Citation
Effects of fuel price shocks in commuting expenditures

Material Information

Title:
Effects of fuel price shocks in commuting expenditures a socio-economic analysis
Creator:
henao, Alejandro ( author )
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
1 electronic file (41 pages). : ;

Subjects

Subjects / Keywords:
Fuel -- Prices ( lcsh )
Transportation -- Analysis ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Review:
Communities are vulnerable to unexpected events, including fuel price shocks. Increases in fuel price lead to changes in people's transportation expenditures and income budgets. The impact of fuel price shocks had received minimal attention in the literature, especially, with respect to immediate individual economic impact for transportation expenditures. This study assesses data from the Denver Regional Council of Governments to examine individual work tours grouped by home transportation analysis zones (TAZ). For all home TAZs, the percent of income spent of commuting expenses increases on average by 2.1% when fuel price doubles, and as high as 11% of individual income. A transportation economic resilient (TER) rating is developed to evaluate home TAZs and measure the additional commuting expense per income when fuel price doubles. Distance, income, transit share, and carpooling share are statistically significant variables to determine TER ratings. Home TAZs that experience the lowest additional percent of income on commuting expenses, or have the best TER ratings, are located in downtown Denver. Three main factors contribute to exhibit good transportation economic resilience: i) high income, ii) good mutli-modal transportation options, and iii) proximity to employment
Thesis:
Thesis (M.S.)--University of Colorado Denver. Civil engineering
Bibliography:
Includes bibliographic references.
General Note:
Department of Civil Engineering
Statement of Responsibility:
by Alejando Henao.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
891655721 ( OCLC )
ocn891655721

Downloads

This item has the following downloads:


Full Text
EFFECTS OF FUEL PRICE SHOCKS IN COMMUTING EXPENDITURES:
A SOCIO-ECONOMIC ANALYSIS
by
ALEJANDRO HEN AO
B.S., University of Colorado Boulder, 2006
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado in partial fulfillment
of the requirements for the degree of
Master of Science
Civil Engineering
2013


This thesis for the Master of Science degree by
Alejandro Henao
has been approved for the
Civil Engineering Program
by
Wesley E. Marshall, Chair
Bruce Janson
Carolyn McAndrews
November 15, 2013


Henao, Alejandro (M.S., Civil Engineering)
Effects of Fuel Price Shocks in Commuting Expenditures: A Socio-economic Analysis
Thesis directed by Assistant Professor Wesley E. Marshall.
ABSTRACT
Communities are vulnerable to unexpected events, including fuel price shocks.
Increases in fuel price lead to changes in peoples transportation expenditures and income
budgets. The impact of fuel price shocks has received minimal attention in the literature,
especially, with respect to immediate individual economic impact for transportation
expenditures. This study assesses data from the Denver Regional Council of
Governments to examine individual work tours grouped by home transportation analysis
zones (TAZ). For all home TAZs, the percent of income spent on commuting expenses
increases on average by 2.1% when fuel price doubles, and as high as 11% of individual
income. A transportation economic resilient (TER) rating is developed to evaluate home
TAZs and measure the additional commuting expense per income when fuel price
doubles. Distance, income, transit share, and carpooling share are statistically significant
variables to determine TER ratings. Home TAZs that experience the lowest additional
percent of income on commuting expenses, or have the best TER ratings, are located in
downtown Denver. Three main factors contribute to exhibit good transportation
economic resilience: i) high income, ii) good multi-modal transportation options, and iii)
proximity to employment centers, such as downtown.
The form and content of this abstract are approved. I recommend its publication.
iii
Approved: Wesley E. Marshall


DEDICATION
I dedicate this work to my spouse, Augusta Henao, who has always supported me
by providing a foundation of love, hard work, and life balance.
To my boys, Tomas and Andres, who are my everyday inspiration.
To my parents, who instilled in me the ummeasurable value of education and have
supported me unconditionally in all aspects of my life.
IV


ACKNOWLEDGMENTS
I would like to thank Dr. Wesley Marshall for mentoring me, sharing his
knowledge, and providing funding and extensive advice for my thesis. I thank Dr. Bruce
Janson and Dr. Carey McAndrews for contributing to my education and serving on my
Thesis Committee. I also would like to thank all the members of the Active Communities
Transportation (ACT) Research Group for their contributions to my research work.
Finally, I thank the National Science Foundation for providing funding for this work
through their Integrative Graduate Education and Research Traineeship (IGERT Award
No. DGE-0654378) program.
v


TABLE OF CONTENTS
CHAPTER
I. INTRODUCTION..................................................1
II. LITERATURE REVIEW.............................................5
III. RESEARCH DESIGN..............................................8
Data.....................................................8
Methodology.............................................10
Commuting Expenditures..................................12
IV. RESULTS......................................................15
Transportation Economic Resilient Rating................17
Multivariate Regression.................................20
V. DISCUSSION AND CONCLUSION....................................26
REFERENCES.......................................................29
vi


LIST OF TABLES
Table
IV. 1 Home TAZ summary statistics variables............................................15
IV.2 Transportation Economic Resilient rating..........................................17
IV.3 Ten best and worst home TAZs TER ratings..........................................19
IV.4 TER rating regression results.....................................................20
vii


LIST OF FIGURES
Figure
III. 1 Transportation Analysis Zones (TAZ-2832)......................................9
IV. 1 Histogram of additional percent of income spent on commuting..................16
IV.2 TER ratings and selected variables for 2,365 home TAZs.........................22
IV.3 TER ratings for 2,365 home TAZs................................................23
IV.4 TER ratings for home TAZs located in the Denver Boulder corridor.............24
IV.5 TER ratings for home TAZs located in CBD and surrounding area...............25
viii


LIST OF ABBREVIATIONS
B Bike
DA Drive Alone
DRCOG Denver Regional Council of Governments
DT Drive to Transit
SR2 Shared Ride 2 people
SR3 Shared Ride 3+ people
TAZ Transportation Analysis Zones
TER Transportation Economic Resilient
W Walk
WT Walk to Transit


CHAPTER I
INTRODUCTION
Transportation is the second largest expense in a typical U.S. household, only
lower than housing expenditures. Between 2010 and 2012, overall spending on
transportation was roughly between 16% and 17% of total annual expenditure (BLS,
2012). With significant car dependency in the U.S., an important portion of household
expenditures is consumed by drivers spending money on purchasing fuel. As fuel price
volatility continues, transportation expenditure varies as well. Consequently, fuel price
changes, travel behavior, and percent of household income spent on transportation have
been topics of interest for research.
Communities are vulnerable to unexpected events such as natural disasters,
terrorist attacks, or geopolitical disruptions. When these events occur, the priority
becomes human survival, and in many cases, transportation is a key component. The
general stand of resilience research studies primarily evaluate the ability of a community
to perform under shock effects (shock-absorption), to avoid the shock altogether
(vulnerability), or the ability to recover quickly from a shock (shockcounteraction)
(Briguglio et al., 2009); transportation resilience has to do with the ability of the
transportation system to maintain a desired level of service or the time it takes to return to
that level of service given a shock to the system (Heaslip et al., 2009). While
transportation resilience research related to natural disasters and terrorist attacks is
extremely important, there has been little research on transportation economics due to
other shocks that focused on issues such as a drastic fuel price increase (Dodson and
Sipe, 2007; Motte-Baumvol et al., 2010; Zheng et al., 2011).
1


Fuel price prediction is a dilemma. Economists and financial institutions have a
poor record when it comes to estimating oil price fluctuations, such as those that occurred
in 2008 (Shafiee and Topal, 2010). One reason for the lack of success in this prediction is
the multitude of factors that contribute to fuel price increases, including diminishing
fossil fuel reserves and dwindling supplies, resurgent demand, a lack of investment in oil,
geopolitical disruptions, natural disasters, and terrorist attacks (Simmons, 2005). Previous
to the run-up of oil prices in 2007-2008, four international events (i.e. Yom Kippur War
on October 6, 1973, Iranian revolution in the fall of 1978, Iraq's invasion of Iran in
September 1980, and Iraq's invasion of Kuwait in August 1990) spaning over the last four
decades had a hand in the dramatic global disruption of oil production and resulting in
price increases (Hamilton, 2009).
The impacts of such global events are difficult to predict, and rarely would be
considered them an impact on the everyday lives of Americans. However, these shocks in
the system do have a local impact, and communities were affected with gasoline
shortages and fuel price shocks. Similar sharp increases in fuel price can also stem from
natural disasters. For example, recent Hurricane Sandy on the U.S. Eastern seaboard led
to fuel scarcity and drastic fuel price increases across the region (Honan, 2012). The
impact of Hurricane Katrina was similar in 2005 (Mouawad and Romero, 2005).
The primary objective of this paper is to investigate the impact of fuel price
shocks such as doubling the cost of gas at the Transportation Analysis Zones (TAZ)
level of geography. Such shocks usually produce changes in travel behavior, car use and
activities, as well as negative impacts on transportation expenditures (Ferdous et al.,
2010). The analysis focuses on home to work tours as they represent travel that people
2


would likely still need to make even under a sharp fuel price shock. The thinking is that
areas with multi-modal options, such as better walking, bicycling, or transit
infrastructure, might be able to better cope with these exogenous influences. In other
words, a community with considerable transit infrastructure even if experiencing
minimal ridership today would theoretically be able to withstand a rising fuel price
shock far better than an auto-dependent region that has not invested in transit.
Additionally, how do the land use patterns impact resilience? What are the roles of socio-
demographics and socio-economics actors? And what are the resulting disparities and
potential equity issues?
To begin to answer these questions, I model a baseline condition and compare that
to the resilience scenario when the fuel price doubles the baseline condition. The
scenarios were derived from the activity-based regional transportation model developed
by the Denver Regional Council, also known as the Focus model. These scenarios are
then assessed with respect to additonal transportation expenditures in terms of the extra
percentage of income spent on transportation. While there are many travel behavior
studies that assess the elasticity of relative minor shift in fuel prices, this
conceptualization of a fuel price shock when the change is sudden and drastic is not
directly comparable. The focus is instead on how different areas within a city would
respond to a major shift in gas prices. This assessment includes the development of a
metric index called Transportation Economic Resilience (TER), which is based on the
percentage of income spent on transportation for work tours. The resulting maps apply
the TER ratings to depict a quantification of home TAZs where I am able to identify the
zones that are resilient or non-resilient to fuel price shocks in the study area. This leads to
3


the identification of several significant variables that contribute to the TER metric index
and transportation resilience.
The findings of this study help increase our understanding of transportation
economic resilience to fuel price shocks in the short-run. These results can be used in
policy implications for automobiles, energy consumption, land use, future transportation
infrastructure investments, as well as developing regional resilience plans. This research
also offers a better understanding of the risks mitigated by diverse transportation
infrastructures that can, in turn, help shape more livable and resilient cities.
4


CHAPTER II
LITERATURE REVIEW
Resilience research has been studied primarily through the lens of natural
disasters such as hurricanes, earthquakes, or tsunamis (Foster, 1995; Chang andNojima,
2001; Bruneau et al., 2003; Pelling, 2003) or terrorist attacks (Battelle, 2007). More
recently, the concept of resilience has become more quantitative and expanded to
transportation (Berdica, 2002; Cova and Conger, 2004; Heaslip et al., 2010; Husdal,
2004; Murray-Tuite, 2006; Serulle et al., 2011).
Several academic research papers have studied the elasticity of petroleum demand
to fuel price increases using several economic models and resulting in different findings
(Dahl and Sterner, 1991; Espey, 1998; Lin and Prince, 2013). Most of these studies have
shown the price elasticity of demand for gasoline to be small (Ferdous et al., 2010; Nicol,
2003; Puller and Greening, 1999; Small and Van Dender, 2007). For example, Hughes et
al. (2008) estimate that the short-run price elasticity of gasoline purchased was between
-0.034 and -0.077 for 2001 through 2006, while the estimated short-run income
elasticities ranged from 0.21 to 0.75. Other studies (Cooper, 2005; Gicheva et al., 2007),
and data from the Bureau of Labor Statistics, have also found that household-level fuel
expenditures increase in proportion to increases in fuel prices, reiterating the notion of
fuel price inelasticity. In addition to the increase expenditures, households adjust their
consumption expenditures (including savings), car use, and activities in response to
increases in fuel prices (Anas, 2007; Dargay and Gately, 1997; Ferdous et al., 2010; Yang
and Timmermans, 2011). A 1975 study investigating travel behavior changes in the U.S.
5


versus those in Europe during the oil embargo revealed that Europeans significantly
increased their transit use while Americans were much more likely to stay home and
forego non-essential travel (Pisarski and Terra, 1975).
As noted above, changes in fuel prices suggest impacts in peoples transportation
expenditures, activities, and travel behavior. Unfortunately, the research aiming to
examine individuals economic impact of extreme fuel price shocks has received minimal
attention in the literature, especially with respect to the immediate increase in individual
transportation expenditures with respect to income. Additionally, the research aiming to
understand the impact variability among areas with different mode shares to absorb fuel
price shocks is very minimal. The presence of transportation options and local
infrastructure could play an important role in the level of transportation economic
resilience. This includes public transit, walking, bicycling, as well as shift trips from
drive alone to carpooling.
Similar to the research of petroleum demand elasticity to fuel prices, there are
some studies that investigate the relationship between gasoline prices and transit ridership
(Currie and Phung, 2007; Haire and Machemehl, 2007; Lane, 2010; Maley and
Weinberger, 2009). The American Public Transportation Association (APTA, 2012)
summarizes these studies with average elasticity values of 0.254 for commuter rail, 0.188
for heavy rail, 0.266 for light rail, 0.139 for buses, and 0.181 for all modes. Each of these
studies is based on the actual ridership change during periods of price change in the past
decade. These studies also focus on elasticities and are sometimes constrained by the
amount of transit service available and the excess capacity of that service. Rather than
6


providing another elasticity analysis, this study examines the different economic response
across a region in terms of the additional income spent on commuting.
This study contributes to the overall body of literature not only by examining the
individual economic impact due to doubling fuel price among different TAZ zones in the
Denver regional area, but also by creating a transportation economic resilience metric
index based on a fuel price shock and the impact on the extra percentage of income spent
on transportation. This facilitates an analysis of zones that are more or less resilient to
fuel price increases.
7


CHAPTER III
RESEARCH DESIGN
The analysis of this paper focuses on the additional transportation expenditures
for home to work tours in the Denver Metropolitan area when the price of fuel doubles
from a baseline condition. Work tours are selected as they represent travel that people
would likely still need to make even under a gas price increase. To assess this
hypothetical mode shift, actual tours are analyzed using a multinomial logistic regression
mode choice model. These trips were extracted from the Denver Regional Council of
Governments (DRCOG) Focus travel model, a regional based activity model. This model
is based on an in-depth, 12,000 household survey of travel behavior in the Denver region,
called Front Range Travel Counts (DRCOG, 2010a).
Data
The source of data for this analysis is the DRCOG Focus travel model with a
baseline year of 2010. The DRCOG area includes the city of Denver and a surrounding
area of approximately 40 miles, for a total of 2,832 TAZ (Figure III. 1).
8


Figure III.l Transportation Analysis Zones (TAZ-2832).
Source: DRCOG Regional Data Catalog (DRCOG, 2010b)
In order to facilitate the analysis and modeling efforts of this study, the data was
further processed using a series of queries in Microsoft SQL Server and PostGIS
(pgAdmin III) in the following manner:
1. Only the tours with a home-based origin were selected;
2. Only the tours with work destinations were selected;
3. Tours with the same home TAZ origin and the same work TAZ destination were
grouped; and
4. Only tours with a total of 10 or more originating at the same home TAZ were
selected.
The final sample of the analysis includes 1,154,673 home to work tours
compromising 654,762 home TAZ to work TAZ combinations. The information included
in the database is as follows:
9


1. Home TAZ ID
2. Work TAZ ID
3. Home to Work Average Di stance
4. Individual Median Income per Home TAZ
5. Numb er of T ours
6. Number of Drive Alone Tours (DA)
7. Number of 2-person Shared Ride Tours (SR2)
8. Number of 3+ people Shared Ride Tours (SR3)
9. Number of Drive to Transit Tours (DT)
10. Number of Walk to Transit Tours (WT)
11. Number of Walk Tours (W)
12. Number of Bike Tours (B)
13. Drive Alone Cost
14. In-Vehicle Travel Time (IVTT)
Methodology
The data analysis process began by first calculating the percentage share of the
seven mode types in the model drive alone, shared a ride 2, shared a ride 3+, drive to
transit, walk to transit, walking, and biking using the processed data.
I then investigated the statistical relationship between mode choice and a drastic
increase in gas price via the multinomial logistic regression model developed for the
Focus model. The intent was to provide an understanding of the mode shift by the
different set of home TAZ to work TAZ tours. I was less interested in absolute numbers
with respect to the mode choice outputs and more interested in the mode shift trends.
10


Thus, the regional multinomial logistic regression mode choice model was a good fit,
despite the testing of a resilience scenario outside of the normal range. Furthermore, this
mode choice model incorporates regional trends from both the 2010 Front Range Travel
Counts as well as a similar survey undertaken in 1997. The development of this model
using longitudinal data, across a time span when gas prices have more than tripled, makes
it an advantageous choice for our purposes.
The basic structure of a multinomial logistic regression mode choice model is
derived from a basic logit model. The following generalized logit equation determines
the probability of choosing a specific mode (Martin and McGuckin, 1998).
where:
Pj = probability of somebody choosing mode i = 1, 2, . k;
ut = utility function describing the relative attractiveness of mode i; and
£f=1 eUi= sum of the functions for all available mode alternatives
The probability of choosing a particular mode depends on the above utility
function relative to the utility functions for all the other mode options. The utility
function of the logit equation is based on the four-step transportation planning model
from DRCOG. It contains variables associated with each mode for a particular type of
tour between two specific zones. For example, the variables of the utility function
describing the relative attractiveness of driving alone include: cost associated from each
home to work zones, income, in-vehicle travel time, out-of-vehicle travel time, AM peak,
PM peak, and tours remaining.
11


Since the intent was to evaluate the immediate impact of doubling the cost of fuel,
the utility function of a particular mode was only affected by the variables containing a
cost component associated to the particular mode. While the utility function for walk to
transit, walking, and biking remains the same after doubling the fuel price; the utility
function for driving alone, shared ride 2, shared ride 3+, and drive to transit was reduced.
The probability of the seven modes was calculated for each of the 654,762 home TAZ to
work TAZ tour combinations.
Commuting Expenditures
In order to calculate the additional annual commuting expenditures due to a fuel
price increase, the assumptions below were implemented. They are derived from the
DRCOG Focus model TAZ to TAZ matrices.
1. Fuel cost baseline for driving alone is $0.15 per mile and $0.30 per mile when fuel
price doubles;
2. Fuel cost for shared ride 2 is equal to 2/3 the fuel cost for driving alone;
3. Fuel cost for shared ride 3+ is equal to 0.5 the fuel cost for driving alone;
4. Transit cost equals $3.20 per tour or 2 times $1.60 per trip; and
5. The cost for walking and biking is negligible.
Resulting average distance and mode shares for a given home TAZ were weighted
based upon the relative number of tours. For example, if a home zone has 100 tours total
and 60 of them are going to destination A with 80% drive alone mode share, 20 to B with
60% drive alone share, 15 to C with 90% drive alone mode share, and 5 to D with 40%
12


drive alone mode share; the home TAZ drive alone mode share would be 75.5%, as
follows:
0.755
0.8(60) + 0.6(20) + 0.9(15) + 0.4(5)
100
The following equation for work tours transportation cost is the total sum of each
mode share multiplied by the corresponding mode cost:
Commuting
Cost
~ AShare ~b SR2share ^ ~b ER3share 2^ ^ cost
+ (DTshare + WTshare) x $3.20 + (DTshare) x DTCost
The additional annual commuting expense was calculated by subtracting the
baseline annual transportation cost from the scenario with fuel price increase. Finally, I
calculated the additional percent of income spent on commuting by dividing the
additional annual commuting expense and the TAZ median of individual annual income.
To assist in evaluating the results, I assessed the additional percent of income
spent on commuting by the amount of standard deviations from the mean for all 2,365
home TAZ zones with the following formula:
TERi =
(% of IncomeMean % of Incomef)
Std. dev.
where:
TERt = transportation economic resilient rating of home TAZ
i = 1, 2, . 2365;
13


% of IncomeMean = mean of additional percent of income spent on
commuting for all i home TAZs;
% of Incomei = additional percent of income spent on commuting for
home TAZ i; and
Std. dev.= standard deviation of additional percent of income spent on
commuting for all i home TAZs
Each home TAZ zone then receives a Transportation Economic Resilient (TER)
rating based on the number of standard deviations from the mean additional percent of
income spent on commuting for all 2,365 home TAZs. A lower value of income spent on
commuting receives a positive rating (i.e. more resilient). In contrast, higher values of
additional income spent on transportation to work tours receive a negative value (i.e. less
resilient).
Finally, to further investigate the relationship between distance, income, mode
share, and TER rating; I used multiple least squares regression to predict TER rating with
distance, income, drive alone, shared ride, and transit percentages for the scenario with
the fuel price shock.
14


CHAPTER IV
RESULTS
The initial results for the 2,365 home TAZ zones listing specific characteristics -
such as the number of tours, average distance home-to-work tours, income, share for each
of the seven modes after the fuel price shock, additional annual commuting expenses per
capita, and additional percent of income spent on commuting are listed in Table IV. 1.
TableIV.1 Home TAZ summary statistics variables
Variable (n=2365) Mean Std. dev. Minimum Maximum
# of Tours 488.23 461.11 10 3,950
Average Tour Distance (miles) 22.18 10.23 2.99 83.83
Annual Individual Median Income per Home TAZ ($) 40,028 15,196 9,698 112,591
Drive Alone 0.679 0.111 0.071 1.000
Shared Ride 2 0.171 0.053 0.000 0.538
t/t j c-u Shared Ride 3+ Mode Share 0.069 0.033 0.000 0.365
After Drive to transit 0.014 0.015 0.000 0.167
Walk to Transit 0.029 0.043 0.000 0.455
Walk 0.034 0.095 0.000 0.681
Bike 0.005 0.009 0.000 0.091
Additional Annual Commuting 745.84 358.80 62.23 2718.26
Expenses per Capita ($) Additional Percent of Median Income Spent on Commuting 0.021 0.011 0.001 0.109
In average, each home TAZ generates approximately 488 home-to-work tours
with a distance mean of 22.2 miles per tour (round trip). The minimal number of tours per
home TAZ is 10 and a maximum of 3,950 tours. The mean individual median income for
all TAZs is approximately $40,000 per year, with a minimum individual medium income
of $9,698 and maximum of $112,591.
15


After the fuel price shock (gas price doubles), the mean proportion for the seven
modes are: 67.9% for drive alone, 17.1% for shared ride 2, 6.9% for shared ride 3+, 1.4%
for drive to transit, 2.9% for walk to transit, 3.4% for walking, and 0.5% for bicycling.
The commuting expense per capita increases on average by a net value of $746 per year.
The average TAZ individual commuting expense ranges from as low as $62 to as high as
$2,718 per year.
When the fuel price doubles, the mean of the additional percent of median income
spent on commuting for the 2,365 TAZ home zones is 2.1% (i.e. on average, individuals
spend an extra 2.1% of their income for home-to-work tours) with a standard deviation of
0.011. Figure IV. 1 displays the frequency histogram of additional percent of income
spent on commuting for the TAZ home zones. The bins are grouped by a range of 0.5
standard deviations (0.56%).
Percent of Income Group
Figure IV.l Histogram of additional percent of income spent on commuting.
16


Transportation Economic Resilient Rating
The Transportation Economic Resilient (TER) rating is generated by normalizing
the additional percent of income spent on commuting dataset from a mean value of 2.1%
to an expected value of 0 and standard deviation 1. Ratings in increments of 0.5 are
shown in Table IV.2. Higher than an additional 2.1% percent of income spent on
commuting receives a negative value, thus representing a relative lack of resilience;
lower than 2.1% percent of income spent on transportation to work tours receive a
positive rating, which corresponds with relative good resilience. The magnitude of the
rating changes negatively or positively proportionally to the increase in percent of income
based on standard deviation.
Table IV.2 Transportation Economic Resilient rating
Additional Percent of Income Spent on Commuting Transportation Economic Resilient Rating
4.84% -2.5
4.29% -2
3.73% -1.5
3.17% -1
2.61% -0.5
2.06% 0
1.50% 0.5
0.94% 1
0.38% 1.5
Table IV.3 shows several characteristics for the TAZ home zones with the ten
worst and best TER scores. For the sake of the shared ride variable, I combined shared
ride 2 and shared ride 3+. Drive to transit and walk to transit were also combined into the
transit variable. Home TAZ ID 963 has the lowest TER rating with a value of -7.89. This
means that the additional 10.9% of income spent on commuting is 7.89 standard
17


deviations higher than the mean value of 2.1%. The TER rating for home TAZ ID 1775 is
1.75, and is the highest of all 2,365 home TAZ zones. On average, a person living in this
home TAZ only spent an extra 0.1% percent of income in transportation. The home TAZ
zones with the worst TER ratings (negative values) tended to have higher than average
tour distance and/or lower incomes. Regarding mode share after the fuel price increase,
the bottom TER rating home TAZ zones tended to have high drive alone percentages as
well as low transit, walking, and bicycling mode shares.
18


Table IV.3 Ten best and worst home TAZs TER ratings.
Home TAZ Tours Average Tour Distance Income Drive Alone Mode Share After ^Ride^ Transit Walking Biking Additional Annual Commuting Expenses per Capita Additional %of Income Spent on Commuting TER Rating
1775 25 3.0 $61,133 19.9% 0.0% 4.0% 68.1% 8.0% $62 0.1% 1.75
1830 11 8.7 $37,213 18.2% 0.0% 45.5% 27.3% 9.1% $65 0.2% 1.69
1745 13 5.4 $62,828 38.5% 7.7% 7.7% 46.2% 0.0% $138 0.2% 1.65
1753 48 5.3 $58,066 22.9% 4.2% 12.5% 60.4% 0.0% $139 0.2% 1.63
1726 89 6.3 $44,818 15.6% 12.5% 15.7% 52.8% 3.4% $113 0.3% 1.62
1779 23 8.3 $69,951 21.7% 4.3% 21.7% 52.2% 0.0% $187 0.3% 1.61
1841 47 7.2 $76,288 27.7% 12.8% 10.6% 46.8% 2.1% $203 0.3% 1.61
1809 45 4.8 $40,853 13.3% 11.1% 13.3% 62.2% 0.0% $108 0.3% 1.61
1728 18 5.8 $40,097 16.7% 11.1% 16.7% 50.0% 5.6% $103 0.3% 1.61
1755 33 5.3 $45,331 15.2% 6.1% 12.1% 63.6% 3.0% $124 0.3% 1.60
1957 308 71.4 $29,644 58.6% 40.7% 0.0% 0.7% 0.0% $2,203 7.4% -4.82
884 51 29.0 $12,681 66.4% 33.6% 0.0% 0.0% 0.0% $968 7.6% -5.00
945 88 24.0 $9,698 62.3% 33.1% 2.3% 2.3% 0.0% $748 7.7% -5.07
883 84 31.2 $13,328 67.5% 32.5% 0.0% 0.0% 0.0% $1,089 8.2% -5.48
3 34 77.3 $32,357 85.3% 14.7% 0.0% 0.0% 0.0% $2,718 8.4% -5.69
958 165 26.4 $10,583 71.1% 28.9% 0.0% 0.0% 0.0% $894 8.5% -5.74
995 105 81.8 $29,828 61.1% 38.9% 0.0% 0.0% 0.0% $2,630 8.8% -6.07
996 439 83.8 $30,381 68.7% 31.3% 0.0% 0.0% 0.0% $2,709 8.9% -6.16
2802 196 82.3 $27,566 68.4% 31.6% 0.0% 0.0% 0.0% $2,680 9.7% -6.88
963 73 31.6 $10,176 65.7% 31.6% 1.4% 1.4% 0.0% $1,105 10.9% -7.89


Multivariate Regression
Table IV.4 presents the results of the multiple least square analysis, including
coefficients of each of the variables, standard errors, t-statistics, p-values, and R2. As
expected, distance is negatively related to home TAZ TER rating, and income is
positively related. The negative and significant coefficient for distance (-0.074) indicates
that home TAZ zones with a higher commute distance are expected to have lower TER
ratings. Income has a positive and significant relationship with the TER rating, with a
coefficient of 3.97xl0"5. Shared ride percentage and transit percentage for the scenario
with a fuel price shock are also statistical significant variables. The positive coefficient
for transit percentage indicates that a 10% increase in transit share is associated with an
additional 0.23 in TER rating when adjusting for distance, income, drive alone, and
shared ride percentages. It is interesting to note the coefficient of drive alone percentage
is not significant. This suggests that distance to work, income, and the availability of
transit impact resilience more than what can be garnered from current driving mode
share.
Table IV.4 TER rating regression results.
Variables Coefficients Std. error tstat p-value
(Constant) -0.13707 0.10313 -1.33 0.184
Distance -0.07444 0.00100 -74.34 0.000
Income 3.97E-05 5.80E-07 68.53 0.000
Drive Alone Percentage -0.13440 0.09999 -1.34 0.179
Shared Ride Percentage 0.80231 0.14252 5.63 0.000
Transit Percentage 2.27818 0.28143 8.09 0.000
R2 0.8522
Observations 2365
20


Figure IV.2 presents individual relationships between home TAZ TER ratings and
each of the following variables: average tour home distance, income, and transit mode
share after. The figure highlights the association of the three variables with respect to the
TER ratings.
Figure IV.3 presents the distribution of TER ratings for 2,365 TAZ home zones.
Results from the regression analysis suggest that longer commuting distances and low
mode share for transit contribute to a negative TER value. From this figure, we see that
lowest ratings are among home zones on the outskirts of the DRCOG region. Figure IV.4
depicts home TAZ zones located along the Denver Boulder corridor in more detail.
This illustrates the positive TER ratings with significant higher values arw within the city
of Denver and the city of Boulder. A final zoom into the Denver central business district
(CBD) reiterates this result. Home TAZ with TER ratings higher than 1.50 are all located
inside the CBD. These are shown in Figure IV.5. In contrast with the home TAZs located
on the perimeter, these high rating TER home TAZ zones seem to be the result of one or
more of the following: low commuting distance, higher income, and/or higher transit,
walking, and biking mode shares.
21


* 20%
V *V**' >*, }
itj*. .V % \ '
30%
40%
50%
Transit Mode Share After
Figure IV.2 TER ratings and selected variables for 2,365 home TAZs.
22


Figure IV.3 TER ratings for 2,365 home TAZs.
K>
OJ


Figure IV.4 TER ratings for home TAZs located in the Denver Boulder corridor.


I
K>
Lt>
Figure IV.5 TER ratings for home TAZs located in CBD and surrounding area.


CHAPTER V
DISCUSSION AND CONCLUSION
Transportation cost is the second highest expenditure in a typical U.S. budget.
With the high dependency on the automobile, a large portion of the variable cost
fluctuates with fuel price volatility. Communities are vulnerable to unexpected events,
including the possibility of fuel price shocks. .
This study has sought to examine the impact of fuel price doubling on peoples
income for commuting-related transportation expenses. To achieve this research goal, I
examined a database of 1,154,673 individual work tours from DRCOG grouped by
home TAZ. I developed a transportation economic resilient (TER) rating to evaluate
2,365 home TAZ to measure the additional commuting expense with respect to income
when fuel price doubles. I used a multiple least square regression to estimate the
relationships between the developed TER rating and work-tour distance, income, and
mode share variables.
The average percent of income spent on commuting, as a result of a doubling of
the fuel price, increases by 2.1%. While this may seem like a relatively low number, the
results also suggested large disparities among home TAZs with some reaching almost
11% and others at only 0.1%. The resulting TER ratings ranged from -7.89 to +1.75.
The worst TER ratings are home TAZs located on the outwards limits of the
geographical area analyzed, and the best ratings are closer to downtown Denver.
Commuting distance, income, shared ride percentage, and transit percentage are all
statistically significant variables to determine TER ratings.
26


Home TAZs whose people commute short distances are very likely to be
economically resilient to fuel price increases. This is because the current cost of
commuting is low, and there also tend to be transportation options available such as
walking and biking. High-income individuals are also considered transportation
economically resilient since their income would not be impacted as much by an increase
in commuting expenditures. In contrast, individuals in low-income home TAZs that
have to drive alone and/or travel long distances for work tours would be the most
impacted by a fuel price increase. Based on the TER rating, these areas are not
considered economically resilient to fuel price shocks. A person living in these areas
would still need to travel to work and without realistic transportation options the
economical impact can be very serious.
When the commute distance is long, home TAZ zones with current high mode
shares for carpooling and transit or the opportunity to shift to high share percentage in
these modes have a very high probability of being transportation economically
resilient. For example, comparing a home TAZ with a transit share 30% higher than
another TAZ, the TER rating is expected to experience a 0.68 higher TER rating when
controlling for other variables via the regression equation. Figures IV.3, IV.4, and IV.5
present the resilient (green TAZ) and non-resilient (red) home TAZs for the area
studied. Areas with a high density of jobs such as Boulder, the Denver CBD, and the
Denver Tech Center tend to be very resilient. Areas with good transit service such
as the Denver-Boulder 36 corridor and along the light rail lines also tend to be more
resilient. The surrounding suburbs with less transit availability tend to be less resilient.
27


These findings introduce a transportation economic resilient index, and the
concept of significant fuel price shocks on transportation expenditures. This work
elaborates on earlier studies regarding the impact of fuel price fluctuations on travel
behavior. In particular, this paper examines regional disparities with respect to
preparedness to a major shift in gas prices for commuting expenses. Since the focus is
on a drastic increase in fuel price, these results are different from elasticity studies that
have considered how relative minor gas price adjustments have impacted travel
behavior. This paper also builds on studies related to transportation cost at specific
areas. For example, CNT (2010) developed a Housing + Transportation Affordability
Index, which measure the cost of housing and transportation at the neighborhood level.
The findings of this study help increase our understanding of transportation
economic resilience with respect to fuel price shocks, and how commute distance,
income, and transit service affect different communities. These results have policy
implications for automobiles (e.g. electric cars, fuel efficiency); land use (e.g.
development of mixed-use, pedestrian, bicycling, and transit oriented communities);
and in facilitating other modes of transportation by investing in transit infrastructure or
other transportation options. This research also offers a better understanding of the risks
mitigated by a diverse transportation infrastructure, and can help shape more resilient
and more livable cities. Such findings can support the development of regional
resilience plans and lead to research on related topics.
28


REFERENCES
Anas, A. (2007) 'A unified theory of consumption, travel and trip chaining', Journal of
urban economics., 62(2), pp. 162
APTA (2012) 'Public Transportation Protects Americans From Gas Price Volatility',
American Public Transportation Association.
Berdica, K. (2002), 'An Introduction to Road Vulnerability: What Has Been Done, Is
Done and Should Be Done', Transport Policy, 9(2): 117-127.
Battelle (2007), 'Evaluation of Systems Available Redundancy to Compensate for Loss
of Transportation Assets Resulting from Natural Disaster or Terrorist Attacks',
National Surface Transportation Policy and Revenue Study Commission,
Washington, D.C.
BLS (2012) 'Economic News Release: Consumer Expenditure', U.S. Department of
Labor: Bureau of Labor Statistics, Washington, D.C.
http://www.bls.gov/news.release/cesan.nrO.htm
Briguglio, L., Cordina, G., Farrugia, N., Vella, S. (2009) 'Economic Vulnerability and
Resilience: Concepts and Measurements', Oxford Development Studies, 37(3), pp.
229-247
Bruneau, M., Chang, S., et al. (2003), 'A Framework to Quantitatively Assess and
Enhance the Seismic Resilience of Communities', Earthquake Spectra, 19(4):
733-752.
Chang, S., Nojima, N. (2001), 'Measuring Post-Disaster Transportation System
Performance: The 1995 Kobe Earthquake in Comparative Perspective',
Transportation Research Part A, 35(6): 475-494.
CNU (2010) 'H+T Affordability Index', Center for Neigbor hood Technology, Chicago,
IL. http://htaindex.cnt.org/
Cooper, M. (2005) 'The Impact of Rising Prices on Household Gasoline Expenditures',
Consumer Federation of America manuscript.
Cova, T. J., Conger, S. (2004), 'Transportation Hazards', Handbook of Transportation
Engineering, M. Kutz. McGrawHill: 17.11-17.24, Washington, D.C.
Currie, G., Phung, J. (2007) 'Transit Ridership, Auto Gas Prices, and World Events: New
Drivers of Change?', Transportation Research Record: Journal of the
Transportation Research Board, 1992(-1), pp. 3-10
29


Dahl, C., Sterner, T. (1991) Analyzing gasoline demand elasticities, a survey, Energy
Econ, 3 (13), 203-210.
Dargay, J., Gately, D. (1997) 'The demand for transportation fuels: Imperfect price-
reversibility?', Transportation Research Part B: Methodological Transportation
Research Part B: Methodological, 31(1), pp. 71-82
Dodson, J., Sipe, N. (2007) 'Oil Vulnerability in the Australian City: Assessing
Socioeconomic Risks from Higher Urban Fuel Prices', Urban Studies, 44(1), pp.
37-62
DRCOG (2010a) 'Front Range Travel Counts', Denver Regional Council of Goverments,
Denver, CO. http://www.drcog.org/index.cfrn7pageMTontRangeTravelCounts
DRCOG (2010b) 'Transportation Analysis Zones', Denver Regional Council of
Governments, Denver, CO.
http://gis.drcog.org/datacatalog/content/transportation-analvsis-zones-taz-2832
Espey, M. (1998) Gasoline demand revisited: an international meta-analysis of
elasticities, Energy Econ, 20, 273-295.
Ferdous, N., Pinjari, A., Bhat, C., Pendyala, R. (2010) 'A comprehensive analysis of
household transportation expenditures relative to other goods and services: an
application to United States consumer expenditure data', Transportation, 37(3),
pp. 363-390
Foster, H. (1995) 'Disaster Mitigation: The Role of Resilience', Proceedings of a Tri-
lateral Workshop on Natural Hazards, D. Etkin. Ontario, Canada, Merrickville:
93-108.
Gicheva, D., Hastings, J.S., Villas-Boas, S., National Bureau of Economic, R. (2007)
'Revisiting the income effect gasoline prices and grocery purchases', National
Bureau of Economic Research, Cambridge, Mass.
http://papers.nber.org/papers/wl3614
Haire, A., Machemehl, R. (2007) 'Impact of Rising Fuel Prices on U.S. Transit
Ridership', Transportation Research Record: Journal of the Transportation
Research Board, 1992(-1), pp. 11-19
Hamilton, J.D. (2009) 'Causes and Consequences of the Oil Shock of 2007-08',
Brookings Papers on Economic Activity, pp. 215-261
Heaslip, K., Louisell, W., Collura, J. (2009) 'A Methodology to Evaluate Transportation
Resiliency for Regional Networks', Transportation Research Board 88th Annual
Meeting, Washington DC.
30


Heaslip, K., W. Louisell, et al. (2010), 'A Sketch Level Method for Assessing
Transportation Network Resiliency to Natural Disasters and Man-Made Events',
Transportation Research Board, Washington, D.C.
Honan, E. (2012) 'Fuel scarce, East Coast struggles to recover from storm', Reuters, New
York.
Hughes, J.E., Knittel, C.R., Sperling, D., Frank, E. (2008) 'Evidence of a shift in the
short-run price elasticity of gasoline demand', The Energy Journal.
Husdal, J. (2004), 'Why Reliability And Vulnerability Should Be An Issue In Road
Development Projects', Samferdsel: Journal of the Norwegian Institute of
Transport Economics, (3).
Lane, B.W. (2010) 'The relationship between recent gasoline price fluctuations and
transit ridership in major US cities', J. Transp. Geogr. Journal of Transport
Geography, 18(2), pp. 214-225
Lin, C.Y.C., Prince, L. (2013) 'Gasoline price volatility and the elasticity of demand for
gasoline', Energy Economics, 38(0), pp. 111-117
Maley, D., Weinberger, R. (2009) 'Rising Gas Price and Transit Ridership',
Transportation Research Record: Journal of the Transportation Research Board,
2139(-l), pp. 183-188
Martin, W.A., McGuckin, N.A. (1998) Report 365: Travel Estimation Techniques for
Urban Planning. National Academy Press, Washington D.C.
Motte-Baumvol, B., Massot, M.-H.l.n., Byrd, A.M. (2010) 'Escaping Car Dependence in
the Outer Suburbs of Paris', Urban Studies, 47(3), pp. 604-619
Mouawad, J., Romero, S. (2005) 'Gas Prices Surge as Supply Drops', The New York
Times.
Murray-Tuite, P. (2006), 'A Comparison of Transportation Network Resilience Under
Simulated System Optimum and User Equilibrium Conditions', Winter Simulation
Conference, Monterey, CA.
Nicol, C.J. (2003) 'Elasticities of demand for gasoline in Canada and the United States',
Energy economics., 25(2), pp. 201-214
Pelling, M. (2003), 'The Vulnerability of Cities: Natural Disasters and Social Resilience',
Earthscan Publications, London.
Pisarski, A., Terra, N. (1975) 'American and European transportation responses to the
1973-74 oil embargo', Transportation, 4(3), pp. 291-312
31


Puller, S.L., Greening, L.A. (1999) 'Household adjustment to gasoline price change: an
analysis using 9 years of US survey data', Energy economics., 21(1), pp. 37
Shafiee, S., Topal, E. (2010) 'A long-term view of worldwide fossil fuel prices', Applied
energy., 87(3), pp. 988-1000
Serulle, N. U., Heaslip, K., et al. (2011), Resiliency of Transportation Network of Santo
Domingo, Dominican Republic Case Study, Transportation Research Record,
2234.
Simmons, M.R. (2005) Twilight in the desert: the coming Saudi oil shock and the world
economy. John Wiley & Sons, Hoboken, N.J.
Small, K.A., Van Dender, K. (2007) 'Fuel Efficiency and Motor Vehicle Travel: The
Declining Rebound Effect', The energy journal /, 28(1), pp. 25
Yang, D., Timmermans, H. (2011) 'Effects of Energy Price Fluctuation on Car-Based
Individual Activity-travel Behavior', SBSPRO Procedia Social and Behavioral
Sciences, 20(pp. 547-557
Zheng, J., Atkinson-Palombo, C., McCahill, C., O'Hara, R., Garrick, N. (2011)
'Quantifying the Economic Domain of Transportation Sustainability',
Transportation Research Record: Journal of the Transportation Research Board,
2242(-l), pp. 19-28
32


Full Text

PAGE 1

EFFECTS OF FUEL PRICE SHOCKS IN COMMUTING EXPENDITURES : A SOCIO ECONOMIC ANALYSIS by ALEJANDRO HENAO B.S., University of Colorado Boulder 2006 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Science Civil Engineering 2013

PAGE 2

ii This thesis for the Master of Science degree by Alejandro Henao has been approved for the Civil Engineering Program by Wesley E. Marshall Chair Bruce Janson Carolyn McAndrews November 1 5 2013

PAGE 3

iii Henao, Alejandro (M.S., Civil Engineering) Effects of Fuel Price Shocks in Commuting Expenditures : A Socio economic Analysis Thesis directed by Assistant Professor Wesley E. Marshall ABSTRACT Communities are vulnerable to unexpected events, including fuel price shocks. Increases in fuel price lead to changes nsportation expenditures and income budgets The impact of fuel price shocks has received minimal attention in the literature, especially, with respect to immediate individual eco nomic impact for transportation expenditures. This study assesses data from t he Denver Regional Council of Governments to examine individual work tours grouped by home tra nsportation analysis zones (TAZ) For all home T AZs, the percent of income spent o n commuting expenses increases on average by 2 .1 % when fuel price doubles, and as high as 11% of individual income. A transportation economic resilient (TER) rating is developed to evaluate home TAZs and measure the additional commuting expense per income when fuel price doubles. D istance, income, trans it share, and carpooling share are statistically significant variables to determine TER ratings. H ome TAZs that experience the lowest additional percent of income on commuting expenses, or have the best TER ratings are located in downtown Denver Three ma in factors contribute to exhibit good transportation economic resilience: i) high income, ii) good multi modal transportation options, and iii) proximity to employment centers, such as downtown. The form and content of this abstract are approved. I recom mend its publication. Approved: Wesley E. Marshall

PAGE 4

iv DEDICATION I dedicate this work to m y spouse, Augusta Henao, who has always supported me by providing a foundation of lo ve, hard work, and life balance. To my boys, Tom s and Andrs, who are my every day inspiration. To my parents, who instilled in me the ummeasurable value of education and have supported me unconditional ly in all aspects of my life.

PAGE 5

v ACKNOWLEDGMENTS I would like to thank Dr. Wesley Marshall for mentoring me, sharing his knowledge, and providing funding and extensive advice for my thesis. I thank Dr. Bruce Janson and Dr. Carey McAndrews for contributing to my education and serving on my Thesis Committee I also would like to thank all the members of the Active Communities Transportation (ACT) Research Group for their contributions to my research work. Finally, I thank the National Science Foundation for providing funding for this work through their Integrative Graduate Education and Research Train eeship (IGERT Award No. DGE 0654378) program

PAGE 6

vi TABLE OF CONTENTS CHAPTER I. INTRODUCTION ................................ ................................ ................................ ........ 1 II. LITERATURE REVIEW ................................ ................................ ............................ 5 III. RESEARCH DESIGN ................................ ................................ ................................ 8 Data ................................ ................................ ................................ ................... 8 Methodology ................................ ................................ ................................ ... 10 Commuting Expenditures ................................ ................................ ............... 12 IV. RESULTS ................................ ................................ ................................ .................. 15 Transportation Economic Resilient Rating ................................ ..................... 17 Multivariate Regression ................................ ................................ .................. 20 V. DISCUSSION AND CONCLUSION ................................ ................................ ....... 26 REFERENCES ................................ ................................ ................................ ................. 29

PAGE 7

vii LIST OF TABLES T able IV.1 Home TAZ summary statistics variables ................................ ................................ 15 IV.2 Transportation Economic Resilient rating ................................ ............................... 17 IV.3 Ten best and worst home TAZs TER ratings. ................................ ......................... 19 IV.4 TER rating regression results ................................ ................................ .................. 20

PAGE 8

viii LIST OF FIGURES Figure III.1 Transportation Analysis Zones (TAZ 2832). ................................ ............................ 9 IV.1 Histogram of additional percent of income spent on commuting. ........................... 16 IV.2 TER ratings and selec ted variables for 2,365 home TAZs ................................ ..... 22 IV.3 TER ratings for 2,365 home TAZs. ................................ ................................ ......... 23 IV.4 TER ratings for home TAZs located in the Denver Boulder corridor ................. 24 IV.5 TER ratings for home TAZs located in CBD and surrounding area. ...................... 25

PAGE 9

ix LIST OF ABBREVIATIONS B Bike DA Drive Alone DRCOG Denver Regional Council of Governments DT Drive to Transit SR2 Share d Ride 2 people SR3 Shared Ride 3 + people TAZ Transportation Analysis Zones TER Transportation Economic Resilient W Walk WT Walk to Transit

PAGE 10

1 CHAPTER I INTRODUCTION Transportation is the second largest expense in a typical U.S. household, only lower than housing expenditures. Between 2010 and 2012, overall spending on transportation was roughly between 16% and 17% of total annual expenditure ( BLS, 2012 ) With significant car dependency in the U.S. an important portion of household expenditures is consumed by drivers spending m oney on purchasing fuel. As fuel price volatility continues transportation expenditure vari es as well. Consequently, fuel price changes, travel behavior, and percent of household income spent on transportation have been topics of interest for research. C ommunities are vulnerable to unexpected events such as natur al disasters, terrorist attacks or geopolitical disruptions. When these events occur the priority becomes human survival, and in many cases, transportation is a key component. The general stand of r esilience research studies primarily evaluate the abilit y of a community to perform under shock effects (shock absorption), to avoid the shock altogether (vulnerability), or the ability to recover quickly from a shock (shockcounteraction) ( Briguglio et al., 2009 ) ; transportation resilience has to do with the ability of the transportation system to maintain a desired level of servic e or the time it takes to return to that level of service given a shock to the system ( Heaslip et al., 2009 ) While transportation resilience research related to natural disasters and terrorist attacks is extremely important, there has been little research on transportation economics due to other shocks that focused on issues such as a drastic fuel price increase ( Dodson and Sipe, 2007 ; Motte Baumvol et al., 2010 ; Zheng et al., 2011 )

PAGE 11

2 Fuel pr ice prediction is a dilemma. Economist s and financial institutions have a poor record when it comes to estimating oil price fluctuations such as those that occurred in 2008 ( Shafiee and Topal, 2010 ) One reason for the lack of success in this prediction is the multitude of factors that contribute to fuel price increases including diminishing fossil fuel reserves and dwindling supplies, resurgent demand, a lack of investment in oil geopolitical disruptions, natural disast ers, and terrorist attacks ( Simmons, 2005 ) Previous to the run up of oil prices in 2007 2008, four inte rnational events ( i.e. Yom Kippur War on October 6, 1973, Iranian revolution in the fall of 1978, Iraq's invasion of Iran in September 1980, and Iraq's invasion of Kuwait in August 1990) spaning over the last four decades had a hand in the dramatic global disruption of oil production and resulting in price increase s ( Hamilton, 2009 ) The impacts of such global events are difficult to predict, and rarely would be considered them an impact on the everyday lives of Americans. However, these shocks in the system do have a local impact, and communities were affected with gasoline shortages and fuel price shocks. Similar sharp increases in fuel price can also stem from natural disasters. For example, recent H urricane Sandy on the U.S. East ern seaboard led to fuel scarcity and drastic fuel price increases across the region ( Honan, 2012 ) The impact of Hurricane Katrina was similar in 2005 ( Mouawad and Romero, 2005 ) The primary objective of this paper is to investigate the impact of fuel price shocks such as doubling the cost of gas at the T ransportation Analysis Zones (TAZ) level of geography. Such shocks usually prod uce changes in travel behavior car use and activities, as well as negative impacts on transportation expenditures ( Ferdous et al., 2010 ) The analysis focus es on home to work tours as they represent trav el that people

PAGE 12

3 would likely still need to make even under a sharp fuel price shock The thinking is that areas with multi modal options, such as better walking, bicycling, or transit infrastructure might be able to better cope with these exogenous influen ces. In other words, a community with consid erable transit infrastructure even if experie ncing minimal ridership today would theoretically be able to withstand a rising fuel price shock far better than an auto dependent region that has not invested in transit. Additionally, how do the land use patterns impact resilience? What are the roles of socio demographics and socio economics actors? And what are the resulting disparities and potential equity issues? To begin to answer the se questions, I model a baseline condition and compare that to the resilience scenario when the fuel price doubles the baseline condition. The scenarios were derived from the activity based regional transportation model developed by the Denver Regional Cou ncil, also known as the Focus model. These scenarios are then assessed with respect to additonal transportation expenditures in terms of the extra percentage of income spent on transportation. While there are many travel behavior studies that assess the el asticity of relative minor shift in fuel prices, this c oncep tualization of a fuel price shock when the change is sudden and drastic is not directly comparable The focus is instead on how different areas within a city would respond to a major shift in gas prices. This assessment includes the development of a metric index called Transportation Economic Resilience (TER), which is based on the percentage of income spent on transportation for work tours. The resulting maps apply the TER ratings to depict a quantification of home TAZs where I am able to identify the zones that are resilient or non resilient to fuel price shocks in the study area. This leads to

PAGE 13

4 the identification of several significant variables that contribute to the TER metric index and tran sportation resilience. The findings of this study help increase our understan d ing of transportation economic resilience to fuel price shocks in the short run These results can be used in policy implications for automobiles, energy consumption, land use, f uture transportation infrastructure investments as well as developing regional resil ience plans. This research also offer s a better understanding of the risks mitigated by diverse transportation infrastructures that can in turn, help shape more livable a nd resilient cities.

PAGE 14

5 CHAPTER II LITERATURE REVIEW Resilience research has been studied primarily through the lens of natural disasters such as hurricanes, earthquakes, or tsunamis (Foster 1995; Chang and Nojima 2001; Bruneau et al. 2003; Pelling 2003) or terrorist attacks (Battelle 2007). More recently, the concept of resilience has become more quantitative and expanded to transportation (Berdica 2002; Cova and Conger 2004; Heaslip et al. 2010; Husdal 2004; Murray Tuite 2006; Serulle et al. 2011). Several academic research papers have studied the elasticity of petroleum demand to fuel price increases using several economic m odels and resulting in differen t findings ( Dahl and Sterner, 1991 ; Espey, 199 8; Lin and Prince, 2013 ) Most o f these studies have shown the price elasticity of demand for gasoline to be small ( Ferdous et al., 2010 ; Nicol, 2003 ; Puller and Greening, 1999 ; Small and Van Dender, 2007 ) For example, Hughes et a l. ( 2008 ) estimate that the short run price elastic ity of gasoline purchased was between 0.034 and 0.077 for 2001 through 20 06, while the estimated short run income elasticities range d from 0.21 to 0.75. Other studies ( Cooper, 2005 ; Gicheva et al., 2007 ) and data from the Bureau of Labor Statistics have also found tha t household level fuel expenditures increase in proportion to increases in fuel prices, reiterating the notion of fuel price inelasticity. In addition to the increase expenditures, households adjust their consumption expenditures (including savings) car u se, and activities in response to increases in fuel prices ( Anas, 2007 ; Dargay and Gately, 1997 ; Ferdous et al., 2010 ; Yang and Timmermans, 2011 ) A 1975 study investigating travel behavior changes in the U.S.

PAGE 15

6 versus those in Europe during the oil embargo revealed that Europeans signif icantly increased their transit use while Americans were much more likely to stay home and forego non essential travel ( Pisarski and Terra, 1975 ) As noted above changes in fuel prices suggest impacts in peopl transportation expenditures, activities and travel behavior. Unfortunately, the research aiming to economic impact of extreme fuel price shocks has received minimal attentio n in the literature, especially with respect to the immediate increase in individual t ransportation expenditures with respect to income. Additionally the research aiming to understand the impact variability among areas with different mode shares to absorb fue l price shocks is very minimal. The presence of transport ation options and local infrastructure could play an important role in the level of tra nsportation economic resilience. This includes public transit, walking, bicycling, as well as shift trips from drive alone to carpooling. Similar to the research of petr oleum demand elasticity to fuel prices, there are some studies that investigate the relationship between gasoline prices and transit ridership ( Currie and Phung, 2007 ; Haire and Machemehl, 2007 ; Lane, 2010 ; Maley and Weinberger, 2009 ) The American Public Transportation Association ( APTA, 2012 ) summarizes these studies with average elasticity values of 0.254 for commuter rail, 0.188 for heavy rail, 0.266 f or light rail, 0.139 for buses, and 0.181 for all modes. Each of these studies is based on the actual ridership change during periods of price change in the past decade. The se studies also focus on elasticities and are sometimes constrained by the amount o f transit service available and the excess capacity of that service Rather than

PAGE 16

7 providing another elasticity analysis, this study e xamin es the different economic response across a region in terms of the a dditional income spent on commuting This study contributes to the overall body of l iterature not only by e xamining the individual economic impact due to doubling fuel price among different TAZ zones in the Denver regional area but also b y creating a transportation economic resilience metric index based on a fuel price shock and the impact on the extra percent age of income spent on transportation This facilitates an analysis of zones that are more or less resilient to fuel price increases.

PAGE 17

8 CHAPTER III RESEARCH DESIGN The analys is of this paper focuses on the additional transportation expenditures for home to work tours i n the Denver Metropolitan area when the price of fuel doubles from a baseline condition Work t ours are selected as they represent trav el that people would likely still need to make even under a gas price increase To assess this hypot hetical mode shift, actual tours are analyzed using a multinomial logistic regression mode choice model. These trips were extracted from the Denver Regional Council of Governments (DRCOG) Focus travel model, a regional based activity model. This model is based on an in depth 12,000 household survey of travel behavior in the Denver region, called Front Range Travel Counts ( DRCOG, 2010a ) Data The source of data for this analysis is the DRCOG Focus travel model with a baseline year of 2010. The DRCOG area includes the city of Denver and a surrounding area of approximately 40 miles for a total of 2,832 TAZ (Figure III.1)

PAGE 18

9 Figure III 1 Transportation Analysis Zones (TAZ 2832) Source: DRCOG Regional Data Catalog ( DRCOG, 2010b ) In order to facilitate the analysis and modeling efforts of this study, the data was further processed using a series of queries in Microsoft SQL Server and PostGIS ( pgAdmin III) in the following manner: 1. Only the tours with a home based origin were selected ; 2. Only the tours with work destination s were selected ; 3. Tours w ith the same home TAZ origin and the same work TAZ destination were g rouped; and 4. Only tours with a total of 10 or more originating at the same home TAZ were selected. The f i nal sample of the analysis includes 1 ,154,673 home to work tours compromising 654,762 home TAZ to work TAZ combination s. The information included in the database is as follows:

PAGE 19

10 1. Home TAZ ID 2. Work TAZ ID 3. Home to Work Average Distance 4. Individual Median Income per Home TAZ 5. Number of Tours 6. Number of Drive Alone Tours (DA) 7. Number of 2 person Share d Ride T ours (SR2) 8. Number of 3 + people Shared Ride Tours (SR3) 9. Number of Drive to Transit Tours (DT) 10. Number of Walk to Transit Tours (WT) 11. Number of Walk Tours (W) 12. Number of Bike Tours (B) 13. Drive Alone Cost 14. In V ehicle Travel Time (IVTT) Methodology The data an alys is process began by first calculating the percentage shar e of the s even mode types in the model drive alone, share d a ride 2, share d a ride 3 + drive to transit, walk to transit, walking, and biking using the processed data. I then investigated t he statistical relationship between mode choice and a drastic increase in gas price via the multinomial logistic regression model developed for the Focus model The intent was to provide an understanding of the mode shift by the different set of home TAZ to work TAZ tours. I was less interested in absolute numbers w i th respect to the mode choice o utputs and more interested in the mode shift trends.

PAGE 20

11 Thus, the regional multinomial logistic regression mode choice model was a good fit, despite the testing of a resilience scenar io outside of the normal range. Furthermore, this mode choice model incorporates regional trends from both the 2010 Front Range Travel Counts as well as a similar survey undertaken in 1997. The development of this model using longitudinal data, across a time span when gas prices have more than triple d, makes it an advantageous choice for our purposes. The basic structure of a multinomial logistic regression mode choice model is derived from a basic logit model. The following generalized logit equation determines the probability of choosing a specif ic mode ( Martin and McGuckin, 1998 ) where: = probability of somebody choosing mode i = 1, 2, . k; = utility function describing the relative attractiveness of m ode i; and = sum of the functions for all available mode alternatives The probability of choosing a particular mode depends on the above utility function relative to the utility functions for all the other mode options. T he utility function of the logit equation is based on the four step transportation planning model from DRCOG. It contains variables associated with each mode for a parti cular type of t our between two specific zones. For example, the variables of the utilit y function describing the relative attractiveness of driving alone include: cost associated from each home to work zones, income, in vehicle travel time, out of vehicle travel time, AM peak, PM peak, and tours remaining.

PAGE 21

12 Since the intent was to evaluate the immediate impact of doubling the cost of fuel, the utility function of a particular mode was only affected by the variables containing a cost component associated to the particular mode. While the utility function for w alk to transit, walking, and biking remains the same after doubling the fuel price; the utility function for driving alone, share d ride 2, share d ride 3+, and drive to transit was reduce d The probability of the seven modes was calculated for each of the 6 54,762 home TAZ to wor k TAZ tour combinations. Commuting Expenditures I n order to calculate the additional annual commuting expenditure s due to a fuel price increase the assumption s be low were implemented. They are derived from the DRCOG Focus model TAZ to TAZ matrices. 1. Fuel cost baseline f or driving alone is $0.15 per mile and $0.30 per mile when fuel price doubles ; 2. Fuel cost for share d ride 2 is equal to 2/3 the fuel cost for driving alone; 3. Fuel cost f or share d ride 3+ is equa l to 0.5 the fuel cost for driving alone; 4. Transit cost equals $3.20 per tour or 2 times $1.60 per trip; and 5. The cost for walking and biking is negligible Resultin g average distance and mode shares for a given home TAZ were weighted based upon the relative number of t ours For example, if a home zone has 100 tours total and 60 of them are going to destination A with 80% drive alone mode share, 20 to B with 60% drive alone share, 15 to C with 90% drive alone mode share, and 5 to D with 40%

PAGE 22

13 drive alone m ode share; the home TAZ drive alone mode share would be 75.5%, as follows: The following equation for work tours transportation cost is the total sum of each mode share multiplied by the c orresponding mode cos t: The additional annual commuting expense was calculated by subtracting t he baseline annual transportation cost from the scenario with fuel price increase Finally, I calculated the additional percent of income spent on commuting by dividing the additional annual c ommuting expense and the TAZ median of individual annual income. To ass ist in evaluating the results, I assessed the additional percent of income spent on commuting by the amount of standard deviations from the mean for all 2,365 home TAZ zones with the following formula: where: = transportation economic resilient r ating of home TAZ i = 1, 2, . 2365;

PAGE 23

14 = mean of additional percent of income spent on commuting for all i home TAZs ; = additional percent of inc ome spent on commuting for home TAZ i ; and = s tandard deviation of additional percent of income spent on commuting for all i home TAZs Each home TAZ zone then receives a Transportation Econom ic Resilient (TER) r ating based on the number of standard deviations from the mean additional percent of income spent on commuting for all 2,365 home TAZs A lower value of income spent on commuting receives a positive rating (i.e. more resilient) In cont rast, higher values of additio nal income spent on transportation to work tours receive a negative value (i.e. less resilient) Finally, to further investigate the relationship between distan ce, income, mode share, and TER rating; I use d multiple least squares regression to predict TER rating with distance, income, drive alone, share d ride, and transit percentages for the scenario with the fuel price shock

PAGE 24

15 CHAPTER IV RESULTS The initial results for the 2,365 home TAZ zones listing specific characteristics su ch as the number of tours, average distance home to work tour s income, share for each of the seven modes after the fuel price shock, additional annual commuting expenses per capita and additional percent of income spent on commuting are listed in Table IV.1. Table IV 1 Home TAZ summary statistics variables Variable (n=2365) Mean Std. dev. Minimum Maximum # of Tours 488.23 461.11 10 3 950 Average Tour Distance (miles) 22.18 10.23 2.99 83.83 Annual Individual Median Income per Home TAZ ($) 40 02 8 15 196 9 698 112 ,591 Mode Share After Drive Alone 0.679 0.111 0.071 1.000 Share d Ride 2 0.171 0.053 0.000 0.538 Share d Ride 3+ 0.069 0.033 0.000 0.365 Drive to transit 0.014 0.015 0.000 0.167 Walk to Transit 0.029 0.043 0.000 0.455 Walk 0.034 0.095 0.000 0.681 Bike 0.005 0.009 0.000 0.091 Additional Annual Commuting Expenses per Capita ($) 745.84 358.80 62.23 2718.26 Additional Percent of Median Income Spent on Commuting 0.021 0.011 0.001 0.109 In average, each home TAZ generates approximately 488 home to work tours with a distance me an of 22.2 miles per tour (round trip). The minimal number of tours per home TAZ is 10 and a maximum of 3,950 tours. The mean individual median income for all TAZs is approximately $40,000 per year, with a minimum individual medium income of $9, 6 98 and maxim um of $112,591.

PAGE 25

16 After the fuel price shock (gas price doubles), the mean proportion for the seven modes are: 67.9% for drive alone, 17.1% for shared ride 2, 6.9% for shared ride 3+, 1.4% for drive to transit, 2.9% for walk to transit, 3.4% for walking, and 0.5% for bicycling. The commuting expense per capita increase s o n average by a net value of $746 per year. The average TAZ individual c ommuting expense ranges from as low as $62 to as high as $2,718 per year. When the fuel price doubles, t he mean of the additional percent of median income spent on commuting for the 2,365 TAZ home zones is 2 1 % (i.e. o n average individuals spend an extra 2.1% of their income for home to work tours) with a standard deviation of 0.011. Figure IV.1 displays the freque ncy histogram of additional percent of income spent on commuting for the TAZ home zones The bins are grouped by a range of 0.5 standard deviations (0.56%). Figure IV 1 Histogram of additional percent of income spent on comm uting 0 150 300 450 600 750 Frequency Percent of Income Group

PAGE 26

17 Transportation Economic Resilient Rating The Transportation Economic R esilient (TER) rating is generated by normalizing the additional percent of income spent on commuting dataset from a mean value of 2.1% to an expected value of 0 and standard deviation 1 Ratings in increments of 0.5 are shown in Table IV.2 Higher than an additional 2.1% percent of income spent on commuting receives a negative value thus representing a relative lack of resilience; lower than 2.1% percent of income spent on tr ansportation to work tours receive a positive rating which corresponds with relative good resilience The magnitude of the rating changes negatively or positively proportionally to the increase in percent of income based on standard deviation. Table IV 2 Tra nsportation Economic Resilient r ating Additional Percent of Income Spent on Commuting Transport ation Economic Resilient Rating 4.84% 2.5 4.29% 2 3.73% 1.5 3.17% 1 2.61% 0.5 2.06% 0 1.50% 0.5 0.94% 1 0.38% 1.5 Table IV.3 shows several characteristics for the TAZ home zones with the ten worst and best TER scores. For the sake of the shared ride variable, I combined share d ride 2 and share d ride 3 + Drive to transit and walk to transit were also combined into the t ransit variable Home TAZ ID 963 has the lowest TER rating with a value of 7.89. This means that the additional 10.9% of income spent on commuting is 7.89 standard

PAGE 27

18 deviations higher than the mean value of 2.1%. The TER rating for home TAZ ID 1775 is 1.75, and is the highes t of all 2,365 home TAZ zones. On average, a person living in this home TAZ only spent an extra 0.1% percent of income in transportation. T he home TAZ zones with the worst TER ratings (negative values) tended to have higher than average tour distance and /or lower income s Regarding mode share after the fuel price increase, the bottom TER rating home TAZ zones tended to have high drive alone p ercentages as well as low transit, walking, and bi cycling mode shares.

PAGE 28

19 Table IV 3 Ten best and worst home TAZs TER r atings H ome TA Z Tours Average Tour Distance Income Mode Share After Additional Annual Commuting Expenses per Capita Additional % of Income Spent on Commuting TER Rating Drive Alone Share d Ride Transit Walking Biking 1775 25 3.0 $61,133 19.9% 0.0% 4.0% 68.1% 8.0% $62 0.1% 1.75 1830 11 8.7 $37,213 18.2% 0.0% 45.5% 27.3% 9.1% $65 0.2% 1.69 1745 13 5.4 $62,828 38.5% 7.7% 7.7% 46.2% 0.0% $138 0.2% 1.65 1753 48 5.3 $58,066 22.9% 4.2% 12.5% 60.4% 0.0% $139 0.2% 1.63 1726 89 6.3 $44,818 15.6% 12.5% 15.7% 52.8% 3.4% $113 0.3% 1.62 1779 23 8.3 $69,951 21.7% 4.3% 21.7% 52.2% 0.0% $187 0.3% 1.61 1841 47 7.2 $76,288 27.7% 12.8% 10.6% 46.8% 2.1% $203 0.3% 1.61 1809 45 4.8 $40,853 13.3% 11.1% 13.3% 62.2% 0.0% $108 0.3% 1.61 1728 18 5.8 $40,097 16.7% 11.1% 16.7% 50.0% 5.6% $103 0.3% 1.61 1755 33 5.3 $45,331 15.2% 6.1% 12.1% 63.6% 3.0% $124 0.3% 1.60 1957 308 71.4 $29,644 58.6% 40.7% 0.0% 0.7% 0.0% $2,203 7.4% 4.82 884 51 29.0 $12,681 66.4% 33.6% 0.0% 0.0% 0.0% $968 7.6% 5.00 945 88 24.0 $9,698 62.3% 33.1% 2.3% 2.3% 0.0% $748 7.7% 5.07 883 84 31.2 $13,328 67.5% 32.5% 0.0% 0.0% 0.0% $1,089 8.2% 5.48 3 34 77.3 $32,357 85.3% 14.7% 0.0% 0.0% 0.0% $2,718 8.4% 5.69 958 165 26.4 $10,583 71.1% 28.9% 0.0% 0.0% 0.0% $894 8.5% 5.74 995 105 81.8 $29,828 61.1% 38.9% 0.0% 0.0% 0.0% $2,630 8.8% 6.07 996 439 83.8 $30,381 68.7% 31.3% 0.0% 0.0% 0.0% $2,709 8.9% 6.16 2802 196 82.3 $27,566 68.4% 31.6% 0.0% 0.0% 0.0% $2,680 9.7% 6.88 963 73 31.6 $10,176 65.7% 31.6% 1.4% 1.4% 0.0% $1,105 10.9% 7.89 19

PAGE 29

20 Multivariate Regression Table IV.4 presents the results of the multiple least square analysis, including coefficients of each of the variables, standard errors, t statistics, p values, and R 2 As expected, distance is negatively related to home TAZ TER rating and income is positively related The negative and significant coefficient for d istance ( 0.074) indicates that home TAZ zones with a higher commut e distance are exp ected to have lower TER ratings. Income has a positive and significant relationship with the TER rating, with a c oefficient of 3.97x10 5 Share d ride percentage and transit percentage for the scenario with a fuel price shock are also statistical significant va riables. The positive coefficient for transit percentage indicates that a 10% increase in transit share is associated with a n additional 0.23 in TER rating when adjusting for distance, income, drive alone, and share d ride percentages. It is interesting to note the coefficient of drive alone percentage is not significant. This suggests that distance to work, income, and the availability of transit impact resilience more than what can be garnered from current driving mode share. T a ble IV 4 T ER rating regression r esults Variables Coefficients Std. error t stat p value (Constant) 0.13707 0.10313 1.33 0.184 Distance 0.07444 0.00100 74.34 0.000 Income 3.97E 05 5.80E 07 68.53 0.000 Drive Alone Percentage 0.13440 0.09999 1.34 0.179 Share d Ride Percentage 0.80231 0.14252 5.63 0.000 Transit Percentage 2.27818 0.28143 8.09 0.000 R 2 0.8522 Observations 2365

PAGE 30

21 Figure IV. 2 presents individual relationships between home TAZ TER ratings and each of the following variables: average tour home distance, income, and transit mode share after. The figure highlights t he association of the three variables with respect to the TER rati ngs. Figure IV.3 presents the distribution of TER ratings for 2,365 TAZ home zones. Results from the regression analysis suggest that longer commuting distances and low mode share for transit contribute to a negative TER value. From this figure, we see tha t lowest ratings are among home zones on the outskirts of the DRCOG region. Figure IV.4 depicts home TAZ zones located along the Denver Boulder corridor in more detail. This illustrates the positive TER ratings with significant higher values arw within t he city of Denver and the city of Boulder. A final zoom into the Denver central business district (CBD) reiterates this result. Home TAZ with TER ratings higher than 1.50 are all located inside the CBD. These are shown in Figure IV.5. In contrast with the home TAZs located on the perimeter, these high rating TER home TAZ zones seem to be the result of one or more of the following: low commuting distance, higher income, and/or higher transit, walking, and biking mode shares.

PAGE 31

22 Figure IV 2 TER r ating s and selected v ariables for 2,365 h ome TAZs -8 -6 -4 -2 0 2 0 15 30 45 60 75 90 TER Rating Average TAZ Tour Distance (miles) -8 -6 -4 -2 0 2 $0 $30,000 $60,000 $90,000 $120,000 TER Rating TAZ Individual Median Income -8 -6 -4 -2 0 2 0% 10% 20% 30% 40% 50% TER Rating Transit Mode Share After

PAGE 32

23 Figure IV 3 TER ratings for 2,365 home TAZs 23

PAGE 33

24 Figure IV 4 T ER r atings for h ome TAZ s located in the Denver Boulder corridor 24

PAGE 34

25 Figure IV 5 TER ratings for home TAZs located in CBD and surrounding are a. 25

PAGE 35

26 CHAPTER V DISCUSSION AND CONCLUSION Transportation cost is the second highest expenditure in a typical U.S. budget. With the high dependency on the automobile, a large portion of the variable cost fluctuates with fuel price volatility Communities are vulnerable to unexpected events includi ng the possibility of fuel price shocks. This study income for commuting related transportation expenses To achieve this research goal, I examined a database of 1,154,673 individual wor k tours from DRCOG grouped by home TAZ I developed a transportation economic resilient (TER) rating to evaluate 2,365 home TAZ to measure the additional commuting expense with respect to income when fuel price doubles. I used a multiple least square regre ssion t o estimate the relationships b etween the developed TER rating and work tour distance, income, and mode share variables. The average percent of income spent on commuting, as a result of a doubling of the fuel price, increases by 2.1%. While this may seem like a relatively low number, the results also suggested large disparities among home TAZs with some reaching almost 11% and others at only 0.1%. The resulting TER ratings ranged from 7.89 to +1.75. The worst TER ratings are home TAZs located on the outwards limits of the geographical area analyzed, and the best ratings are closer to downtown Denver. Commuting distance, income, shared ride percentage, and transit percentage are all statistically significant variables to determine TER ratings.

PAGE 36

27 Home TAZ s whose people commute short distances are very likely to be economical ly re silient to fuel price increases. This is because the current cost of commuting is low, and there also tend to be transportation options available such as walking and biking High i ncome individuals are also considered transportation economic ally resilient since their income would not be impacted as much by an increase in commuting expenditures. In contrast, i ndividuals in low income home TAZs that have to drive alone and /or travel long distance s for work tours would be the most impacted by a fuel price increase. Based on the TER rating, these areas are not considered economical ly resilient to fuel price shocks A person living in these area s would still need to travel to wor k and without realistic transportation options the ec onomical impact can be very serious. When the commute distance is long, h ome TAZ zones with current high mode shares for carpooling and transit or the opportunity to shift to h igh share percentage in these modes have a very high probability of being transportation economic ally resilient. For example, comparing a home TAZ with a transit share 30% higher than another TAZ the TER rating is expected to experience a 0.68 higher TER rating when controllin g for other variables via the regression equation Figures IV.3, IV.4, and IV.5 present the resilient (green TAZ) and non resilient (red) home TAZs for the area studied. A reas with a high density of jobs such as Boulder, the Denver CBD, and the Denver Te ch Center tend to be very resilient. Areas with good transit service such as the Denver Boulder 36 corridor and along the light rail lines also tend to be more resilient. The surrounding suburbs with less transit availability tend to be less resilien t.

PAGE 37

28 These findings introduce a tra nsportation economic resilient index and the concept of significant fuel price shocks on transportation expenditures This work el aborate s on earlier studies regarding the impact of fuel price fluctuations on travel behavi or. In particular, this paper examines regional disparities with respect to preparedness to a major shift in gas prices for commuting expenses Since the focus is on a drastic increase in fuel price, these results are di fferent from elasticity studies that have considered how relative minor gas price adjustments have impacted travel behavior. This paper also build s on studies related to trans portation cost at specific area s For example, CNT ( 2010 ) developed a Housing + Transportation Affordability Index, which measure the cost of housing and transportation at the neighborhood level. The findings of this study help increase our understanding of transportation economic resilience with respect to fuel price shocks and how commute distance, income, and transit service affect different communities. T hese results have policy implications for automobiles (e.g. electric cars, fuel efficiency); land use (e.g. d evelopment of mixed use, pedestrian, bicycling, a n d transit oriented communities); and in facilitat ing other modes of transportation by investing i n transit infrastructure or other transportation options This research also offer s a better understanding of the risks mitigated by a diverse transportation infrastructure and can help shape more resilient and more livable cities. Such findings can support the developme nt of regional resilience plans and lead to research on related topics.

PAGE 38

29 REFERENCES Anas, A. (2007) 'A unified theory of consumption, travel and trip chaining', Journal of urban economics. 62(2), pp. 162 APTA (2012) 'Public Transportation Protects Americans From Gas Price Volatility', American Public Transportation Association Berdica, K. (2002), 'An Introduction to Road Vulnerability: What Has Been Done, Is Done and Should Be Done', Transport Policy 9(2): 117 127. of Transportation Assets Resulting from Natural Disaster or Terrorist Attacks', National Surface Transportation Policy and Revenue Study Commission Washington, D.C. BLS (2012) 'Economic News Release: Consumer Expenditure', U.S. Department of Labor: Bureau of Labor Statistics Washington, D.C. http://www.bls.gov/news.release/cesan.nr0.htm Briguglio, L., Cordina, G., Farrugia, N., Vella, S. (2009) 'Economic Vulnerability and Resilience: Concepts and Measurements', Oxford Development Studies 37(3), pp. 229 247 Bruneau, M., Chang, S., et al. (2003), 'A Framework to Quantitatively Assess and E nhance the Seismic Resilience of Communities', Earthquake Spectra 19(4): 733 752. Chang, S., Nojima, N. (2001), 'Measuring Post Disaster Transportation System Performance: The 1995 Kobe Earthquake in Comparative Perspective', Transportation Research Part A 35(6): 475 494. CNU (2010) 'H+T Affordability Index', Center for Neigborhood Technology Chicago, IL. http://htaindex.cnt.org/ Cooper, M. (2005) 'The Impact of Rising Prices on Household Gasoline Expenditures' Consumer Federation of America manuscript Cova, T. J., Conger, S. (2004), 'Transportation Hazards', Handbook of Transportation Engineering M. Kutz. McGraw Hill: 17.11 17.24, Washington, D.C. Currie, G., Phung, J. (2007) 'Transit Ridership, Auto Gas Prices, and World Events: New Drivers of Change?', Transportation Research Record: Journal of the Transportation Research Board 1992( 1), pp. 3 10

PAGE 39

30 Dahl, C., Sterner, T. ( 1991 Analyzing gasoline demand elasticities, a survey Energy Econ 3 (13), 203 210. Dargay, J., Gately, D. (1997) 'The demand for transportation fuels: Imperfect price reversibility?', Transportation Research Part B: Methodological Transportation Research Part B: Methodologica l 31(1), pp. 71 82 Dodson, J., Sipe, N. (2007) 'Oil Vulnerability in the Australian City: Assessing Socioeconomic Risks from Higher Urban Fuel Prices', Urban Studies 44(1), pp. 37 62 DRCOG (2010a) 'Front Range Travel Counts', Denver Regional Council of G overments Denver, CO. http://www.drcog.org/index.cfm?page=FrontRangeTravelCounts DRCOG (2010b) 'Transportation Analysis Zones', Denver Regional Council of Governments Denver, CO. http://gis.drcog.org/datacatalog/content/transportation analysis zones taz 2832 Espey, M. (1 998 Gasoline demand revisited: an international meta analysis of elasticities Energy Econ 20, 273 295. Ferdous, N., Pinjari, A., Bhat, C., Pendyala, R. (2010) 'A comprehensive analysis of household transportation expenditures relative to other goods and services: an application to United States consumer expenditure data', Transportation 37(3), pp. 363 390 Foster, H. (1995) 'Disaster Mitigation: The Role of Resilience', Proceedings of a Tri lateral Workshop on Natural Hazards D. Etkin. Ontario, Canada, Merrickville: 93 108. Gicheva, D., Hastings, J.S., Villas Boas, S., National Bureau of Economic, R. (2007) 'Revisiting the income effect gasoline prices and grocery purchases', National Bureau of Economic Research, Cambridge, Mass. http://papers.nber.org/papers/w13614 Haire, A., Machemehl, R. (2007) 'Impact of Rising Fuel Prices on U.S. Transit Ridership', Transportation Research Record: Journal of the Transportation Research Board 1992( 1), pp. 11 19 Hamilton, J.D. (2009) 'Causes and Consequences of the Oil Shock of 2007 08', Brookings Papers on Economic Activity pp. 215 261 Heaslip, K., Louisell, W., Collura, J. (2009) 'A Methodology to Evaluate Transportation Resiliency for Regional Networks', Trans portation Research Board 88th Annual Meeting Washington DC.

PAGE 40

31 Heaslip, K., W. Louisell, et al. (2010), 'A Sketch Level Method for Assessing Transportation Network Resiliency to Natural Disasters and Man Made Events', Transportation Research Board Washingt on, D.C. Honan, E. (2012) 'Fuel scarce, East Coast struggles to recover from storm', Reuters New York. Hughes, J.E., Knittel, C.R., Sperling, D., Frank, E. (2008) 'Evidence of a shift in the short run price elasticity of gasoline demand', The Energy Jour nal Husdal, J. (2004), 'Why Reliability And Vulnerability Should Be An Issue In Road Development Projects', Samferdsel: Journal of the Norwegian Institute of Transport Economics (3). Lane, B.W. (2010) 'The relationship between recent gasoline price fluc tuations and transit ridership in major US cities', J. Transp. Geogr. Journal of Transport Geography 18(2), pp. 214 225 Lin, C.Y.C., Prince, L. (2013) 'Gasoline price volatility and the elasticity of demand for gasoline', Energy Economics 38(0), pp. 111 117 Maley, D., Weinberger, R. (2009) 'Rising Gas Price and Transit Ridership', Transportation Research Record: Journal of the Transportation Research Board 2139( 1), pp. 183 188 Martin, W.A., McGuckin, N.A. (1998) Report 365: Travel Estimation Techniques for Urban Planning National Academy Press, Washington D.C. Motte Baumvol, B., Massot, M. H.l.n., Byrd, A.M. (2010) 'Escaping Car Dependence in the Outer Suburbs of Paris', Urban Studies 47(3), pp. 604 619 Mouawad, J., Romero, S. (2005) 'Gas Prices Surge as Supply Drops', The New York Times Murray Tuite, P. (2006), 'A Comparison of Transportation Network Resilience Under Simulated System Optimum and User Equilibrium Conditions', Winter Simulation Conference Monterey, CA. Nicol, C.J. (2003) 'Elasticities of demand for gasoline in Canada and the United States', Energy economics. 25(2), pp. 201 214 Pelling, M. (2003), 'The Vulnerability of Cities: Natural Disasters and Social Resilience', Earthscan Publications London. Pisarski, A., Terra, N. (1975) 'American and European transpo rtation responses to the 1973 74 oil embargo', Transportation 4(3), pp. 291 312

PAGE 41

32 Puller, S.L., Greening, L.A. (1999) 'Household adjustment to gasoline price change: an anal ysis using 9 years of US survey data', Energy economics. 21(1), pp. 37 Shafiee, S., Topal, E. (2010) 'A long term view of worldwide fossil fuel prices', Applied energy. 87(3), pp. 988 1000 ortation Network of Santo Transportation Research Record 2234. Simmons, M.R. (2005) Twilight in the desert : the coming Saudi oil shock and the world economy John Wiley & Sons, Hoboken, N.J. Small, K.A., Van Dender, K. (2007) 'Fuel Efficiency and Motor Vehicle Travel: The Declining Rebound Effect', The energy journal / 28(1), pp. 25 Yang, D., Timmermans, H. (2011) 'Effects of Energy Price Fluctuation on Car Based Individual Activity travel Be havior', SBSPRO Procedia Social and Behavioral Sciences 20(pp. 547 557 Zheng, J., Atkinson Palombo, C., McCahill, C., O'Hara, R., Garrick, N. (2011) 'Quantifying the Economic Domain of Transportation Sustainability', Transportation Research Record: Jour nal of the Transportation Research Board 2242( 1), pp. 19 28