability of infrastructure. Summary and Analysis of Civil Society Engagement Cases This section provides a summary and analysis of information presente d in both case studies, beginning with Table 9.1 that lists key developments for each civil society engagement construct, by country, along with general observations. Afterwards, a written analysis draw s conclusions about what can be learned from the i mpl ementation of these programs that would be applicable to other countries.

PAGE 224

211 Table 9.1 Civil Society Engagement Case Study Findings by Construct Kenya South Africa General Observations Engaged in Problem Articulation CRC in 2007 helped identify urban sanitation policy and program weaknesses and advocated for additional resources at the national level WAGs piloted in 2010 to engage public in id entifying and solving problems 2005 Civil Society Strategy Action Plan for water and sanitation calls for municipalities to engage civil society Local protests and campaigns serve to highlight poor water and sanitation services in urban areas While only in a pilot stage, t he CRC and WAG approaches in Kenya show promise as a means to identify problems and motivate resolution Openness of society in South Africa allows for protests that highlight probl ems, but ideally there would be institutionalized mechanisms for service d issatisfaction to be registered Engaged in Policy Development and Planning Water Act of 2002 required government to: 1) consult with public on major decisions, and 2) have public serve on governing boards WAGs charged to disseminate information to consumers CSOs invited to participate in national Sector Working Groups in 2012, leading up to National Medium Term Plan KEWASNET (outcome of the CRC) serves as a CSO umbrella and networking organization to influence policy in the sector Municipal Systems Act of 2000 requires municipalities to consult with the community about service options 2003 Strategic Framework creates a Water Services Sector Leadership Group of stakeholders provides policy and strategic input to the regulatory process although it may be dominated by powerful interests National government tends to work with consultants and established CSOs; new CSOs may experience difficulty in influencing decisions Kenya shows several examples of CSOs working at the national level with policymakers, as well as more grassroots outreach through the WAGs While national level CSOs exist in South Africa, the extent to which they are included in policy development discussions with the government is unclear

PAGE 225

212 Kenya South Africa General Observations KEWASNET Learning Forums inform key stakeholders about sector issues and provide a forum for discussion with government officials A 2005 study found CSOs to be a relatively subdued presence in policy discussions and not well organized National level CSOs, su ch as the Socio Economic Rights Institute, advocate for policy reform and supports CSOs on issues such as water and sanitation Power disparities at the local level can impede the way that consultation works in practice Centre for Civil Society has been critical of policies that promote commercialization of the sector, saying that needs of poor residents are not adequately considered when these policies are developed SAWC is working both at the national level on policy issues and organizing loca l Water Dialogues to highlight local problems Engaged in Implementation and Management WASREB advises water providers on complaints handling, tariff setting, and consumer protection WAGs facilitate meetings between the community and service providers The 2003 Strategic Framework has one of its goal s to promote CSOs in provision of sustainable and affordable water services Both the WASREB and the WAGs in Kenya service as intermediaries between service providers and consumers on management issues, whereas no comparable bodies seem to exist in South Africa

PAGE 226

213 Kenya South Africa General Observations The Water Service Trust Fund that finances projects was considered to have some of the best practices for citizen participation among various sector funds Many of the service implementation problems in South Africa are attributed to weak local government performance, however recent introduction of incentive based programs are improving performance There are many examples of non profit or community associations organizing to provide water and sanitation services in communities The Centre for Environmental Rights calls for government to support better networking and information sharing among CSOs working in the sect or and to develop a CSO toolkit CSO s are directly involved in service provision in South Africa, although the extent to which they are supported by the government is unclear Incentive based certification program for municipalities is effective for improving services in South Africa Tran sparency and Information Disclosure Water Law of 2002 requires that a national water management strategy be published in the Kenya Gazette and that public consultations be conducted for the strategy New legislation is given a public comment period publicized on website and in popular daily newspaper Annual public expenditure review conducted for the sector The 2003 Strategic Framework has a goal of encouraging CSO participation in sector monitoring A national information and reporting system tracks performance against targets identified in the Masibambane Plan, although reliability and usefulness of the data are questioned; quality, access and functionality of infrastructure not well addressed in the reporting system Kenya demonstrates several examples of information sharing, both at the national and local level through WAGs; a great deal of i nformation is available to the public on line South Africa has policies to promote information sharing and monitoring, but shortcomings have been ident ified in what is reported

PAGE 227

214 Kenya South Africa General Observations Sector agencies publish newslet ters and maintain websites; there are several technology based tools th at consumers can access on line WAGs provide the community with information on policy development through annual reports and other mechanisms The WARIS program reports on key indicators for water service boards and service providers Blue Drop ratings for specific municipalities provide information on service quality Both the WARIS reports in Kenya and the Blue Drop reports in South Africa are good examples of information sharing that promote accountability for service providers Feedback/Complaint Mechanisms CRC program is a means for consumers to rate service quality WAGs provide feedback to service providers, push for complaints to be addressed, and monitor complaint resolution MajiVoice website allows for feedback and complaints to be registered on line or via cell phone; system ha s improved complaint resolution Bl ue Drop incentive based program for municipalities incorporates feedback from NGOs and consumers when rating municipalities Community based water forums are being established to receive feedback and complaints about service quality The CRC, WAG, and Maji Voice programs in Kenya both have been shown to be effective as feedback mechanisms and to improve complaint resolution Only recently in South Africa has the Blue Drop program institutionalized any type of formal feedback mechanism, and this does not inc lude individual service issues Decentralization Mandate given to eight regional water service boards in 2002 to contract for service provision; policy and regulatory functions stay with the central government 1996 Constitution mandates local government responsib ility for water supply and sew age disposal 2003 Strategic Framework reiterates local service provision responsibility Kenya supported sector decentralization since the early 2000s through the regional Water Service Boards and the Water Service Trust Fund; this is furthered with the 2010 political decentralization

PAGE 228

215 Kenya South Africa General Observations Water Service Trust Fund was established in 2002 to mobilize funds and help reach underserved populations; local and regional entities are able to access these funds directly New constitution in 2010 gives newly formed counties respo nsibility for service delivery A central government agency, DWA, develops and implements policy Revenu e and investment decisions are made by local governments Weak local government capacity is an obstacle to effective local service provision; lack of skills, poor responsiveness, and corruption are cited as issues A 2011 concept paper recognizes the need for enhanced support and training for local governments; this is the focus of the most recent Masibambane plan In South Africa, local government service provision is legally mandated, but many deficiencies are noted in practice Sector Planning In 2006, a SWAp is developed and the first annual sector review meeting is held, producing Partnership Principles In 2009, a Sector Investment Plan is developed A 2011 study finds that not all donors are aware of or engage with the Sector Investment Plan T here are examples of coordinated planning occurring at the municipal level, such as the NaMSIP project for Nairobi The Masibambane sector wide plan developed in the early 2000s with input from stakeh olders and donors; plan was periodically updated until 2 011 Masibambane cited as a major factor in promoting collaborative processes and in improving reporting from municipalities Sanitation is the focus of a 2011 policy framework, bringing needed attention to the issue Both countries have benefited from sect or wide planning Masibambane plan made civil society engagement a policy priority beginning in 2004 Sector Investment Plan and Partnership Principles may be weakened if not all donors are participating in the exercis e

PAGE 229

216 Engaging Civil Society in Problem Articulation Comparing the cases, we see two very different mechanisms by which the public registered dissatisfaction with poor service qua lity T he CRC is Kenya gave citizens a voice in identifying problems and led to development of a related advocacy agenda for urban sanitation. By contrast, in South Africa, where no such institutionalized feedback mechanisms was in place, citizens tur ned to protests to prompt action on issues of importance, including urban san itation. The Kenya n approach, by involving diverse s takeholders from the start, laid the groundwork for greater coo peration in addressing problems; i n South Africa, the situation was often characterized by tensio n and distrust among consumers, the governm ent, and service providers. Engaging Civil Society in Policy Development and Planning Both countries have in place legal requirements for government to consult with civil society, but the working practices in Kenya, such as dialogues and exchanges facil itated by the WAGs and the impact of KEWASNET, seem to have had resulted in more meaningful interactions. Further, inclusio n of CSOs in working groups for Medium Term Plan development in 2012 allowed them to contribute to long range policy and planning. The lack of such formalized inclusion in South Africa has the effect of more subjective CSO interactions, favoring groups the national government has historically worked with, or subject to power disparities at the local level. Ideally, for maximum input, countries w ould have in place both methods: 1) public discourse and information sharing, through CSO networks and the media, to foster a well infor med public, coupled with 2) formal mechanisms for CSO input to policy development and planning, both at the national and local level.

PAGE 230

217 Engaging Civil Society in Implementation and Management CSOs are involved in implem entation and management both as small scale service providers at the commun ity level, and as g overnment partners in overseeing water service pro viders. As service providers, it is most often in the form of CBOs, whose role has been facilitated in some countries by local funding made available through sector decentralization. These CBOs often have a demonstration value in showing how services ca n be adapted to local needs and within local affordability standards. However, these activities are usually initiated outside of the realm of government. On a broader level, when CSOs ar e engaged in oversight of service providers, they take on a more in fluential role. By being part of rating systems, such as the Blue Drop pro gram in South Africa, or having dialogues with service providers, such as through the WAGs in Kenya, civil society helps to bring about positive adaptations and improvements in serv ice. These benefits were found to be present in both situations of weak and strong local government capacity, as they are more related to local government priorities than local government capacity. The earlier civil society involved in the process, such as during selection of service options and management approaches, the more likely that local projects will reflect consumer preferences from the start. In is unclear that civil society is engaged at this stage in either coun try. Transparency and Information Disclosure Comprehensive, national monitoring and evaluation systems to track progress on sector targets are commendable, but need sufficient resources to produce quality data and go beyond basic measures, to incorporate qualitative issues of service reliability and

PAGE 231

218 sustainability. Evaluation programs, such as Blue Drop in South Africa, get more at these qualitative issues and are likely of more value to civil society in pursuing advocacy agendas and pressuring local gov ernments on specific service issues. Opportunities for in person meetings and hearings, where service providers respond directly to questions, such as the WAG forums in Kenya, provide an additional layer of transparency. In these types of forums, a ra nge of issues may be raised to which service providers are compelled to respond. For example, these forums are a means by which government and service providers may be asked ab out allegations of improper conduct. The newly available technology based tools in Kenya demonstrate cutting edge techniques to share information and receive feedback from consumers with a wider reach than either in person meetings or written rep orts. Such increased knowledge and input on the part of the public can only strengthen t he role of CSOs in various sector activities. public expenditure review, which publicizes how government funds allocated for the sector are used. Feedback and Complaint Me chanisms A range of possible consumer feedback and complaint mechanisms are demonstrated in Kenya, including the MajiVoice website for real time feedback and complaints f rom consumers, the periodic CRC, and WAGs acting as watchdogs to see that complaint s are responded to in a timely manner. One can see the advantage s of having thi s suite of mechanisms in place so that : 1) indivi dual feedback can be registered

PAGE 232

219 and tracked ; 2) consume r experiences can be synthesized and documented; and 3) there is a means by which service providers are held accountable for responding to complaints. While the annual Blue Drop rating process in South Africa incorporates consumer input, there doe s not appear to be an ongoing program for consumers to provide feedback and register service compla ints T he extent to which a municipality or service provider address es deficiencies may not become apparent until the next annual reporting cycle. The advantage of having a more current open feedback and respon se sys tem is that ongoing issues, such as service interruptions, maintenance problems, and rep orts of corruption can be tracked by CSOs to encourage timely respon se actions. Conclusion of Case Studies In addition to researching hypothesis H3 regarding the relationship between civil society engagement and country progress, this chapt er also tried to answer one of the follow up questions posed at the end of the quantitative analysis chapter : Given that civil society engagement was positively associated with country achievement across all four outcome variables, and mean composite values of the variable, (0.48 to 0.58), were modest relative to the maximum possible value of 1.0, what can be lear ned from countries employing new forms of civil society engagement that might be beneficial to other countries seeking to heighten their level of engagement? The case s showe d that civil society can help gain the trust of co nsumers by for example, u sing community leaders as intermediaries between ser vice providers and consumers hosting public hearings and serving as an intermediary between government and local communities when conflict develops. Both Kenya and South Africa

PAGE 233

220 demonstrate transparency with service provider performance data now reported to the public on a regular basis. Published data can be used by government and civil society to monitor service progress to different areas and to highlight equity concerns. Customer feedback system s were a lso shown as a way of improving dialogue between service providers and consumers. C ivil society engagement has given a greater voice to marginalized groups by drawing attention to service problem s The experience of South Africa shows that having forums for civil society engagement is important even in relatively high income countries, and even in relatively open societies. In less developed countries, CSOs have informed the general public on issues su rrounding water and sanitation includi ng how local funding systems work, and l arger CSOs have provided sector education to local g overnment personnel. The case studies demonstrated a number of ways that countries with mid range composite values for civil society engagement can increase their e ngagement, not only for problem articulation and policy development, but also for disclosure, monitoring, and consumer feedback. These practices are important not only for extending new services, but also for sustaining service quality over the long term.

PAGE 234

221 CHAPTE R X SUMMARY AND CONCLUSIONS This chapter will summarize key elements of my dissertation, including background information, resear ch methods, quantitative and case study results, and theoretical contributions. It will also present limitations of the analysis and potential areas for further research. Introduction Rapid urbanization over the last two decades has intensified the challenge of providing adequate water and sanitation services to residents of cities, particularly those living in sub standard housing conditions in urban slums. B ased on the most recently available global data (2015), 9% of th access to safe drinking water and 32% lack access to basic sanitation (WHO & UNICEF, 2000 2015). Health risks posed to society from a lack of clean water and sanitation create a rationale for government and the global community to promote these services. D evelopment efforts in recent years have focused on this challenge, including M illennium Developm ent Goal (M DG ) Target 7.C 37 set in the year 2000, and subsequent UN recognition of the human right to water and sanitation (UN General Assembly, 2010). Institutions and governance are both key components in the provision of water and sanitation services, as recognized by international forums and s cholars in recent years 37 The t arget is to reduce by half the percentage of people who lack access to safe and sustainable water and basic sani tation services from the base year of 1990 to 2015 (UN Millennium Development Goals list).

PAGE 235

222 ( Saleth & Dinar, 2005; Koundouri, Saleth et al ., 2008). Institutions structure human exchanges and provide incentive st ructures (North 1990), such as who makes decisions about the sector and who allocates available resources. I nstitutions are important i n motivating government performance in the delivery o f public goods (Cremer, et al., 1994 ; Frank & Martinez Vazquez, 2014 ). Governance includes institutions, while also encompassing the values and processes by which government authority is exercised, including whose voices are heard in decision making (UNDP, 1997: UNDP, 2015). Effective water sector governance includes approaches that ar e open and transparent, as well as operations that are accountable, efficient, responsive and sustainable ( Rogers & Hall 2003 ). The goal of this research was to better understand if certain national institutional and governance attributes generally assoc iated with good governance have had a positive impact on government performance in the water and sanitation sector, specifically for providing urban w ater and sanitation services Institutions and governance influence all models of gover nment service provision including those that engage the private sector. To study the institutional and governance attributes, specific r esear ch questions and hypotheses were developed as follows: RQ1. What is the role of decentralization of water and sanitation serv progress on providing urban wat er and sanitation access ? H1 A high level of decentralization in water and sanitation services is associated with greater country progress on providing urban wat er and sanitation access.

PAGE 236

223 RQ2. What is the role of sector wide strategy and investment coordination progress on providing urban wat er and sanitation access ? H2 A high level of sector wide strategy and investment coordination in water and sanitation services is associated wit h greater country progress on providing urban wat er and sanitation access. RQ3. What is the role of civil society engagement providing urban wat er and sanitation access ? H3 A high level of civil society engagement in water and sanitation services is associated with greater country progress on providing urban wat er and sanitation access. These three attributes were selected because : a) they have been identified in the literat ure as having potential to contribute to effective government service provision and donor aid effectiveness; b) they have not been studied extensively for the water and sanitation sector, particularly on a large scale basis; c) there was an observed gap of research on potential interactions among the three attributes; d) they represent a departure from previous approaches to water and sanitation service provision in many developing countries; and e) relatively new UN data collection programs provide compara ble data on these attributes for multiple countries. The unit of analysis for this study w a s a country and the population of interest wa s low and middle income countries as classified by the World Bank (World Bank website, 2015 ). The timeframe studied wa s approximately 1990 to 2012. Theoretical Background Chapter III presented an overview of relevant literature on government service provision W ater and sanitation service access exhibits the characteristics of a public

PAGE 237

224 good characterized by low subtractability and low excludability Arguments can be made that the government has advantages over the priv ate sector market for providing public goods, based on improved s ervice efficiency, reduced undersupply problems, and greater social welfare with equity of distribution The private market can fail to produce an adequate amount of the good due to information asymmetry, monopoly power, externalities, or allocative inefficiency that leaves some consumers underserved. Inequity in who benefits fr om the good can exist with private provision, but can be addressed through redistributive government policies and targeted programs under government provision (Weimer & Vining, 2005) For water and sanitation s ervice, an adequate supply is especially important because it can also be considered a merit good, one whose consumption has a positive overall impact on society (Musgrave & Musgrave, 1989 ), by avoiding negative health externalities. The f indings from this dissertation will s how, among other things, the extent to which these arguments for government provision are supported based on recent country experience with urban water and sanitation service provision. In particular, the potential of the three institutional and governance attributes studied to overcome u n dersupply and equity problems will be evaluated. Research Methods Research was conducted in two phases: quantitative and quali tative, with the research methodology discussed in C hapters IV and VII, respectively. Phase one In phase one, regression tests were run for a sample of 75 low and middle income countries ( World Bank website, 2015 ) to assess the impact of the three key

PAGE 238

225 institutional and governance attributes (explanatory variables) on country progress in providing water and sanitation access in urban areas (outcome variables). S everal control variables were also included to account for countr y background conditions that the literature indicated could be important factors in country achievement, either positive or neg ative Table 10.1 overviews the four outcome variables, six explanatory variables, and seven control variables that w ere included in the regression model s. Note that the outcome and explanatory variables are specific to either the water sub sector or the sanitation sub sector. The country progress outcome variables were developed to roughly c orrespond with the MDG target a nd time p eriod while also dependent on which years had sufficient data availability. Table 10.1 Overview o f All Variables Outcome Variables Key Explanatory Variables Control Variables 1. Water Access 38 in 2012 2. Water 12 Year Progress 39 1. Water Decentralization 2. Water Sector Planning 3. Water Civil Society Engagement 1. Urban Population Change 2. Corruption Level 3. Political Fragility 4. Security Fragility 5. Literacy Rate 6. GDP per capita 7. Renewable Water Resources 3. Sanitation Access in 2012 4. Sanitation 12 Year Progress 4. Sanitation Decentralization 5. Sanitation Sector Planning 6. Sanitation Civil Society Engagement Note: All variables are at the country level. 38 Access means meeting JMP definitions for improved sources; see Table 4.5. 39 Percent c hange in access over the 12 year period of 2000 to 2012 relative to the percentage of urban dwellers that lacked access in 2000

PAGE 239

226 There are two outcome variables each for the water sub sector and the sanitation sub sector Outcome variables were derived from the UN Joint Monitoring Programme for Water Supply and Sanitation ( JMP ) dataset 40 that annually measures the percent of the po pulation in a country that lack water and sanitation services meeting minimal international standards, broken down by urban and rural areas ( WHO & UNICEF, 2000 2015 ). For this research, JMP data on urban access rates were used for all outcome variables. Because country access rates e xpressed as a percent do not produce a normally distributed variable, the country percentages were not used directly, but instead used to classify individual country achievement level s as low, medium or high for each of the four outcome variables. This pr oduced ordered categorical outcome variables, therefore the specific statistical test used was ordered logistic regression (ologit). The key explanatory variables representing each of the three attributes were derived from individual country responses to select questions from the 2011 UN Global Analysis and Assessment of Sanitation and Drinking Water ( GLAAS ) survey 41 Responses took the form of a number reflecting the degree to which a country had implemented a specific item referred to in a question ( WHO, 2011 ) The average of a usually asked 40 JMP is the official UN mechanism for monitoring country progre ss towards MDG Target 7.C and has tracked individual country water and sanitation access progress each year since 1990. 41 GLAAS is a UN Water data collection initiative implemented by the World Health Organization (WHO). The GLAAS survey covers governance and institutional environments in individual countries specific to water and sanitation.

PAGE 240

227 separately for water and sanitation, became composite value for a particular explanatory variable, such as water decentralizati on Country c ontrol variable values were derived from widely recognized global s Index (see Appendix C). Each outcome variable was th e focus of its own ologit model, so two models were run for water and two for the sanitation, each of which included all three explanatory variables related to the sub sector, as well as all the control variables. A series of correlation tables and frequency tables were also generated (see Appendices D and E). Ologit model results, (see Table 5.9), were evaluated primarily for the statistical significance and the direction, (positive o r negative), of the coefficients for each of the three explanatory variables. Of secondary interest was the overall st rength of the model, as indicated by the pseudo R 2 and the p rob > chi 2 In addition, odds ratios and predicted probability graphs, (wher e all other explanatory and control variables are taken at their means), were generated for each explanatory and outcome variable pairing, for a total of twelve graphs (see Figures 5.1, 5.2, and 5.3 ). Phase Two For the phase two research, f our country c ase stu dies were completed : two for decentralization and two for c ivil society engagement. The case studies chronicled country history with the attribute of interest and further probed questions raised in the quantitative analysis. No country case studies were conducted for sector planning, as questions emerging from the quant itative analysis were better addressed by other follow

PAGE 241

228 up research including comparison with results from other studies. Also, most countries do not have a long history of sector planning suitable for a detailed analysis. Case study countries were selected based on having : an extreme high end value on the key explanatory vari able of intere st; a depth of experience with implementing the attribute ; and a re asonable amount of written material available about the country experience. The case studies allowed for examination of some aspects of the attribute that were not well represented in the quantitative data. T o further study decentralization, I chose Colum bia and Mali, two countries with sector decentralization experience dating back to the early 1990s Th e two countries differ in some key background condi tions, such as income level and urban access rates for water and sanitation in the year 2000 To further explore civil society engagement, Kenya and South Africa were selected for case studies because they introduced, during the 2000 to 2012 timeframe, some innovative civil society initiatives at the national level with country wide influence Metho ds for data collection and analysis included document review and consultations with country experts. Documents originated from international organizations, academic journals, national governments, and non profit organizations They were reviewed and code d for evidence of the attribute of interest using construct level definitions (see Appendix A), as well as for information on general questi ons emerging from the phase one analysis For the consultations with country experts, persons from diverse profess ional affiliations and backgrounds (two to five per country, see Appendix G), were consulted via Skype for interviews or via email for questionnaires, while some sent relevant written

PAGE 242

229 material. General guides were developed for collecting information fro m consultants, ( see Appendix H), although more specific questions were also asked based on the background of the consultant. Interview summaries, along with email responses and written materials supplie d by consultants, were coded similar to the process d escribed above for document review. The following three sections will summarize phase one and phase two findings for each attribute, as detail ed in Chapters VI, VIII, and IX, and also present theoretical implications. When applicable, inter relation shi ps between attributes are noted, as the successful realization of one is often a building block for effective implementation of ano ther. Key literature that was reviewed in Chapter III is referenced throughout the discussion. A subsequent section will summarize overall findings on government service provision and how this compares to theory on this topic. Decentralized Services While the literature presents several qualifying conditions for successful implementation of de centralization, the theoretical arguments referenced below describe the potential for decentralization to have a positive impact on government service provision. For this reason, I hypothesized that: H1. A high level of decentralization in water and sanitation services is associated with greater country progress on providing urban water and sanitation access. Th e concept of decentralization was defined for thi planning, decision making, or management functions from the central national Many articles also emphasize that fiscal decentralization, which include s revenue raising

PAGE 243

230 authority and expenditure decisions, is a key sub component of decentralization ( Bardhan & Mookherjee, 2000; Mugabi & Njiru, 2006; Gille spie, 2005; Rodinelli, 1990). The qualifying conditions for successful decentralization were incorporated into the constructs that define decentralization for this research: a) Decision making authority for service provision is delegated to sub national levels of government. b) Authority to manage expenditures and raise revenues is delegated to sub national levels of government. c) There are clearly defined rights, responsibilities, and roles for sector management actors at the national and sub national levels. d) Sub national level authorit ies are monitored by communities and/or the central government including, for example, regarding public funds expenditures and equality in service provision. e) Support is provided to build managerial and technical capacity of sub national levels of governme nt. Quantitative Results The quantitative analysis presented in Chapter V found that decentralization as represented by the data sourced from the UN GLAAS survey, is not associated with country achievement for urban water access. This runs counter to h ypothesis H1. However, it is important to note that not all constructs of decentralization were represented by the UN GLASS survey data on which the country decentralization values were based 42 so certain aspects of decentralization such as clearly defin ed roles for sector management actors, may have been absent even in countries with high decentralization values. Therefore, the observed statistical relationship should be viewed with caution. 42 Represented in the quantita tive data were constructs a) and b).

PAGE 244

231 F or urban sanitation the quantitative analysis showed that higher decentralization is associated with higher country progress which is consistent with hypothesis H1, although subject to the same cautionary note about construct representation mentioned above. This divergent relationship between decentralizat ion impact on water and sanitation was apparent in several ways, including the correlation direction, the ologit coefficients, the odds ratios, a nd the predicted probabilities. In exploring why opposing decentralization impacts were observed for water and sanitation, I drew p rimarily on f indings from the case studies to develop three propositions that are presented below. The actual explanation may depend on one or more of the following, which are not mutually exclusive: Proposition 1: Incomplete Realization of Decentralization: In many countries, including some case study countries, decentralizat ion has not been fully realized to encompass all aspects such as fiscal decentralization, consequently limiting the potential for theorized benefits. So, rather than concluding that decentralization is associated with a la ck of urban water progress, we might say that incomplete decentralization is associ ated with a lack of urban water progress Proposition 2: Sanitation Only R ecently Became a Development Priority Sanitation service has historically received less attention and been viewed as less of a priority than water, both globally and in a domestic context. In recent years sanitation has been promoted in the global develop ment community, so country progress o n urban sanitation from 2000 to 2012 could have been due to evolving development priorities as well as to decentralization.

PAGE 245

232 Proposition 3: Sanitation is Easier to Address at the Local Level of Government San itation most often appears in national budgets on the same budget line as water, and it is often dealt with as an add on to water planning. The nature of sanitation service may be such that local governments, once given the authority to manage it, could d evelop appropriate infrastructure options without a great deal of central government financial, t echnical and managerial support, to include supporting small scale sanitation service providers that employ relatively simple technologies suited to local sett ings. Theoretical Implications Aside from the fundamental question of why decentralization was shown to impact water and sanitation progres s differently, a number of other findin gs related to decentralization are discussed below along with their theoreti cal implications. Theory points to several benefits associated with de centralized government services. A llocative efficiency is thought to be improved because of increased competition and the ability to capture diverse preferences for public goods (Oates, 197 2; Musgrave 1997; Tiebout 1956 ; Rond inelli et al., 1989 ) However, fiscal decentralization requires sufficient central government authority to moderate excessive financial demands from local governments and to avoid inefficient spending (Oates, 2005). By example, spending inefficiency was a problem with local government s i n Colombia in the 1990s There, inefficiencies developed due to weak fiscal discipline at sub national levels of government, and national level fiscal reforms wer e ultimately implemented to modify the incentive structure for local governments and to mitigate moral hazard problems. However, some see the oversig ht provisions currently in place as too restrictive.

PAGE 246

233 Complete fiscal decentralization depends not only on an adequate local tax base but also on the ability of local government to enforce tax laws The reality in many countries, including case study c ountries, is that local funding for sector services comes from a combination of central government transfers and locally generated funds. New taxation rules designed to fund lower levels of government under decentralization must be clearly understood by t he public, which may require particular effort in low literacy countries In Mali, the taxation system was confusing at first, but was eventually reformed to clarify roles and communication. This is an area where civil society can help with understand ing of new tax systems as th e public is more likely to comply wi th tax rules if they understand how funds are directed to wards community services In situations where local governments rely on central government transfers for sector funding predictability in transfers allows for more effective investment planning The Colombia case showed more pred ictability in government transfer s than Mali, which allowed for more long term planning of investments by local government. Efficiency gains un der decentralization were negatively impact ed by institutional arrangements at the national level in both Colombia and Mali. A n array of sector agenc ies at the national level was confusing for municipalities in exercising their responsibilities, as they s ometimes received conflicting directives from different agencies. This underscores the importance of role clarification for different government agencies with sector authority as mentioned in the decentralization literature (Marcus and Onjala, 2008 ). A related benefit is that c larity of agency roles and responsibilities supports and streamlines sector planning activities.

PAGE 247

234 Another benefit of decentralization discussed in the literature is that local authorities can be more responsive to local needs and preferences than centralized systems ( Blair, 2000; Wunsch & Olowu, 1990; Wolman, 1990; Azfar et al., 2004). However, this cannot be fully realized in countries largely dependent on donor projects that operate outside the government budget and management system. In Mali, most sector projects in 2006 occurred outside of government budgets, therefore elements of local decision making and investme nt management were likely not in place for a majority of projects, which can hurt sustainability. Also, donor pr oject s may not have refelcted local comm unity priorities. S maller municipalities face particular obstacles in being responsive to community needs under decentralization This was a consistent theme fr om both Colom bia and Mali since smaller municipalities often have less knowledge of technical options, as well as a lower income population and lower overall commercial tax base. H aving a system of inter subsidies within a municipality is often in sufficient for funding as many communities do n ot have enough high income consumers to subsidize low income consumers. However, the national government can provide programs to overcome these obstacles. Examples from Colombia are royalty redistribution and the system of cross subsidies between cities and an example from South Africa is the Water Ser vices Trust Fund Another benefit of decentralization presented in the literature is better representation of good governance values, including equity in government services, positive government interactions with citizens, and improved political equity ( Wolman,

PAGE 248

235 1990; Pius Kulipossa, 2004) C itizen interaction is discussed more in a later sec tion but equity a nd political training are discussed here. In Colombia, equity was enhanced by decentralization related legislation that mandated a transfer system from the central government to local governments that included wider royalty distribution and transfers sp ecifically targeted to poorer and less populated areas of the country ( Granados & Snchez, 2014; World Bank, 2012a) In Mali, poorer municipalities are given little in the way of supplemental financial support ( World Bank, 2008 ), likely due to the limited overall economic base of the country. While t he case studies showed that capacity building is crucial for local governments to undertake new sector responsibilities, it was rarely undertaken prior to the start of decentralization. Tools for capacity building were eventually developed in Colombia, to include : software for financial manag ement and technology evaluation; certi f ication programs for operators; and, demonstration p rojects and exchanges between municipalities, focused on skills for private sector contracting and joint ventures. Recommendations for Practitioners This sub section lists select recommendations for practitioners regarding decentralization that emerged from this analysis: While altering traditional central government authority s tructures poses particular political challenges, it is worth investing the time and effort to clarify roles and responsibilities under decentralization. Local governments and do nors will find it easier to interact with fewer agencies with clear guidelines and procedures.

PAGE 249

236 Capacity building for local government is important prior t o the start of decentralization as well as periodic trainings for newly elected local officials. Donors can support this activity and in can be included in sector plans Systems for tracking municipal progress under decentralization, including baseline data, should be developed from the early stages of reform They can be valuable for measuring efficiency gains and progress in achieving service equity. This research show s there are many complex as pects to plan for and support when a country transitions from centralized service provision to a decentralized approach. Decentralization may place responsibility for water and sanitation service at sub national levels of govern ment, however their ability to exercise this authority effectively depends on a number of other factors th at are not always in place, even several years after dec entralization is put into law. Initiating decentralization reforms without first addressing i ssues such as funding mechanisms, roles and authorities for different agencies and government levels, and capacities needed on the part of local government, can create confusion and inefficiencies that limit the potential benefits Sector wide Strategy and Investment Coordinatio n Drawing on the literature about foreign aid effectiveness and coordinated sector planning I dev eloped the second hypothesis relating to sector wide strategy and investment coordination : H2. A high level of sector wide strategy and investment coordination in water and sanitation services is associated with greater country progress on providing urban water and sanitation access. Sector wide strategy a nd investment coordination wa s define d for this research as: mechanisms for transparent, country led partnering with donors as part of coordinating,

PAGE 250

237 directing, and utilizing international aid towards priority areas for the sector (adapted fro m Tropp, 2007). Elements detailed in the litera tur e for successful implementation of this attribute were reflected in the construct level definitions: a) The recipient country identifies priority projects and develops a realistic investment plan for the sector b) Development of partnering agreements between donors and recipient countries is driven by the recipient country c) A single agency is charged with coordinating all donor aid received in conjunction with a sector strategy d) Sector wide strategies are prepared in collaboration with stakeholders and don ors, who are provided with information about the sector e) A system of monitoring is in place for compliance with the terms of aid partnering agreements Quantitative Results Of the three attributes studied the one for which it was most difficult to draw conclusions was sector wide strategy and investment coordination (sector planning), as the phase one findin gs from the statistical models were mostly inconclusi ve. While correlations show ed a slight positive relationship between sector planning and a ll the outcome variables, the ologit results and predicted probability graphs show ed varied directional relationships. Only one outcome variable, Water 12 Year P rogress had a statistically significant relationship (positive) to sector planning. A follow on attempt to relate my outcome variables to sector planning data from two small scale outside studies 43 also s howed positive assoc iations between the sector planning measures and the Water 12 Year Pr ogress outcome variable (see Appendix F) So there is some 43 One study on PRSP inclusion was for ten countries (WSP, 2003), and another study on sector plans was for thirteen countries (Water Partnership Program, 2010).

PAGE 251

238 limited information in support of hypothesis H2 from the quantitative analysis, specific to the water sector. The four case studies, while not focused specifically on sector planning, did provide some information by which to further evaluate this a ttribute, since it was examined as a secondary topic for each case. The findings are discussed below. Theoretical Implications Research has brought attention to the shortcomings of traditional approaches to international development aid, which is often characterized by information, incentive and ownership problems (Gibson, Andersson et al., 2005 ; Riddell, 2007; Boone, 1996; Pedersen, 2001; Buchanan, 1977). According to theory, d onor aid is more impactful and sustainable when there is ownership and coordination on the part of t he recipient country, because country o wnership can support fo llow through on plans and build capacity for program management. Also on the topic of donor aid, p roblem s of asymmetric information can exist between donors and a id recipient countries (Martens et al., 2002: Rauchhaus, 2009; Pedersen, 2001). These problems have prompted scholars to suggest new modalities of foreign aid that foster more collaborative partnerships a mong donors and aid recipient countries so that financial resources can be used more ef ficiently towards country goals. Exemplified by sector wide approach (SWAp) and P overty Reduction Strategy Paper (P RSP) exercises, these modalities are theorized to increase the overall effectiveness of donor aid, by providing more predictabi lity and efficiency in aid allocation and also by allowing donors to monitor actions taken or not taken by the aid recipient country (WHO, 2013; Harpe, undated ) For secto r wide plans to be most

PAGE 252

239 ef fective, donors should be involved with plan development and their projects should be incorporated int o the plans. While more data are needed to study country d onor aid coordination practices, some anecdo tal evidence emerged from the case studies In Mali, d onor collaboration was enhanced with PROSEA in 2006 the national Water and Sanitation Sector Programme which led to a specific do nor coordination plan in 2008, illustrating the potential of planning exercises to improve information sharing. In Kenya, a 2009 Sector Investment Plan formed the basis for dev elopmen t of Partnership Principles, although a 2011 study found that not all donors were aware of or engaged with the p lan (WSP & AMCOW, 2011a). The case studies along with outside research also identified circumstances that can impede country ownershi p and cohesive planning. First, when responsibilities for water and sanitation at the national level are fragmented and complex, as has previously been noted for Colombia and Mali, this creates obstacles for cohesive sector planning. A recent study found that sector plan preparation is still seen largely as a donor led event in some countries rather than a country organized initiative ( EuropeAid Development and Cooperation Directorate, 2011 ). When different go vernment agencies perceive they are competing for the same pool of financial resources, the incentive to cooperate under one lead agency for plan development may be compromised (Slaymaker & Newborne, 2004). P lannin g exercises can be more effective if sensitive to the role of differ ent domestic agencies in contributing to overall sector goals, to counter any impression that the lead agency will receive credit for achievement of plan goals ( EuropeAid Development and Cooperation

PAGE 253

240 Directorate, 2011 ) This includes making sure that indiv idual contributions are clearly delineated in plans and that agencies are recognized accordingly for intermediary accomplishments, extending to those agencies that are not the lead agency for plan development On a related topic, i nvestment planning is fa cilitated when there are consistent budgeting procedures in place for different domestic agencies, a s well as for donor funded projects. R egular data collection for tracking investments and progress on plan goals is enhanced if t he modalities for data tracking and collection are well articulated to all participants. Results tracking efforts are most effective when focused on outcomes in line with plan g oals, rather than outputs, as may be customary for donor projects ( Mehta, 20 02 ) This approach helps with adjusting priorities and reallocating resources to needed areas during plan updates. The annual sector review process in Kenya and periodically updated Masibambane sector plans 44 in South Africa a re examples of this in practi ce. Experience in Kenya demonstrated that when civil society participates in sector plan development, collaborative processes and avenues of engagement initiated for planning purposes can ultimately strengthen ongoing dialogue betwe en government and civil society, increasing the i nclusiveness of government. Efficiency in Donor Aid Allocation Problems of adverse selection and moral hazard can exist with foreign aid allocation when sufficient information about country needs is not shared among all 44 The Masibambane sector plan was regularly updated until 2011, when donor funds that supported the program were shifted to other countries.

PAGE 254

241 pa rties, particularly between the aid recipient country and donors (Calmette and Kelkenny, 2002). Theories about foreign aid say that potential problems can be minimized needs and investment efforts. The JMP data collection process itself is a step towards addressing information problems and improving the efficiency of international aid. As JM P applies a consistent methodology across all countries to develop access rates, it guards against country misreprese ntation of domestic needs. P art of the JMP process is to research and reconcile differences with national access figures, as warranted, so this brings a degree of both consistency and fairness to the process. By referring to the JMP data, donors have the information necessary to direct aid to countries with the greatest need. Recommendations for Practitioners This sub section highlights select recommendations for practitioners regarding sector planning that emerged from this analysis: Considering that sanitation progress has lagged behind wat er, and that inclusion of sanitation was minimal in many PRSPs and sector plans, (WSP, 2003; Mehta 2002; Newborne, 2004), sanitation specific national plans, such as those developed in South Africa (2011) and Mali (2008 ), can help accelerat e progress in the sub sector by clarifying roles and directing donor support. Civil society has an ongoing role to play in monitoring and updating sector plans, even in countries that have achieved high access rates, so that advances and service quality can be sustained over time. Donors should support these activities, particularly in low inco me countries.

PAGE 255

242 Experience in Nairobi, Kenya shows there are potential benefits of SWAp type planning exercises at the municipal level, particularly for large urban areas. Donor funding of planning exercises directly at this level can enhance local owners hip potential. While results from both this analysis and the outside studies reviewed are not entirely clear on the impact of sector planning to date, there is recognition t hat SWAps can play an important role in the long term for securing water and sani tation access rates and addressing system maintenance needs. More data are needed to track not only sector plan development, but also donor coordination systems and plan m aintenance over time. This information would allow for a more complete evaluation o f hypothesis H2 for both water and sanitation. Civil Society Engagement The hypotheses I drew from the literature on civil society engagement is : H3. A high level of civil society engagement in water and sanitation services is associated with greater country progress on providing urban water and sanitation access. For purposes of this resea rch, civil society engagement was defined as: the existence of opportunities for the public and consumers to become involved in decision making, including program planning and implementation (adapted from deLeon and deLeon, 2002). Construct level definitions for civil society engagement are drawn from the literature and presented below: a) Opportunities exist for consumers and civil society groups to become involved in problem articulation. b) Opportunities exist for consumers and civil society groups to become involved in decision making about policy development and p lanning.

PAGE 256

243 c) Opportunities exist for consumers and civil society groups to become involved in decision making about implementation and management. d) Sector operates with transparency and information disclosure including for use of funds and natural resources. e) There is a feedback/complaint system in place. Quantitative Results Civil society engagement was consistently shown to be ass ociated with country progress for both water and sanitation demonstrating statistical significance across all four outcome variables in the ologit analyses, and showing similar relationships in th e bivariate tests. This offers strong support for hypothesis H3 for both sub sectors Building on this finding, the case studies se t out to illustrate ways t hat civil society engagement has been shown to be effective at different stages of water and sanitation service provision. Theoretical Implication s Civil society engagement is theorized to enhance the effectiveness of government a nd to improve government outcom es. Among other things, civil society engagement is said to promote government transparency and accountability (Thomas, 1995 ; E. Ostrom, 2005; Hooper, 2011; World Bank, 2004). Here, I highlight some examples from the cases studies that show how civil society engagement in the sector has helped to achieve these goals. Both Kenya and South Africa demonstrate sector transparency with service provider performance data now reporte d to the public on a regular basis. In Kenya, the WARIS program reports on key indicators for regional water service boards and service providers, and in South Africa, the Blue Drop program rates municipal service providers

PAGE 257

244 on a number of indicators link ed to an incentive based certification program for municipalities. T hese data collection efforts extend ing to all public and publically contracted private service providers, increase government transparency and accountability While data quality and rel iability can continually be improved, information on service provider performance is already prompting improvements among low scoring providers. This performance data can also be used by the central government and civil society to monitor service levels in different urban areas and to highlight equity concerns during sector planning activities The case study research also showed that civil society engagement can help gain and maintain the trust of consumers. In Kenya, u sing community leaders as int ermediaries between ser vice providers and consumers bridge s the trust gap between these two groups, through events such as public forums or ganized by a civil society organization (CSO) Similarly, public hearings hosted by the Water Action Groups (WAG) in Kenya bring together government, service providers and consumers, creating avenues of communication that might not otherwise develop. In South Africa, members of the South African Water Caucus network often serves as an intermediary between government an d local communities when conflict develops. Civil society engagement is also theorized to promote redistributional equity because the interests of poorer populations, particularly in urban areas, can be better represented through open and inclusive proc esses (Wijava, 2005; Sohail et al., 2005; McGranahan & Satterthwaite, 2006 ). While specific data on service provision equity was not collected or analyzed for this research, t he case studies show examples of civil society activities giving a greater voice to marginalized groups In Kenya, the MajiData

PAGE 258

245 website specifically reports low income area service indicators and provides opportunities for information exchange among interested parties. In S outh Africa CSO s play a significant role in highlighting problems of social injustice and giving voic e to disadvantaged populations, sometim es through protests and civil unrest. The value of civil society engagement for problem articulation was emphasiz ed in the literature for its value in helping to identify potential solutions before problems persist ( Sohail et al., 2005; McGranahan & Satterthwaite, 2006 ). It can also be a means for consumers to express preferences about service delivery in the local context (Mirumachi & Van Wyk, 2010; Lynn, 1994), and it can be help to identify and addr ess service problems. The merit of these ideas is reinforced by the experience in South Africa, where a lack of effective mechanisms for engaging the pub lic in problem recognition in some communities led to civil unrest being viewed as the only means to highlight deficiencies in sanitation service. This situation aggravated a climate of distrust and potential hostility betw een government and the public an d might have been avoided with greater engagement. The experience of South Africa shows that having regular forums for civil society engagement is important even in relatively high income countries, and even in relatively open societies. The literature showed that civil society engagement can be valuable at various stages of the policy development and implementation process (deLeon & deLeon, 2002; Saleth & Dinar, 2004). There are in fact, several case examples of c ivil society engagement improving gove rnment service outcomes at v arious stages. In Mali and Kenya, civil society contributed ideas in national forums that influenced the formation of national policies that addressed for example, sector financing and the recognition of

PAGE 259

246 sanitation as a priori ty issue. The MajiVoice customer feedback system in Kenya was proposed during an annual meeting to address the problem of consumer dissatisfaction with the ability to register complaints. This feedback tool, readily accessible to the public by phone and internet, has streamlined registering complaints and providing feedback to service providers in Kenya and also improved complaint resolution rates Civil society engagement in one stage of the process often br anches effectively into another. T he residents association s that participated in the C ard ( CRC ) exercise in Nairobi Kenya eventually formed KEWASNET, now act ive in national level advocacy. I n Mali, consultations held annually with stakeholders since 2007 have led to in creased CSO advocacy and policy impact, including creation of a watchdog group whose recommendations have been adopted by the government. While touting the many benefits of civil society engageme nt, the literature also cautioned about the difficulties of realizing some theorized benefits in a developing country setting, due to institutional and cultural barriers to full participation and education and time constraints among the general population (Research Triangle Institute, 200 4; Tevelow, 2004; Funke et al., 2007; Mirumachi & Van Wyk, 2010). My research found that i n less developed countries, CSOs have helped to overcome these challenges by educating the general public on issues surrounding water and sanitation. Examples ar e the Dialogue Day public hearings in Mali and public awareness campaigns by CSOs in South Africa. Larger CSOs have also helped educate local g overnment personnel and strengthened the capacity of local CSOs by hosting events such as the KEWASNET Learning Forums in Kenya and the S outh A frican W ater C aucus community based Water Dialogues in South Africa. S uch events may be of

PAGE 260

247 particular value in c ountries with low access rates and low literacy rates. A side benefit is that t he more the public understands issues surrounding decentralization such as how local revenues are raised and expended for the water sector the more likely they are to comply wit h taxation and tariff programs. Recommendations for Practitioners This sub section highlights select recommendations for practitioners regarding civil society engagement that emerged from this analysis: A starting small approach can be one way of promoting civil society engagement in countries that have not historically adopted participation principles, either culturally or politically. Community c onsultations through WAGs in Kenya led to CSOs being invited to participa te in a national plan sector working group, and a national Dialogue Day in Mali eventually was expanded to 25 communities. Larger CSOs can help educate local g overnment personnel and build the capacity of local CSOs, by hos ting events such as the Learni ng Forums in Kenya and the Water Dialogues in South Africa. Such events may be of particular value in countries with low access rates and low literacy rates. Smaller CSOs at the local level may need support from more powerful CSO groups in the country or outside donors to b uild their capacity, so they can raise their profile and be more impactful in smaller communities. Also, supporting newer CSOs helps guards against possible government tendency to interact primarily with larger, more established CSOs. The cases provide evidence of how civil society engagement has contributed to good government per formance, illustrating some reasons why the quantitative analyses

PAGE 261

248 provide strong support for hypothesis H3. While many countries have plans and legal requirements for civil society participation in the sector, these are not always realized in practice if government resources a nd motivation are not present both at the national and local government levels. The donor community can help support meaningful civil society engagement and draw from these positive examples of civil society impact. Limitations of this Study This section will discuss key limitatio ns of the analysis and results presented in this disse rtation. First, there is the inherent diffi culty of find ing meaningful measures for big picture concepts like those represented by the key explanatory variables. The GLAAS data used to derive the key explanatory variables did not capture all the constructs of interest, therefore, the content valid ity for the quantitative analyses was less than complete, and some constructs were only evaluated through the case studies Also, because GLAAS survey responses are self reported, attempts were made to triangulate the variables derived from GLAAS data wit h other independent data sources that measure similar concepts (see Chapter IV ). Co mparisons done for dec entralization and civil society engagement generally supported the self reported data by revealing moderate levels of positive correlation with outside indices. A potential th reat to internal validity for this analysis is history that is, events outside of the institutional and governance attribute s being studied that have impacted country progress on water and sanitation. The use of contr ol variables in the statistical models addressed this to some degree, by recognizing differences among countries on

PAGE 262

249 key background conditions, 45 and the case studies allowed for more in depth examination of count ry context and other events. Another limita tion is that the collection time period for country institutional and governance data, (2011 GLAAS Survey), does not directly correspond with th e time period represented in the 12 Year Progress outcome variables, (2000 through 2012), and there may have bee n changes in the institutional and g overnance environment during this period that are not reflected. Th is temporal problem was less of an issue in the case studies which chronicled events over several years with attention to the sequencing of events. Co ncerning external validity, the 75 country sample 46 consisted of those countries who responded to the 2011 GLAAS survey. On average, t he sample group countries are higher in population, lower in GDP per capita, and lower in water and sanitation acc ess rates than the entire group of 139 low and middle income countries Areas for Further Research The statistical model developed for this research can be adapted to analyze more recent GLAAS and JMP data, as well as alternative measures of institutions and governance to see how the results compare to those presented here for the 2000 to 2012 time period. In addition, several other lines of research merit further exploration, including: 45 One shortcoming was the inability to identify and include a foreign aid indicator for the sector that distinguished among aid to service related programs, as opposed to w ater resources programs, and was available for a large number of countries. 46 The actual number of countries included in different bivariate and multivariate tests ranged from 66 to 75, due to outliers and other data issues that caused some countries t o be excluded from c ertain tests (see Appendix B).

PAGE 263

250 Why did the quantitative results show that decentralization impacts water and sanitation progress differently? The three propositions put for th to answer this question could be further explored with additional data collected from multiple countries. What additional data would be useful to collect about sector pl ans to better assess their impact? Examples would be degree of investment planning, frequency of plan updates, extent to which aid partnering agreements are driven by recipient countries and if monitoring is in place for compliance with aid agreements. Given the pressing need for advancement in urban sanitation access, h ow can data collection efforts be enhanced to specifically focus on sanitation indicators, such as measures of service access, quality, and equity and also to learn more about how local governments are managing these services under decentralized systems? How are national level performance tracki ng systems for in dividual service providers being used by civil society for third party monitoring? Is there additional data that could be collected to better promote transparency and accountability? Conclusion With a focus on low and middle income countries t his dissertation evaluated if three specific institutional and governance attributes, theorized as ben eficial in public affairs literature have had a positive impact on government performance, specifically for urban w ater and sanitation service provision For the water sub sector, regression test results showed that decentralization was negatively associated with the country progress outcome variables, perhaps due to

PAGE 264

251 incomplete implementation of all aspects of decentralization, as was observed in the case studies. For th e sanitation sub sector, decentra lization was positively associated with the outcome variables, although other factors may have contributed to sanitation achievement in recent years, including increased donor attentio n to the sub sector. Regression results for sector wide strategy and investment coordination were mostly i nconclusive, except that it showed modest significance as a positive factor for the Water 12 Year Progress outcome variable. Civil society engagement was statistically significant as a positive factor for all water and sanitation outcome variables. The se quantitative findings coupled with insights from the four country case studies, have implications for theories about government provision of public goods (Weimer & Vining, 2005) which center on : a) addressing efficiency and undersupply problems ; and b) im pr oving social welfare and equity of distribution How the findings of this research relate to each of these topics is discussed below. Efficiency and Undersupply Problems The con tribution of decentralization towards addressing efficiency and undersupply problems in urban water and sanitation provision is not evident. The case studies showed that many aspects of decentralization, as outlined in the constructs for this research, are not fully implemented thus creating several impediments to realizing theorized efficiency benefits. Local governments faced difficulty in being more responsive to local needs when gu idelines originating from central government agenci es were confusing and when local decisions were subject to approval by central government authorities. Potential for improvement in resource allocation was also shown to be compromised when some local governments subject to soft budget constraints from the

PAGE 265

252 central government demonstrated spending inefficiencies. Effective f iscal rules, predictability of transfers from the central government, local tax collection ability, and the extent to which local governments have i nfluence over donor projects are all important for realizing improved resource allocation and efficiency under de centralization, and these are not always present. In addition, a lack of local government capacity to, for instance, select technologies and effectively negotiate private sector contracts, compromises potential efficiency gains. The unders upply problem in small and low income cities is most effectively addressed under decentralization when they are given addition al support in the form of local manager trainings and targeted financial support programs. Sector planning was shown, through limited examples to improve efficiency and undersupply problems by minimizing conflicting efforts on the part of donors and domestic agencies, to the extent that all parties participate in and share information during sector planning exercises. Monitoring sector outcome s and regularly updating plans are important for setting new goals and maintaining efficiency gains. Plans specifically focused on sanitation can bring coordinated effort to address ing the significant undersupply problems in that sub sector. Ci vil socie ty engagement was shown to have improved sector efficiency by yielding good programmatic ideas, in both public forums and sector planning exercises, as well as by highlighting deficiencies and service issue s. Public input on service quality is increasingl y impactful through widely accessible technology based feedback and complaint mechanisms. Further, when civil society educates the public on the financial aspect s of water and sanitation services improvement in tax revenues can result making

PAGE 266

253 more resou rces available for service provision. When civil society helps educate local government managers, this improves the potential for efficient and expanded service delivery at the local level Even in relatively high income high education countries civil society has value for buildi ng trust between service providers and the public and serving as a watchdog on local government spending and power abuse. Social Welfare and Equity of Distribution Higher urban water and sanitati on access rates, which have occurred in almost all sample countries since the year 2000, represen t an improvement in overall social welfare by reducing the potential for health problems that stem from poor water supply and sanitation. The contribution of the three attributes studie d in this research towards these improvements is discussed here. ity of distribution is ambiguous particularly when considering the capacity and financing challenges faced by many small and low income municipalities, where the greatest service needs often exist. Generally speaking, it was when targeted financial support pro grams were made available to these municipalities that they were able to make stride s in reaching underserved areas, as well as when overall GDP per capita improved. Comprehensive r eporting on local service provider performance is a decentralization related practice that has been shown to improve overall societal welfare and improve equity in distribution, because it m otivates municipalities to address identified deficiencies. Finally, one c annot entirely evaluate equity and societal welfare impacts of decentralization in co untries where donor led projects dominate So, w hile the overall contribution of decentralizati on towards increased equity a nd social welfare is unclear

PAGE 267

254 the potential for centralized systems or the private sector market to perform better is also uncertain. The varied stakeholder i nput afforded by sector planning exercis es has been shown to help d irect both domestic and donor resources toward programmatic are as of need, such as sanitation, thereby contributing to overall social welfare. The impact of sector planning specifically for help ing to enhance service equity in low income areas is less cle ar from the available data. Government engagement with civil society has improved the distributional equity of urban water and sani tation services. Community outreach activities and broad access feedback tools give voice to a wide range of urban residents, even those living in informal settlements, and provide some protection against ignoring their needs. Also, i nitiatives geared towards gr eater transparency, including the publishing of service provider performance data have allowed civil societ y to draw attention to service deficiencies in specific geographic a reas This oversight role was found to be relevant in situations of both weak and strong local government capacity, as performance often relates more to local government priorities than l ocal government capacity. The more opportunities for en gagement that exist between central and local governments and civil society, the more civil society can monitor equity in service provision, express concerns about social inequality and suggest servi ce solutions suitable for low income areas Coordination within Government In reviewing findings from this research, an overall theme emerged about the value of coordination in government services For example, s tandardized approaches to reporting on service provider performance, where local governments collect and report

PAGE 268

255 these data to a central government agency helps with service monitoring and quality. By contrast, when coordination is lacking among central g overnment agencies under decentralization, the confusion of roles obstructs the ability of local governments to implement their responsibilities efficiently. A la ck of coordination among government agencies can also create obstacles for transparent and s treamlined sector planning. Nonetheless, any degree of information sharing and coordinated effort among parties engaged in sector planning, including government agencies service providers, donors and civil society, reduces problems of information asymmet ry and moves toward greater efficiency and supply. Similarly, coordinated governme nt effort to engage with civil society, particularly for events occurring on a regular basis, yields positive results in the form of increased trust and improved ser vice qua lity Closing On balance, this dissertation provides a variety of evidence of how instit utions and governance can enhance government provision of public goods such as water and sanitation service provided that attention is paid to how institutions are o rganiz ed and how governance is put into practice The findings and conclusions detailed above have implications for all low and middle income countries, even those that have already achieved significant urban water and sanitation access rates. The continuing goal for these countries and the global community is to reach the remaining urban population without adequate service and to sustain the quality of existing servi ces through ongoing sector development efforts.

PAGE 269

256 BIBLIOGRAPHY Abrahamsen, R. (2004). The power of partnerships in global governance. Third World Quarterly 25 (8), 1453 1467. African Development Fund. (2007). Mali Drinking Water Supply and Sanitation Projects in Gao, Koulikoro, and Segou Regions: Appraisal Report. Tu nis, Tunisia. Agyeman, K. (2007). Privatization of water in Ghana: stopped in its tracks or a strategic pause?. International Journal of Environmental Studies 64 (5), 525 536. Alesina, A. and B. Weder (2002). American Economic Review, 92, 1126 1137. Anbarci, N., Escaleras, M., & Register, C. A. (2009). The ill effects of public sector corruption in the water and sanitation sector. Land economics 85 (2), 363 377 Andres, L. A., D. Sislen, and P. Marin, eds. (2011). Charting a New Course: Structural Bogot: PPIAF/World Bank. Anzagi, B. (2015). Response to author's questionnaire, received July 7, 2015. Araral J. (2008). Public provision for urban water: getting prices and governance right. Governance 21 (4), 527 549. Asthana, A. N. (2003). Decentralisation and supply efficiency: the case of rural water supply in Central India. Journal of Development S tudies 39 (4), 148 159. Auer, M. R. (2007). More aid, better institutions, or both?. Sustainability Science 2 (2), 179 187. Azfar, O., Khknen, S., & Meagher, P. (2001). Conditions for effective decentralized governance: A synthesis of research finding s University of Maryland, Center for Institutional Reform and the Informal Sector. Azfar, O., Kahkonen, S., Lanyi, A., Meagher, P., & Rutherford, D. (2004). Decentralization, governance and public services: The impact of institutional arrangements. Devol ution and development: Governance prospects in decentralising states 19 62. Baer, M. (2009). Turning the tide on water privatization: civil society and the World Bank. Conference Papers -International Studies Association 1. Bahl, R., & Linn, J. (1992) Urban public finance in developing countries. Oxford University Press. New York

PAGE 270

257 Bahl, R., & Wallace, S. (2005). Public financing in developing and transition countries. Public Budgeting & Finance 25 (4s), 83 98. Bahl, R. W., Linn, J. F., & Wetzel, D. L. (2013). Financing metropolitan governments in developing countries. Bakker, K., Kooy, M., Shofiani, N. E., & Martijn, E. J. (2008). Governance failure: rethinking the institutional dimensions of urban water supply to poor households. World Development 36 (10), 1891 1915. Bardhan, P. (1997). Corruption and Development: A Review of Issues Journal of Economic Literature 35(3), pp. 1310 1346. \ Bardhan, P. K., & Mookherjee, D. (2000). Corruption and decentralization of infrastructure delivery in developing countries Boston University, Institute for Economic Development. Bardhan, P., & Mookherjee, D. (2005). Decentralizing antipoverty program delivery in developing countries. Journal of Public Economics 89 (4), 675 704. Barrera Osorio, F., & Olivera, M. (2007). Does Society Win or Lose as a Result of Privatization?: Provision of Public Services and Welfare of the Poor: The Case of Water Sector Privatization in Colombia Inter American Development Bank. British Journal of Political Science. Volume 18, 1988, pp. 353 395. Blair, H. (2000). Participation and accountability at the periphery: democratic local governance in six countries. World Development 28 (1), 21 39. Bond, P. (2013). Water rights, commons and advocacy narratives. South African Journal of Human Rights 29 (1), 126 144. Bond, P. (2014). The political economy of water management: Neoliberalism and social resistance in South Africa. Presented to the Seminar on Water Governance, Allocation and Regulation, School of Oriental and African Studies, University of London, May 30, 2014. Boone, P. (1996). Politics and the effectiveness of foreign aid. European Economic Review 40 (2), 289 329. Botting, M. J., Porbeni, E. O., Joffres, M. R., Johnston, B. C., Black, R. E., & Mills, E. J. (2010). Water and sanitation infrastructure for health: The impact of foreign aid. Globalization and Health 6 12.

PAGE 271

258 Briscoe, J ., & Garn, H. A. (1995 ). Financing water supply and sanitation under Agenda 21. Natural Resources Forum (Vol. 19, No. 1, pp. 59 70). Blackwell Publishing Ltd. Buchanan, J. M. (1977). "The Samaritan's Dilemma." In Freedom in constitutional contract: Perspectives of a political economist Texas A & M University Press College Station and London. Calmette, M. F., & Kilkenny, M. (2002). International charity: For the poor?. Economic Inquiry 40 (3), 497 507. Centre for Environmental Rights. ( 2012 ). Stop Treading Wat er: What civil society can do to get water governance in South Africa back on track. Cape Town, South Africa. Chacon, M. L. (2013). In the Line of Fire: Political Violence and Decentralization in Colombia. Available at SSRN 2386667 Chang, M., Memon, M. A., & Imura, H. (2004). Public Private Partnerships for Urban Environmental Infrastructure, Learning from International Experiences and Its Application to China. China Council for International Cooperation on Environment and Developmen t, Institute for Global Environmental Strategies, Financial Mechanisms for Environmental Protection in China, China Environmental Science Publisher, Beijing 175 238. Cheru, F. (2006). Building and supporting prsps in Africa: What has worked well so far? What needs changing?. Third world quarterly 27 (2), 355 376. CIA (2011). CIA World Factbook. Coulibaly, C. (2010). Decentralization in Mali: A Constrained Responsibi lity Transfer Process. Cremer, J., Estache, A., & Seabright, P. (1994). The decentralization of public services: lessons from the theory of the firm (Vol. 1345). World Bank Publications. Davison, C. A. (2001 ). Urban governance and the effective delivery and management of infrastructure services in urban areas in Zimbabwe: An appraisal of water and sewerage services delivered in Ruwa. In Urban Forum (Vol. 12, No. 2, pp. 139 170). Springer Netherlands. Inter Press Service August 23, 2006 deLeon, L. and P. deL eon (2002). "The Democratic Ethos and Public Management." Administration & Society 34(2): 229.

PAGE 272

259 Department of Water and Environmental Affairs, Republic of South Africa. (2008). Masibambane: Annua l Report 2007/08. Department of Water and Sanitation, Republic of South Africa. (2004). Civil Society Strategy to achieve the meaningful involvement of civil society in the Masibambane Prgoramme. Department of Water and Sanitation, Republic of South Afr ica. (2005). Civil Society Advocacy and Communication Strategy: Action and Implementation Plan. Department of Water and Sanitation, Republic of South Africa. (2011). Final Evaluation of the Masibambane III Programme FY 2007/2008 FY 2010/11. Department of Water and Sanitation, Republic of South Africa. (2013). Blue Drop Progress Report: Executive Summary. Retrieved from https://www.dwa.gov.za/dir_ws/DWQR/ on June 2, 2015. Development Assist ance Committee of the OECD. (2005). Harmonizing donor practices for effective aid delivery. Vol. 3. Strengthening procurement capacities in developing countries. Direction Nationale des Collectivits Territoriales (Mali), & Lux Dev (Luxembourg). (2013). Recueil des textes de Base de la Dcentralisation. Dudley, R. (2000). The Rotten Mango: The Effect of Corruption on International Development Projects Part 1: Building a System Dynamics Basis for Examining Corruption Oregon, USA. Eales, K. (2008). Par tnerships for sanitation for the urban poor: Is it time to shift paradigm? IRC Symposium: Sanitation for the Urban Poor: Partnerships and Governance. Delft, Netherlands. Easterly, W. (2009). How the millennium development goals are unfair to Africa. Worl d Development 37 (1), 26 35. Echavarra, J. J., Rentera, C., & Steiner, R. (2002). Decentralization and bailouts in Colombia. Inter American Development Bank, Latin American Research Network, Red de Centros de Investigacin Research Network Working p aper #R 442 Environmental Monitoring Group. (2005). The Millennium Development Goals: A guide for South African civil society. Cape Town, South Africa. Estache, A. and Kouassi, E. (2002). Sector organization, governance, and inefficiency of African water utilities. Washington, U.S., The World Bank Institute. World Bank Policy Research Working Paper No. 2890

PAGE 273

260 EuropeAid Development and Cooperation Directorate General European Commission. (2 011). Study of Sector wide Approach in the Water Sector. Brussels Luxemburg. Evans, B. (1995). Institutional incentives and the provision of urban sanitation: Can the community increase the chances of success?. London School of Economics, London Evans, supply. Summit, Bellagio, Italy, July. Everatt, D., P. Ravenscroft, et al. (2007). Summative Evaluation o f the Masibambane II Programme. Department of Water Affairs and Forestry, South Africa. Faguet, J. P., & Snchez, F. (2014). Decentralization and access to social services in Colombia. Public choice 160 (1 2), 227 249. Falleti, T. G. (2010). Decentralization and subnational politics in Latin America Cambridge University Press. FAO, AQUASTAT [Data file]. Retrieved May 7, 2014 from http://www.fao.org/nr/water/aquastat/main/i ndex.stm FAO, AQUASTAT. (2015). Water Resources Important Note. Retrieved November 18, 2015 from http://www.fao.org/nr/water/aquastat/water_res/index.stm Farvacque Vitkovic, C., Casalis, A., Diop, M., & Eghoff, C. (2007). Development of the Cities of Mali. World Bank. Washington, D.C. Fischer, F. (2000). Citizens, Experts, and the Environment Durham and London Duke University Press. Fischer, F. (2003). Reframing Public Policy: Discursive Politics and Deliberative Practic es Oxford, Oxford University Press. Fogden, J. (2009). Access to Safe Drinking Water and Its Impact on Global Economic Growth. Forbes, M., & Lynn, L. E. (2005). How does public management affect government performance? Findings from international researc h. Journal of Public Administration Research and Theory 15 (4), 559 584. Ground. World Bank.

PAGE 274

261 Frangano, F., Linares, C., Lockwood, H., Rivera, D., Trevett, A., & Yepes, G. (2001). Strategic Paper No. 1: Case Studies on Decentralization of Water Supply and Sanitation Services in Latin America. Prepared for the USAID Bureau for Latin America and the Caribbean. Washington, DC: Environmental Health Project. Frank, J ., & Martinez Vazquez, J. (2014). Decentralization and infrastructure: from gaps to solutions. International Center for Public Policy. Funke, N., Nortje, K., Findlater, K., Burns, M., Turton, A., Weaver, A., & Hattingh, H. (2007). Redressing inequality: South Africa's new water policy. Environment: Science and Policy for Sustainable Development 49 (3), 10 23. Gathuru, N.H. (2014). Factors Influencing Sustainability of Donor Funded Projects: The Case of Water and sanitation Projects in Laikipia East District, Laikipia County, Kenya. Gerring, J. (2007). Case study research: principles and practices, Cambridge University Press Cambridge. Gibson, C. C., K. Andersson, et al. (2005). The Samaritan's Dilemma, Oxford University Press. Gillespie, B. (2005). Financing urban water supply and sanitation. International Review for Environmental Strategies 5 (2), 449. Giugale, M., Lafourcade, O., & Luff, C. (Eds.). (2003). Colombia: the economic foundation of peace World Bank Publications. Global Water Partnership. (2002). Introducing Effective Water Governance mimeo, April 2002. Global Water Partnership Technical Committee (2003). Effective Water Governance. Gomez, C. F. (1991). Regulating death: euthanasia and the case of the Netherlands Free Pr. Granados, C., & Snchez, F. (2014). Water reforms, decentralization and child mortality in Colombia, 1990 2005. World Development 53 68 79. Gupta, S., H. Davoodi, & R. Alonso Economics of Governance 3, 23 45. Habib, M., & Zurawicki, L. (2002). Corruption and foreign direct investment. Journal of international business studies 291 307.

PAGE 275

262 Hall, D. Lobina, E., & Motte, R. D. L. (2005). Public resistance to privatisation in water and energy. Development in Practice 15 (3 4), 286 301. Hamilton, L. (2012). Statistics with STATA: Version 12 Cengage Learning. Handoussa, H. (2009). Lessons from the MD Gs in Africa. African Development Review 21 (2), 213 223. Hanjahanja, R. (2011). Thinking about Scale from the Start. (PowerPoint presentation). US Agency for International Development. Washington, D.C. Harpe, Jean de la ( undated wide approach 20, 2013 from http://www.irc.nl/page/49510 Hill, C. J., & Lynn, L. E. (2005). Is hierarchical governance in decline? Evidence from empirical research. Journal of Public Administration Research and Theory 15 (2), 173 195. Hooper, B. P. (2011). Integrated water resources management and river basin governance. Journal of Contemporary Water Research and Education 126 (1), 3. iHub Research. (2012). Water governance in Kenya: Ensuring accessibility, service delivery, an d citizen participation. Innes, J. E., & Booher, D. E. (1999). Consensus building and complex adaptive systems: A framework for evaluating collaborative planning. Journal of the American Planning Association 65 (4), 412 423. Inter American Development Bank. (1997). Latin America after a Decade of Reforms. Economic and Social Progress in Latin America Washington, D.C International Monetary Fund. ( 2002 ). Mali: Heavily Indebted Poor Countries Initiative Assessment and Action Plan. Washington, D.C. International Monetary Fund. (2008). Mali: Poverty Reduction Strategy Paper, Issues 8 121 of IMF Country Report. Washington, D.C. International Monetary Fund (2013). "Debt Relief Under the Heavily Indebted Poor Countries (HIPC) Initiative." Retrieved Oct. 10, 2013 from http://www.imf.org/external/np/exr/facts/hipc.htm International Monetary Fund (2015). Poverty Reduction Strategy in IMF supported Programs Factsheet dated July 16, 2015 Washington, D.C.

PAGE 276

263 Iribarnegaray, M. A., & Seghezzo, L. (20 12). Governance, Sustainability and Decision Making in Water and Sanitation Management Systems. Sustainability 4 (11), 2922 2945. ISF UTS. (2011). South Africa Water, Sanitation and Hygiene Sector Brief Prepared for AusAID by the Institute for Sustainable Futures, University of Technology, Sydney, Australia. Ivanyna, M., & Shah, A. (2014). How close is your government to its people? Worldwide indicators on localization and decentralization. Economics: The Open Access, Open Assessment E Journal 8 (2014 3), 1 61. http://dx.doi.org/10.5018/economics ejournal.ja.2014 3 ; and related dataset al : http://dx.doi.org/10.7910/DVN/24566. Izaguirre, A. K. (2002). Private infrastructure: a review of projects with private participation, 1990 2001. World Bank, Washington, D.C. Jones, D., & Williamson, T. (2005). How collaboration is sharping the water services sector in South Africa. BPD Occasional Paper. Water Information Network, South Africa. Jtting, J., Corsi, E., Kauffmann, C., McDonnell, I., Osterrieder, H., Pinaud, N., & Wegner, L (2005). What makes decentralisation in developing countries pro poor?. The European Journal of Development Research 17 (4), 626 648. Kaufmann, D.; Kraay, A.; Mastruzzi, M. (2006). Measuring Corruption: Myths and Realities The World Bank. Keefer, P., & Khemani, S. (2005). Democracy, public expenditures, and the poor: understanding political incentives for providing public services. The World Bank Research Observer 20 (1), 1 27. Kenya Human Rights Commission, Social and Public Accountability Network. (2010). Harmonization of Decentralized Development in Kenya: Towards Alignment, Citizen Engagement and Enhanced Accountability. Kessides, I. (2004). Reforming Infrastructure: Priva tization, Regulation, and Competition; Chapter 5: Reforming the Water Sector. The World Bank and Oxford University Press. King, G., Keohane, R. O., & Verba, S. (1994). Designing social inquiry: Scientific inference in qualitative research Princeton University Press. Klitgaard, R.E. (1988). Controlling Corruption Berkeley, California. University of California Press

PAGE 277

264 Koundouri, P., Saleth, R.M., et al. (2008). Evaluating the Institution Impact Interactions in the Context of Millennium Development Goals. Coping with Water Deficiency Springer Netherlands. 48: 189 212. Krause, M. ( 2007 ) Countr ies: Cross Dissertation for the degree of Doctor of Economics, Department of Economics, Justus Liebig Universitt Gieen Lane, J. (2004). Ghana, Lesotho, and South Africa: Regional Expansion of Water Supply in Rural Areas. Prepared for Scaling Up Poverty Reduction: A Global Learning Process and Conference, Shanghai, China, May 25 27, 2004. Lenton, R., A. M. Wright, et al. (2005). "Health, dignity and development: what will it take? UN Millennium Project Task Force on Water and Sanitation." London, UK: United Nations Development Program Li, M. (2011). Walking on the Tightrope" -Can Water TNC Tackle Drinking Water Crisis in Developing Countries?. Asian Social Science 7 (5), p122. Litvack, J., & Seddon, J. (Eds.). (1999). Decentralization briefing notes Washington DC: World Bank Institute. Lobina, E., & Hall, D. (2008). The comparative advantage of the public sector in the development of urban water supply. Progress in Development Studies 8 (1), 85 101. Lopez Murcia, J. (2015a). Water supply a nd sanitation in Colombia: From decentralisation to recentralisation. Posted March 16, 2015 to www.GlobalWaterForum.org under Development, Governance. Lopez Murcia, J. (2015 b ). Interview Summary and follow up correspondence. Interview conducted with author by Skype, May 18, 2015. Lynn, L. E. (1994). Public management research: The triumph of art over science. Journal of Policy Analysis and Management 13 (2), 231 259 Madrigal, R., Alpzar, F., & Schlter, A. (2011). Determinants of performance of community based drinking water orga nizations. World Development 39 (9), 1663 1675. MajiData website, http://www.majidata.go.ke/information.php?MID=MTA = ( accessed June 8, 2015 ) Water Services Regulatory Board, Kenya.

PAGE 278

265 Marcus, R. R., & Onjala, J. (2008). Exit the State: Decentralization and the need for local social, political, and economic considerations in water resource allocation in Madagascar and Kenya. Journal of Human Development 9 (1), 23 45. Marshall, M. G., & Cole, B. R. (2011). Global Report 2011: Conflict, Governance, and State Fragility, Center for Systemic Peace. Martens, B., Mummert, U., Murrell, P., & Seabright, P. (2002). The institutional economics of foreign aid Cambridge University Press. Mauro, P. (1995). Corruption and growth. The quarterly journal of economics 681 712 McGinnis, M. D. (1999). Polycentric governance and development: Readings from the workshop in political theory and policy analysis University of Michigan Press. McGinn is, M. D. (2011). Networks of adjacent action situations in polycentric governance. Policy Studies Journal 39 (1): 51 78. McGranahan, G., & Satterthwaite, D. (2006). Governance and getting the private sector to provide better water and sanitation services to the urban poor International Institute for Environment and Development (IIED). Mehta, Meera. (2002). Water Supply and Sanitation in PRSP Initiatives: A Desk Review of Emerging Experience in Sub Saharan Africa (SSA). Water and Sanitation Program. Nairobi, Kenya. Meng, X., Zhao, Q., & Shen, Q. (2011). Critical success factors for tra nsfer operate transfer urban water supply projects in China. Journal of Management in Engineering 27 (4), 243 251. Miller, Gary J. Managerial Dilemmas. New York: Cambridge University, 1992. Mirumachi, N., & Van Wyk, E. (2010). Cooperation at different scales: challenges for local and international water resource governance in South Africa. The Geographical Journal 176 (1), 25 38. Mjoli, N. (2010). Review of Sanitation Policy and Practice in South Africa fro m 2001 2008. Water Research Commission. Moe, T. M. (1984). The New Economics of Organization. American Journal of Political Science 28 (4), 739 777. Mookherjee, D., & Bardhan, P. (2005). Decentralization, Corruption a nd Government Accountability: An Overview (No. WP2005 023). Boston University Department of Economics.

PAGE 279

266 Moremong, H. (2015). Response to author's questionnaire, received August 5, 2015. Morrison, L. (undated). Raising the Citizens' Voice in the Regulation of Water Services. Mvula Trust. Johannesburg, South Africa. Mtolo, J. (2015). Response to author's questionnaire, received August 19, 2015. Mugabi, J., & Njiru, C. (2006). Managing water services in small towns: Challenges and reform issues for low income countries. Journal of Urban Planning and Development 132 (4), 187 192. Munnik, V., Wilson, J., Pereira, T. (2014). Assessing the role of civil society in South African water policy within the context of legal requirements and democratic discourse. Environmental Monitoring Group. Cape Town, South Africa. Musgrave, R. A. (1997). Devolution, grants, and fiscal competition. The Journal of Economic Perspectives 11 (4), 65 72. Musgrave, R., & Musgrave, P. P.(1989) Public Finance in Theory And Practice International Edition McGraw hill Book Company Nabli, M. K., & Nugent, J. B. (1989). The new institutional economics and its applicability to development. World Development 17 (9), 1333 1347. Ndii, D. (2010). Decentralization in Kenya: Background Note Ndun g'u, M (2009). Kenya Water and Sanitation Programme and the Water Sector Reform Programme: A Joint Sida, GTZ and Government of Kenya Mid Term Evaluation Mission Report Newborne P. (2004). Making the Case for Water: A Review of Poverty Reduction in Ten Countries in Africa, Asia and Latin America Overseas Development Institute. London, UK. Newborne, P., Zuleta, C. M., & Lleras, G. R. (2010). Review of the Sector wide approach in environment in Colombia: 2007 2010. North, D. C. (1990). Institutions, Institutional Change and Economic Performance Cambridge, UK Cambridge University Press Nyamu Musembi, C., Musyoki, S. (2004). Kenyan civil society perspectives on rights, rights based approaches to development, and participation. Institute of Development Studies. Brighton, Sussex, England. Nyangena, K.O. (2008). Privatization of water an d sanitation services in Kenya: Challenges and prospects. Africa Development 33 (4).

PAGE 280

267 Nye, J. S. and Donahue, J.D. (2000). Governance in a Globalizing World Brookings Institution. Washington, D. C. Oakerson, R. J., & Parks, R. B. (1999). Citizen Voice and Public Entrepreneurship: The Organizational Dynamic of a Complex Metropolitan County. In Polycentricity and local public economies: readings from the Workshop in Political Theory and Policy Analysis (p. 306). Univ of Michigan Press. Oakley, P. (1991). The concept of participation in development. Landscape and Urban Planning 20 (1), 115 122. Oates, W. E. (1972). Fiscal decentralization. Oates, W.E. (2005). Toward a second generation theory of fiscal federalism, Inter national Tax and Public Finance, 12, 349 373. O E C D. (2002 2012). Water Supply and Sanitation: Overseas Development Assistance [Data file query]. Available from http://www.oecd.org/dac/stats/idsonline.htm r and Sanitation Services in Development Co operation Report 2002: Efforts and Policies of the Members of the Development Assistance Committee OECD Publishing. OECD. (2006). DAC guidelines and reference series. Applying strategic environmental assessment: Good practice guidance for devel opment co operation, OECD Publishing. OECD. (2009). Private Sector Participation in Water Infrastructure: OECD Checklist for Public Action OECD Publishing. Onda, K., LoBuglio, J., & Bartram, J. (2012). Global access to safe water: accounting for water quality and the resulting impact on MDG progress. International journal of environmental research and public health 9 (3), 880 894. Opiyo, R. (2015). Response to author's questionnaire received July 27, 2015. ORGUT Consulting (AB) Kenya, Ltd. (2010). Lessons Learned and Good Practices from Support to the Kenyan Water Sector. Royal Dutch Embassy/DANIDA. Ostrom, E. (1986). "An Agenda for the Study of Institutions." Public Choice 48(1): 3 25.

PAGE 281

268 Ostrom, E. (2003). How types of goods and property rights jointly affect collective action. Journal of Theoretical Politics 15 (3), 239 270. Ostrom, E. (2005). Understanding Institutional Diversity Princeton University Press. Ostrom, E. (2011). Background on the Institutional Analysis and Development Framework. Policy Studies Journal 39 (1): 7+. Ostrom, V. (1999 ). Cryptoimperialism, predatory states, and self governance. In Polycentric governance and development: readings from the Workshop in Political Theory and Policy Analysis (pp. 166 85). Ostrom, V., & Ostrom, E. (1972). Legal and political conditions of water resource development. Land Economics 48 (1), 1 14. Ostrom, V., & Ostrom, E. (1977). A theory for institutional analysis of common pool problems. Managing the commons 157 72. Ostrom, V., Tiebout, C. M., & Warren, R. (1961). The organization of government in metropolitan areas: a theoretical inquiry. The American Polit ical Science Review 55 (4), 831 842. Overseas Development Institute. (2004). Report on Water and Poverty Reduction for the UK. Overy, N. (2013). The Social Justice Coalition and A ccess to Basic Sanitation in Informal Settlements in Cape Town, South Africa. Study No. 11. International Budget Partnership. Pedersen, K. R. (2001). The Samaritan's dilemma and the effectiveness of development aid. International Tax and Public Finance 8 (5 6), 693 703. Peters, G. B, & Pierre, J., (1998). Journal of Public Administration Research and Theory. Volume 8: 2, pp. 223 243. Pilgrim, N., Roche, B., Revels, C., & Kalbermatten, J. (2004). Town Water Supply and Sanitation. Pius Kulipossa, F. (2004). Decentralisation and democracy in developing countries: an overview. Development in Practice 14 (6), 768 779. Plummer, J., & Slaymaker, T. (2007). Rethinking governance in water services Overseas Development Institute.

PAGE 282

269 Priscoli, J. D. (2004). What is public participation in water resources management and why is it important?. Water International 29 (2), 221 227. Ramirez, J.M., Bedoya, J.G., Diaz, Y. (2014). Convergencia Social en Colombia : El Rol de la Geografia Ecoonomica Y de la Descentralizacion. Fedesarrollo. Bogata, Colombia. Ramrez, J.M., Daz, Y., & Bedoya, J. G. (2014). Decentralization in Colombia: Searching for social equity in a bumpy economic geography. Working paper No.62, Fedesarrollo, centro de investigacion Economica y Social Ramirez, J.M. (2015). Interview Summary. Interview conducted with author by Skype, April 9, 2015. Rauchhaus, R. W. (2009). Principal agent problems in humanitarian i ntervention: Moral hazards, adverse selection, and the commitment dilemma. International Studies Quarterly 53 (4), 871 884. Rawls, J. (1999). A theory of justice Belknap Press. Research Triangle Institute (2004). Public Participation and Local Government: an Analysis of four U.S. models, Resesarch Triangle Institute Richards, T., E. Doering, et al. (2008). Water supply and sanitation sector reforms in Kenya, Tanzania, Uganda, and Zambia. GTZ. Riddell, R. C. (2007). Does foreign aid really work? OUP Oxford. Riley, S. (2004). The political economy of anti corruption strategies in Af rica in Robinson, M. 2004, Corruption and Development: An Introduction Routledge, London. Rogers, P., & Hall, A. W. (2003). Effective water governance (Vol. 7). Stockholm: Global Water Partnership. Rojas, F., & Frank, J. (2002). Assessment of the Impa ct of Decentralization: The Case of Colombia (1991 2001). Public Sector Group/Decentralization Thematic Group, Funding 2003 Rojas, J.H. (2015). Interview Summary. Interview conducted with author by Skype, March 31, 2015. Rondinelli, D. A. (1990). Financing the decentralization of urban services in developing countries: Administrative requirements for fiscal improvements. Studies in Comparative International Development 25 (2), 43 59.

PAGE 283

270 Rondinelli, D. A., McCullough, J. S., & Johnson, R. W. (1989). Analysing decentralization policies in developing countries: a political economy framework. Development and change 20 (1), 57 87. Sabie, O. (2012). Corruption in the Water Sector. Revista" Administratie si Management Public"(RAMP) (19), 85 96. Saleth, R. M., & Dinar, A. (2004). The institutional economics of water: a cross country analysis of institutions and performance Edward Elgar Publishing. Saleth, R. M., & Dinar, A. (2005). Water institutional reforms: theory and practice. Water Policy 7 (2005), 1 19. Saleth, R. M., Kadushkin, A., & Ali, T. (2005). Water sector reforms for catalyzing the MDGs: To wards an evaluation framework. Saleth, R. M., Dinar, A., Ne ubert, S., Kamaiah, B., Manoh aran, S., Abayawardana, S., & Silva, S. D. (2007). Institutions, impact synergies and food security: a methodology with results from the Kala Oya Basin, Sri Lanka. Research Report International Water Management Institute (124) Samad, T., Lozano Gracia, N., & Panman, A. (Eds.). (2012). Colombia Urbanization Review: Amplifying the Gains from the Urban Transition World Bank Publications. Snchez Torres, F., & Pachn, M. (2013). Decentralization, Fiscal Effort and Social Progre ss in Colombia at the Municipal Level, 1994 2009: Why Does National Politics Matter? Inter American Development Bank. Sanitation and Water for All. ( undated ) Mali: Commitments from the 2014 High Level Meeting. http://sanitationandwaterforall.org/report_card/mali Retrieved March 22, 2015. Sanitation and Water for All website, http://sanitationandwaterforall.org/about#sthash.JUkX34Vb.dpuf Schneider, A., & Ingram, H. (1993). Social construction of target populations: Implications for politics and policy. American Political Science Review 334 347. Schneider, A. L., & Ingram, H. M. (1997). Policy design for democracy Lawrence, KS: Universit y Press of Kansas. Schreiner, B., Hassan, R. (2011). Transforming water management in South Africa: designing and implementing a new policy framework. Springer publishing. Sebetao, O. (2015). Response to author's questionnaire, received July 21, 2015.

PAGE 284

271 Seppl, O. T. (2002). Effective water and sanitation policy reform implementation: need for systemic approach and stakeholder participation. Water Policy 4 (4), 367 388. Shivaii, S. (undated). An overview of the Kenya water and sanitation CSOs network. PowerPoint presentation, KEWASNET. Shleifer, A.; Vishny, R.W. (1993), Corruption Quarterly Journal of Economics, Vol. 108 No. 3, pp. 599 617. Shoemaker, P. J., J. E. Tankard, et al. (2004). How To Build Social Science Theories Thousand Oaks, CA, Sage Publications, Inc. transparency in the water and sanitation sector. International Water and Sanitation Centre (IRC), Thematic Overview Paper 16 Sidibe, N. (undated). Recherche Transnationale. Dcentralisation en Afrique de L'ouest et du Centre: Apprendre des Expriences Locales et Sectorielles : Cas du Mali, Rapport Final. Education Research Network for West and Central Africa. Bamako, Mali. Silva, J., & Andia, T. (2006). Agua para la vida. La Infancia, el agua y el saneamiento ba General de la Nacio n, Ministerio Ambiente, Vivienda y Desarrollo Territorial, Superintendencia d e Servicios Pu blicos Domiciliarios. Departamento Nacional de Planeacion, Fondo de las Naciones Unidas para la Infancia, Bogota. Slaymaker, T., & Newborne, P. (2004). Implementation of Water Supply & Sanitation Programmes under PRSPs: Synthesis of Re search Findings from sub Saharan Africa. Overseas Development Institute and WaterAid. London, UK. Sohail, M., Cavill, S., & Cotton, A. P. (2005). Sustainable operation and maintenance of urban infrastructure: Myth or reality?. Journal of urban planning and development 131 (1), 39 49. Stata Annotated Output: Ordered Logistic Regression. UCLA: Institute for Digital Research and Education from http://www.ats.ucla.edu/stat/sta ta/output/stata_ologit_output.htm (accessed May 11, 2015). Stata Data Analysis Examples: Ordered Logistic Regression. UCLA: Institute for Digital Research and Education, from http://www.ats.ucla.edu/stat/stata/dae/ologit.htm. (accessed August 11, 2014).

PAGE 285

272 partnerships for water and sanitation in Africa: When is a stakeholder a partner?. Environmental Politics 15 (03), 362 378. Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics. Tevelow, A. (2004). Global Governance and Social Capital: Mapping NGO Capacities in Different Institutional Contexts. Communication and Public Participation in Environmental Decision Making S. P. Depoe, J. W. Delicath and M. F. A Eslenbeer. Albany, NY, State Univeristy of New York Press. The Cities Alliance & UCLG Africa. (2013). Assessing the Institutional Environment of Local Governments in Africa. The Hunger Project. ( undated .) Participatory Local Democracy Fact Sheet: Mali. http://localdemocracy.net/countries/africa/mali/ retrieved March 13, 2015. Thomas, J. C. (1995). Public Participation in Public Decisions Sa n Francisco Jossey Bass Publishers. Thompson, F. (2008). The political economy of public administration. Handbook of Public Administration 1063 1100. Tiebout, C. M. (1956). A pure theory of local expenditures. The Journal of Political Economy 64 (5), 41 6 424. Tissington, K. (2011). Basic Sanitation in South Africa: A Guide to Legislation, Policy and Practice. Socio Economic Rights Institute of South Africa, Johannesburg, South Africa. Torres Reyna, O. (2014). Predicted probabilities and marginal effe cts after (ordered) logit probit using margins in Stata (v2.0). Princeton University, from http://dss.princeton.edu/training/ Transparency International (Ed.). (2008). Global corruption report 2008: corruption in the water sector Cambridge University Press. Transparency International (2013). Corruption Perception Index [Data file]. Retrieved July 13, 2014 from http://cpi.transparency.org/cpi2013/results/ Treisman, D. (2000). Journal of Public Econom ics 76, 399 457. Tropp, H. (2007). Water governance: trends and needs for new capacity development. Water Policy 9 19.

PAGE 286

273 Tukahirwa, J. T., Mol, A. P. J., & Oosterveer, P. (2013). Comparing urban sanitation and solid waste management in East African metr opolises: The role of civil society organizations. Cities 30 204 211. UN (2012). "The Millenium Development Goals Report: 2012." UN (2014). "The Millennium Development Goals Report: 2014." UN Committee on Economic, Social and Cultural Rights (2002). "General Comment No. 15. The Right to Water." UN General Assembly (2010). "Resolution A/RES/64/292. UN Millennium Development Goals list. R etrieved October 19, 2012 from http://www.un.org/mill enniumgoals/environ.shtml UN Office of the High Commissioner for Human Rights (2010). "A landmark decision to make the right to water and sanitation legally binding." Retrieved Sept. 20, 2013 from http://www.ohchr.org/EN/NewsEvents/Pages/RightToWaterAndSanitation.aspx UNCHS (2003). Local Action for Global Goals: Water and San Cities 2003. London: Earthscan. UNCHS (2006). Meeting Development Goals in Small Urban Centres: Water and UNDP. (1997). Governance for Sustainable Human Development: UNDP Po licy D ocument. New York. UNDP. (1980 2013). Table 2: Human Development Index Trends. http://hdr.undp.org/en/content/table 2 human development index trends 1980 2013 ( accessed April 7, 2015 ) UNDP (2006). Beyond Scarcity: Power, Poverty, and the Global Water Crisis. Human Development Report 2006. United Nations Development Program, New York. UNDP. (2009). Country Sector Assessments: UNDP GoAL WaSH Programme, Vol ume 1, Mali. New York. UNDP. (2015). Governance for Sustainable Development: Integrating Governance in the Post 2015 Development Framework. Discussion Paper. New York. U NICEF ( undated ). "Terms of Reference: WHO/UNICEF Joint Monitoring Program (JMP) for Water Supply and Sanitation."

PAGE 287

274 UNWWAP (2003). Water for people, water for life. United Nations World Water Assessment Program. Rome: FAO. UNWWAP (2006). Water: A Share d Responsibility. United Nations World Water Assessment Program. US Agency for International Development. (1998). Democratic Decentralization in Mali: A Work in Progress. Washington, D.C. US Agency for International Development. (2010a ). Mali: Water and Sanitation Profile. Washington, D.C. US Agency for International Development. ( 2010b ). Comparative Assessment of Decentralization in Africa: Mali In Country Assessment Report. Washington, D.C. US Agency for International Development (2011). The 2011 CSO Sustainability Index for Sub Saharan Africa. Washington, D.C. US Agency for International Development. (2012). The 2012 CSO Sustainability Index for Sub Saharan Africa. Uyterwaal, E., Dietvorst, C. (2007). Inventory of available information on SWAps in the water sector. IRC International Water and Sanitation Centre. Vittinghoff, E., & McCulloch, C. E. (2007). Relaxing the Rule of Ten Events per Variable in Logistic and Cox Regression. American journal of epidemiology 165 (6), 710 718. Water and Sanitation Program. (2002). The National Water and Sanitation Programme in Water and Sanitation Program. (2003). Water Supply and Sanitation in Poverty Reduction Strategy Papers: Developing a Benchmarking Review and Exploring the Way Forward. Nairobi, Kenya. Water and Sanitation Program. (2006a). Getting Africa on Track to Meet the MDGs on Water and Sanitat ion: A Status Review of Sixteen African Countries. Nairobi, Kenya. Water and Sanitation Program. (2006b). Water Supply and Sanitation in Poverty Reduction Strategy Papers (PRSP): Benchmarking Performance, Field Note. Nairobi, Kenya.

PAGE 288

275 Water and Sanita tion Program. (2011). Promoting voice and accountability in urban water and sanitation services: Lessons on introducing demand side accountability tools in Kenya. Field Note 63347. (2011a). Water Supply and Sanitation in Kenya: Turning Finance into Services for 2015 and Beyond. (2011b). Water Supply and Sanitation in South Africa: Turning Finance into Serv ices for 2015 and Beyond. Water and Sanitation Program. (2014). Mali Rural Water Supply Sector fact sheet. Water Integrity Network. (2012). Water Action Groups (Kenya). Case Information Sheet. Water Partnership Program (of the African Development Bank). (2010). Water Sector Governance in Africa: Volume 2, Assessment Guidelines. Tunis, Tunisia. Water Services Regulatory Board. (2010). Ensuring consumer participation in water service delivery through Water Action Groups. Nairobi, Kenya. Water Services Sector 2012 2013. Issue No. 7. Nairobi, Kenya WaterAid. (2008). Financing Water and Sanitation at Local Levels : Synthesis Report. London, England. WaterAid & Freshwater Action Network. ( 2011 ). Learning from Experience : Rights and G overnance A dvocacy in the Water and Sanitation S ector London, England. Watkins, K. (2006). Human Development Report 2006 Beyond scarcity: Power, poverty and the global water crisis. UNDP Human Development Reports (2006) Weimer, D. L. and A. R. Vining (2005). Policy Analysis: Concepts and Practice Upper Saddle River, New Jersey Pearson Prentice Hall WHO (1992). The International Drinking Water Supply and Sanitation Decade: End of Decade Review. Geneva: World Health Organization WHO (2000). The world health report 2000: health systems: improving performance WHO (2011). Country Sanitation, Drinking Water Sectors, and Hygiene Promotion Questionnaire of the Global Analysis and Assessment of Sanitation and Drinking Water.

PAGE 289

276 WHO (2013). "Sector wide Approaches." Retrieved Sept. 20, 2013 from http://www.who.int/trade/glossary/story081/en/ WHO (2014 a ). Investing in Water and Sanitation: Increasing Access, Reducing Inequalities. WHO (2014b). Preventing diarrhoea through better water, sanitation and hygiene: exposures and impacts in low and middle income countries WHO (2015). "WHO strengthe ns focus on water, sanitation and hygiene to accelerate elimination of neglected tropical diseases." Retrieved September 11, 2015 from http://www.who.int/water_sanitation_health/events/wash and ntd strategy/en/ WHO and UN Water (2012). "UN Water Global Analysis and Assessment of Sanitation and Drinking Water: The Challenge of Extending and Sustaining Services." WHO and UNICEF (2000 2015 ). Joint Monitoring Programme (JMP) for Water Supply and Sanitation [Data file query]. Available at http://www.wssinfo.org/data estimates/tables/. WHO WHO and UNICEF (2011). "Report: Third Meeting, Strategic Advis ory Group, 13 14 December 2011." WHO and UNICEF (2012). "Progress on Drinking Water and Sanitation: 2012 Update." Wijaya, A. F. (2005). Water as a public good in Indonesia Flinders University, Faculty of Social Sciences.. Wilder, M., & Romero Lankao, P. (2006). Paradoxes of decentralization: Water reform and social implications in Mexico. World Development 34 (11), 1977 1995. Wolman, H. (1990). Decentralization: what it is and why we should care. BENETT RJ, op. cit 29 41. World Alliance for Citizen Participation Enabling Environment Index [Data file]. Retrieved February 16, 2014 from http://civicus.org/eei/. World Alliance for Citizen Participation (2013b). Background Note: CIVICUS Enabling Environment Index. Retrieved February 27, 2014 from http://civicus.org/eei/downloads/CIVICUS_EEI_BACKGROUND_NOTE_23_Se pt_2013.pdf )

PAGE 290

277 World Bank. (undated/a). Citizen Feedback Drives Performance Improvements in World Bank. (undated/b). Expanding Access to Water and Sanitation: Colombia. Partnerships in Development program brief. Wor ld Bank. (undated/c). Implementation Status & Results: Mali Governance and Budget Decentralization Technical Assistance Project (P112821). Washington, D.C. World Bank. (1995). Public Policy f or the Private Sector, Note No. 59. Concessions: the Way to Privatize Infrastructure Sector Monopolies Washington, D.C. World Bank (2000 2012a). World Development Indicators, GINI [Data file] Retrieved January 8, 2015 from http://data.worldbank.org/indicator/SI.POV.GINI. World Bank (2000 2012 b ). World De velopment Indicators, Urban Population [Data file] Retrieved April 8, 2014 from http://data.worldbank.org/indicator/SP.URB.TOTL. World Bank. (2003). World Development Report 2004: Making Services Work for Poor People World Bank. World Bank. (2005). Colombia : Public Expenditure Review Washington, DC. World Bank. https://openknowledge.worldbank.org/handle/10986/8559 License: CC BY World Bank. (2006). Local solutions improve water supply and sanitation services in Colombia Water feat ure stories ; issue no. 8. Washington, DC: World Bank. World Bank. (2008). Mali Public Expenditures Review: Rural Water and Sanitation Sector Washington, DC. World Bank. (2011). Kenya Nairobi Metropolitan Services Improvement Project Washington, DC: World Bank. World Bank. (2012a). Colombia Water and Sanitation Sector Support Project : Implementation Completion and Results Report Washington, DC: World Bank. t World Bank ( 2012b ). World Development Indicators Country Population Total [Data file]. Retrieved March 19, 2014 from http://data.worldbank.org/indicator/SP.POP.TOTL. World Bank ( 2012c ). World Development Indicators GDP per Capita [Data file]. Retrieved March 2, 2014 from http://data.worldbank.org/indicator/NY.GDP.PCAP.CD.

PAGE 291

278 World Bank. (2013). Six Cases of Local Participation in Kenya: Executive Summary. World Bank website. (2015). Country and Lending Group Data. http://data.worldbank.org/about/country and lending groups (accessed June 10, 2015). World Bank Private Participation in Infrastructure Database. (2015). http://ppi.worldbank.org (accessed September 14, 2015). Wunsch, J. (1999). 'Institutional Analysis and Decentralization: Developing and Analytical Framework for Effective Third World Administration Reform,' in McGinnis, M. (ed.), Polycentric Governance and Development University of Michigan press: Ann Arbor, MI. Wunsch, J. S., & Olowu, D. (Eds.). (1990). The failure of the centralized state: Institutions and self governance in Africa Westview Press.

PAGE 292

279 APPENDIX A THEORETICAL CONCEPTS AND CONSTRUCTS FOR RESEARCH QUESTIONS Research Question, Hypothesis, Definition of Key Term Construct Level Definition (Source Literature) Probes for Construct Measurement (Source Material Code ) 1 RQ1: Decentralized Services RQ1: What is the role of decentralization of water and sanitation services in a extending water and sanitation access in urban areas, as per the objective of MDG 7.C? H1: A high level of decentralization in water and sanitation services is associated with greater country progress on providing urban water and sanitation access Decentralization: The transfer of planning, decision making, or m anagement functions from th e central national government to organizations at the regional or local levels (Handoussa, 2009, p.217) a) Decision making authority for service provision is delegated to sub national levels of government ( Handoussa, 2009; Jutting, 2005; Mugabi and Njiru 2 006; Davison, 2001) To what degree has decentralisation of service delivery been carried out? (G) b) Authority to manage expenditures and raise revenues is delegated to sub national levels of government ( Bahl & Linn, 1992; Bardhan & Mookherjee, 2000; Mugabi and Njiru 2006; Davison, 2001 ; Gillespie, 2005; Rodinelli, 1990) Is funding available at local level from the national level? (i.e. is funding in line with decentralization policies?) (G) Do utilities have operational decision making autonomy in i nvestment planning, human resources, finance and procurement management? (G water only) c) There are clearly defined rights, responsibilities, and roles for sector management actors at the national and sub national levels ( Marcus and Onjala 2008; Mugab i and Njiru 2006) Are the institutional roles of national & local governments, utilities, water boards, regulator etc. clearly defined and operational? (D/I) d) Sub national level authorities are monitored by communities and/or the central government including, for example, public funds expenditures and equality in service provision ( Keefer & Khemani, 2005; Bardhan &Mookherjee, 2005; Mookherjee & Bardhan, 2005; Azfar, Kahkonen et al., 2004; Bardhan & Mookherjee, 2000; Mexico, Wilder and Romero 2006) A re there mechanisms in place for monitoring the performance of sub national level authorities, for example public funds expenditures and equality in service provision? (D/I)

PAGE 293

280 e) Support is provided to build managerial and technical capacity of sub nationa l levels of government ( Marcus and Onjala 2008; Funke et al., 2007; Mugabi and Njiru 2006; Seppala, 2002; Pilgrim et al., 2004) Are there programs in place to build the managerial and technical capacity of sub national levels of government who are given a uthority for service provision? (D/I) RQ2: Sector w ide Strategy and Investment Coordination RQ2. What is the role of sector wide strategy and investment coordination by a recipient country in its progress on extending water and sanitation access in urban areas, as per the objective of MDG 7.C? H2: A high level of sector wide strategy and investment coordination in the water and sanitation sectors is associated wit h greater country progress on providing urban water and sanitation access Sector Wide Strategy and Investment Coordination : Mechanisms for transparent, country led partnering with donors as part of coordinating, directing, and utilizing international aid towards priority areas for the sector (adapted from Tropp, 2007) a) The recipient country identifies priority projects and develops a realistic investment plan for the sector (Gibson, Andersson et al., 2005; Riddell 2007; V. Ostrom 1999a; Auer, 2007) Do es the government have a sector wide approach (SWAp) or another similar sectoral framework that involves all development partners? (G) Is there an investment program that is agreed and published? (G) b) Development of partnering agreements between donors and recipient countries is driven by the recipient country (Gibson, Andersson et al., 2005; V. Ostrom 1999a; Auer, 2007) Is the process used to develop donor partner agreements led primarily by the recipient country? (D/I) c) A single agency is c harged with coordinating all donor aid received in conjunction with a sector strategy (Auer, 2007; Riddell 2007) Is there a government agency with a clear mandate to lead and coordinate the policy development and planning of the sector with donors, other g overnmental institutions or non state actors? (D/I ) d) Sector wide strategies are prepared in collaboration with stakeholders and donors, who are provided wi th information about the sector (Tropp 2007; WHO, 2000; Abrahamsen, 2004) Is there a national i nformation system that covers the sector that is used to inform decisions/strategy and resource allocation? (G) Is there an annual or biennial review in place to monitor performance of systems and services, and to set new ta rgets and undertakings? (G) e) A system of monitoring is in place for compliance with the terms of aid partnering agreements (Calmette & Kelkenny, 2002; Buchanan 1977) Are there mechanisms in use to monitor compliance of donors and recipient countries with the agreed terms of aid par tnering agreements? (D/I)

PAGE 294

281 RQ3: Civil Society Engagement RQ3. What is the role of civil society engagement in a extending water and sanitation access in urban areas, as per the objective of MDG 7.C? H3: A high level of civil society engagement in water and sanitation services is associated with greater country progress on providing urban water and sanitation access Civil Society Engagement: The existence of opportunities for the public and consumers to become involved in decision making, including program planning an d implementation (adapted from deLeon and deL eon, 2002) a) Opportunities exist for consumers and civil society groups to become involved in problem articulation (Fischer, 2003; Oakley, 1991; E. Ostrom, 2005; Oakerson & Parks, 1999) b) Opportunities exist for consumers and civil society groups to become involved in decision making about policy development and planning (Gillespie, 2005: Lynn, 1994; deLeon & deL eon, 2002; McGranahan & Satterthwaite, 2006; E. Ostrom, 2005; Oakerson & Parks, 1999) c) Opportunities exist for consumers and civil society groups to become involved in decision making about implementation and management (deLeon & deL eon, 2002; Mugabi & Njiru, 2006; McGranahan & Satterthwaite, 2006; E. Ostrom, 2005; Oakerson & Parks, 1999) Are there clearly defined procedures in laws, policies or plans for informing, consulting with and supporting participation of citizens and communities in planning, budgeting and implementing for the sector at the national and local level? (G) How are consumers and civil society groups engaged at different stages of policy development and planning, program implementation and management? (D/I) d) Sector operates with transparency and information disclosure, including for use of funds and natural resources (Gillespie, 2005; Mugabi & Njiru, 2006; E. Ostrom, 2005; Hooper, 2011; Madrigal, Alpizar et al., 2011) Does the government publish information on its operations, including expenditures, and use of natural resources? (D/I) e) There is a feedback/complaint mechanism system in place ( Li, 2011; Meng et al., 2011; Araral, 2008) Are there accessible, affordable, timely and effective complaint mech anisms in place for people who have unsatisfactory access to services? (G) 1 (G) = GLAAS Survey Question; (D/I) = Documents/Interviews Source for GLAAS Questions: WHO (2011)

PAGE 295

282 APPENDIX B LIST OF SAMPLE COUNTRIES WITH DATA NOTATIONS Codes: 0= no GLAAS data available; 1= partial GLAAS data available; X= not used due to outlier pairings or incompatible data Water Variables Sanitation Variables Control Variables Country Decentr alization Sector Planning Civil Society Access in 2012 12 Year Progress Decentr alization Sector Planning Civil Society Access in 2012 12 Year Progress Afghanistan X X Angola Azerbaijan X X Bangladesh Benin Bhutan Bolivia (Plurinational State) Brazil 1 1 X X 1 Burkina Faso 1 Burundi 1 X X Cambodia Cameroon Central African Republic 1 X X Chad 1 Colombia 1 Congo 1 Cte d'Ivoire Democratic Republic of the Congo Dominican Republic 1 Egypt El Salvador 1 1 Equatorial Guinea 0 0 0 0 Ethiopia Fiji 1 See Note 1

PAGE 296

283 Gabon 1 Gambia Ghana Guinea 1 Guinea Bissau Haiti 0 0 1 0 0 Honduras India 1 0 0 0 0 0 Indonesia Iran (Islamic Republic of) X X Jordan Kenya Kyrgyzstan 1 Lao People's Democratic Republic 1 Lebanon X X Lesotho Liberia Madagascar Malawi Maldives See Note 1 Mali X X Mauritania Mongolia Morocco Mozambique Myanmar 1 Nepal Niger Nigeria Oman Pakistan 1 Panama 1 1 Paraguay Philippines Rwanda

PAGE 297

284 Samoa 1 1 See Note 2 Senegal X X Sierra Leone South Africa South Sudan 0 0 Sri Lanka Sudan Tajikistan 1 Thailand 1 Timor Leste 1 See Note 3 Togo Uganda Uzbekistan Viet Nam 1 Yemen 1 Zimbabwe Note 1: Average value used for corruption. Note 2: Average values used for corruption, political fragility, and security fragility. Note 3: Average values used for political fra gility and security fragility.

PAGE 298

285 APPENDIX C VARIABLE OPERATIONALIZATION AND SOURCE INFORMATION Outcome Variable Operationalization Source Water Access in 2012 ) Percentage of the urban population lacking access to an improved source in 2012 was transposed to derive the percentage of the urban population with access to an improved source in 2012. WHO/UNICEF Joint Monitoring Program for Water Supply and Sanitation Water 12 Year Progress ( Water Progress ) Percentage of the population with access to an improved source in the year 2000 was subtracted from the percentage of the population with access to an improved source in the year 2012 to measure the 12 year change. The result was then divided by the perce ntage of the population that lacked access in the year 2000. WHO/UNICEF Joint Monitoring Program for Water Supply and Sanitation Sanitation Access in 2012 ) Percentage of the urban population lacking access to an improved s ource in 2012 was transposed to derive the percentage of the urban population with access to an improved source in 2012. WHO/UNICEF Joint Monitoring Program for Water Supply and Sanitation Sanitation 12 Year Progress ( Sanitation Progress ) Percentage of t he population with access to an improved source in the year 2000 was subtracted from the percentage of the population with access to an improved source in the year 2012 to measure the 12 year change. The result was then divided by the percentage of the WHO/UNICEF Joint Monitoring Program for Water Supply and Sanitation

PAGE 299

286 po pulation that lacked access in the year 2000. Key Explanatory Variable Operationalization Source Water Decentralization ( W decentralization ) Average of responses to survey questions 4i, 8b, and 9j 2011 GLAAS survey Water Sector Planning ( W sector planning ) Average of responses to survey questions 5a, 5b, 5c, and 5f 2011 GLAAS survey Water Civil Society Engagement ( W civil society ) Average of responses to survey questions 5d and 7a 2011 GLAAS survey Sanitation Decentralization ( S decentralization ) Average of responses to survey questions 4n and 8b 2011 GLAAS survey Sanitation Sector Planning ( S sector planning ) Average of responses to survey q uestions 5a, 5b, 5c, and 5d 2011 GLAAS surv ey Sanitation Civil Society Engagement ( S civil society ) Average of responses to survey questions 5e and 7a 2011 GLAAS survey Control Variable Operationalization Source Urban Population Change Difference in the urban populations (as a percent of total country population) in the year 2012 and 2000, divided by the 2000 figure. World Bank Development Indicator Corruption Level Corruption Perceptions Index (ranging from 0 for very corrupt to 100 for very clean) was transposed so that a higher score represents a more corrupt country. Transparency International (2013 data) Political Fragility Political Effectiveness index (ranging from 0 for effective to 3 for ineffective) was retitled to better correspond with the measurement scale. Center for System atic Peace (2012 data) Security Fragility Security Effectiveness index (ranging from 0 for effective to 3 for ineffective) was retitled to better correspond with the measurement scale. Center for Systematic Peace (2012 data)

PAGE 300

287 Literacy Rate Percent of the population age 15 and over who can read and write. CIA World Factbook (2011) GDP per capita G ross domestic product divided by mid year population (in current US dollars) World Bank Development Indicators (2012) Renewable Water Resources Cubic meters per year/ per capita UN Food and Agriculture Organization AQUASTAT global water information system (2011) Foreign Aid to Water and Sanitation Cumulative total development assistance for the water and sanitation sectors from 2002 to 2012 divided by an average of the 2002 and 2012 total country population figures Organization for Economic Cooperation and Development (development assistance) and World Bank Development Indicators (country population)

PAGE 301

288 APPENDIX D CORRELATION TABLES Outcome Variable Correlatio ns N=72 1 2 3 4 1 Water Access in 2012 1.00 2 Water 12 Year Progress 0.38 1.00 3 Sanitation Access in 2012 0.35 0.09 1.00 4 Sanitation 12 Year Progress 0.18 0.33 0.53 1.00 Key Explanatory Variable Correlations N=73 1 2 3 4 5 6 1 Water Decentralization 1.00 2 Water Sector Planning 0.25 1.00 3 Water Civil Society Engagement 0.32 0.51 1.00 4 Sanitation Decentralization 0.53 0.29 0.34 1.00 5 Sanitation Sector Planning 0.21 0.61 0.56 0.25 1.00 6 Sanitation Civil Society Engagement 0.38 0.35 0.64 0.30 0.55 1.00 Control Variable Correlations N=75 1 2 3 4 5 6 7 8 1 Urban Population Change 1.00 2 Corruption Level 0.12 1.00 3 Political Fragility 0.05 0.39 1.00 4 Security Fragility 0.03 0.23 0.05 1.00 5 Literacy Rate 0.26 0.15 0.32 0.17 1.00 6 GDP per capita 0.28 0.23 0.38 0.14 0.62 1.00 7 Renewable Water Resources 0.00 0.02 0.02 0.11 0.01 0.02 1.00 8 Foreign Aid to Water and Sanitation 0.14 0.17 0.07 0.33 0.07 0.11 0.19 1.00

PAGE 302

289 Correlations among the Key Water Variables and Control Variables Correlations among the Key Sanitation Variables and Control Variables N=69 1 2 3 1 Water Decentralization 1.00 2 Water Sector Planning 0.30 1.00 3 Water Civil Society Engagement 0.44 0.50 1.00 4 Urban Population Change 0.03 0.11 0.01 5 Corruption Level 0.15 0.33 0.36 6 Political Fragility 0.25 0.03 0.22 7 Security Fragility 0.12 0.15 0.17 8 Literacy Rate 0.17 0.11 0.23 9 GDP per capita 0.20 0.04 0.32 10 Renewable Water Resources 0.07 0.20 0.21 11 Foreign Aid Water & Sanitation 0.23 0.02 0.09 N=69 1 2 3 1 Sanitation Decentralization 1.00 2 Sanitation Sector Planning 0.22 1.00 3 Sanitation Civil Society Engagement 0.23 0.53 1.00 4 Urban Population Change 0.10 0.09 0.10 5 Corruption Level 0.19 0.19 0.11 6 Political Fragility 0.17 0.21 0.14 7 Security Fragility 0.22 0.05 0.04 8 Literacy Rate 0.10 0.05 0.35 9 GDP per capita 0.09 0.16 0.37 10 Renewable Water Resources 0.09 0.32 0.29 11 Foreign Aid to Water and Sanitation 0.23 0.13 0.02

PAGE 303

290 APPENDIX E FREQUENCY TABLES FOR KEY EXPLANATORY VARIABLES AND OUTCOME GROUPS Decentralization Water Decentralization Composite Value Water A 2012 Outcome Group Total by Score Low Medium High .17 0 5 0 5 .25 0 0 2 2 .33 1 2 3 6 .5 2 7 9 18 .67 4 7 7 18 .75 0 1 3 4 .83 7 3 5 15 1 1 0 4 5 Total by Group 15 25 33 73 Water Decentralization Composite Value Water Progress Outcome Group Total by Score Low Medium High .17 1 3 1 5 .25 0 0 2 2 .33 2 1 2 5 .5 1 12 4 17 .67 5 7 6 18 .75 1 2 1 4 .83 7 5 3 15 1 2 1 2 5 Total by Group 19 31 21 71 Sanitation Decentralization Composite Value Sanitation A 2012 Outcome Group Total by Score Low Medium High 0 3 4 2 9 .25 5 6 4 15 .5 7 7 5 19 .75 1 5 10 16 1 3 4 6 13 Total by Group 19 26 27 72

PAGE 304

291 Sanitation Decentralization Composite Value Sanitation Progress Outcome Group Total by Score Low Medium High 0 2 6 1 9 .25 7 5 2 14 .5 6 8 4 18 .75 5 3 8 16 1 3 5 5 13 Total by Group 23 27 20 70 Sector Planning Water Sector Planning Composite Value Water A 2012 Outcome Group Total by Score Low Medium High .13 0 1 1 2 .25 1 1 3 5 .38 1 4 2 7 .5 4 2 1 7 .63 2 1 2 5 .67 0 0 1 1 .75 0 7 5 12 .88 3 5 5 13 1 4 4 12 20 Total by Group 15 25 32 72 Water Sector Planning Composite Value Water Progress Outcome Group Total by Score Low Medium High .13 0 1 1 2 .25 3 0 2 5 .38 0 5 1 6 .5 3 4 0 7 .63 3 1 1 5 .67 1 0 0 1 .75 1 7 3 11 .88 4 5 4 13 1 4 8 8 20 Total by Group 19 31 20 70

PAGE 305

292 Sanitation Sector Planning Composite Value Sanitation A 2012 Outcome Group Total by Score Low Medium High 0 1 0 0 1 .13 1 3 1 5 .17 0 1 1 2 .25 2 1 2 5 .38 0 5 1 6 .5 2 4 2 8 .63 3 6 4 13 .75 4 3 4 11 .88 3 0 4 7 1 3 3 8 14 Total by Group 19 26 27 72 Sanitation Sector Planning Composite Value Sanitation Progress Outcome Group Total by Score Low Medium High 0 1 0 0 1 .13 2 2 1 5 .17 2 0 0 2 .25 2 1 2 5 .38 2 1 3 6 .5 2 4 2 8 .63 2 5 5 12 .75 4 3 3 10 .88 2 5 0 7 1 4 6 4 14 Total by Group 23 27 20 70 Civil Society Engagement Water Civil Society Composite Value Water A 2012 Outcome Group Total by Score Low Medium High 0 1 2 1 4 .25 4 7 1 12 .5 7 10 11 28 .75 0 2 12 14 1 4 4 7 15 Total by Group 16 25 32 73

PAGE 306

293 Water Civil Society Composite Value Water Progress Outcome Group Total by Score Low Medium High 0 1 2 1 4 .25 2 6 3 11 .5 11 13 3 27 .75 1 3 10 14 1 5 7 3 15 Total by Group 20 31 20 71 Sanitation Civil Society Composite Value Sanitation A 2012 Outcome Group Total by Score Low Medium High 0 5 6 0 11 .25 5 7 2 14 .5 6 9 11 26 .75 3 3 12 18 1 0 1 2 3 Total by Group 19 26 27 72 Sanitation Civil Society Composite Value Sanitation Progress Outcome Group Total by Score Low Medium High 0 4 5 1 10 .25 7 2 5 14 .5 7 11 7 25 .75 4 8 6 18 1 1 1 1 3 Total by Group 23 27 20 70

PAGE 307

294 APPENDIX F GRAPHS RELATING SECTOR PLANNING TO COUNTRY PROGRESS Series 1: Outcome variables related to PRSP water and sanitation inclusion (N=10) PRSPscor = Extent to which water and sanitation are included in country full PRSP, derived from Water and Sanitation Program, 2003, p.10 Wat2012G = Water A 2012

PAGE 308

295 WatProgR = Water Progress

PAGE 309

296 Sanit2012G = Sanitation A 2012

PAGE 310

297 SanProgR = Sanitation Progress

PAGE 311

298 Series 2: Outcome variables related to water sector strategies (N = 13) SectPlan = Extent to which country has a water sector strategy, derived from Water Partnership Program, 2010, p.11 12 Note: The number of points plotted may appear as less than 13 because of overlapping dots for countries with the same x and y variable pairings. Wat2012G = Water A 2012

PAGE 312

299 WatProgR = Water Progress

PAGE 313

300 Sanit2012G = Sanitation A 2012

PAGE 314

301 SanProgR = Sanitation Progress

PAGE 315

302 APPENDIX G PERSONS CONSULTED FOR CASE STUDIES Name Professional Background Input 1 Johnny Rojas Economist and Researcher CINARAS, research and development institute for the water sector University of Valle Cali, Colombia I, D Juan Mauricio Ramirez Economist and Researcher focused on local and regional development and decentralization Fedesarrollo, social and economic policy research institute Bogota, Colombia Formerly, Deputy Director of the National Planning Department I, D Julian Lopez Murcia PhD Student at Oxford University Studies on the government recentralization processes in Colombia, particularly for water supply and sanitation Oxford, United Kingdom I, D Prof. Chibane Coulibaly Professor and author of articles on decentralization University Mande Bukari Bamako, Mali D Yaya Boubacar Deputy Director National Directorate of Water Bamako, Mali D Oumar Sebetao Water Specialist National Directorate of Water Koulikoro, Mali Q Mahamadou Sidibe Former Secretary General of the Ministry in charge of Drinking Water Supply and Sanitation Bamako, Mali D Romanus Opiyo, PhD Lecturer and author of articles on water governance Department of Urban and Regional Planning University of Nairobi Nairobi, Kenya Q Brenda Anzagi ICT Officer Water Services Regulatory Board Nairobi, Kenya Q, D Helen Moremong Deputy Director Water Sector Support Q, D

PAGE 316

303 Department of Water and Sanitation Pretoria, South Africa Prof. Patrick Bond, PhD Director Centre for Civil Society University of KwaZulu Natal Durban, South Africa D Jessica Wilson Programme Manager, Water and Climate Change Environmental Monitoring Group Cape Town, South Africa D Jabu Mtolo Deputy Director, Provincial Collaboration Department of Water and Sanitation Pretoria, South Africa Q Catherine Mgangira Project Officer, Official Development Assistance Programme Delegation of the European Union to South Africa D 1 Input Codes: I = Interviewed via Skype; Q = Responded to questions via email; D = Provided relevant documents

PAGE 317

304 APPENDIX H CASE STUDY GENERAL CONSULTATION GUIDES Decentralization 1. Tell me about your experience with or knowledge of water and sanitat ion policy in country ? 2. What has been the impact of sector decentralization on country progress in extending water and sanitation access? 3. What differences, if any, were there in the challenges faced by the water relative to the sanitation sector, both in ge neral and specifically with regard to decentralization? 4. Have there been changes to the roles of national level agencies for water and sanitation since the start of decentralization in the 1990s? How has this impacted municipalities? 5. Are you aware of any capacity building or technical assistance programs for municipalities that accompanied decentralization reforms? If so, please explain. 6. Are you aware of any national efforts that monitored t he impact of decentralization of water and sanitation access in urban areas? Please explain. 7. To what extent was there fiscal decentralization of water and sanitation services and how did this work in practice? 8. Were there any national programs designed to promote coverage for low income residents, such as subsidies? Have these been successful in helping to increase coverage? 9. What has been the impact of private sector involvement on increasing water and sanitation access is urban areas? 10. Has country prepared sector wide plans for the water or sanitation sector and if so, when? To what extent do you feel that these plans contributed to progress? 11. How engaged was civil society with the development of national water and sanitation policy during this period and in what ways? To wh at extent do you feel that civil society engagement contributed to progress? 12. Do you have any other information to add regarding influential activities in the sector? 13. Are there reports, datasets, or other documentation that you think might be useful for me to review as related to my research? 14. Is there any information we discussed today that you prefer not be attributed to you in documents I will prepare reporting my research findings? 15. Please confirm your name, position, and contact information for my records.

PAGE 318

305 Civil Society Engagement 1. Tell me about your experience with or knowledge of water and sanitation policy in country ? 2. In what ways has civil society been engaged with the development of national water and sanitation policy since about the year 2000? How do you feel this engagement has impacted country progress in extending water and sanitation access? 3. What are some of the primary ways that the national government reaches out to civil society when developing policies and plans, for example, pub lic forums or committees and do you feel these are effective mechanisms in practice? 4. To what extent is civil society involved in sector oversight and management at the national level? 5. Is information about pending policies, programs, and results in the s ector published by the national government? How is it published and made available? 6. Does the national policy support programs where the public can provide feedback or issue complaints about service delivery? How do they work? 7. In general, what are the different challenges faced in increasing sanitation access relative to water access for people in urban areas? 8. What differences, if any, were there in the way that civil society has been engaged with water and sanitation? 9. Who are the major civil society or ganizations working on water and sanitation at the national level? 10. Has country prepared sector wide plans for the water or sanitation sector and if so, when? How engaged was civil society in its development and ongoing revision? To what extent do you feel that these plans contributed to progress? 11. What has been the impact of sector decentralization (if applicable)? 12. Do you have any other information to add regarding influential activities in the sector? 13. Are there reports, datasets, or other documenta tion that you think might be useful for me to review as related to my research? 14. Is there any information we discussed today that you prefer not be attributed to you in documents I will prepare reporting my research findings? 15. Please confirm your name, position, and contact information for my records.