M. B. A., Tribhuvan University, Nepal, 1994
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Arts
This thesis for the Master of Arts
has been approved
Rimal, Nilu (M. A., Sociology)
Thesis directed by Associate Professor Richard H. Anderson
This study presents an analysis of the causes and consequences of fertility and
mortality transition in contemporary developing countries. Childrens role in the
family, aspiration of the parents, employment in non-agricultural sector, and
urbanization account for 72.2% of variation in total fertility rate. Contraceptive
prevalence, income, female secondary school enrollment, infant mortality, and
womens labor force participation account for 98.4% of total variation in total
fertility rate. Female literacy, income, public health expenditure, and external debt
explain 73.4% of total variation in infant mortality rate. Fertility and infant
mortality account for 93.1% of total variation in the proportion of children age 0-
14. Infant mortality accounts 48.6% of total variation in unemployment rate and
the latter accounts for 78.4% of total variation in fertility in cross country model
and 38.4% of total variation within country.
This abstract accurately represents the content of the candidates thesis. I
recommend its publication.
To my husband Michael Lewandowski for his support and confidence during long
hours of writing.
I wish to thank the members of the committee, Richard Anderson, Candan-
Duran Aydintug, and Yili Xu, for the time they dedicated to advising me.
Especially, I am grateful to Dr. Richard Anderson, advisor of the
committee, for his invaluable suggestions throughout all chapters and earlier
drafts of the study.
Purpose of the Study.........................6
Arrangement of the Thesis....................6
2. LITERATURE REVIEW
Theory of Demographic Transition........... 8
Theories of Fertility Decline...............23
Causes of Mortality Decline.................45
Population Components and Age Structure Dynamics.53
Nature of Data Set..........................59
Variables and Indicators....................61
Method of Analysis..........................67
4.1 Summary of Regression Analysis for..........................75
Variables Predicting Fertility (VOC)
4.2 Summary of Regression Analysis for
Variables Predicting Fertility (Proxi).....................78
4.3 Summary of Regression Analysis for
Variables Predicting Infant Mortality......................81
4.4 Summary of Regression Analysis for Variables.................83
Predicting Proportion of Children
4.5 Summary of Regression Analysis for Infant....................84
Mortality Predicting Unemployment
4.6 Summary of Regression Analysis for..........................85
Unemployment Predicting Fertility
4.7 Correlation Table ..........................................86
As fertility and mortality are the major components of vicious poverty
cycle of currently developing countries, it is important to make an
intensive effort to accelerate progress and more effectively address the
causes and consequences of fertility and mortality at country and global
levels. In order to do so, a thorough understanding of process of fertility
and mortality in relation to socioeconomic factors is required (Lopez,
2000). In the past, several studies have been conducted and hypotheses are
developed linking process of fertility and mortality with the pattern of
socioeconomic changes. In this thesis I propose to examine the linkage of
these socioeconomic conditions with fertility and mortality and will try to
determine the relative importance of each. The data I analyze for the
purpose of this study comes from World Bank. The World Bank 2000 data
set is the latest version and no research has be conducted so far using this
version of data set to test the interaction between fertility and mortality
with socioeconomic variables.
The vicious cycle of poverty related with fertility and mortality circles
around the demographic and socioeconomic components. Persistent high
fertility contributes to a very young age-structure. Higher percentage of
younger age cohorts not only creates the momentum for future population
growth but also contributes significantly to increasing unemployment.
Moreover, the governments of developing countries are unable to extend
their resources and opportunities in education, health, and other areas of
human necessities to meet the demand of rapidly growing population
(World Bank, 2000).
As the majority of the population lack access to education their
aspiration for a higher living standard become insignificant. As a result, the
couples do not perceive children as an investment but as productive assets
or security for the old age. Such perceptions lead to high fertility, which
further reduces the level well being of people in general. Limited access to
education also affects the ability of the poor to get jobs and to obtain
information that could improve the quality of their lives. Similarly, lack of
access to healthy living contributes to increase infant mortality which has a
significant positive association with fertility.
Two popular theories of fertility: theory of value and cost of children
and theory of proximate determinants are tested in this study. The theory
of value and cost of children emphasizes that children are viewed as
productive assets in traditional societies and costly liabilities in modem
societies. With industrialization and overall modernization the economic
value of childrens labor shifts from positive to negative. Qualitative studies
have been performed in Nigeria (Caldwell, 1982), Bangladesh (Caldwell,
1999), and Java and Nepal (Nag et al 1978) to explain how value and cost
of children are determined based on level of socioeconomic development,
performed qualitative study within a country context and concluded that
child education, industrialization, urbanization, and parents aspiration for
high living are the principal determinants of value and cost of children,
which finally influence the fertility. In this thesis I aim to test the
combined effect of these variables in a cross-country framework by using
Another theory known as theory of proximate determinants views fertility
as the function of exposure factors, natural fertility factors, and deliberate
fertility control factors. Exposure factors and natural fertility factors
determine the supply and demand of children. Supply and demand of
children further influence the deliberate fertility control factors. Exposure
factors, natural fertility factors, and deliberate fertility control factors are
known as proximate determinants because these variables act as mediator
between modernization and fertility. Modernization variables are found to
have profound impact on proximate determinants of fertility. Richards,
1983; Robinson, 1992; and Mahadavi, 1990 conducted quantitative studies
to test the impact of proximate determinants on fertility. These studies
suggest that family planning programs, infant mortality, income, female
education, and womens labor force participation are the major predictors
of fertility. I aim to test how much variation in fertility is caused by the
combined effect of these variables and compare the contribution of
proximate determinant variables with the contribution of value and cost of
children variables in determining fertility.
Both theories emphasize modernity variables as the principal determinants
of fertility. The point of difference between these two theories lies in the
nature of their explanation. Value and cost of children perspective explains
fertility in terms of economic benefit of children. Theory of proximate
determinants explains fertility in terms of supply and demand of children.
Now the question remains how much variation in fertility is explained by
each set of variables? Further, within each set of variables, which variables
are more influential in determining fertility?
Many studies have been conducted over the past decades in order to
provide an understanding of the causes of declining mortality in developing
countries (Levine et al., 1991; Lena and London, 1993; Hojman, 1996;
Lopez, 2000). The most common explanation for declining mortality as
provided by these studies rests on public health measures and female
literacy. Hojman (1996) studied the impact of external debt on overall
mortality. In this thesis I will measures the combined affect of external
debt, public expenditure on health, and female literacy on mortality.
Past studies of Freeman, 1979; Easterlin, 1978; Korenman and Neumark,
1997 suggest that a higher proportion of younger cohorts causes wage
deterioration of younger males which further causes fertility reduction
among younger cohorts. Easterlins (1978) hypothesis assumes that declining
infant mortality is responsible for increasing the size of younger cohorts.
An increase in the size of younger cohorts leads to the deterioration of the
wage of younger males. Wages deteriorate because the supply of employees
exceeds the supply of jobs. If infant mortality is the cause for increased
size of younger cohorts, and increased size of younger cohorts is the cause
for unemployment, then infant mortality should have an impact in the
employment conditions. Obviously, the survival chances of todays infant
will not influence the job market until these infants reach working age. In
many countries legal working age is 16. Thus, my question is Is the
infant mortality rate of 1960, 1961, 1962 to 1978 associated with
unemployment rate of 1976, 1977, 1978 to 1994 respectively? Another
question is does fertility decline with the increase in unemployment rate?
Purpose of the Study
Examine how much variation in fertility is explained by the combined
effect of socioeconomic variables that determine value and cost of
Examine how much variation in fertility is caused by the combined
effect of socioeconomic variables that influence proximate determinants.
Identify the most significant variable/s that determine the value and cost
Identify the most significant variable/s that affect the proximate
Identify the most significant variable/s that determine the level of infant
Examine how much variation in age-structure is caused by fertility and
Examine how much variation in unemployment rate is explained by
infant mortality rate.
Arrangement of the Thesis
Theories of demographic transition and past research in the field are
examined in Chapter 2 Literature Review. This section is divided into 4
subsections. Subsection Theory of Demographic Transition includes a
brief overview of the theory and perspectives of different theorists. In the
subsection, Theory of Fertility Decline, the causes of fertility decline is
explained. The impact of socioeconomic variables on mortality is examined
under subsection Causes of Mortality Decline. The interdependent effect
of age structure, fertility, and mortality is examined in the subsection
Population Components and Age-Structure dynamics.
The method of study is explained in chapter 3. The contents of
Methodology section are Nature of Data Set, Identification of Variables,
Definition of Variables, Methods of Analysis and Data Limitations. Chapter
4 and 5 include Findings and Conclusions of the study.
Theory of Demographic Transition
The theory of demographic transition provides a link between
development perspectives and population dynamics. The basic paradigm of
demographic transition theory is based on the model of modernization that
describes the demographic changes. The theory states that rise in standard
of living engendered by advancement of technology in the early 18th
century Europe improved general hygiene and health care facilities which
in turn reduced the level of mortality. While mortality declined, fertility
remained at pre-industrial level for several years. Prolonged decline in
mortality, especially infant or child mortality, had an ultimate impact on
fertility. Fertility started to decline in the late 19th century and the
equilibrium between fertility and mortality was reestablished during the 20th
century. In addition to the reduction in infant or child mortality, other
modernity variables had important roles in reducing fertility in 18th century
Europe (Szreter, 1993).
In most versions of transition theory, the general character of the process
experienced by the industrialized countries is depicted as a series of three
stages at which mortality and fertility behaved in particular ways in
accordance with fundamental economic and social changes of development
usually viewed as synonymous with modernization with low mortality and
fertility prevailing at the point at which development was achieved and the
demographic transition thus completed (Leibenstein, 1974).
In the first stage, equilibrium of population size is maintained by high
birth and death rates. Because of high mortality in the absence of the
health and sanitation services and other features that characterize modem
society, societies in the first stage of demographic transition employed a
series of pronatalist props to keep fertility high (Caldwell, 1976). Otherwise
the high mortality rate would have led to population decline and eventual
extinction. Over time, mortality began to decline in country after country
as methods of death control became available, creating a situation in which
both high fertility and low mortality were valued. In addition to health and
sanitation, other development related factors contributed to mortality decline.
Prominent among these factors were improved transportation and
communication facilities, the shift from home to factory based employment,
and the advent of scientific agriculture (Freedman, 1979).
The equilibrium of the first stage became potentially unstable and this
instability led to the second stage, in which there was rapid population
growth resulting from an imbalance between continuing high fertility and
declining mortality rates. The onset of the third stage occurred when
individual members of the population consciously began to control their
fertility and the birth rate gradually declined toward equilibrium with the
low death rate (Teitelbaum, 1987).
Central to the thinking of the proponents of this demographic transition
model was the view that fertility decline lags behind mortality decline
because it cannot occur until the traditional social and economic
arrangements and institutions that support high fertility are eroded by the
advent of the transformations shaped by modernization, which favors a
reduction in fertility to levels that correspond with lower levels of
mortality. The same process of modernization, largely equated with
industrialization and urbanization, which brought down the death rates, was
seen as destroying the pronatalist props (Caldwell, 1976).
Subsequent historical research on fertility and mortality in 19th century
Europe, particularly the work of the Princeton group suggests that overall
fertility levels in Europe before the demographic transition varied
considerably across countries and regions (Teitelbaum, 1987). Moreover the
patterns of development conditions varied. It appears that in the European
case, cultural and linguistic factors are more closely associated with
variations in fertility decline than is the cluster of modernization factors
posited as central to transition theory (Freedman, 1976).
Coale (as cited in Teitelbaum, 1987) summarized the transition theory in
the form of three broad preconditions for sustained decline in marital
1. Fertility must be within the calculus of conscious choice. Potential
parents must consider it an acceptable mode of thought and form of
behavior to balance advantages and disadvantages before deciding to
have another child.
2. Reduced fertility must be advantageous. Perceived social and economic
circumstances must make reduced fertility seem an advantage to
3. Effective techniques of fertility reduction must be available. Procedures
that will in fact prevent births must be known and there must be
sufficient communication between spouses and sufficient sustained will
in both to employ them successfully.
The European data show that a high level of development was ultimately
sufficient to establish these three preconditions for a decline in marital
fertility across Europe. There is no evidence, however, of any threshold
levels of development that were necessary for this to happen, and it is
evident that the preconditions for fertility decline could exist in situations
of little economic development (Teitelbaum, 1987). The most striking
findings to emerge from the research on the fertility transition in Europe is
that it occurred under remarkably diverse socioeconomic and demographic
conditions and this diversity is to be encountered in the less developed
countries as well (Knodel and Walle, 1979).
Fertility started to decline in England only around 1870s, half the
century after industrial revolution. While in France, fertility started to
decline as early as 1780 long before the start of industrial revolution
(Caldwell, 1999). Family structure and social practices dictated fertility
level. Pronatalist props and the extended family system supported a high
fertility regime in England. Conversely, in France, delayed age at marriage,
high proportion of never married, and coitus interrupts played important
roles in maintaining low fertility (Wrigley, 1969). Social structure was
another cause of fertility differentiation between England and France. In
England, wealth was highly capitalized by few, leaving behind the great
majority of property-less unskilled workers. A majority of urban workers
did not secure the improvement in living standard until the 1860s and
1870s. Hence the lack of improvement in the standard of living restricted
peoples aspiration to grow high, which in turn conspired to retain the high
fertility during the first 5 decades of industrial revolution. In France, on
the other hand, wealth was dispersed more equally among the people
throughout late 18th and early 19th century French societies (Goldsheider,
1971). The aspiration to have better life was not limited to a small
segment of population. The falling mortality rate was another cause of
fertility decline (Wrigley, 1969). Because of the system of property
inheritance and slow pace of industrialization in the cities, rural to urban
migration remained at low levels (Caldwell, 1999). Therefore, rural people
had to make a choice between reducing fertility and maintaining high
living standard or maintaining high fertility and reducing standard of living.
In the given circumstance the French people chose to reduce fertility
Another reason for retaining high fertility in England during the early
decades after industrial revolution was the rural-urban migration. Rural
England had higher births rates and lower mortality rates than those of
urban England. The natural increase of population in rural England was
compensated by outward migration to urban England. Because of migration,
rural England did not feel a pressure to restrict fertility (Goldsheider,
1971). Fertility decline in England that started around 1870s was caused by
increased levels of education especially among females, secularization of
the society, and the extensive change in peoples opinion towards use of
contraception as being legitimate and morally acceptable (Woods, 1986).
The change was rapid and was catalyzed by massive family planning
programs during the 1870s, 1880s, and 1890s (Caldwell, 1999).
Thus the major weakness of demographic transition theory is that it
could not define a precise threshold of modernization that would reliably
identify a population in which fertility is ready to fall. The empirical
evidence from developing countries suggests that different countries entered
the transition at different level of development. Human Development Index
(HDI) for Singapore and Hong Kong was relatively higher (0.69 and 0.67
respectively) at the onset of transition. Singapore entered the transition in
1959 and Hong Kong in 1960. Conversely, countries like India,
Bangladesh, and Nepal had relatively low HDI at the onset of transition.
The HDI at the onset of transition was 0.36 for India, 0.32 for
Bangladesh, and 0.33 for Nepal. India entered the transition in 1973,
Bangladesh entered in 1978, and Nepal in 1988. Also there is a notable
variation in the pace of fertility decline across countries. In Bolivia and
India there was only 10-15% decline in fertility for the first 10 years as
opposed to Thailand and Singapore where fertility declined by more than
30% every decade. The pace of socio-economic change after the onset of
transition might be the cause for variation in the pace of the transition
across countries (Bongaarts and Watkins 1996). The pace of transition
might be rapid for countries that have a rapid pace of socio-economic
Examples from Bangladesh, Pakistan, and Philippines show that a small
change in socioeconomic factors leads to fertility decline. Given the high
infant mortality rate of 82 per thousand and low income per capita of
US$270 in 1997, Bangladesh is the only country in the world to have
fertility of less than five births per woman. This is a low rate of fertility
as compared to other low income countries where fertility rate is more
than 6 births per woman (Caldwell, et al., 1999). In the early 1970s
Bangladeshs fertility was seven births per woman (Arthur and McNicoll,
1978). It is clear that within two decades Bangladesh was able to reduce
its total fertility rate by more than two births per woman.
Findings of a survey conducted in Bangladesh shows that the fertility
decline in Bangladesh is a direct response to decline in mortality rates,
changing status of marginalized women, commercialization of farming, and
aspiration for upward mobility (Arthur and McNicoll 1978). Another study
conducted in two rural agrarian districts of Bangladesh provides evidence
that the fertility decline is attributable more to rising living standard,
education intervention programs, and opportunities for upward mobility and
less to family planning program and girls education (Caldwell et al., 1999).
The evidence from different studies conducted over a period of eight
years indicates that Pakistan finally has begun to show a trend of fertility
decline. As one of the poorest country in the world, Pakistan is the 7th
most populous country. Pakistan continued to have more than 6.0 births per
woman throughout the 1980s. In early 1990s, Pakistans fertility started to
decline and total fertility rate dropped to six children per woman (Sathar
and Casterline 1998).
During the 1970s and 1980s Pakistan had somewhat better economy than
the economy of 1990s. The economic conditions of 1980s were capable to
support the governance of high fertility regime. As the economy started to
decline beginning in the 1990s, unemployment increased, the real wage of
workers decreased, and inequality of income increased. The overall impact
of these macroeconomic forces was that the couples were no more in a
condition to afford to have many children. Beside economic stress,
migration to urban areas and exposure to mass media played a significant
role in reducing fertility in Pakistan. Both exposure to mass media and
migration to urban areas introduced a new but costly life style that would
not support the governance of high fertility regime. Another important
cause of fertility decline in Pakistan is massive change in family structure.
Now 50% of the households are nuclear. The decision to reproduce is in
the hands of couples not in the hands of elders of the family (Sathar and
Casterline 1998). The fertility decline of Pakistan captures the essence of
both growing poverty and modernity. Thus we can conclude that fertility
decline is not only associated with positive economic development.
The -case of Philippines demonstrates that fertility decline is related to
modernization. The electrification of one of the rural areas of Philippines
caused substantial fertility decline. Electrification increased agricultural and
non-agricultural production, created new employment in both agricultural
and non-agricultural sectors, improved quality of education, improved
sanitation, and improved the condition of health centers (Herrin 1979).
Creation of employment opportunities opened the door for women to get
jobs outside home. Increased production and employment opportunities
together increased family income. Increase in family income changed
consumption habits. Families started to buy durable consumer goods (TV,
sewing machines, radio etc.). Along with consumption habits, the family
style also changed. Parents started to give importance to childrens
education and healthy living of themselves and their children. Change in
consumption habits and life style was expensive. In order to keep up with
high living standard, couples decided to reduce their fertility (Herrin 1979).
Historical experience of developed countries suggests that_the shift from
agriculture to industry does not automatically reduce fertility. Institutional
changes like breakup of kinship domination, change in family structure
from extended family to nuclear family, increase in living standards, and
dependence on social network of institution rather than on family bring the
pressure for fertility reduction. The experience of developing countries
suggests that conditions like urbanization, labor force participation of
women, aspiration for upward mobility, jobs in non-agricultural sector,
infant mortality decline, and education of women and children may explain
significant decline in fertility. In the case of developing countries, not only
the positive socio-economic forces but also the negative forces such as
growing poverty contribute to fertility decline as in case of Pakistan.
In short, demographic transition theory relates type of population growth
to the level of socio-economic and technological development of a society.
The theory suggests that in traditional societies fertility and mortality rates
are high and in modem societies fertility and mortality rates are low. In
between there is a demographic transition.
Development of Demographic
In 1929, Warren Thompson (as cited in Nam and Gustavus, 1976),
developed the idea of demographic transition and denoted industrialization
as the necessary condition to accomplish shift from high fertility and
mortality to low fertility and mortality. Thompson classified countries into
three categories based on their fertility and mortality pattern. In the first
category were the industrialized countries of Northern and Western Europe
and the United States where high rates of fertility and mortality were
transformed into low rates of fertility and mortality. In the second category
were the Southern and Central European countries (Italy, Spain and Slavic
countries) where both fertility and mortality were declining. But there was
a gap in between the rates of fertility and mortality decline. Fertility was
declining at a very slow, pace while mortality was declining rapidly. The
third category was the rest of the world where both fertility and mortality
Thompsons classification of countries entails the conditions of stationary
and growing population. A country will have a stationary population under
following two conditions:
High fertility and high mortality: High fertility is offset by high
mortality that restricts the growth of population.
Low fertility and low mortality: Both fertility and mortality are low
there is little difference between fertility and mortality. Low mortality
prevents population from declining.
A country will have a growing population under the following two
High fertility and low mortality: This will lead to population increase
since the level of fertility far exceeds the level of mortality.
Low fertility and stationary mortality: Mortality remains at the previous
level but fertility declines further. In this condition population declines.
In 1945, Notestein (as cited in Nam and Gustavus, 1976) refined
Thompsons classification and coined the term Demographic Transition.
Notestein classified stages of population growth as the phase of incipient
decline, transitional growth, and growth potential. Incipient decline is the
condition of low fertility and low mortality. Transitional growth is the
condition of rapid mortality decline followed by slow fertility decilne.
Growth potential is the condition of high fertility and high mortality.
By taking the Western European experience of demographic transition,
we can formulate that modernization of society is the important route to
achieve the transition from high fertility and mortality to low fertility and
mortality. Fertility decisions become rational in modem societies due to
education, industrialization, high living standard, and individualism. In
agrarian societies, pronatalist props act as catalyst to maintain high fertility
regime (Notestein, 1945, as cited in Caldwell, 1982).
In 1963, Davis (as cited in Woods, 1986) elaborated the work of
Thompson and Notestein and supported the viewpoints that fertility decline
is the response to modernization. In addition, Davis emphasized that
fertility is rational in all kind of societies. The high fertility of traditional
societies is rational because in such societies children are the net
producers, education of children is not deemed important, and the cost of
children is shared among the family members of the extended family
system. Conversely, low fertility of modem societies is because children are
the net consumers, education of children is deemed important, and couples,
not the whole family is responsible for the cost of children.
Stage Theories of
In 1947, Blacker (as cited in Shrivastava, 1983) identified five stages of
population growth: high stationary stage, early expansion stage, late
expansion stage, low stationary stage, and declining stage. The high
stationary stage is prevalent in primitive societies where famines and
epidemics cause deaths. The early expansion stage is characterized by
declining mortality rates due to innovation in medical technology and food
production. In late expanding stage, fertility starts to decline as a response
to declining mortality. In the low stationary stage, both fertility and
mortality are at their lowest level. In the stage of declining population,
fertility declines even further but there is no further decline in mortality. In
1930, of all the countries in the world, 22% were in the high stationary
stage, 40% were in the early expanding stage, 22% in the late expanding
stage, 14% in the low stationary stage, and 2% in the declining stage
(Blacker, 1967, as cited in Shrivastava, 1983).
Cowgill (As cited in Nam and Gustavus, 1976) introduced four cycles of
growth: the primitive cycle, the modem cycle, the future cycle, and the
probable cycle. In the primitive cycle, fertility rates are high but the
mortality rates vary. Mortality varies inversely with the food production of
society. The rate of mortality is also associated with the epidemics of
diseases and natural disasters. Modem cycle is the period of transition
between high stationary and low stationary population. In this period both
fertility and mortality decline but mortality declines at a faster rate than
does fertility. The future cycle is the period when mortality declines to the
level of biological minimum. In this cycle population growth is just the
function of fertility. The imaginary cycle has not taken place in the history
of humanity yet.
Laundry (as cited in Shrivastava, 1983) envisioned three stages of
population growth: primitive population regime, intermediate population
regime, and modem population regime. In primitive population regime,
increase or decrease of population is solely determined by the availability
of food supply. In intermediate population regime, population increase is
not only the function of food availability, but also, the function of
economic development. Once the pattern of continuous food supply is
established, people will want to increase the standard of living. In order to
increase living standard people will start controlling their fertility. In
modem population regime, people control births even more than in
intermediate population regime. People do so in order to be able to derive
maximum benefit from the booming economy.
Theories of Fertility Decline
There are two broad theoretical camps that explain the link between
modernity and fertility behavior. One camp provides an explanation of how
the direction of wealth flow effects value and cost of children in pre-
industrial and industrial societies and the other provides an explanation of
how modernity variables influence supply and demand of children via
proximate determinants (Caldwell, 1982).
In this section I will review the literature on how modernity affects
fertility behavior from the perspective of these two theoretical camps.
Direction of Wealth Flow: Value
and Cost of Children
There are two different types of fertility regimes. One is high fertility
regime and the other is low fertility regime. Fertility behavior is rational
under both regimes. Non-industrial societies have high fertility because
wealth flows from children to parents. Conversely, industrial societies have
low fertility because wealth flows from parents to children (Caldwell,
1976). As the direction of wealth flow changes from children-parents to
parent-children the value and cost of children changes. In the former,
benefits from children exceed the cost. In the latter, the cost of children
exceeds the benefits.
Childrens role in family is one of the main predictors of fertility. In
agrarian societies very young children perform such tasks as gathering
firewood, sweeping floors, grinding and pounding grains, taking food to
adults in the field, carrying water, and running errands. Older children are
involved in cooking meals, hunting, herding, working in the fields, fishing,
and making pots, containers, mats, and nets (Harris and Rose, 1987). The
study conducted in one of the villages of Bangladesh demonstrates that by
the age 12 a boy produces enough to pay for his food and other necessary
expenses. By 15 he pays for all his past consumption. By the age of 21
he pays for the consumption of his own and of his sister. The income
earned by the boy beyond age 21 and until the boy gets married becomes
the surplus for parents. The average age at marriage for boys in that
village is mid twenties (Cain, 1977). Studies in Java and Nepal indicate
that approximately 80% of old people are supported by their children.
These studies show that the average boy and girl gets involved in
productive activities by age five. By the ages of 15-19, these boys and
girls spend more than 55 hours per week in different types of activities,
ranging from agriculturally productive activities to household maintenance
(Nag et al, 1978). Thus, in agrarian societies children are the net producers
and wealth flows from children to parents and hence fertility is high.
In modem societies universal child education undermines the value of
child labor. Child education also contributes to increasing the cost of
children. Thus, in modem societies, the childrens role as a producer of
family income is passive. Therefore, fertility is low in modem societies
(Harris and Rose, 1987). Towards the late 19th century, mass education in
Australia allowed parents to withdraw their children from wage employment
and send them to school, which was one of the major causes of the onset
of fertility decline (Caldwell, 1982). Withdrawal of children from the labor
force to start schooling changed the direction of wealth flow. Before the
introduction of mass education system, the direction of wealth flow was
from children to parents. Afterward, the direction of wealth flow changed
from parents to children. The change in direction of wealth flow also
changed the relationship between parents and children. Consequently, the
value and cost of children were altered. The educational cost for children,
perhaps, was the greatest among all the costs of children in industrialized
countries. As the phase of industrialization shifted from textile
manufacturing to chemical and steel manufacturing, the demand for skilled
workers increased enormously. Middle class and working class parents
immediately sensed the importance of education for their children. As a
result, not only did enrollment increase, but there was also an increase in
the number of years of schooling and the cost of education (Becker, 1991).
The U.S. data from 1900-1956 shows that the percentage increase in the
cost of education was approximately three and a half times as large as the
percentage increase in consumer income (Schultz, 1974).
Parents taste is another important variable that predicts the level of
fertility. Evidence suggests that in agrarian societies parents do not have
demand for consumer goods and the cost of children is low. Conversely, in
modem societies, taste for consumer goods engendered by the aspiration for
higher living standard tend to increase the direct cost of children (Bulatao
and Lee, 1983). Increased aspiration of parents to acquire durable goods
was one of the reasons for massive fertility decline in Philippines (Herrin,
1979). In Australia, department stores were established during thel880s
with a notable increase in the sale of items like childrens toys (Caldwell,
and Ruzicka, 1978). This indicates the rise in the emotional value of
children with the rise in use of consumer goods. Thus change in
consumption habits changed both the emotional value as well as the cost
Occupation is one of the factors that predict level of fertility. The
evidence from 28 developing countries confirms that individuals involved in
agriculture have the highest fertility rate of 8.38 births per woman. The
lowest fertility rate of 6.74 was found among professionals, managers, and
clerical employees. People involved in manual work had a fertility rate of
7.63. The restructuring of the labor market has caused fertility to decline
in many developing countries (Boserup, 1981). Thus, in modem societies a
higher percentage of people are employed in the non-agricultural sector
hence the fertility is low as opposed to agrarian societies where the
majority of people are employed in agricultural sector and hence the
fertility is high.
Urban or rural residency effects the perception of value and cost of
children. Empirical evidence from Nigeria suggests that there is a notable
difference between the perception regarding value and cost of children of
rural and urban residents. On the benefit side, middle class urban parents
emphasize emotional and psychological benefits from children as opposed
to the rural parents who emphasize economic benefits to be generated from
the children. On the cost side, urban parents are more conscious of the
opportunity cost and direct cost of rearing children while rural parents are
not conscious of such costs. Low class urban parents share the feature of
middle class urban parents and rural parents. The low class urban parent
has high fertility with higher economic burden and higher cost of urban
living (Caldwell, 1982). These findings suggest that children are valued
differently in different socio-economic groups even within the same
country. Among higher socio-economic groups the economic cost of
children is higher but they are more valuable emotionally. Therefore, as the
socio-economic status improves, the economic cost of raising children
increases and demand for children decreases.
From the theory of value and cost of children we can conclude that
many modernization variables play a significant role in determining fertility
level in a society. In traditional societies the value of children is high and
cost is low. This high value and low cost of children leads to high
fertility. In modem societies, value of children is low and cost is high.
Thus, in modem societies fertility is at lower level. Higher value and
lower cost of children in traditional societies is because of high
concentration of people in agricultural occupation, higher percentage of
children in the labor force, less people in urban areas, and low aspiration
of parents. Conversely, lower value and higher cost of children in modem
society is due to higher importance of child education, high aspiration,
fewer children in the labor force, more people in urban areas, and fewer
people in agricultural occupation. Thus, childrens role in producing family
income, urbanization, parents aspiration, and employment in agriculture
sector exert combined force to determine the level of fertility. It is much
more rational for couples to limit their fertility.
Hypothesis I: Fertility varies inversely with the variation in urbanization
and parents aspiration and positively with the variation in childrens role
in producing family income and employment in agricultural sector.
Theory of Proximate Determinants
Bongaarts theory of proximate determinants and Easterlins supply
demand framework has become fairly dominant in the contemporary debate
of fertility change (Caldwell, 1982). Easterlin identifies three determinants
that influence fertility decision-making. The supply of children, the demand
for children, and cost of fertility regulation. The proximate determinants of
fertility are exposure factors, deliberate fertility factors, and natural fertility
factors (Bongaarts 1978).
1. Exposure factors: proportion of married women, age at marriage, and
frequency of intercourse.
2. Deliberate marital fertility control factors: Contraception use and
3. Natural marital fertility factors: postpartum, fecundability, infecundability,
and permanent sterility.
Socio-economic factors have direct impact on exposure factors and
natural marital fertility factors, and indirect impact on deliberate marital
fertility control factors. The three intervening variables that come in
between socio-economic variables and deliberate fertility control factors are
supply of children, demand for children, and the regulating cost of fertility
(Easterlin and Crimmins, 1985).
Supply of children determines the potential family size. This shows how
many surviving children a family could have in the absence of deliberate
efforts to limit fertility. Potential family size depends on the chance of
child survival and the levels of natural fertility. Both child survival and
natural fertility are determined by cultural and socio-economic factors.
Demand for children determines the desired family size. This shows how
many surviving children a family would want if the costs of fertility
regulation were negligible. The desired family size is the function of four
variables; income and wealth, tastes and norms, time costs of children, and
direct costs and benefits of children (Easterlin, 1975).
There are two types of costs involved in fertility regulation. First are the
psychological costs such as aversion for the general concept of family
planning and the shortcomings of abortion. Second are the economic costs,
such as money and time required to acquire information about the new
techniques of family control (Easterlin, and Crimmins, 1985).
Deliberate control of fertility depends bn the difference between potential
and desired family size. If potential family size is smaller than the desired
family size, then there is encouragement for having more children. If
desired family size is smaller than the potential family size, households are
motivated to regulate their fertility in order to stay away from the
prospects of surplus children (Easterlin and Crimmins, 1985).
Empirical evidence suggests that in developed nations, the desired family
size is much smaller than the potential family size. The smaller desired
family is the principal determinant of low fertility in developed nations in
the given scenario of negligible cost of fertility regulation (Richards, 1983).
According to the same studies, in many poor developing countries, the
situation is opposite; fertility is determined primarily by the supply factors.
The following paragraphs illustrate how education, infant or child
mortality, family planning programs, urbanization, and income affect fertility
via proximate determinants.
Education and Fertility. Until the mid-1970s, it was commonly believed
that increased education caused fertility to decline uniformly. However, a
broad analysis of the existing empirical evidence up to the late 1970s
shows some inconsistencies of these widely accepted assumptions. In least
developed countries, among widely illiterate rural societies this expected
inverse relationship between education and fertility was not found.
Inconsistencies across countries and regions suggest that inverse relationship
between education and fertility cannot be established and that other related
variables should be studied in relation to fertility (Cochrane, 1979).
The evidence suggests that there is a threshold level of education below
which fertility has positive association with education. Usually the threshold
level is beyond primary education (Bulatao and Lee, 1983). Below the
threshold level, education is positively associated with fertility. Below the
threshold level, education increases mothers awareness on hygiene and
nutrition care, which in turn increases the likelihood of child survival, thus,
raising the potential family size (Cochrane 1979). This occurrence is
usually credited to the impact of primary education on child survival.
Beyond threshold level education is inversely associated with fertility.
Education beyond threshold level decreases the desired family size. A
better-educated woman has more possibilities to improve her earning
capabilities. Under such conditions, the cost of children increases due to
the involvement of the opportunity cost of the mothers time in rearing
children. Thus an educated woman restricts her fertility (Cochrane, 1979).
Different studies of relationship between fertility and education reveal
varied patterns. Twenty-six of such studies found an inverse relationship
between education and fertility. Thirteen studies found a reverse U-shaped
relationship or positive relationship. Seven studies found either positive
relationship or no relationship (Jejeebhoy, 1995). What appears to be
happening is that on the one hand, societies at very low levels of
development, small amounts of education do not produce a decline in
fertility or in some cases actually lead to an increase in fertility. On the
other hand when a greater amount of education are obtained fertility then
starts a continual decline. Thus education could have both positive as well
as negative impacts on fertility.
Negative Impact of Education on Proximate Determinants. A cross-
country comparison based on World Fertility Survey data shows that
education has an inverse impact on fertility. However, the strength of the
relationship was dependent on social structure, cultural environment, and
economic development (Cochrane, 1979).
Female education has tremendous impact in determining the age at
marriage The findings from Demographic Health Survey (DHS) data
suggest that age at marriage of educated women is higher than that of
uneducated women (Shapiro, 1996). A considerable cross countries variation
has been found in the age at marriage even at the same level of female
education reflecting the impact of socio-cultural norms and traditions.
Similarly, within country analysis shows a significant gap between the age
at marriage of educated and uneducated women even in the regions like
Sub-Saharan and North Africa where the early marriage is rewarded.
In Taiwan, during the transition, increased age at marriage explained
67% decline in total fertility rate (Freedman et al, 1994). Marital fertility
rate decreases with the increased age at marriage. Reduced marital fertility
rate ultimately brings down the total fertility rate (Schultz, 1980). The
fertility data from 1950 to 1973 suggest that 86% of fertility decline of
Japan was due to decline in marital fertility rate (Ogawa and Retherford,
1993). Similarly, decline in marital fertility rate in Taiwan explained 33%
decline in total fertility rate (Freedman et al, 1994)
In the case of Africa this inverse relationship between schooling and
age at marriage might not be valid, since in many African societies,
marriage is not necessarily a long-term unification. Thus the unification
does not necessarily mean marriage. Therefore, in African society women
might marry at a later age but still might bear children at an earlier age.
In Latin America and Asia, women with no schooling marry at an early
age. In Africa, however, the marriage index for those with all levels of
schooling is always lower than that of the unschooled (Cochrane, 1979).
Education reduces the fertility regulation costs by altering attitudes
towards the use of contraception and increasing the know how of effective
use of contraceptives. By increasing income, education also compensates for
the economic costs of fertility regulation (Richards, 1983)
Positive Impact of Education on Proximate Determinants. Education has a
positive relationship with the proximate determinants in various ways. First,
it increases parents knowledge on health care, which in turn increases the
chance of child survival. The relationship between a mothers education
and childs survival is stronger than the relationship between a fathers
education and childs survival (Cochrane, 1979). Evidence from Mexico
shows that relatively weaker relationship between a fathers education and
health care of children, in comparison to the relationship between a
mothers education and health care of children usually signifies a gender
based division of labor within a household (Becker, 1991).
Education leads to less breast-feeding which increases the chance of
conceiving by shortening postpartum infecundability. Postpartum
infecundability shortens as the duration of breast-feeding decreases (Bulatao
and Lee, 1983). Postpartum infecundability is the period between the child
delivery and return to ovulation, which depends mainly on the intensity
and the duration of breast-feeding (Nag, 1983). Thus, education increases
the natural fertility potential.
Empirical evidence suggests that in developing countries schooling has
led to less breast-feeding. A recent study in Cote DIvoire shows that
duration of breast-feeding decreases with the increase in secondary
education, which enhances the chance of getting pregnant. However,
education also increases the age of co-habitation. These two processes are
offsetting, but the delayed age of cohabitation exceeds the postpartum
infecundability, except among the oldest cohorts (Easterlin and Crimmins,
Education also changes traditional beliefs and values and, therefore,
challenges cultural restrains like intercourse prohibitions. Some parts of
Sub-Saharan Africa have a culture that prohibits sexual contact during
lactation (Cochrane, 1979). Education makes individuals challenge such
intercourse prohibition. Such challenges in turn increases natural fertility.
Infant or Child Mortality and Fertility. Significant impact of infant or
child mortality on proximate determinants of fertility has been found.
However, the cause-effect relationship between fertility and child or infant
mortality is not that simple. Not only is mortality the cause for fertility
fluctuation, but fertility is the cause for mortality fluctuation as well. A
higher rate of fertility causes higher rate of mortality, since high fertility is
associated with short birth intervals, large family size, low nutrition and
calorie intake, less care for children etc., all of which shows a combined
effect in increasing the likelihood of child mortality (Caldwell, 1986). The
following paragraphs illustrate positive and negative impact of infant
mortality on proximate determinants.
Infant or child mortality has an inverse association with deliberate
fertility control factors and positive association with exposure factors.
Demand for children decreases with a decrease in infant and child
mortality. A decline in infant or child mortality leads parents to believe
that almost all the children ever bom to them will survive. So they do not
feel a need to hoard extra births (Hojman, 1996). Increased chance of child
survival makes parents believe that they will get their return from the
investment in children. This belief will lead parents to invest in child
quality, rather than in child quantity. Investment in child quality motivates
parents to control their natural fertility (Jejeebhoy, 1995).
Low infant mortality also has impact on postpartum infecundity, as long
as the infant is breast-fed. Survival of infants extends the postpartum
infecundity because of breast-feeding. In other words, decline in infant
mortality reduces the natural fertility potential (Nag, 1983). Evidence from
poor agricultural societies suggest that a woman will have 20% more births
throughout her childbearing age, if all of her babies die in infancy, as
opposed to a woman whose babies do not die in infancy (Caldwell, 1986).
Conversely, in the condition of high rate of infant or child mortality,
parents may try to bear more children than what they have desired for just
to be insured against the risk of potential deaths. If the death of a child
occurs too late in the reproductive cycle, the parents may not able to
replace the child. Thus the number of children in a household is
determined by an individuals own experience of mortality, as well as their
perception of societal level mortality (Cochrane, 1979).
Womens Labor Force Participation. Empirical findings suggest that
demand for children decreases with womens participation in the labor
force. If women participate in the labor force, the opportunity cost of their
time increases. Increase in the opportunity cost of time leads to lower
fertility (Mahadavi, 1990). Womens participation in the labor force was
one of the major forces that led to fertility decline in early 19 century
Europe and other industrialized countries. As women stay longer in the
labor force before marrying and their wages continue to rise, the
opportunity costs of leaving the labor force also continue to rise, making
women less inclined to resign from their jobs. In Japan the fall in
proportion married, and consequently, the fall in age-specific marital
fertility after 1950 is attributable to womens participation to labor force
engendered by womens educational attainment (Ogawa and Retherford,
Women tend to have fewer children when they have better opportunities
for education and employment in non-agricultural sectors (Hirschman and
Guest 1990). Cross-country studies of 83 countries show that women's
participation in labor force explain 91.24% of the variation in fertility. The
explanation of this finding was that womens labor force participation
increases the opportunity costs of their time which motivates women to
control their fertility (Cain 1993). The study of four Southeast Asian
countries reveal that fertility is relatively lower in those countries where
the proportion of women employed outside agriculture sector is higher
(Hirschman and Guest 1990). Since 1960s, fertility in Ecuador started to
decline. The present fertility level of Ecuador is about four children per
women. This fertility rate is moderate in relation to other Latin American
countries. Womens labor force participation along with contraception use
and female education has been found to be the principle determinant of
fertility decline of Ecuador (Lobao and Brown 1998). In the case of
developing countries, rural areas are predominantly based on domestic
production. Therefore, childcare may not necessarily be a constraint in
labor force participation. Women can work and take care of their children
at the same time. In urban industrial areas, the availability of childcare
services undermines the relationship between womens labor force
participation and fertility (Bulatao and Lee, 1983).
Family Planning Programs. Family planning programs have very
insignificant impact on fertility. The comparative analysis of the data from
1960s and 1970s of 83 countries confirms that family planning programs
account for 3 to 10 percent of variation in fertility level (Hernandez,
1984). However, the findings from other cross-country studies suggest that
well managed extensive family programs have sustained impact on fertility,
independent of other socio-economic variables (Mauldin, 1986). A
significant regional variation has been found in use of contraception. The
rates of contraception use vary from more than 60% in Sri Lanka, Brazil,
Thailand, and Colombia to less than 10 percent in Mali, Uganda, Burundi,
and Liberia. The rate of contraception use is lowest in Sub-Saharan
regions. (Hernandez, 1984).
During the 1980s, many developing countries experienced a significant
decline in fertility with an increased application of family planning
practices (Mauldin 1983). However, the ratio of reproductive age females
using family planning programs varies significantly across developing
countries. Empirical evidence shows that the ratio of reproductive age
women practicing family planning program is higher in countries like India,
China, and Indonesia and lower in countries like Bangladesh, Nigeria, and
Pakistan. Despite this fact the effects of the national family planning
programs in countries like Bangladesh, Nigeria, and Pakistan have seemed
to be negligible for several years (Bulatao and Lee, 1983).
Family planning programs are found to have significant impact on
reducing cost of fertility regulation. Reduction in cost of fertility regulation
are the strong motivators to control fertility (Robinson, 1992). The
establishment of family planning clinics providing services at prices well
below prevailing prices in the market leads to fertility decline. A family
planning program may also reduce the psychological cost related with the
use of methods of fertility control by providing legitimization, via
advertising and demonstration, to customs that might otherwise would have
been considered as unacceptable to traditional culture (de Graft, 1991).
Income. There are two main schools of thought that explain the impact
of income on fertility. First school is the Chicago school, or home
economists, and the second school is the Easterlins school.
Chicago School or Home Economists. Under this school of thought,
children are viewed as consumer goods. Children as consumer goods are
explained by income effects and substitution effects. Income effect takes
into account only mens income and ignores womens income. On the
other hand substitution effect considers womens income as the major
determinant of fertility (Becker, 1991).
Increase in mens income increases the demand for children, given the
condition that value of children is constant and women are the only
responsible persons for child rearing. Such correlation between mens
income and demand for children is called the pure income effect (Becker,
1991). Increase in demand for children have negative impact on deliberate
fertility control factors.
Increases in womens income, on the other hand not only increases the
family income but also increases the price of children by increasing the
opportunity cost of mothers time (Heerink, 1994). Therefore, womens
income has a substitution effect in terms of increased opportunity cost of
The substitution effect arises when the increased income enhances
aspiration for quality children (Becker, 1991). Under such circumstances,
parents prefer fewer children of higher quality to many children of inferior
quality. The substitution effect is engendered when increase in income
makes parents acquire other insurances as the alternatives to child services
for old age security. Substitution of child services by other insurance
alternatives decreases the demand for children (Caldwell, 1982). Substitution
effects in the desired family size may be viewed as indirect income
Easterlins School. Income influences fertility through intervening
variables (Easterlin and Crimmins, 1985). Evidence suggests that income
seems to have a positive effect on nutrition, education, and health and a
negative effect on infant mortality and child labor force participation,
whereas the effect of income on female labor force participation can be
positive or negative depending on the country context (Heerink, 1994).
Income can have an inverse impact on exposure factors in the societies
where marriages are postponed by the dowry requirement or by the setting
up cost of a new household (Repeto, 1979). In such circumstances, the age
at marriage may be lowered due to increasing income, although wealthier
parents tend to have high aspirations for their children in terms of
education and better employment.
Increase in income can enhance natural fertility potential since an
increase in income may increase the practice of breast-feeding substitutes.
As a result, the duration and the intensity of breast-feeding decreases,
which in turn shortens the period of postpartum infecundability. Shortening
of the period of postpartum infecundability increases the natural marital
fertility (Nag, 1983).
In developing countries, fertility regulation cost forms a considerably big
portion of family resources of poor households. In such a situation, a rise
in income can enhance the access to fertility control and the use of
contraceptive methods among poor households (Easterlin and Crimmins,
1985). For example, in Bangladesh, participation in income generation
programs increased the use of contraception among poor household women
(Caldwell et al., 1999)
The studies of cross-countries and country specific macro-economic data
provide an insight that in the course of economic development, the higher
educational expenses and higher opportunity cost of mothers time tends to
increase the relative costs of children. The perceived benefit of children
decreases due to non-participation of children in the labor force and
availability of other alternatives of child services for old age security
At the onset of economic development the fertility increasing effect
(resulted from reduced breastfeeding) surpasses the fertility-decreasing effect
while in the later phase of economic development fertility-decreasing effect
surpasses the fertility increasing effect (Repeto, 1979). Therefore, we can
assume that with increasing income, fertility first goes up to a certain level
and then decreases.
Theory of proximate determinants assumes that modernization variables
have direct impact on exposure and natural fertility factors, which in turn
determine the supply and demand of children. Deliberate fertility control
factors are determined by the supply and demand of children.
Modernization variables have significant influence on exposure and natural
fertility factors. Thus, modernization variables have ultimate impact on
fertility. Hypothesis II: Fertility varies negatively with income, womens
labor force participation, female education, and family planning programs
and positively with infant mortality.
Causes of Mortality Decline
As a country approaches modernity, mortality starts to decline. Better
life conditions and breakthrough in health care technologies are the major
variables that affect mortality conditions. Highly effective vaccines have
been invented to prevent and cure for various communicable diseases like
tetanus, measles, diphtheria, polio, tuberculosis, and so forth. Availability of
antibiotics has reduced the risk of mortality form pneumonia. Similarly,
diarrhea can be treated effectively through Rehydration-Oral-Therapy.
Effective insecticides have been invented to control insects like mosquito
which spreads malaria (World Health Organization, 1999).
Mortality started to decline after World War I in many developing
countries. In some countries it started only after World War II (Lopez,
2000). While in all of the developed countries mortality decline began in
early 19th century (Wrigley, 1969). For many developing countries, average
life expectancy at birth did not exceed 30 years in the early 20th century.
The average life expectancy hit 55 years by 1970 (Lopez, 2000). The
average life expectancy further increased from 55 years to 65 years in
between 1970 and 1997 (World Bank 2000).
A significant cross country variation in the pattern of mortality decline
in developing countries has been noticed. The variation also occurs within
a country. Rural people as opposed to urban people and poor as opposed
to rich are more exposed to life-threatening diseases. With the few
exceptions in some South Asian countries, females have longer life
expectancy than males (World Bank, 2000). Findings of past research
indicate that mortality is largely the function of public health. Other
variables like female literacy and external debt have significant impact on
Public Health Measures
Nature of Diseases. In developing countries, infectious diseases have
major role in shortening life expectancies below the western standard. In
more developed countries, deaths from cardio-vascular diseases outnumbers
the deaths from communicable diseases. The situation is reversed in case
of less developed countries (World Health Organization, 1999).
Understanding of increased life expectancy in developing countries depends
on the nature of diseases that caused deaths. The communicable diseases
which are prevalent in developing countries are within the control of
human made technologies. Eradication of such diseases in developing
countries plays a major role in increasing life expectancy (Lopez, 2000).
Over the past 15 years approximately 10 million people have been cured
for Leprosy. The number of episodes from Guinea Worm disease declined
by 90% over past 10 years. More than 100 countries are free from Guinea
Worm disease and Neo-natal Tetanus. Neo-natal Tetanus used to kill more
than 300,000 infants and 40,000 mothers every year (World Health
Organization, 1999). The research findings suggest that the public health
programs aimed at controlling infectious diseases alone can reduce mortality
to the level at which life expectancy can reach 50-55 years within a
decade. The analysis of 10 health programs in developing countries
^confirms that child mortality can be reduced to a half within a decade
through extensive use of health technologies such as vaccination, antibiotics,
and Rehydration-Oral-Therapy (Lena and London, 1993).
Causes of Diseases. Lack of access to sanitation and safe drinking water
is the main cause for prevalence of infectious diseases in developing
countries. More than 400 million people in middle-income countries and
about 1.4 billion people_in low-income countries did not have access to
sanitation in 1996 (World Bank, 2000).
In developing countries as a whole, about 1.5 billion people did not
have access to potable water in 1997. In Sub-Saharan Africa, more than
half of the entire population lacked access to safe drinking water. However,
some countries like Pakistan, Cote dIvoire, Cambodia etc. gained access to
safe drinking water for more people in between 1986 and 1996 (World
Bank, 2000). Lack of access to sanitation and clean drinking water makes
individuals especially children under 5-years-old and infant more vulnerable
to infectious diseases like Diarrhea (Synder and Merson, 1982).
The empirical evidence suggest that in less developed countries as a
whole, 1000 million incidence of diarrhea are reported every year out of
which 3.3 millions die. Diarrhea is prevalent among infants and children
under 5 years old (Bern et al., 1992). Despite the prevalence of Diarrheal
mortality in many less developed countries, the evidence from various
studies suggest that the death due to diarrhoea has been reduced in many
parts of the world over the past decades (World Health Organization,
Public Expenditure on Health. Launching of widespread public health
programs targeted at the prevention and cure of infectious diseases such as
tuberculosis, malaria, smallpox, cholera, measles, and yellow fever since
late 1940s has been the major cause of mortality decline in many low-
income countries (Gage and OConnor, 1994).
Thus, increase in life expectancy is heavily based on the public
expenditure on health care. Survival chance of infants and children can be
increased with relatively little investment in public health (Hojman, 1996).
Analysis of data from 83 countries reveals that world child mortality
declined by 16% in between 1988 and 1999. Of this decline approximately
99% occurred in low and middle-income countries. The chance of survival
has been increased in almost all parts of the world. Chance of child
survival of newborn up to the age of 5 is 85% in Africa, 90% in
Southeast Asia, and 96% in Western Pacific Regions, Latin America, and
the Eastern Mediterranean region. Achievement in public health program
implementation is the main reason for this remarkable improvement in child
survival rate (Lopez, 2000).
Income has tremendous impact on the inequalities of death across socio-
economic classes within a country. Children in better-off families have
relatively better chance of survival than do the children in poor families.
However, the gap of this inequality varies across countries. In Brazil, under
5 mortality rate for the richest 20% is 90 times less than that of the
poorest 20% while in Pakistan and Ghana under 5 mortality for richest
20% is only 1.1-1.2 times less than that of poorest 20% (World Bank,
The gap of mortality between poorest and richest within a country is
because of less access to health care facilities to the poor. For instance,
the data from ten less developing countries suggest that in between 1992
and 1997, 78% of births among poorest 20% were not attended by health
care professionals as opposed to only 24% among the richest 20%.
Similarly, in 1990, in Peru, only 35% of urban poor and 20% of rural
poor received treatment when sick as opposed to 57% of urban rich and
39% or rural rich (World Bank, 2000).
Another point of inequality in mortality between poor and rich is in the
nature of disease. The Poor are more likely to be infected by
communicable diseases than the rich. For instance, in the world as a
whole, of all the deaths in 1990, 59% were caused by communicable
diseases among the bottom 20% compared to only 8% for the highest 20%
(World Bank, 2000).
Fetal growth and birth weight are strongly affected by mothers
nutritional status during pregnancy. Food supplement intake during
pregnancy reduces the chance of neo-natal mortality and low-birth-weight
infants (Lena and London, 1993). Mothers nutritional status is highly
dependent on family income (Nag, 1983). After birth, breast-feeding is the
optimum source of nutrition for the first 4-6 months. After 6 months, the
child needs external food supplements although breast-feeding supplies
three-fourths of necessary protein need. Eventually, the child will shift to
regular adult diet. In developing countries this is the critical stage for
infants. First, because of food scarcity, the infant does not get enough
calories and second, the entire process of food preparation, handling, and
feeding is unsafe. In such conditions a child is more likely to get infected
by diseases. Adequacy of food in the family is mainly dependent on
family income. Thus children in high-income family are subjected to lesser
risk of mortality (Gage and OConnor, 1994).
Empirical evidence shows that the probability of using modem health
care facilities is higher among women with some years of schooling. A
few studies have emphasized that education enhances women's
communication skills which in turn improve the health conditions. In
rural Bangladesh, women with some level of education listened to health
programs in the radio that helped these women to provide better care to
their infants and children (Lindenbaum et.al 1989). Similarly, evidence from
Kenya suggests that schooling helps mothers to understand the causes of
diarrheal disease and follow the printed instructions to prepare oral
rehydration therapy solutions (Eismon and Patel 1987).
These studies evidence that education increases mothers access to public
information about hygiene and nutrition which in turn increases the chance
of child survival.
External Debt Situation
External debt explains 50% variation in mortality level. After 1982, the
external debt crisis prevented the governments of less developed countries
from spending more on expensive modernized hospital and forced these
governments to focus on primary health care facilities. Primary health care
is more effective than modernized hospitals in reducing child mortality.
Switching of resources to primary health care after 1982 increased the
probabilities of child survival which influenced the overall rate of mortality
(Hojman, 1996). Another study suggests that high external debt of
developing countries has negatively affected immunization coverage,
adequate nutrition, balanced urbanization, prevalence of health attendants,
and economic growth which in turn have inverse relationship with infant
mortality (Brudshaw et al, 1993). Similarly, Logie (1992) found positive
impact of external debt on infant mortality.
Based on the available literature, public expenditure on health, female
literacy, household income, and external debt are the major determinant of
mortality. Hypothesis III: Mortality varies inversely with the variation in
public expenditure on health, external debt, income, and positively with the
variation in female illiteracy.
Population Components and
Age Structure Dynamics
An age structure corresponds to the number of people of a given age in
a population. Interaction between fertility and mortality determines the age
structure of a closed population. A population is considered to be old or
young depending upon the proportion of people at different ages (Keyfitz
and Flieger, 1971). Developing countries have younger age structure than
have the developed countries. In 1998, in developing countries as a whole,
approximately 34% of the population were under 15 years of age as
compared to only 20% in developed countries as a whole (Population
Reference Bureau, 1999). In this section I will review literature on how
fertility and mortality are interrelated with age structure of a population.
Impacts of Mortality Changes
In all societies, the youngest and the oldest age groups are most
susceptible to death and in modem societies where maternal mortality rates
are fairly low, males are more likely than females to die at any given age
(Woods, 1986). When mortality levels change, all ages tend to be affected,
even though some more than others. Thus, improved health conditions in
society lower death rates at all ages. However, the youngest age group
benefits the most from the declining mortality pattern (Caldwell, 1986).
Therefore, the decrease in infant or child mortality can substantially
increase the number of young people. Examination of the population data
of 11 Latin American countries reveals that of 27 million people alive in
all the eleven countries in the 1960s who would not have been alive if
there had not been a mortality decline since the 1930s, sixteen millions
(59%) were under age 15 (Arriaga, 1970). In relative terms a lowering of
mortality especially, infant or child mortality in Latin America noticeably
raised the proportion of people at the young ages, slightly elevated the
proportion at old ages, and lowered the proportion at middle ages (14-64).
However, in absolute terms the number of people at all ages increased
(Arriaga, 1970). Thus, infant or child mortality is the biggest factor that is
responsible in creating younger age structure.
Fertility and Age Structure
Effects of fertility in a population stay age after age. If the birth rate
were to drop suddenly in one year, then as those birth cohorts get older,
there will always be fewer of them than there are people just older and
just younger. If fertility goes up, then there will be more people in each
younger age group. In these situations; rising and declining fertility have
had strong influences on the age structure of the United States
(Shrivastava, 1983). In 1975, in the United states, only 25% of the
population was under age 15 and only 10% of the population was above
65. The majority of the people were in their 20s. This was the impact of
past fertility trends. High concentration of people in their 20s was the
impact of baby boom of 1940s and 1950s. Fewer people under 15-age
category were due to declining fertility since 1960s (Nam and Gustavus,
The number of births increases if the number of reproductive age women
increases. During 1969 and 1970, the birth rates increased in the United
States when baby-boom babies reached their childbearing age (Coale, 1987).
In Taiwan, in between 1983 and 1991, decreased proportion of childbearing
age women followed by increased proportion of older women caused
fertility to decline (Freedman et al, 1994). Decline in fertility in high
fertility country decreases the proportion of children but increases the
proportion of childbearing age women. Because of higher proportion of
women in childbearing age the number of births will increase. Taiwan
achieved below replacement level fertility after 1983. However, the fertility
data from 1983 to 1991 indicate that crude birth rates still exceed crude
death rates regardless of the below replacement level fertility (Herrink,
1994). This is the impact of high fertility in the past. Because of higher
fertility in the past, the proportion of women in childbearing age was high
and so the number of births. Between 1956 and 1983, the proportion of
women aged 20-34 increased by 2.3% (Griffith, 1994).
In developing countries, ongoing high fertility and declining infant or
child mortality have created a very young age structure. In Sub-Saharan
Africa while infant or child mortality has started to decline, fertility rates
are at the previous high level. Based on 1990 estimates, Sub-Saharan
Africa still has highest total fertility rate of 6.0 children per woman. In the
1950s TFR was 6.6 children. These persisting high birth rates in Sub-
Saharan Africa have created a very young structure. In 1997, 45% of the
population was under age 15 (World Bank, 2000).
Higher proportion of people in childbearing age increases the dependency
ratio' In the United States, the youth dependency ratio decreased
continuously from 1900 until 1940-1950. High fertility rates of 1940-1950
increased the youth dependency ratio that continued until 1960-1970 (Coale,
1987). The younger age structure causes population momentum. For
Bangladesh it is predicted that 82 percent of the population growth over
the next 50 years will be caused by momentum. In other words even if
every Bangladeshi couple were to limit their fertility to two children, the
countrys population would still grow by 80 millions by 2050 simply
because of the young population structure (Cleland and Streatfield, 1992).
Thus, declining infant mortality and high fertility together exert a
significant pressure to increase the size of younger cohorts. Hypothesie-IV:
The proportion of younger age people varies inversely with the variation in
infant mortality rate and positively with the variation in fertility rate.
Relative wage of younger males deteriorates in relation to that of older
males with the increase in supply of younger cohorts. Decline in wage of
younger cohorts reduces their earning potential in relation to their
aspirations as shaped in their parental households. The decline in cohorts
wage in relation to those of older cohorts, especially, the parents may
encourage fertility reduction among the younger cohorts through delayed
marriage and increased female labor force participation as they struggle to
maintain their economic conditions (Easterlin, 1993). Studies of Korenman
and Neumark, 1997 supports this theory of Easterlin (1993) that increased
size of younger cohorts exerts a significant pressure on the relative wage
of younger males and wage decline causes fertility reduction .
From the preceding discussion we know that declines in infant mortality
tend to increase the size of younger cohorts. Bigger size of younger
cohorts exerts pressure that tend to reduce the wage of younger male.
Wage deteriorates because the supply of younger cohorts (the workers)
exceeds the total number of available jobs. Thus, infant mortality of time
(t) should have significant impact on unemployment conditions of time t-
16. Hypothesis-V: Unemployment rate at time (t) varies inversely with the
variation in infant mortality rate of time (t-16)*.
Lack of access to a continuous source of stable income that is heightened
by prevailing higher unemployment rate leads younger cohorts to reduce
fertility. Thus, fertility should decline with the increase in rate of
unemployment. Hypothesis VI: Fertility varies inversely with the
*(t-16) is 16 years before the given year, if the given year is 1979, t-16 is 1963
Nature of Data Set
I chose CD-Rom database of World Bank because it provides estimates
of socioeconomic indicators for 207 countries covering the period 1960
through 1998. The dataset consists of data gathered from various sources;
World Health Organization, International Monitory Fund, World Bank itself,
and the United Nations (World Bank, 2000).
The dataset consists of aggregate data of all countries, economic regions,
and geographic regions. Economic regions are classified into four regions,
low income, lower middle income, upper middle income, and high-income
countries. The countries having annual income less than U.S.$ 675 are
categorized as low income countries, between U.S.$ 675 and below 2700
are categorized as lower middle income countries, between U.S.$2700 and
U.S.$ 8000 are categorized as upper middle income countries, and over
U.S.$ 8000 are categorized as high income countries (World Bank, 2000).
In addition to each economic region, it also gives aggregates of low and
middle-income countries as a single unit. High-income countries are further
classified into OECD (Organization for Economic Cooperation and
Development) member countries and non-OECD countries. Geographic
regions are classified into 7 regions namely East Asia & Pacific, Europe &
Central Asia, Latin America & Caribbean, Middle East & North Africa,
South Asia, sub-Saharan Africa, and North America. Geographic
classification of country is based on geographical territory.
The World Bank collects data from low and middle-income countries as
a part of its funding projects to provide economic and technical assistance
in those countries (World Bank, 2000). The data set consists of six
domains: world overview, people, environment, economy, states and market,
and global links. World overview includes indicators prevalence of child
malnutrition, life expectancy, child mortality, adult illiteracy, urban
population, access to sanitation, and growth of consumption per capita. The
peoples domain includes data on population and labor force structure,
education, health, and income distribution. The environment domain consists
of data on land use and agricultural productivity, water use, deforestation,
and energy use and emissions. The economy domain includes data on
structure of GDP (Gross Domestic Product), structure of demand,
government finances, balance of payments, and current accounts. The
domain states and market include data on private sector finance,
government finance, power and transportation, and communication. The
domain global trade has data on export, import, aid and financial flows
(World Bank, 2000).
Variables and Indicators
1. Total Fertility Rate (TFR): Total fertility rate is the number of
children who would be bom to a woman if she were to live to the
end of her childbearing years and bear children at each age in
accordance with prevailing age-specific fertility rates (World Bank,
2000, p. 278). Total fertility rate is used as an indicator of fertility.
Total fertility is a dependent variable in Hypotheses -1, Hypothesis II
and Hypothesis VI.. It is an independent variable in Hypothesis V.
2. Infant Mortality Rate (IMR): Infant mortality rate is the number of
infants who die before reaching one year of age, expressed per 1,000
live births in a given year (World Bank, 2000, p. 278). Infants and
children face the greatest risk of death if the health conditions are
poor. On the other hand infants and children are the greatest
beneficiaries of improved health conditions. Thus, improvement in infant
mortality is reflected by increased life expectancy. Infant mortality is
one of the well known indicator of overall mortality level of a country.
Thus, improvements in infant mortality rate (IMR) represent the
improvement in overall mortality. Infant mortality is an independent
variable in Hypothesis II, Hypothesis IV, and Hypothesis V. It is a
dependent variable in Hypothesis HI.
3. Agriculture value added % share of GDP (AGGDP): Agriculture
value added % share of GDP measures the output of the agriculture
sector. The Agriculture includes the production from forestry, hunting,
fishing, crop cultivation, and livestock production (World Bank, 2000,
p. 280). Share of agriculture in GDP shows the level of
industrialization. Shift from agricultural job to industrial job takes place
when the society moves from agricultural production system to
industrial production system. Higher share of agriculture in GDP
represents the domination of agricultural jobs. Agriculture value added
is an independent variable in Hypothesis -1.
4. GNP per capita (GNPpc): GNP is the sum of value added by all
resident producers plus any taxes (less subsidies) that are not included
in the valuation of output plus net receipts of primary income
(employee compensation and property income) from non-resident
sources. GNP per capita is gross national product divided by midyear
population (World Bank, 2000, p. 274). Gross national product per
capita can be used as an indicator of average national income per
capita because it consists of income earned by citizens of the country
including wages and profits earned by them outside the country. GNPpc
is an independent variable in Hypothesis II and Hypothesis III.
5. Female illiteracy rate (ILLIF): Female illiteracy rate is the proportion
of female adults aged 15 and above who cannot, with understanding,
read and write a short simple statement of their everyday life (World
Bank, 2000, p. 275). Female illiteracy rate is an independent variable in
6. Female as a percentage of labor force (FELAB): Female as a
percentage of labor force shows the extent to which women are active
in the labor force (World Bank, 2000, p. 275). Female as a percentage
of labor force is an independent variable in Hypothesis IV.
7. Percentage of urban population (URBPOP): Urban population is the
share of population living in areas defined as urban in each country
(World Bank, 2000, p. 275). Urbanization is a well-known concomitant
of development. In low-income countries, the great majority of people
live in rural areas: in middle and high income countries, most live in
town and cities. Percentage of urban population is an independent
variable in Hypothesis -1.
8. Contraceptive, prevalence rate (CONTRA): Contraceptive prevalence rate
is the percentage of women age 15-49 who are practicing, or whose
sexual partners are practicing any form of contraception (World Bank,
2000, 278). Family planning programs do two things; they make one or
more forms of contraceptive more widely and cheaply available and
they undertake information and propaganda activities to urge people to
use them. Thus, contraceptive prevalence can be treated as an indicator
of family planning program. This variable is treated as an independent
variable in Hypothesis II
9. Public health expenditure (PEH): Public expenditure on health consists
of recurrent and capital spending from government (central and local)
budgets, external borrowings and grants (including donations from
international agencies and non-governmental organizations), and social
health insurance funds, organizations), and social (or compulsory) health
insurance funds (World Bank, 2000, p. 277). The development and
dissemination of simple cures for widespread health problems must
accompany the spread of medical services into the rural areas.
Governments in the poorest countries typically spend only a few dollars
per capita. The inadequacy becomes clearer when we compare public
health expenditure of developed countries with that of developing
countries. Public health expenditure is treated as an independent variable
in Hypothesis III.
10. Un-enrolled primary percentage (UNENEOLL): Un-enrolled primary
percentage are the number of school-age children not enrolled in
primary school as a share of all primary school-age children (World
Bank, 2000). Childrens role in producing family income is largely the
function of childrens education. We can assume that a society that
emphasize child education will have lower percentage of children in the
labor force and consequently, lower economic contribution of children.
This inverse relationship between child education and child labor can be
explained based on the fact that if child education is emphasized then
children are sent to school. If children are sent to school they are less
likely to be put to productive activities. Thus, Percentage of un-enrolled
primary school age children can be used as an indicator of childrens
role in producing family income. This is an independent variable in
11. Unemployment (UNEMP): Unemployment refers to the share of the
labor force that is without work but available for and seeking
employment (World Bank, 2000). This is a dependent variable in
Hypothesis V and an independent variable in Hypothesis VI.
12. Female secondary school enrollment (FESECED): Female secondary
school enrollment is the ratio of girls of all ages enrolled in secondary
school to the total number of girls in the age group (World Bank,
2000). Percentage of female in secondary school in time (t) will effect
fertility after some time. Female secondary school enrollment can be
used as an indicator of female education. The countries having higher
female secondary enrollment at time (t) must have lower rate of fertility
after some time. In order to allow for time lags to capture the fact that
female students become mothers after a number of years, the data of
female secondary school enrollment is based on year 1990 and data on
fertility is based on the year 1997. the inclusion of this variable aims
to capture possible threshold effects of education on fertility.
13. Total population between the ages 0 to 14 (CHILDPROP): Total
population between the ages 0-14 refers to all residents regardless of
legal status or citizenship except for refugees not permanently settled in
the country of asylum, who are generally considered part of the
population of the country of origin (World Bank, 2000). Proportion of
children aged 0-14 was derived by dividing the population of 0-14 age
group of each country by total population of the respective countries.
Total population follows the same definition as the total population
between ages 0-14.
14. Total external debt (EXDEBT): Total external debt is the sum of
public, publicly guaranteed, and private non-guaranteed long term, use
of IMF credit, and short-term debt (World Bank, 2000, p. 278 ).
Many developing countries are unable to generate sufficient domestic
saving to finance economic growth. These countries get resources from
other countries in the form of foreign loan that requires repayment by
the recipient country.
15. TV: Television sets are the estimated number of television sets in use,
per 1,000 people (World Bank, 2000, p. 278 ). Increasing aspiration of
parents brings changes in consumption habits. In low income countries,
poor families are struggling to survive. In such families, the aspiration
to acquire consumer goods is out of question. However, in better-off
families, there is a growing aspiration to lead an advanced life style.
Consequently, their consumption patterns are changing from basic
necessities of survival to consumer goods like radio, TV, computer,
telephone etc. The countries that have higher percentage of people using
consumer goods can be interpreted as the country of people having
higher aspirations. Such aspiration leads parents to trade-off between the
number of children and number of consumer goods. For the purpose of
this study, prevalence of TV is taken as an indicator of parental
Method of Analysis
Aggregate data of low and middle-income countries for the year 1997 is
used to estimate the impact of explanatory variables in fertility and
mortality. The sample consists of 156 countries out of which 63 are low-
income countries, 57 are lower middle-income countries, and 36 are upper
middle-income countries. It was assumed that different countries represent
the different level of development. Aggregate data of low and middle-
income countries over time were analyzed to test the impact of fertility
and mortality in the proportion of children age 0-14.
In order to test the impact of fertility and infant mortality in proportion
of children, the data of low and middle income countries as a group was
analyzed over a period of 14 years from 1983-1997.
Likewise, data from Brazil over time were analyzed to measure the
impact of infant mortality on unemployment and impact of unemployment
on fertility. Since the effect of infant mortality of today doesnt have an
impact in unemployment until next 15 years (tentatively), infant mortality
data from 1960-1978 and unemployment data since 1979-1998 were taken
for the regression analysis. It was assumed that infant mortality rate of
1960 will have effect in unemployment condition of 1975, infant mortality
rate of 1965 will have an impact in unemployment rate of 1980, and so
on. Brazils data were taken as a sample because of its completeness.
Regression analysis was performed to test the impact of the set of
independent variables in dependent variable.
There are two main types of problems. The first problem is posed by
poor quality of data. The data from less developed countries depend
entirely for estimations based on religious records, undeveloped censuses,
and special sample surveys (World Bank, 2000). Even the demography of
the statistically advanced industrialized countries has certain major
limitations to produce reliable data (Donaldson, 1991).
The second problem is inherent in methodology itself. The risk of
ecological fallacy restricts the value of conclusions drawn with the aid of
country-based causal models. Even though there is a strong positive
association between employment in agriculture sector and marital fertility, it
does not mean that farmers wives produce large families; rather, it means
that large families are common in countries where agriculture is an
Another method inherent problem is based on the assumptions necessary
for regression analysis. First, the regression analysis requires that all the
variables should be measured at interval/ratio level. There are certain
factors like socio-cultural beliefs which have significant impact in fertility
but are measured at nominal or ordinal levels. Another problem is of co-
linearity between independent variables that undermines the significance of
one or more independent variables. To overcome this problem, I ran the
second test by dropping one of the highly correlated variable.
The theory of demographic transition is macro since the forces that
affect both mortality and fertility are defined outside of the family units
and the theory implies that the key variables should be defined at a higher
level of aggregation (Keyfitz and Flieger, 1971). The detail is lost when
data are aggregated by geographical or economic regions. Furthermore, the
unit of observation is geographical and economic regions or countries
which are different from the micro level unit of observation (individuals
and families) upon which the theory is based. Another limitation of macro
level data is that it ignores socio-cultural value and belief systems.
This study deals with the causes and consequences of fertility and
mortality variations. The causes of fertility variations are explained in terms
of two popular theories: theory of value and cost of children and theory of
proximate determinants. Percentage of urban population, share of agriculture
in GDP, percentage of un-enrolled school age children, and percentage of
people using TV are chosen as indicators of the variables of value and
cost of children. Contraceptive prevalence, infant mortality rate, female
secondary school enrollment, GNPpc, and womens labor force participation
are chosen as indicators of proximate determinants variables.
External debt, public health expenditure, percentage of female illiteracy,
and GNPpc are the indicators for the variables affecting mortality change.
Consequences of fertility and mortality change are reflected in age-structure
and unemployment conditions. Following paragraphs summarize the findings
of this study.
Findings on Fertility Hypotheses
Theory of Value
and Cost of Children
Theory of value and cost of children is based on the assumption that
value and cost of children that determine the fertility level is different in
modem and traditional societies. The difference in value and cost of
children in modem and traditional societies is due to different role of
children in producing family income, difference in degree of urbanization,
difference in the degree of industrialization, and difference in the degree of
aspiration level for higher living. These variables that determine value and
cost of children, have ultimate impact in determining fertility level. Thus,
the Hypothesis I as developed in chapter 2 states Fertility varies
inversely with the variation in urbanization and parents aspiration and
positively with the variation in childrens role in producing family income
and employment in agricultural sector. As shown in table 4.1, the
independent variables together account 72.2% of total variation in fertility.
Use of TV and percentage of un-enrolled primary school age children have
significant impact while share of agriculture in GDP and percentage of
urban population have insignificant impact on fertility. As shown in table
4.7, percentage of un-enrolled primary school age children is highly
correlated with share of agriculture in GDP. This high correlation between
share of agriculture in GDP and percentage of un-enrolled primary age
school children could have undermined the significance of share of
agriculture in GDP and percentage of urban population in determining
fertility. Thus, the second test was run by dropping percentage of un-
enrolled primary school age children. In the second test, prevalence of TV,
share of agriculture in GDP, and urban population explained 46.8% of total
variation in fertility.
Before controlling un-enrolled primary school age children, beta slopes
indicate that fertility declines by .092 standardized unit with a standardized
unit increase in percentage of urban population and by .306 standardized
units with a standardized unit increase in prevalence of TV. A standardized
unit increase in the percentage of un-enrolled primary school age children
causes .578 standardized unit increase in fertility and a standardized unit
increase in the percentage share of agriculture in GDP causes fertility
increase by .039 standardized unit. Beta slopes after controlling percentage
of un-enrolled primary school age children suggest that fertility declines by
.262 standardized units with one standardized unit increase in urban
population and by .369 standardized units with one standardized unit
increase in prevalence of TV. Fertility increases by .204 standardized units
with one standardized unit increase in share of agriculture in GDP.
Value of F 52.53 (Before controlling un-enrolled primary school age
children) and 52.476 (after controlling un-enrolled primary school age
children) are significant at 1% level. This indicates that the magnitude of
the impact of percentage of urban population, share of agriculture in GDP,
percentage of population using TV, and percentage of un-enrolled primary
school age children have diverse impact in different countries in
determining the value and cost of children.
Summary of Regression Analysis
for Value and Cost of Children
Variables Predicting Fertility
Variables R2 P t Partial F
All Variables Constant (3.660)** .722 8.218 52.53
Urban Population -.092 -.1.072 -.118
Share of agriculture in GDP .039 .494 .055
Prevalence of TV -.306** -3.936 l O
Un-enrolled primary school age children .578** 8.400 .682
Controlling Highly Correlated Variables Constant (4.900)** .468 11.349 52.476
Share of agriculture in GDP .204* 2.850 .299
Percentage of Urban Population -.262** -3.746 -.380
Prevalence of TV -.369** -5.498 -.517
N = 86 **p<.01, *p<.05
Theory of Proximate Determinants
Theory of proximate determinants is based on the assumption that
womens labor force participation, womens education, income, infant
mortality rate, and family planning programs affect exposure factors, natural
fertility factors, and deliberate fertility control factors which in turn affect
overall fertility level. Based on this assumption Hypothesis II as
developed in Chapter 2 is stated as Fertility varies negatively with
income, womens labor force participation, female education, and family
planning programs and positively with infant mortality. As shown in table
4.2, these independent variables together suggest 98.4% of total variation in
fertility. Percentage of female in the labor force, contraceptive prevalence,
and infant mortality have significant impact on fertility while female
secondary school enrollment and GNPpc have insignificant impact on
fertility. As shown in Table 4.7, Female labor force participation is highly
correlated with female secondary school enrollment. The second test was
run by controlling percentage of the females in the labor force. In the
second test, contraceptive prevalence, female secondary education, infant
mortality rate, and GNPpc explained 95.1% of total variation in fertility.
Beta slopes before controlling female labor force participation suggest
that fertility declines by .234 standardized units with a standardized unit
increase in percentage of female in the labor force, .693 standardized units
with a standardized unit increase in contraceptives prevalence, .011
standardized units with a standardized unit increase in the percentage of
female secondary enrollment, and .003 standardized units with a
standardized unit increase in GNPpc. A standardized unit increase in infant
mortality causes fertility to increase by .247 standardized units. Beta slopes
after controlling highly correlated variables suggest that fertility declines by
.349 standardized units with a standardized unit increase in female
secondary school enrollment, by .289 standardized units with a standardized
unit increase in GNPpc, and .639 units with a standardized unit increase in
contraceptive prevalence. Fertility increases by .327 standardized units with
a standardized unit increase in infant mortality.
Summary of Regression Analysis for
Variables Predicting Fertility
R2 P t Partial F
All Variables .984 50.02
Constant (8.358)** 10.174
Female labor participation -.234* -2.932 -.326
Contraceptive prevalence -.693** -10.136 -.785
Female secondary school -.011 -.098 -.049
GNPpc -.003 -.035 -.017
Infant mortality rate .247* 3.335 .352
Controlling Highly Correlated Variables .951 102.205
Constant (6.301)** 24.951
Contraceptive prevalence -.639 -7.517 -.684
Female secondary school -.349 -6.687 -.419
GNPpc -.289 4.639 -.335
Infant mortality rate .327 5.746 .421
N = 85 **p<.01 *p<.05
Thus from Hypothesis -1, use of TV and percentage of un-enrolled
primary school age children contribute more in determining fertility.
From Hypothesis 2, contraceptive prevalence, infant mortality, and
percentage of female in the labor force contribute more in determining
fertility. Value of R square from table 4.1 and table 4.2 suggest that
proximate determinant variables explain higher percentage of variation in
fertility than do value and cost of children variables.
Findings on Mortality Hypothesis
Public health expenditure, female illiteracy, income, and external debt are
known to be the major determinants that determine mortality. Based on this
assumption, Hypothesis III as developed in Chapter 2 is stated as
Mortality varies inversely with the variation in public expenditure on
health, external debt, income, and positively with the variation in female
illiteracy. As shown in table 4.3, these independent variables together
account for 73.4 % of total variation in infant mortality. Of the four
independent variables, two variables female illiteracy rate and GNPpc have
significant impact. External debt and public expenditure on health are found
to have insignificant impact on infant mortality. Beta slopes suggest that
infant mortality increases by .698 standardized units with a standardized
unit increase in female illiteracy. Infant mortality decreases by .193
standardized unit with a standardized unit increase in GNPpc, by .084
standardized units with a standardized unit increase in public expenditure
on health, and by .040 standardized unit with a standardized unit increase
in external debt. As shown in Table 4.7, GNPps is highly correlated with
female illiteracy and external debt is highly correlated with income. The
impact of external debt and public health expenditure was found to be
significant when GNPpc and female illiteracy were controlled. After
controlling highly correlated variables, GNPpc and public health expenditure
explain 33.2% of variation on infant mortality. Beta slopes suggest that
fertility declines by .539 standardized units with a standardized unit
increase in public expenditure on health and by .214 standardized units
with a unit standardized unit increase in external debt.
Value of F 56.382 (before controlling highly correlated variables) and
25.556 (after controlling highly correlated variables) which are significant at
1% and 5% level respectively suggests that public expenditure on health,
external debt, GNPpc, and female illiteracy have varied impact on infant
mortality across countries.
Summary of Regression Analysis
for Variables Predicting Mortality
P t Partial F
All Variables .734 56.382
Constant (33.548)** 4.452
Public exp. on health -.084 -1.144 -.140
Female illiteracy .698** 8.991 .732
GNPpc -.193* -2.169 -.301
External debt -.040 -.049 -.076
After Controlling Highly Correlated Variables .332 25.556
Constant (87.035)** 14.442
Public Expenditure on Health -.539** -6.684 -.550
External Debt -.214** -2.651 -.253
N = 83 **p<.01 *p<.05
Findings on the Consequences of
Fertility and Mortality
Consequences of mortality and fertility changes are reflected in the age
structure of a population. Age structure tends to be younger with declining
infant mortality and persistent high fertility. Based on this assumption,
Hypothesis IV as developed in Chapter 2 is stated as The proportion of
younger age people varies inversely with the variation in infant mortality
rate and positively with the variation in fertility rate As shown in table
4.4, these independent variables together suggests 93.1% of total variation
in the percentage of children aged 0-14. Beta slopes indicate that
percentage of children age 0-14 increases by .576 standardized units with a
standardized unit increase in fertility. Percentage of children age 0-14
decreases by .681 standardized units with a standardized unit increase in
infant mortality. Both independent variables have significant impact in the
proportion of children age 0-14. Value of F (113.289) suggests that the
magnitude of effect of fertility rate and infant mortality rate in the
proportion of children age 0-14 varies over time.
Summary of Regression Analysis
for Variables Predicting
Proportionof Children Age 0-14
P t Partial F
Constant (28.101)** 34.997
FERT .576** 5.798 .601
IMR -.681** -8.074 -.752
Number of years = 14 **p<.01 *p<.05
Younger age structure causes many impediments in a social system.
Unemployment is one of them. Easterlins hypothesis states that declining
infant mortality causes the size of younger age structure to increase and
increased size of younger cohorts tend to deteriorate the wage of younger
cohorts. Based on this theory, we can say that infant mortality affects
employment conditions in a country. Wage deterioration is caused when
demand for job exceeds the supply of job. Unemployment rate indicates
the demand and supply of job. Since declining infant mortality tends to
increase the size of younger cohorts we may conclude, that declining infant
mortality rate today will causes higher unemployment after some time say
after 16 years. Based on this assumption the Hypothesis V as developed
in Chapter 2 is stated as Unemployment rate at time (t) varies inversely
with the variation in infant mortality rate of time (t-16) As shown in
table in 4.5, infant mortality rate explains 48.6% of total variation in
unemployment rate. Beta slope suggests that unemployment increases as the
infant mortality declines (-.697).
Summary of Regression Analysis
for infant mortality rate Predicting
Unemployment Over Time
Variable R" P t F
Constant (11.619)** 28.685
Infant mortality -.697** 6.558
Number of years =16 **p<.01 *p<.05
Easterlines hypothesis also states that wage deterioration causes
demographic adjustments among the younger cohorts. This leads to the
assumption that unemployment rate affects fertility rate. Thus, Hypothesis-
VI as developed in Chapter 2 is stated as Fertility varies inversely with
the unemployment rate. As shown in table 4.6, unemployment rate
accounts for 78.4% variation in fertility in cross country model.
Unexpectedly, this relationship is positive. Beta slope suggests that fertility
increases with the increase in unemployment (.886).
The association was found to be negative when the regression was run
within a country (Brazil) over time 1965-1998. In this case unemployment
accounts for only 38.4% variation in fertility. Beta slope suggests that
fertility declines with the increase in unemployment -.620.
Summary of Regression Analysis
for Unemployment Predicting Fertility
R" P t F
Cross Country .784 158.65
Constant (10.536)** 8.235
Within Country .384 .886** 5.412
Constant (15.689)** 10.528 69.76
N = 83 (cross country) Number of Years = 14 (within country)
Correlation Coefficients for Predicting
Variables of Fertility and Mortality
AGGDP* UNENROLL .807*
FEL AB CONTRA .547**
FESECED* CONTRA 821**
IMR* CONTRA .459*
CONTRA* GNP .256*
N=189 **p<.01 *p<.05
This study emphasizes the causes and consequences of fertility and
mortality in a social system. Findings on the hypothesis of value and cost
of children suggest that the countries that have higher rate of children out
of school and lower percentage of people possessing TV have higher total
fertility rate. Higher rate of children out of school indicates higher
contribution of children in producing family income. In such conditions
parents value children. Likewise, in the countries where child education is
not emphasized, the educational cost is saved that reduces the overall cost
of children. TV as an indicator of consumer goods reflects the aspiration
of higher living. Lower percentage of people possessing TV reflects that
people are unable to afford or less motivated to have higher living
standard. In such conditions, the cost of children becomes insignificant.
Fertility tends to be high in societies where cost of children is negligible
and value of children is high. This finding confirms the conclusion drawn
by Caldwell, 1982 and Nag et al, 1978.
Findings on the hypothesis of proximate determinants suggest that infant
mortality and contraceptive prevalence are the major determinants of
fertility. This finding confirms the findings of Richards, 1983; Robinson,
1992; Mahadavi, 1990. It should be noted that a fall in infant mortality
rate would only decrease the demand for children. For fertility to actually
decrease, the means to curtail fertility must be readily available. It is here
that family planning services play an important role in developing
countries. In the absence of such services, a decline in the child mortality
rate will not reduce fertility but merely result in an unmet demand for
contraceptives. In the absence of decline in infant mortality rates, on the
other hand, the availability of family planning services will have little
effect on fertility. In other words, decline in infant mortality rate is a
necessary but not sufficient condition for a reduction in fertility.
Finding on mortality hypothesis suggest that female illiteracy is the
major variable that has profound impact in infant mortality of developing
countries. A literate woman can provide better health and nutritional care
to her baby than does an illiterate woman. After controlling for income
and female illiteracy, this study support the past findings (Hojman, 1994;
Brudshaw et al, 1993; and Logie, 1992) that public expenditure on health
and external debt have significant negative impact on infant mortality.
Hojman (1994) argued that increase in external debt makes a country
invest less in big hospital and consequently these countries spend in small
health facilities that reach grass root people.
A significant variation in the proportion of children age 0-14 was
explained by infant mortality and fertility this finding validates the findings
of Arriaga, 1970; Griffith, 1994.
It was found that infant mortality has significant negative impact on
unemployment conditions. The logic behind this relationship is that decrease
in infant mortality increases the size of younger cohorts seeking jobs after
15 years. As the number of jobseekers exceeds the number of available
jobs, the rate of unemployment goes up. This finding validates the finding
of Freeman, 1979; Easterlin, 1978; Korenman and Neumark, 1997 that
infant mortality increases the size of younger cohorts and size of younger
cohorts influence the relative wage.
Unemployment was found to be positively related to fertility in cross-
country model and negatively related to fertility in within country model.
The later finding supports the finding of Easterlin (1993). Explanation for
the positive relationship rests on the cross-sectional model and symmetrical
relationship between unemployment and fertility. Better-off countries have
lower unemployment rate and so is the fertility rate and vice versa. The
relationship between unemployment and fertility is such that increase in
unemployment rate lowers the likelihood of womens labor force
participation. Because women are not involved in labor force, the
opportunity cost of their time is lower. Thus fertility increases.
These findings lead to following research questions.
From the theory of value and cost of children, percentage of un-
enrolled primary school age children and percentage of people
possessing TV were found to have profound impact on fertility. From,
theory of proximate determinant, contraceptive prevalence, infant
mortality, and percentage of female in the labor force were found to
have major impact on fertility. The question is how much variation in
fertility will be explained with the combined effect of percentage of un-
enrolled primary school age children, use of TV, contraceptive
prevalence, percentage of female in the labor force and infant mortality.
Why income is found to have insignificant impact on fertility?
Why unemployment has positive impact in cross-sectional model and
negative impact in within country model?
Does decline in fertility helps in alleviating poverty of developing
This study identifies the major determinants of fertility, mortality, and
age-structure changes. Also, the study highlights the impact of fertility and
mortality over unemployment conditions. Identification of major causes and
consequences of fertility and mortality is important in formulating national
and international level policies.
At national level, identification of cause of population processes help
policymakers develop appropriate policies in order to achieve maximum
benefits from the available resources by properly allocating budget and
implementing suitable human welfare programs (such as family planning,
nutrition, vaccination, education, etc). Thus, findings of this study suggest
that in order to achieve lower level of fertility and mortality in currently
developing countries, effective health, education, and poverty alleviation
programs should be launched. Health programs should focus on reducing
infant or child mortality and family planning. Following strategies would be
helpful to make such health programs effective.
Protect infants and children from disease, and nutritional deficiency.
Provide reproductive age population with the knowledge for responsible
Provide adults with means and services to make informed choices in
family planning and reduce the risks of pregnancy and childbirth.
Promote womens literacy and health awareness programs.
Education of women and children is another key variable that affect
fertility and mortality. Thus, effective education programs for women and
children should be launched. Such programs should
emphasize universal education for children.
ensure that those who did not benefit from early education learn basic
skills through non-formal education.
promote the continuation to secondary education especially of women.
teach adults, particularly women, the basic learning and life skills to
address family, community, and livelihood needs.
Poverty is another cause of high fertility and mortality. Thus poverty
alleviation programs will help to reduce the fertility and mortality to some
extent. Such poverty alleviation programs should focus on
raising household incomes through improving agricultural production and
providing access to credit.
promoting training and vocational education to enable adults to obtain
employment or start up enterprises.
helping communities provide and manage sustainable economic and
social services to meet their basic livelihood needs.
preventing excessive child labor.
At international level, donor agencies may find this study useful to
identify the most influential variables that contribute in fertility and
mortality decline of developing countries. Identification of such variables
helps international donor agencies provide support for international public
goods, notably for research and dissemination of new means of fertility
control and vaccines for communicable diseases such as HIV/AIDS, TB,