<%BANNER%>

Comparison of Socioeconomic Impacts of Market-Based Instruments for Mobility Management under Uncertainty

Permanent Link: http://ufdc.ufl.edu/UFE0045459/00001

Material Information

Title: Comparison of Socioeconomic Impacts of Market-Based Instruments for Mobility Management under Uncertainty
Physical Description: 1 online resource (51 p.)
Language: english
Creator: Bekoe, Patrick A
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2013

Subjects

Subjects / Keywords: gasoline-tax -- market-based-instruments -- mileage-fee -- mobility-management -- tax -- tradable-credit -- transportation-economics
Economics -- Dissertations, Academic -- UF
Genre: Economics thesis, M.A.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: This study compared the socioeconomic impacts of three market-based instruments: gasoline tax, mileage fee and tradable credit schemes, for mobility management. The National Household Travel Survey (NHTS) data for 2009 was used for the analysis. A hypothetical case study where FDOT intends to reduce the total vehicle miles by 15% in Florida was targeted. A deterministic household travel demand function was developed and used to examine the socioeconomic impact of the three market-based instruments. It was found that all three instruments are capable of achieving the hypothetical 15%reduction of total travel demand in Florida. However, they generate different amounts of revenue and impose different socioeconomic impacts on Florida residents. Gasoline tax and mileage fee schemes charge travelers more to discourage their travels to achieve the control target. Consequently, the government receives much more revenue. At the same time, the schemes hurt residents more and do more harm to the poor than the rich. The tradable credit scheme generates the least revenue for the government but has a less regressive impact on residents. Secondly,we assumed that the household travel demand function is uncertain and conducted similar analysis to determine the socioeconomic impact of the three instruments. The three policy instruments were all capable of reducing the VMT by 15% on average. However, the gasoline tax and mileage fee policies were found to have a lower success rate and the revenue generated was variable. For the tradable credit scheme, the success rate was found to be 100% and the revenue generated fixed; the socio economic impact was found to be similar to that obtained under the deterministic travel demand.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Patrick A Bekoe.
Thesis: Thesis (M.A.)--University of Florida, 2013.
Local: Adviser: Slutsky, Steven M.
Local: Co-adviser: Yin, Yafeng.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2013
System ID: UFE0045459:00001

Permanent Link: http://ufdc.ufl.edu/UFE0045459/00001

Material Information

Title: Comparison of Socioeconomic Impacts of Market-Based Instruments for Mobility Management under Uncertainty
Physical Description: 1 online resource (51 p.)
Language: english
Creator: Bekoe, Patrick A
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2013

Subjects

Subjects / Keywords: gasoline-tax -- market-based-instruments -- mileage-fee -- mobility-management -- tax -- tradable-credit -- transportation-economics
Economics -- Dissertations, Academic -- UF
Genre: Economics thesis, M.A.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: This study compared the socioeconomic impacts of three market-based instruments: gasoline tax, mileage fee and tradable credit schemes, for mobility management. The National Household Travel Survey (NHTS) data for 2009 was used for the analysis. A hypothetical case study where FDOT intends to reduce the total vehicle miles by 15% in Florida was targeted. A deterministic household travel demand function was developed and used to examine the socioeconomic impact of the three market-based instruments. It was found that all three instruments are capable of achieving the hypothetical 15%reduction of total travel demand in Florida. However, they generate different amounts of revenue and impose different socioeconomic impacts on Florida residents. Gasoline tax and mileage fee schemes charge travelers more to discourage their travels to achieve the control target. Consequently, the government receives much more revenue. At the same time, the schemes hurt residents more and do more harm to the poor than the rich. The tradable credit scheme generates the least revenue for the government but has a less regressive impact on residents. Secondly,we assumed that the household travel demand function is uncertain and conducted similar analysis to determine the socioeconomic impact of the three instruments. The three policy instruments were all capable of reducing the VMT by 15% on average. However, the gasoline tax and mileage fee policies were found to have a lower success rate and the revenue generated was variable. For the tradable credit scheme, the success rate was found to be 100% and the revenue generated fixed; the socio economic impact was found to be similar to that obtained under the deterministic travel demand.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Patrick A Bekoe.
Thesis: Thesis (M.A.)--University of Florida, 2013.
Local: Adviser: Slutsky, Steven M.
Local: Co-adviser: Yin, Yafeng.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2013
System ID: UFE0045459:00001


This item has the following downloads:


Full Text

PAGE 1

1 COMPARISON OF SOCIOECONOMIC IMPACTS OF MARKET BASED INSTRUMENTS FOR MOBILITY MANAGEMENT UNDER UNCERTAINTY By PATRICK AMOAH BEKOE A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS UNIVERSITY OF FLORIDA 2013

PAGE 2

2 2013 Patrick Amoah Bekoe

PAGE 3

3 To my Mom, the Late Bernice Oye Sakyi, who first taught me the basic principles of economics

PAGE 4

4 ACKNOWLEDGMENTS To Go d be the Glory for great things He has done. I owe all I am and have achieved to the Almighty God who has been my buckler and sword during my adventure to pursue a master of arts in economics alongside my PhD in civil engineering. My deepest a ppreciation goes to my committee chair, Prof. Steven M. Slutsky and co chair Prof. Yafeng Yin for their remarkable guidance and support to bring this dream to fruition. Mention is made of my nonpareil supervisor, Prof Mang Tia, who agreed to all ow me to pursue a du al degree Prof Tia I thank God for your life; you have been a great blessing to me and my family. To the staff in Economics Department Shawn Lee and Martha Shaw I am thankful to you for your assistance, support and deep sense of du ty. To the staff at the graduate school, school of engineering and business school I say a big thank you to you all. I also want to thank my friend and mentor Mr. Robert Nii Ayettey who advis ed me to pursue this program if I ever get the opportunity My dearest wife h as been supportive during this period honey thanks for always been there I will always love you. To my friends who encouraged, advised, supported and lifted my soul when you heard about this adventur e, I owe all of you I big thank you. Fina lly I want to thank my late mom, Bernice Oye Sakyi, who t aught me the basics of economics when I was young Sisi, your effort has not been in vain, God bless you and keep you till we meet again, I love you.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ................................ ........................... 7 LIST OF FIGURES ................................ ................................ ................................ ......................... 8 ABSTRACT ................................ ................................ ................................ ................................ ..... 9 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .................. 11 1.1 Pr oblem Statement ................................ ................................ ................................ ............ 11 1.2 Objective ................................ ................................ ................................ ........................... 12 1.3 Outline of Research ................................ ................................ ................................ .......... 12 2 LI TERATURE REVIEW ................................ ................................ ................................ ....... 13 3 METHODOLOGY ................................ ................................ ................................ ................. 17 4 DATA AND MODEL CALIBRATION ................................ ................................ ................ 21 4.0 Data Description ................................ ................................ ................................ ............... 21 4.1 Model Calibration ................................ ................................ ................................ ............. 21 5 IMPACT ANALYSIS ................................ ................................ ................................ ............ 25 5.1 Comparison of Schemes under Deterministic Household Travel Demand ...................... 25 5.1.1 Gasoline Tax ................................ ................................ ................................ ..... 25 5.1.2 Mileage Fee ................................ ................................ ................................ ...... 26 5.1.3 Mileage Fee of 1.61 Cents/Mile (Revenue Neutral Fee) ................................ 27 5.1.4 Mileage Fee of 3.70 Cents/Mile ................................ ................................ ....... 27 5.1.5 Tradable Credit ................................ ................................ ................................ 28 5.1.6 Overall Comparison under Deterministic Household Travel Demand ............ 29 5.2 Comparison of Schemes under Stochastic Travel Demand Function .............................. 30 5.2.1 Reduction of VMT under Uncertainty ................................ .............................. 31 5.2.2 Price Variation under Uncertainty ................................ ................................ .... 31 5.2.2 Changes in Socio Economic Indicators under Uncertainty .............................. 31 6 SUMMARY AND CONCLUSIONS ................................ ................................ ..................... 34 APPENDIX A MATLAB PROGRAM FOR DETERMINISTIC ANALYSIS ................................ ............. 36

PAGE 6

6 B MATLAB PROGRAM FOR STOCHASTIC ANALYSIS ................................ ................... 40 LIST OF REFERENCES ................................ ................................ ................................ ............... 48 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ......... 51

PAGE 7

7 LIST OF TABLES Table page 4 1 Descriptive Statistics by Income Group ................................ ................................ ............ 22 4 2 Estimated Model ................................ ................................ ................................ ................ 23 4 3 Elasticity by Income Group Based on Average Income ................................ .................... 24 5 1 ine Tax) ................................ ................................ ................................ ................................ .... 26 5 2 Mileage Fees ................................ ................................ ................................ ...................... 27 5 3 Fee=1.61cents/mile) ................................ ................................ ................................ ........... 27 5 4 Fee=3.70 cents/mile) ................................ ................................ ................................ .......... 28 5 5 Tradable Credit ) ................................ ................................ ................................ ................................ 29 5 6 Instruments (Deterministic) ................................ ................................ ............................... 30 5 7 ................... 30 5 8 Instruments (Stochastic) ................................ ................................ ................................ ..... 32

PAGE 8

8 LIST OF FIGURES Figure page 3 1 ................................ .. 20 5 1 Changes in total VMT reduction with number samples for different policies. ................. 32 5 2 .......... 32 5 3 Changes in total revenue with number samples for different policies. .............................. 33 5 4 Changes in total social welfare with number samples for different policies. .................... 33

PAGE 9

9 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Arts COMPARISON OF SOCIOECONOMIC IMPACTS OF MARKET BASED INSTRUMENTS FOR MOBILITY MANAGEMENT UNDER UNCERTAINTY By Patrick Amoah Bekoe May 2013 Chair: Steven Slutsky Cochair: Yafeng Yin Major: Economics This study compared the socioeconomic impacts of three market based instruments: gasoline tax, mileage fee and tradable credit schemes, for mobility management. The National Household Travel Survey ( NHTS ) data for 2009 was used for the analysis. A hypothetical case study where FDOT intends to reduce the total vehicle miles by 15% in Florida was targeted. A deterministic household travel demand function was developed and used to exam ine the socioeconomic impact of the three market based instruments. It was found that all three instruments are capable of achieving the hypothetical 15% reduction of total travel demand in Florida. However, they generate different amounts of r evenue and impose different socioeconomic impacts on Florida residents. Gasoline tax and mileage fee schemes charge travelers more to discourage their travels to achieve the control target. Consequently, the government receives much more revenue. At the sa me time, the schemes hurt residents more and do more harm to the poor than the rich. The tradable credit scheme generates the least revenue for the government but has a less regressive impact on residen ts Secondly, we assumed that the household travel de mand function is uncertain and conducted similar analysis to determine the socioeconomic impact of the three instruments. The

PAGE 10

10 three policy instruments were all capable of reducing the VMT by 15% on average However, the gasoline tax and mileage fee policie s were found to have a lower success rate and the revenue generated was variable. For the tradable credit scheme, the success rate was found to be 100% a nd the revenue generated fixed; t he socio economic impact was found to be similar to that obtained unde r the deterministic travel demand.

PAGE 11

11 CHAPTER 1 INTRODUCTION 1.1 Problem Statement Traffic congestion is perhaps one of the most severe problems threatening the economic and social wellbeing of many societies. According to the Texas Transportation Institute (TTI), in 2011, congestion caused urban Americans to travel an extra 5.5 b illion hours and to purchase an extra 2.9 billion gallons of fuel amounting to a total congestion cost of $121 billion. If this trend is not curbed, it is estimated that by 2020, travelers will be delayed an extra 8.4 billion hours and purchase an extra 4.5 billion gallons amounting to a tot al congestion cost of $ 199 billion. To avert this situation from degenerating, transportation planners and engineers need a concerted effort. To this end, the Federal Highway Administration (FHWA) has been embarking on different strategies to reduce t improving services on existing roads, p ricing, adding capacity, better work zones, travel options, and traveler information. This research focuses on reducing congestion using pricing scheme s. Congestion pricing also called value pricing is a way of harnessing the power of the market to reduce waste associated with traffic congestion (FHWA, 2006). There are four main types of pricing strategies that the FHWA have adopted, namely, variably pr iced lanes, variable tolls on entire roadways, cordon charges, and area wide charges. The variably priced lanes includes express tolls and High Occupancy Toll (HOT) and the latter involves charging low occupancy vehicles tolls for u sing HOT lanes while high occupancy vehicles (HOVs), public transit buses and emergency vehicles are allowed to use the HOT lanes free of charge or at a reduced rate. The vari able tolls on entire roadways involve changing a flat toll rate on existing roads to a variable toll schedule such that tolls are higher during peak travel

PAGE 12

12 hours and lower during off peak or shoulder hours. Cordon pricing involves charging a fee to enter a congested area. The a rea wide pricing scheme (al so called the mileage fee pricing scheme) involves per mile charges Other methods of reducing congestion are by increasing fuel or gasoline charges and by using tradable credits. The gasoline charges, mileage fee, and tradable credits are generally marke t based instruments. This research compares the effectiveness of the se market based instruments in reducing the total travel demand of travelers in Florida State by 15% The comparison is done by assuming a deterministic household travel demand followed by a stochastic household travel demand. 1.2 Objective The main objectives of this research are to determine: The social economic impacts of market based mobility management instruments: i.e. gasoline tax, mileage fee and tradable credits in regulati ng Vehicle Miles Travel (VMT) in Florida. To determine the most effective market based mobility management instrument in regulating VMT when the individual household demand functions are assumed to be stochastic 1.3 Outline of Research To achieve the above o bjectives, Chapter 2 present s a literature review on market based mobility management instruments in curbing congestion Chapter 3 gives a detailed methodology, Chapter 4 gives the details of the data and model calibration for our analysis, Chapter 5 prese nts the impact analysis, and Chapter 6 gives the conclusion s and recommendation s

PAGE 13

13 CHAPTER 2 LITERATURE REVIEW There are generally two approaches transportation planners/engineers employ to ameliorate congestion : increasing road capacity (supply) or reducing traffic (demand). The Antony (1962, 1992), increasing capacities usually induce new demands; which ultimately re verts the highway to its originally congested condition. In view of this, current focus has been on the demand m anagement of traffic congestion ; th is involve s using market based instruments for congestion mitigation. Broadly, market based instruments can b e classified into two classes, i.e. price and quantity based. The price based i.e. congestion pricing which forms the basis of the seminal work by Pigou (1920), charges vehicles using congested roads to bear a tax equal to the difference between marginal social and marginal private cost involved. Economic theory suggests congestion pricing as an efficient pricing strategy that requires the users to pay more for a public good, thus increasing the welfare gain or net benefit for society. The main idea is t o charge travelers the marginal external costs that their trips impose to the society to reduce traffic congestion or increase social welfare Policy makers have used several pricing strategies in curbing congestion. The most commonly used is the gasoline tax, which serves both as a road user fee and a means to ameliorate congestion. Morrison ( 1986 ) and Small et al. ( 1989) provide an extensive literature review on optimal road user fees both under congested and uncongested circumstances. It must be noted t hat r oad maintenance cost is the primary component of road user fees while the congestion cost is a results of flow exceeding capacity ( Small et al. ( 1989 ) ). Increasing the g asoline tax is mostly unappealing given how s ociety sees the gasoline tax and thus it receives a lot

PAGE 14

14 of resistance in implementation. The resistance has been mainly focused on perceived inequality or unfairness of the tax. The Federal Highway Administration (FHWA) has been focusing on high priority effort s to help reduce congestion in the United States There are four main types of pricing strategies that the FHWA ha s adopted, namely, variably priced lanes, variable tolls on entire roadways, cordon charges, and area wide charges. Th e variably priced lanes include express tolls and High Occu pancy Toll (HOT) and the later involve charging low occupancy vehicles tolls for using HOT lanes while high occupancy vehicles (HOVs), public transit buses and emergency vehicles are allowed to use the HOT lanes free of charge o r at a reduced rate. The variable toll s on entire roadways is involve changing a flat toll rate on existing roads to a variable toll schedule such that tolls are higher during peak travel hours and lower during off peak or shoulder hours. Cordon pricing in volves charging a fee to enter a congested area. Places that have operationalized such pricing schemes include Singapore in 1975, Central London in 2003, and central Stoc kholm on a trial basis in 2006. The a rea wide pricing scheme (also called the mileage fee pricing scheme) involves per mile charges In the United Sates, there is growing push for the implementation of the mileage fee policy mainly because of the low revenue generated from the fuel tax to fund road projects Oregon tried the mileage fee pol icy in 2003, under this scheme truck operators reported their in and out of state mileage and they were exempted from the state fuel tax (TRB Committee for the Study of Long Term Viability of Fuel Ta xes for Transportation Finance ( 2006 ) ). Other studies tha t focused on Oregon include Whitty and Imholt (2005), Whitty et al. (2006), and Zhang et al. (2009) Due to the perception of inequity associated with congestion pricing, attempts have been made to develop a more equitable pricing scheme. Major proposals have been on developing an

PAGE 15

15 appealing Pareto improving congestion pricing with revenue redistribution schemes. Key researchers that have worked in this area include Lawphonpanich and Yin (2010), Song et al., (2009), Lui et al. (2009), Nie and Liu (2010), an d Guo and Yang (2010). Despite these proposals, there are still resistances since the government is seen as an objectionable tax collector, and proving their revenue neutrality is difficult for people to believe. In view of this, focus is now turning to the demand management of congestion. In this proposition, the government fixes the quantity of travel demand, and then assigns mobility rights equally to all individual travelers or inhabitants ensuring that equity is revealed. The two common forms of quan tity based instruments are the road space rationing and cap and trade schemes. In the road space rationing, the government restricts private cars from using the road network on certain days with the aim of ensuring fairness and reducing congestion. Example of cities that have implemented such a scheme is Mexico and Sao Paulo w h ere the number of vehicles is controlled through plate number based space rationing. In this system, the quantity of cars on the road is controlled by the authority allowing certain number plates to use the road facility on a specified day. However this scheme has been found to be short lived, has a perverse incentive of second car ownership, and has been proven to be unsustainable (Davis ( 2008 ); Mahendra ( 2008 ); Wang et al. ( 2010 ) ) In the cap and trade schemes i.e. tradable permit schemes and tradable credit schemes, the state agency or government distribute a specified quantity credits (cap) by first selling them to potential travelers. Thereafter, credits can be traded among travelers and the price is determined by the market through free trading. By deciding the initial credit distribution and the subsequent credit charging scheme, the agenc y can achieve its policy goal. Such a scheme has been recently studied by Yang and Wang (2011), Wang et al. (2012), Goddard ( 1997) Verhoef

PAGE 16

16 et al. (1997) Viegas (2001), Kockelman and Kalmanje (2005) and Raux (2007) Tradable credit schemes offer several advantages. First, the credit market allows those who value travel time savings less to be directly compensated by selling credits to those who value them more. This mechanism promises simpler and fairer distribution of the benefits from congestion reducti on. Secondly, as no transfer of wealth takes places between travelers and the authority, the payment made to acquire credits is less likely to be perceived as a tax. Finally, when justified, the welfare effects of the schemes on individuals may be controll ed by the way the credits are distributed.

PAGE 17

17 CHAPTER 3 METHODOLOGY To achieve the objective s of this study, a hypothetical sc enario where Florida DOT plans to reduc e the total vehicle miles traveled by Florida residents by 15% while maintaining at least the current level of revenue from the existing state gasoline tax is considered The 2009 NHTS data for Florida was used for the analysis; it was first cleaned and then a travel demand function which relates the total annual VMT by a household to its social characteristics the attributes of transportation services and the cost/price of travel was established with the aid of Statistical Product and Service Solutions ( SPSS ) version 17 software. Thereafter, a scenario in controlling th e VMT by implementing three policy instruments i.e. increasing state gasoline tax, replacing the state tax with mileage fees and the use of tradable credit scheme was considered under a deterministic household travel demand function The socioeconomic impl ications of these policy instruments were then accessed. A similar analysis was further done considering that the individual travel demand functions are stochastic and FDOT still desires to reduce the total VMT by 15%. Implementation of the socioeconomic i ndicators was all done with the aid of MATLAB version R2011b software The i mplementation of the state gasoline tax and replacement of the state tax with mileage fees are fairly straightforward However, the use of a tradable credit scheme requires further ex planation The f ollowing is an explanation of how the tradable credit scheme was implemented. It is first assumed that the state agency first allocates travel credits equiv alent to 85% of the current VMT to eligibl e individual households and then collects one credit for each mile traveled. To maintain the current level of revenue, the initial price of the credit is equal to the current revenue divided by 85% of the current VMT. Subsequently, the state agency will cr eate a credit market that enables credit trading among households If a household travel need is

PAGE 18

18 more than the credits allocated, they will purchase more from the market or give up their travel. Similarly, households with excess credits will either sell th eir credits in exchange for money or give them to environmental agencies. It must be noted that during this stage the government or stage agency do es not interfere with the credit market but only act as a manager in monitoring the system. It is a ssumed th at the transaction cost is negligible; thus the price of the credit is mainly determined by the supply and de mand of credits on the market (Yang and Wang, ( 2011 ) ). Base d on the travel demand fu nction, the changes in consumer s surplus, revenue and social welfare across income groups was estimated The and social welfare are estimated approximately as follows: (For Gas Tax) (For Mileage Tax) Where, = Change in Revenue ($) = Change in Social Welfare ($) =Current equivalent price or cost of travel $/mile =New equivalent price or cost of travel $/mile =Annual miles driven by household under current price (mile) =Annual miles driven by household under new price (mile) =Current state (and local) gasoline tax in $/gallon =New state (and local) gasoline tax in $/gallon

PAGE 19

19 =Average miles per gallon in mile/gallon = The mileage fee in $/mile For the tradable scheme, the change in consumer surplus is explained in the following example. Assuming in a mileage fee policy a flat fee of 3 7 cents/mile is required to achieve a 15 % VMT reduction, then using Figure 1(a), the area abcd surplus where 1.6 cents/mile is the current equivalent travel price. Suppose government intends to distribute credits per households at a price of 1 7 cents/credit to maintain the current revenue, and the market clearing price for each credit is 3 7 cents/credit (this is determined from the market demand function i.e. the sum of the individual household demand functions). Now, assuming each household is allocated 10 00 credits and considering a household whose demand is 8 00 vehicle miles at 3 7 cents/mile. The surplus this household receives from travel in this case is shown by area oefc in Figure 1(b). Additionally, the household sells the extra 200 cred its to the market and earns an income of $4. 0 (i.e. (0.0 37 0.017 )x 200). Therefore, the change in der the credit scheme will be Area oefc + $4.4 Area oab For a household with a demand of more than 10 00 vehicle mil es, it will be better off similarly to the forgoing with the only difference been that it incurs additional credit expense instead of earnings. The calculation o f the socioeconomic measures under uncertainty in the travel demand function follows a similar procedure. However, for the tradable credit scheme, it must be noted that the market clearing price changes a t each sample. Details are provided in Chapter 5

PAGE 20

20 (a) (b) Figure 3 1.

PAGE 21

21 CHAPTER 4 DATA AND MODEL CALIBRATION 4 .0 Data D escription The 2009 NHTS data for Florida was used for this study It consists of four files: household, person, trip and vehicle files i.e. 15,884 household entries, 30,952 person entries, 114,910 trip entries and 29,457 vehicle entries in the dataset. Analysis was carried out at the household le vel; therefore, some attributes from the vehicle and person data files had to be integrated into the household data file which resulted in a total of 13,086 household data used for the analysis. The original data set does not provide specific income value s thus, an averag ing value of income range of each income category was used for the model calibration and analysis. The BESTMILE (best estimate of annual miles) and EIA fuel efficiency (measures from Energy Information Administration) were used instead of ANNMILES (self reported annualized mile estimate) and EPA (Environmental Protection Agency) fuel efficiency. Mamun ( 2012 ) provides details of the data cleaning in his dissertation. Following the data cleaning, a statistical analysis was run with the aid of SPSS version 17 statistical software Table 4 1 prese nts a summary of the descriptive statistics ; f rom the table, we can observe the following: the average household vehicle ownership increases with increase in income, the average fuel efficienc y of vehic les are similar among different income groups, and the total annual miles driven per household increases with increase in income. 4 .1 Model Calibration Multiple regression models have been found to be c apable of capturing real life behavior, and are easy to execute and effective for policy analyses. A multiple regression model was adopted in this research following a similar model adopted by McMullen et al. ( 2010 ) In their model, total annual miles driven by a household, i.e., travel demand of the h ousehold, is assumed

PAGE 22

22 to be a function of social characteristics of the household and attributes of transportation services. The household vehicle ownership is assumed to be fixed and a log log linear regression model is used to avoid a negative value of ve hic le miles driven by a household. The functional form below was used for the analysis : Whe re, TM = Total annual miles driven by all vehicles in a household (mile) EQ = Equivalent travel price or cost per mile ($/mile) estimated as follows: where ( i represents individual vehicles in a household) hh tinc = Total annual household income ($) hhv cnt = Household vehicle count U = dummy variable 1 if the household is located in urban area, 0 otherwise SUB = dummy variable 1 if the househol d has different types of vehicles, i.e. car, van, SUV, truck and RV, 0 otherwise wkrcnt = Number of workers in the household hhchild = Number of children in the household Table 4 1. Descriptive Statistics by Income Group Income Group Income range Household No. Veh. Per HH Avg. veh. MPG per HH Total annual VMT per HH Total HH Rural HH Urban HH Group 1 $0 $19,999 2119 475 1644 1.40 20.64 12187 Group 2 $20,000 $39,999 3288 737 2551 1.64 21.04 15926 Group 3 $40,000 $59,999 2468 537 1931 1.91 21.49 21035 Group 4 $60,000 $79,999 1800 373 1427 2.16 21.78 24650 Group 5 $80,000 $200,000 3411 653 2758 2.42 21.46 29629 Overall Avg. 1.93 21.27 21056

PAGE 23

23 Table 4 2 Estimated Model Variable Name Coefficient Std. error t statistic Constant 2.4787 0.6472 3.8298 In(EQ ) 5.4067 0.3416 15.8274 In(hhtinc) 0.7612 0.0620 12.2745 In(hhv cnt) 0.9188 0.0196 46.9485 U 0.1821 0.0147 12.3627 In(hh tinc)*In( EQ ) 0.3449 0.0327 10.5518 SUB 1.6881 0.1163 14.5149 In(EQ )*SUB 0.7499 0.0607 12.3453 wrkcnt 0.1509 0.0084 18.0693 hhchild 0.1023 0.0079 12.8797 For model calibration, the equivalent travel price or cost per mile, i.e., EQ in the above is estimated as the net gasoline price plus the federal and state taxes per gallon divided by the miles per gallon of the vehicle in a household. If the household owns multiple vehicles, the cost per mile is a weighted average using the total annual miles driven by each vehicle as the weight. The model is calibrated with the clean data set, and the coefficients of the model are presented in Table 4 2. The adjusted R square is 0.56, and all the coefficients have the correct sign and are statistically significant at the 99% confidence interval The calibrated demand model can be used to predict new demands via the changes in the equivalent travel cost caused by those three instruments. More specifically, the equivalent travel or cost per mile will increase with the increased state gasoline tax. Under the mileage fee policy, the state tax becomes zero and the mileage fee will be directly added to the equivalent travel cost per mile. With the tradable credit scheme, the state tax is also zero and the market clearing credit price is added to the equivalent travel cost per mile. The demand elasticity is calculated using the expression below and the results shown in Table 4 3 From the elasticity values, we can observe that lower income people are more sensitive to fuel cost than those with higher incomes, and households with only one type of vehicle are more sensit ive than households with multiple types of vehicles.

PAGE 24

24 Table 4 3 Elasticity by Income Group Based on Average Income Income groups ($) Avg. Income ($) Elasticity with SUB Elasticity without SUB $0 $19,999 12705 2.02 2.19 $20,000 $39,999 30290 1.59 1.85 $40,000 $59,999 50365 1.32 1.67 $60,000 $79,999 70847 1.10 1.56 $80,000 $200,000 130457 0.85 1.35

PAGE 25

25 CHAPTER 5 IMPACT ANALYSIS The focus of the impact analysis will be on changes in consumer surplus, revenue and social welfare. The change in consumer surplus captures the impact of a policy on the household whereas the change in revenue gives an estimate of the feasibility of t he policy. The total change in social welfare is the sum of the change in consumer surplus and change in revenue. The foregoing is the analysis of the impact of the three policies on the household with different income groups assuming the state agency in tends to reduce the total VMT by 15%. 5.1 Comparison of Schemes under D eterministic H ousehold T ravel D emand A comparison between the three policy instruments is made assuming that the individual household demand function is deterministic Appendix A shows a MATLAB program written to conduct the impact analysis 5 .1 .1 Gasoline T ax In 2009, the average state (and local) gasoline tax in Florida was 3 4.5 cents/gallon. To reduce the VMT by 15%, the tax needs to be 58.5 cents/gallon, this may be considered as too high and politically unacceptable. However, many OECD countries (Wikipedia, ( 201 2) ) have comparatively higher taxes. From the analysis, government will generate additional $2.28 million revenue if such a tax is implemented. Furthermore, the total change in consumer surplus will be $2.70 million while the total change in social welfare will be $0.42 million. The impact on households in the different income category is summaries in Table 5 1 From Table 5 1 it can be seen that all the changes in consumer s surpluses for the various income groups are negative. Furthermore, it can be seen from the percentage change in

PAGE 26

26 consumer surplus that the lower income groups suffer more than those in the higher income groups. It must however be noted that although the model predicts that an increase in tax by about 24 cents/gallon will c ause a 15% reduction in VMT, this may not be the real case as other factors not captured in the model may cause a lesser reduction in VMT. Table 5 1 Average Changes in Revenue and Social Welfare ( Gasoline Tax ) Income group Avg. change in consumer surplus ($) Avg. change in consumer surplus as % of avg. income Avg. change in revenue ($) Avg. change in social welfare ( $) $0 $19,999 106.44 0.84 78.01 28.44 $20,000 $39,999 150.32 0.50 119.45 30.87 $40,000 $59,999 200.26 0.40 166.53 33.73 $60,000 $79,999 247.25 0.35 212.23 35.02 $80,000 $200,000 306.57 0.23 273.31 33.25 5 1. 2 Mileage F ee The flat fee structure was implemented in our mileage fee analysis. Currently, the aver age gasoline tax in Florida is 52.9 cents/ gallon with the federal tax and 34.5cents/gallon without the federal tax, i.e., the sum of the state and county taxes. Using 21 miles per ga llon (MPG) as the average fuel efficiency of vehicles in Florida, the current state gas tax of 34.5 cents/gallon is equivalent to 1.64 cents/mile, if all other factors remain the same. To obtain a revenue neutral impact fee and fees for other purposes, the model is executed multiple times, and the resulting socioeconomic impacts are summarized in Table 5 2 From Table 5 2 it can be seen that the revenue neutral flat mileage fee is 1.61 cents/mile. Also the mileage fee required to reduce VMT by 15% is 3.70cents/mile. Details of the socioeconomic impacts of 1.61 and 3.70 cents/mile across different income groups are presented in section 5.1.3 and section 5.1.4 respectively

PAGE 27

27 5. 1. 3 Mileage Fee of 1.61 Cents/M ile ( R evenue N eutral F ee) For the revenue neutral mileage fee i.e. 1.61 cents/mile, it can be observed from Table 5 3 that the average percentage change in consumer surplus as a percentage of income across the across the various income group is negligible. Although there is a general increase in the average change in consumer surplus as the income levels increases, it can be observed that those in the highest income groups do have negative values. Table 5 2 Changes in Revenue Social Welfare and VMT under different Mileage Fees Mileage fee (cents/mile) Total change in consumer surplus ($) Total change in revenue ($) Total change in social welfare ($) % VMT reduction 1.60 3,818 18,364 14,546 0.92 1.61 28,989 3,518 32,508 1.00 1.62 54,141 25,366 79,507 1.08 1.63 79,272 47,179 126,450 1.16 1.65 129,480 90,701 220,180 1.32 1.70 254,640 198,910 453,550 1.71 2.00 995,610 830,760 1,826,400 4.02 2.50 2,195,100 1,822,300 4,017,400 7.60 3.00 3,354,900 2,744,800 6,099,700 10.89 3.50 4,479,500 3,606,100 8,085,600 13.93 3.70 4,920,300 3,935,100 8,855,400 15.08 Table 5 3 Average Changes in Revenue and Social Welfare ( Mileage Fee=1.61cents/mile ) Income group Avg. change in consumer surplus ($) Avg. change in consumer surplus as % of avg. income Avg. change in revenue ($) Avg. change in social welfare ($) $0 $19,999 1.92 0.02 0.35 2.28 $20,000 $39,999 3.21 0.01 1.09 4.30 $40,000 $59,999 5.25 0.01 2.85 8.10 $60,000 $79,999 8.30 0.01 5.41 13.71 $80,000 $200,000 3.97 0.00 5.15 9.13 5 1. 4 Mileage Fee of 3.70 C ents/ M ile The impact of a flat mileage fee of 3.70 cents/mile is presented in Table 5 4 It can be seen that the average change in consumer surplus for such a fee will generally lead to negative changes in consumer surpluses for all inco me groups i.e. they are regressive From the average

PAGE 28

28 change in consumer surplus as a percentage of income, it can be seen that the lower income groups are more impacted by such a fee than the high income groups. Table 5 4 Average Changes in Surplus Revenue and Social Welfare (Mileage Fee=3.70 cents/mile) Income group Avg. change in consumer surplus ($) Avg. change in consumer surplus as % of avg. income Avg. change in revenue ($) Avg. change in social welfare ($) $0 $19,999 190.99 1.50 1 26.57 64.42 $20,000 $39,999 273.56 0.90 201.10 72.46 $40,000 $59,999 369.38 0.73 289.01 80.36 $60,000 $79,999 462.59 0.65 377.18 85.41 $80,000 $200,000 548.78 0.42 473.02 75.76 Compared to the gasoline tax policy the mileage fee policy achieves the control target with slightly higher impacts. Moreover, although its distributional effects look similar, the mileage fee policy is slightly more regressive. 5. 1. 5 Tradable C redit The explanation to the tradable credit system given in chapter 3 assumed a linear demand function; however, the demand function for our analysis is nonlinear. A linear approximation of the demand function was used and found to be a good approximation for t he foregoing analysis. Consider ing a tradable credit policy where the total credits equal to 85% of the current VMT are uniformly distributed among 13086 households in our sample E ach household therefore need s to pay 1.7 3 cents of credits in order to allo w the state to maintain its current level of revenue. The distributional impacts of this scheme are shown in Table 5 5 It can be observed from the average change in consumer surplus that those in the lower income bracket have positive and higher values. Also form the average change in consumer surplus as percentage of average income, there is a similar trend where those in the lower income group do have higher positive valu es.

PAGE 29

29 Table 5 5 Tradable Credit ) Income group Avg. change in consumer surplus ($) Avg. change in consumer surplus as % of avg. income Avg. change in revenue ($) Avg. change in social welfar e ($) $0 $19,999 133.84 1.05 46.90 180.74 $20,000 $39,999 65.93 0.22 47.48 113.40 $40,000 $59,999 17.82 0.04 62.87 45.05 $60,000 $79,999 91.91 0.13 57.27 34.64 $80,000 $200,000 202.86 0.16 58.63 144.23 5 1.6 Overall C omparison under D eterministic H ousehold T ravel D emand Table 5 6 shows their standard deviation, and VMT under the three instruments for the dif ferent schemes. W hile Table 5 7 shows the average changes in con different schemes. From Table 5 6, it can be seen that all the instruments successfully achieve the control target, verifying the traditional axiom in economics literature that the use of prices or q uantities as management instruments achieves the same level of efficiency in an idealized environment. Although they achieve the same control target it can be seen from Table 5 7 that the magnitude of their socioeconomic impacts is variable. Specifically the flat mileage fee leads to the most adverse percentage as a percentage of income and is regressive, i.e. it affect s the lower income groups more adversely than the higher income groups. Conversely the tradable credit schemes lead to much minor changes and is more progressive i,e its adverse effect increases as income levels increases This observation is also consistent with the literature where a tradable credit scheme is known to minimize the total co st of reaching a pre determined environment standard regardless of the initial allocation of credits, provid ed there is no transaction cost.

PAGE 30

30 Table 5 6. Total Changes in (Deterministic) Instru ment Total Change Consumer Surplus ($) Total change in revenue ($) Total Change in social welfare ($) % VMT reduction Gasoline Tax 15.08 Mileage fee 15.08 Tradable Credits 15.00 Table 5 7 Income group Gasoline Tax (%) Mileage Fee (%) Tradable Credits (%) $0 $19,999 0.84 1.50 1.05 $20,000 $39,999 0.50 0.90 0.22 $40,000 $59,999 0.40 0.73 0.04 $60,000 $79,999 0.35 0.65 0.13 $80,000 $200,000 0.23 0.42 0.16 5 2 Comparison of Schemes under Stochastic T ravel D emand F unction In the previous section, the household demand function was assumed to be deterministic however, individual household s will respond differently to these policy schemes. This section deals with the situation where there are uncertainties in the individual household travel demand functi on Firstly, a comparison of the effectiveness of the three policy schemes in reducing the total VMT by 15% is analyzed. Secondly, the socioeconomic impacts of the three policy schemes are evaluated under such conditions. Below is a typical travel demand f unction under stochastic degradation following a uniform distribution between the nominal deman d used for the analysis : w 0. 15 0. 15 )

PAGE 31

31 To achieve this, a Monte Carlo simulation is conducted; different number of samples was tried and one thousand samples were found to be enough for the purpose of this study. Appendix B shows details of the MATLAB p rogram written for the analysis. 5.2.1 Reduction of VMT under U ncertainty Figure 5 1 shows the results of the percentage reduction i n total VMT with number of samples for the three policy schemes. Als o shown, are the success rate in achieving 15% reduction of VM T for the three policy schemes. It can be observed that the tradable credit scheme is 100% successful in achieving the desired percent reduction whiles the succ ess rate for the gasoline tax and mileage fee policy schemes are 83.5% and 84.3% respectively. 5.2.2 Price Variation under U ncertainty In the gasoline tax policy the tax is fixed whiles in the mileage fee policy the mileage fee is fixed H owever as seen in Figure 5 1 under uncertainty the market clearing price under the tradable credit scheme fluctuates. I t was found to range between 1.49 cents to 5.97 cents; people in the market may not be pleased with such wide variation in the price of credit s. 5.2.2 Changes in Socio E conomic I ndicators under U ncertainty Figure 5 2 to Figure 5 4 shows the changes in surplus, revenue and social welfare each for the three policy schemes. Table 5 8 shows a summary of the changes in the socioeconomic i ndicators under uncertainty. It can be seen from Figure 5 3 that the revenue generated under the tradable credit scheme is constant and for policy and budgeting purposes, the government knows the revenue before the planning year. However for the other two policy schemes i.e. gasoline tax and mileage fee policy, the government will not know the exact revenue in the preceding year when budgeting.

PAGE 32

32 Table 5 8 Total (Stochastic) Instrument Total Change Consumer Surplus ($) Total change in revenue ($) Total Change in social welfare ($) % VMT reduction Gasoline Tax 2,886,600 2,428,700 457,910 15.08 Mileage fee 4,694,400 3,660,100 1,034,300 15.08 Tradable Credits 6,868 713,740 720,600 15.00 Figure 5 1 C hanges in total VMT reduction with number samples for different policies. Figure 5 2 Changes in total consumer surplus wi th number samples for different p olicies.

PAGE 33

33 Figure 5 3 Changes in total revenue with number samples for different policies. Figure 5 4 Changes in total social welfare with number samples for different policies.

PAGE 34

34 CHAPTER 6 SUMMARY AND CONCLUSIONS This study compared the socioeconomic impacts of three market based instruments, gasoline tax, mileage fee and tradable credit schemes, for mobility management. The NHTS 2009 data for was used for the analysis. A hypothetical case study where FDOT intends to reduce the total vehicle miles by 15% in Florida was considered Firstly, we developed a d eterministic household travel demand function and examined the socioeconomic impact of the three market based instruments. It was found that the three instruments are capable of achieving the hypothetical 15% reduction of total travel demand in Florida. Ho wever, they generate different amounts of revenue and impose different socioeconomic impacts on Florida residents. Gasoline tax and mileage fee schemes charge travelers more to discourage their travels to achieve the control target. Consequently, the gover nment receives much more revenue. At the same time, the schemes hurt residents more and do more harm to the poor than the rich. The tradable credit scheme generates the least revenue for the government but has a less regressive impact on residence. Second ly, we assumed that the household travel demand function is uncertain and conducted similar analysis to determine the socioeconomic impact of the three instruments. The three policy instruments were all capable of reducing the VMT by 15%. However, the gaso line tax and mileage fee policies were found to have a lower success rate and the revenue generated was variable; this may not be go o d for the government in terms of budgeting. For the tradable credit the success rate was found to be 100% and the revenue generated fixed hence helps the government for budgeting purposes H owever, we observed price volatility which may not be appealing to people. The socio economic impact was found to be similar to that obtained under the deterministic travel demand.

PAGE 35

35 The co mparative analysis does not consider many practical issues such as the implementation and administration cost s A mileage fee scheme is likely to be more costly to implement and operate than the gasoline tax. The implementation cost for a tradable credit s cheme is probably the highest among these three instruments. Moreover, the efficiency of the scheme may be adversely affected by transaction costs and speculation behavior in the credit market, which are ignored in our analysis. The NHTS data do not contain any information on how residents react to a mileage fee or tradable credit policy. Our analysis simply assumed that residents react to them the same way as an increase in gas price, which is not necessarily the case. Moreover, we also assumed that vehicle ownership and land use patterns would remain the same after the implementation of these new instruments. This assumption may be valid for short term assessment. For a long term assessment, more advanced model s should be used

PAGE 36

36 APPENDIX A MATLAB PROGRAM FOR DETERMINISTIC ANALYSIS clear all ; close all ; clc; data=xlsread( 'finaldata.xlsx' ,1, 'A2:k13087' ); % Assigmnent of variables hhtotinc=data(:,1); hhvehcnt=data(:,2); hhchild=data(:,3); wrkcnt=data(:,4); vmt=data(:,5); rawpm= data(:,6); sub=data(:,7); u=data(:,8); avgprice=data(:,9); mpg=data(:,10); net_gas_price=data(:,11); %creation of new variables initial conditions gastax=0.345; vmtfee=0.037; % used for mileage fee calculations anngallons=vmt./mpg; gasprice=net_gas_price +gastax; exp_gas_tax=gasprice.*anngallons; exp_vmt=net_gas_price.*anngallons+vmt.*vmtfee; chng_exp=exp_gas_tax exp_vmt; pm=rawpm./vmt; % Regression Parameters b0= 2.4787; b1= 5.4067; b2=0.7612; b3=0.9188; b4= 0.1821; b5=0.3449; b6=1.6881; b7=0.7499; b8=0.1509; b9=0.1023; % ELASTICITY BY INCOME e=b1+b5.*log(hhtotinc)+b7.*sub; % with sub e2=b1+b5.*log(hhtotinc); % without sub % Calculate a new PM pm_wotax=(net_gas_price)./(mpg); pm_tax=(net_gas_price+gastax)./(mpg); pm_vmt=((net_gas_price)./(mpg )+vmtfee); % current VMT tot_vmt=sum(vmt) % Predicted VMT without gas tax

PAGE 37

37 lnvmt_wotax=b0 + b1.*log(pm_wotax)+ b2.*log(hhtotinc)+ b3.*log(hhvehcnt)+ b4.*u + b5.*log(hhtotinc).*log(pm_wotax)+ b6.*sub+b7.*log(pm_wotax).*sub + b8.*wrkcnt + b9.*hhchild; vmt_wotax=exp(lnvmt_wotax); totvmt_wottax=sum(vmt_wotax) %Predicted VMT with gas tax lnvmt_tax=b0 + b1.*log(pm_tax)+ b2.*log(hhtotinc)+ b3.*log(hhvehcnt)+ b4.*u + b5.*log(hhtotinc).*log(pm_tax)+ b6.*sub+b7.*log(pm_tax).*sub + b8.*wrkcnt + b9.*hhchild; vm t_tax=exp(lnvmt_tax); % Detailed Gastax vmt. Omit when necessary totalvmt_tax=sum(vmt_tax) %Predicted VMT with mileage fee lnvmt_fee=b0 + b1.*log(pm_vmt)+ b2.*log(hhtotinc)+ b3.*log(hhvehcnt)+ b4.*u + b5.*log(hhtotinc).*log(pm_vmt)+ b6.*sub+b7.*log(pm_vmt).*sub + b8.*wrkcnt + b9.*hhchild; vmt_fee=exp(lnvmt_fee); % PREDICTION OF SOCIOECONOMIC PARAMETERS % gas tax gastax1=0.345; pm_tax1=(net_gas_price+gastax1)./(mpg); lnvmt_tax1=b0 + b1.*log(pm_tax1)+ b2.*log(hhtotinc)+ b3.*lo g(hhvehcnt)+ b4.*u + b5.*log(hhtotinc).*log(pm_tax1)+ b6.*sub+b7.*log(pm_tax1).*sub + b8.*wrkcnt + b9.*hhchild; vmt_tax1=exp(lnvmt_tax1); gastax2=0.585; pm_tax2=(net_gas_price+gastax2)./(mpg); lnvmt_tax2=b0 + b1.*log(pm_tax2)+ b2.*log(hhtotinc)+ b3.*lo g(hhvehcnt)+ b4.*u + b5.*log(hhtotinc).*log(pm_tax2)+ b6.*sub+b7.*log(pm_tax2).*sub + b8.*wrkcnt + b9.*hhchild; vmt_tax2=exp(lnvmt_tax2); chng_csg=(pm_tax1 pm_tax2).*(vmt_tax1+vmt_tax2).*0.5; chng_revg=((gastax2./mpg).*vmt_tax2) ((gastax1./mpg).*vmt_tax1 ); chng_swg=chng_csg+chng_revg; tot_csg=sum(chng_csg) tot_revg=sum(chng_revg) tot_swg=sum(chng_swg) per_chng_vmt_gt=(sum(vmt_tax2) sum(vmt))/(sum(vmt)) % Mileage fee %(note the Pm values used are from the first pm stated) if vmtfee>0.0164 chng_csfee=(pm_tax pm_vmt).*(vmt_tax+vmt_fee).*0.5; % remember to change to difference in PM when the mileage fee is greater than 1.64(in this case) else chng_csfee=(pm_vmt pm_tax).*(vmt_tax+vmt_fee).*0.5;

PAGE 38

38 end chng_revfee=vmtfee.*vmt_fee (gastax./mpg ).*vmt_tax; chng_swfee=chng_csfee+chng_revfee; % Calculation of total Social Economic Parameters (based on mileage fee) tot_csfee=sum(chng_csfee) tot_revfee=sum(chng_revfee) tot_swfee=sum(chng_swfee) % Changes in VMT diff_vmt=sum(vmt_tax) sum(vmt_fee); per_chng_vmt=diff_vmt*100/sum(vmt_tax) % IMPLEMENTATION OF TRADABLE CREDITS % Total Revenue under current Conditions rev_current=(gastax./mpg).*vmt_tax; tot_rev_current=sum(rev_current) % Allocation of credits % By household curr_vmt=sum(vmt)*0.85; % this is assuming that we intend to reduce the VMT by 15%. cred_hh=tot_rev_current/curr_vmt % Total Miles credit allocated based on current household no_hh=13086; mil_credit_hh=(tot_rev_current)/(no_h h*cred_hh) % Socio economic parameter % Individual Demand functions ind_cred_demand=vmt.*0.85; vmtfeem=0.037; %market clearing credit pm_vmtm=((net_gas_price)./(mpg)+vmtfeem); lnvmt_feen=b0 + b1.*log(pm_vmtm)+ b2.*log(hhtotinc)+ b3.*log(hhvehcnt) + b4.*u + b5.*log(hhtotinc).*log(pm_vmtm)+ b6.*sub+b7.*log(pm_vmtm).*sub + b8.*wrkcnt + b9.*hhchild; vmt_feen=exp(lnvmt_feen); unit_cred_0=(vmt_tax.*vmtfee vmt_feen.*0.0161)./(vmt_tax vmt_feen); unit_cred_1=(vmt_tax.*vmtfee vmt_feen.*cred_hh)./(vm t_tax vmt_feen); chng_demand=mil_credit_hh vmt_tax; chng_cs_cred=0.5*(unit_cred_1+vmtfee).*vmt_feen+chng_demand.*(vmtfee cred_hh) (0.5*unit_cred_0.*vmt_tax); rev_cred=vmt.*cred_hh; % Revenue

PAGE 39

39 chng_rev_cred=rev_cred rev_current; chng_sw_cred =chng_cs_cred+chng_rev_cred; tot_cs_cred=sum(chng_cs_cred) tot_rev_cred=sum(chng_rev_cred) tot_sw_cred=tot_cs_cred+tot_rev_cred

PAGE 40

40 APPENDIX B MATLAB PROGRAM FOR STOCHASTIC ANALYSIS clear all ; close all ; clc; data= xlsread( 'finaldata.xlsx' ,1, 'A2:j13087' ); % Assigmnent of variables hhtotinc=data(:,1); hhvehcnt=data(:,2); hhchild=data(:,3); wrkcnt=data(:,4); vmt=data(:,5); rawpm=data(:,6); sub=data(:,7); u=data(:,8); avgprice=data(:,9); mpg=data(:,10); %creation of new variables gastax=0.345; vmtfee=0.037; anngallons=vmt./mpg; exp_gas_tax=avgprice.*anngallons; net_gas_price=avgprice gastax; exp_vmt=net_gas_price.*anngallons+vmt.*vmtfee; chng_exp=exp_gas_tax exp_vmt; pm=rawpm./vmt; % Regression Parameters b0= 2.4787; b1= 5.4067; b2=0.7612; b3=0.9188; b4= 0.1821; b5=0.3449; b6=1.6881; b7=0.7499; b8=0.1509; b9=0.1023; % PREDICTION OF SOCIOECONOMIC PARAMETERS % gas tax gastax1=0.345; pm_tax1=(net_gas_price+gastax1)./(mpg); lnvmt_tax1=b0 + b1.*log(pm_tax1)+ b2.*log(hhtotinc)+ b3.*log(hhvehcnt)+ b4.*u + b5.*log(hhtotinc).*log(pm_tax1)+ b6.*sub+b7.*log(pm_tax1).*sub + b8.*wrkcnt + b9.*hhchild; vmt_tax1=exp(lnvmt_tax1); gastax2=0.585; pm_tax2=(net_gas_price+gastax2)./(mpg); lnvmt_tax2=b0 + b1.*log(pm_tax2)+ b2.*log(hhtotinc)+ b3.*log(hhvehcnt)+ b4.*u + b5.*log(hhtotinc).*log(pm_tax2)+ b6.*sub+b7.*log(pm_tax2).*sub + b8.*wrkcnt + b9.*hhchild;

PAGE 41

41 vmt_tax2=exp(lnvmt_tax2); % Mileage fee % Calculate a new PM pm_wotax=(net_gas_price)./(mpg); pm_tax=( net_gas_price+gastax)./(mpg); pm_vmt=((net_gas_price)./(mpg)+vmtfee); % current VMT tot_vmt=sum(vmt) %Predicted VMT with gas tax lnvmt_tax=b0 + b1.*log(pm_tax)+ b2.*log(hhtotinc)+ b3.*log(hhvehcnt)+ b4.*u + b5.*log(hhtotinc).*log(pm_tax)+ b6.*sub+b7.*log(pm_tax).*sub + b8.*wrkcnt + b9.*hhchild; vmt_tax=exp(lnvmt_tax); % Detailed Gastax vmt. Omit when necessary totalvmt_tax=sum(vmt_tax) %Predicted VMT with mileage fee lnvmt_fee=b0 + b1.*log(pm_vmt)+ b2.*log(hhtotinc)+ b3.*log(hhvehcn t)+ b4.*u + b5.*log(hhtotinc).*log(pm_vmt)+ b6.*sub+b7.*log(pm_vmt).*sub + b8.*wrkcnt + b9.*hhchild; vmt_fee=exp(lnvmt_fee); % IMPLEMENTATION OF TRADABLE CREDITS % Total Revenue under current Conditions rev_current=(gastax./mpg).*vmt_tax; t ot_rev_current=sum(rev_current) ind_cred_demand=vmt.*0.85; vmtfeem=0.037; %market clearing credit pm_vmtm=((net_gas_price)./(mpg)+vmtfeem); lnvmt_feen=b0 + b1.*log(pm_vmtm)+ b2.*log(hhtotinc)+ b3.*log(hhvehcnt)+ b4.*u + b5.*log( hhtotinc).*log(pm_vmtm)+ b6.*sub+b7.*log(pm_vmtm).*sub + b8.*wrkcnt + b9.*hhchild; vmt_feen=exp(lnvmt_feen); % Allocation of credits % By household curr_vmt=sum(vmt)*0.85; % this is assuming that we intend to reduce the VMT by 15%. cred_hh=tot_rev_current/curr_vmt % Total Miles credit allocated based on current household no_hh=13086; mil_credit_hh=(tot_rev_current)/(no_hh*cred_hh); % Socio economic parameter % Individual Demand functions ind_cred_demand=vmt .*0.85;

PAGE 42

42 N=1000; %for z=1:2 %for j=1:10 for i=1:N r(:,i)= 0.15+0.3*rand(1,13086); % Generate random numbers r(:,i)=reshape(r(:,i),13086,1); % this converts it into a column vector vmt_taxn(:,i)=(1+r(: ,i)).*exp(lnvmt_tax2); sumvmt_taxn(:,i)=sum(vmt_taxn(:,i)); % VMT Gas Tax vmt_feem(:,i)=(1+r(:,i)).*exp(lnvmt_fee); sumvmt_feem1(:,i)=sum(vmt_feem(:,i)); % VMT Mileage fee vmt_feen(:,i)=(1+r(:,i)).*exp (lnvmt_feen); sumvmt_feen(:,i)=sum(vmt_feen(:,i)); % Changes in socioeconomic measures chng_csg(:,i)=(pm_tax1 pm_tax2).*(sum(vmt_tax1)+sumvmt_taxn(:,i)).*0.5; chng_revg(:,i)=((gastax2./mpg).*sumvmt_taxn(:,i)) ((gas tax1./mpg).*sum(vmt_tax1)); chng_swg(:,i)=chng_csg(:,i)+chng_revg(:,i); tot_csg(:,i)=mean(chng_csg(:,i)); tot_revg(:,i)=mean(chng_revg(:,i)); tot_swg(:,i)=mean(chng_swg(:,i)); % Changes in VMT per_chng_vmt_gt(:,i)=abs(((sumvmt_taxn(:,i)) sum(vmt))*100/(sum(vmt))); % mileage Fee % Mileage Fee if vmtfee>0.0164 chng_cs_fee(:,i)=(pm_tax pm_vmt).*(sum(vmt_tax)+sumvmt_feem1(i)).*0.5; else chng_cs_f ee(:,i)=(pm_vmt pm_tax).*(sum(vmt_tax)+sumvmt_feem1(i)).*0.5; end chng_rev_fee(:,i)=vmtfee.*sumvmt_feem1(i) (gastax./mpg).*sum(vmt_tax); chng_sw_fee(:,i)=chng_cs_fee(:,i)+chng_rev_fee(:,i); % Calculation of total Social Economic Parameters (based on mileage fee) tot_cs_fee(:,i)=mean(chng_cs_fee(:,i)); tot_rev_fee(:,i)=mean(chng_rev_fee(:,i)); tot_sw_fee(:,i)=mean(chng_sw_fee(:,i)); % Changes in VMT per_chng_vmtfee(:,i)=abs(((sumvmt_feem1(:,i)) sum( vmt_tax))*100/(sum(vmt_tax))); % Bisection Method vmtfee1(i)=0.01; % start of interval

PAGE 43

43 vmtfee2(i)=0.080; % end of interval n=1000; pm_vmt1(:,i)=(( net_gas_price)./(mpg)+vmtfee1(:,i)); lnvmt_fee1(:,i)=b0 + b1.*log(pm_vmt1(:,i))+ b2.*log(hhtotinc)+ b3.*log(hhvehcnt)+ b4.*u + b5.*log(hhtotinc).*log(pm_vmt1(:,i))+ b6.*sub+b7.*log(pm_vmt1(:,i)).*sub + b8.*wrkcnt + b9.*hhchild; %vmt_fee1(i) =exp(lnvmt_fee1(i)); vmt_fee1(:,i)=(1+r(i)).*exp(lnvmt_fee1(:,i)); sumvmt_fee1(:,i)=sum(vmt_fee1(:,i)); % Changes in VMT diff_vmt1(i)=sum(vmt_fee1(:,i)) sum(vmt_tax); f_a(i)=(diff_vmt1(i)*100/sum(vmt_tax))+15; pm_vmt2(:,i)=((net_gas_price)./(mpg)+vmtfee2(:,i)); lnvmt_fee2(:,i)=b0 + b1.*log(pm_vmt2(:,i))+ b2.*log(hhtotinc)+ b3.*log(hhvehcnt)+ b4.*u + b5.*log(hhtotinc).*log(pm_vmt2(:,i))+ b6.*sub+b7.*log(pm_vmt2(:,i)).*sub + b8.*wrkcnt + b9.*hhchild; %vmt_fee2=exp(lnvmt_fee2); vmt_fee2(:,i)=(1+r(i)).*exp(lnvmt_fee2(:,i)); sumvmt_fee2(i)=sum(vmt_fee2(:,i)); % Changes in VMT diff_vmt2(i)=sum(vmt_fee2(:,i)) sum(vmt_tax); f_b(i)=(diff_vmt2(i)*100/sum(vmt_tax ))+15; if f_a(i)*f_b(i)> 0.0 error ( 'same end points.' ) end for j = 1:n c(i)=(vmtfee2(i)+vmtfee1(i))/2; pm_vmtc(:,i)=((net_gas_price)./(mpg)+c(:,i)); lnvmt_feec(:,i)=b0 + b1.*log(pm_vmtc(:,i))+ b2.*log(hhtotinc)+ b3.*log(hhvehcnt)+ b4.*u + b5.*log(hhtotinc).*log(pm_vmtc(:,i))+ b6.*sub+b7.*log(pm_vmtc(:,i)).*sub + b8.*wrkcnt + b9.*hhchild; %vmt_fee2=exp(lnvmt_fee2); vmt_feec(:,i)=(1+r(i)).*exp(lnvmt_feec(:,i) ); sumvmt_feec(i)=sum(vmt_feec(:,i)); % Changes in VMT diff_vmtc(i)=sum(vmt_feec(:,i)) sum(vmt_tax); f_c(i)=(diff_vmtc(i)*100/sum(vmt_tax))+15; %disp([ c f_c]) if f_c(i) == 0.0 % solved the equa tion exactly e = 0.0001; break % jumps out of the for loop end if c(i)*f_c(i) < 0 vmtfee2(i)=c(i); else vmtfee1(i)=c(i); end end e(i) = (vmtfee2(i) vmtfee1(i))/2; c(i); fee(i)=c(i);

PAGE 44

44 % Tradable credits % VMT tradable credits unit_cred_0(:,i)=((vmt_tax).*fee(i) (vmt_feen(i)).*0.0161)./((vmt_tax) (vmt_feen(:,i))); unit_cred_1(:,i)=((vmt_tax).*fee( i) (vmt_feen(:,i)).*cred_hh)./((vmt_tax) (vmt_feen(:,i))); chng_demand=mil_credit_hh vmt_tax; chng_cs_cred(:,i)=0.5.*(unit_cred_1(:,i))+fee(i).*vmt_feen(:,i)+chng_demand.*(fee(i) cred_hh) (0.5.*unit_cred_0(:,i).*(vmt_tax)); rev_cr ed=vmt.*cred_hh; % Revenue chng_rev_cred=rev_cred rev_current; chng_sw_cred(:,i)=chng_cs_cred(:,i)+chng_rev_cred; % Tradable tot_cs_cred(:,i)=mean(chng_cs_cred(:,i)); tot_rev_cred(i)=sum(chng_rev_cred); tot_sw_cred(:,i)=tot_cs_cred(:,i)+tot_rev_cred(i); per_chng_vmt_cred(:,i)=abs(((sumvmt_feec(:,i)) sum(vmt_tax))*100/(sum(vmt_tax))); end x=1:10:1000; z=1:1:1000; %Means %Perc entage Changes mean(per_chng_vmt_gt) mean(per_chng_vmtfee) mean(per_chng_vmt_cred) % Changes in Consumer surplus mean(tot_csg) mean(tot_cs_fee) mean(tot_cs_cred) % Changes in Revenue mean(tot_revg) mean(tot_rev_fee) mean(tot_rev_cred) % Changes in social welfare mean(tot_swg(x)) mean(tot_sw_fee) mean(tot_sw_cred) % Ploting Percentage Reduction in VMT figure subplot( 221);plot(x,per_chng_vmt_gt(x), 'b' 'LineWidth' ,1.5);

PAGE 45

45 hold on plot(z,mean(per_chng_vmt_gt), 'r' 'LineWidth' ,2); xlabel( 'Number of Samples' ); ylabel( 'Percent Reduction of VMT' ); legend( 'Gasoline Tax' 'Mean' ); g rid on ; hold on subplot(222);plot(x,per_chng_vmtfee(x), -r' 'LineWidth' ,1.5); hold on plot(z,mean(per_chng_vmtfee), 'b' 'LineWidth' ,2); xlabel( 'Number of Samples' ); ylabel( 'Percent Reduction of VMT' ); legend( 'Mileage Fee' 'Mean' ); grid on ; hold on subplot(223);plot(x,per_chng_vmt_cred(x), ':m' 'LineWidth' ,1.5); hold on plot(z,mean(per_chng_vmt_cred), 'b' 'LineWidth' ,2); xlabel( 'Numb er of Samples' ); ylabel( 'Percent Reduction of VMT' ); legend( 'Tradable Credit' 'Mean' ); grid on ; hold off hold on subplot(224);plot(x,c(x), ':m' 'LineWidth' ,1.5); hold on plot(z,mean(c), 'b' 'LineWidth' ,2); xlabel( 'Number of Samples' ); ylabel( 'Market Clearing Price($)' ); legend( 'Tradable Credit' 'Mean' ); grid on ; hold off % Plotting Changes in Consumer Surplus figure subplot(221);plot(x,tot_csg(x), 'b' 'LineWidth' ,1.5); hold on plot(z,mean(tot_csg), 'r' 'LineWidth' ,2); xlabel( 'Number of Samples' ); ylabel( 'Change in Consumer Surplus' ); legend( 'Gasoline Tax' 'Mean' ); grid on ; hold on subplot(222);plot(x,tot_cs_fee(x), -r' 'LineWidth' ,1.5); hold on plot(z,mean(tot_cs_fee), 'b' 'LineWidth' ,2); xlabel( 'Number of Samples' ); ylabel( 'Change in Consumer Surplus' ); legend( 'Mileage Fee' 'Mean' ); grid on ;

PAGE 46

46 hold on subplot(223);plot(x,tot_cs_cred(x), ':m' 'LineWidth' ,1.5); hold on plot(z,mean(tot_cs_cred), 'b' 'LineWidth' ,2); xlabel( 'Number of Samples' ); ylabel( 'Change in Consumer Surplus' ); legend( 'Tradable Credit' 'Mean' ); grid on ; hold off % Plotting Changes in Revenue figure subplot( 221);plot(x,tot_revg(x), 'b' 'LineWidth' ,1.5); hold on plot(z,mean(tot_revg), 'r' 'LineWidth' ,2); xlabel( 'Number of Samples' ); ylabel( 'Change in Revenue' ); legend( 'Gasoline Tax' 'Mean' ); grid on ; hold on subplot(222);plot(x,tot_rev_fee(x), -r' 'LineWidth' ,1.5); hold on plot(z,mean(tot_rev_fee), 'b' 'LineWidth' ,2); xlabel( 'Number of Samples' ); ylabel( 'Change in Revenue' ); legend( 'Mileage Fee' 'Mean' ); grid on ; hold on subplot(223);plot(x,tot_rev_cred(x), ':m' 'LineWidth' ,1.5); hold on plot(z,mean(tot_rev_cred), 'b' 'LineWidth' ,2); xlabel( 'Number of Samples' ); ylabel( 'Change in Rev enue' ); legend( 'Tradable Credit' 'Mean' ); grid on ; hold off % Plotting Changes in Social Welfare figure subplot(221);plot(x,tot_swg(x), 'b' 'LineWidth' ,1.5); hold on plot(z,mean(tot_swg(x)), 'r' 'LineWidth' ,2); xlabel( 'Number of Samples' ); ylabel( 'Change in Social Welfare' ); legend( 'Gasoline Tax' 'Mean' ); grid on ; hold on subplot(222);plot(x,tot_sw_fee(x), -r' 'LineWidth' ,1.5); hold on

PAGE 47

47 plot(z,mean(tot_sw_fee), 'b' 'LineWidth' ,2); xlabel( 'Number of Samples' ); ylabel( 'Change in Social Welfare' ); legend( 'Mileage Fee' 'Mean' ); grid on ; hold on subplot(223);plot(x,tot_sw_cred(x), ':m' 'LineWidth' ,1.5); hold on plot(z,mean(tot_sw_cred), 'b' 'LineWidth' ,2); xlabel( 'Number of Samples' ); ylabel( 'Change in Social Welfare' ); legend( 'Tradable Credit' 'Mean' ) ; grid on ; hold off

PAGE 48

48 LIST OF REFERENCES Davis, L.W., 2008. The effect of driving restrictions on air quality in Mexico City. Journal of Political Economy 116 (1), 38 81. Downs, Anthony, 1962. 16(3): 393 409. Downs, Anthony, 1992. Stuck in Traffic: coping with Peak hour Traffic congestion. Washington, DC: Brookings Institution Press. Federal Highway Administration, 2006. Congest ion Pricing: A Primer, Publication No. FHWA HOP 07 074 Available online: http://www.ops.fhwa.dot.gov/publications/congestionpricing/congestionpricing.pdf Upd ated December 8, 2006, Accessed March 28, 2013. Goddard, H.C., 1997. Using tradable permits to achieve sustainabili policy design issues and efficiency conditions for controlling ve hicle emissions, congestion and urban decen tralization with an application to Mexico C ity. Environmental and Resource Economics 10 (1), 63 99. Guo, X.L., Yang, H., 2010. Pareto improving congestion pricing and revenue refunding with multiple user classes. Transportation Research Part B 44 (8 9), 9 72 982. Kockelman, K.M., Kalmanje, S., 2005. Credit based congestion pricing: a policy proposal and 9), 671 690. Lawphongpanich, S., Yin, Y., 2010. Solving the Pareto impr oving toll problem via manifold suboptimization. Transportation Research Part C 18 (2), 234 246. Liu, Y., Guo, X., Yang, H., 2009. Pareto improving and rev enue neutral congestion pricing schemes in bi modal traffic networks. Netnomics 10 (1), 123 140. Ma hendra, A., 2008. Vehicle restrictions in four Latin American cities: is congestion pricing possible? Transport Review 28 (1), 105 133. Mamun S. MD. 2012 Impact Analysis of Site Development and Mileage Fee. PhD Dissertation, University of Florida, Gaines ville. McMullen, B. S., L. Zhang and K. Nakahara., 2010. Distributional Impacts of Changing from a Gasoline Tax to a Vehicle mile Tax for Light Vehicles: A case study of Oregon. Transport Policy Vol. 17, 2010, pp. 359 366. Morrison, Steven A. 1986. A Survey of Road Pricing. Transportation Research. Vol. 20A, pp. 89 97.

PAGE 49

49 Nie, Y., Liu, Y., 2010. Existence of self financing and Pareto improving congestion pricing: impact of value of time distribution. Transportation Research Part A 44 (1), 39 51. Pigou, A. 1920. The Economics of Welfare. London: Macmillan Raux, C., 2007. Tradable Driving Rights in Urban Area s: Their Potential for Tackling Congestion and Traffic Related Pollution, Working Paper, Laboratoire transports CNRS, Universit d e Lyon, ENTPE. Small, Kenneth A., Clifford Winston, and Carol A. Evans. 1989. Road Work: A New Highway Pricing & Investment Policy. Brookings Institution. Song, Z., Yin, Y., Lawphongpanich, S., 2009. Nonnegat ive Pareto improving tolls with multiclass net work equilibria. Transportation Research Record 2091, 70 78. Texas A&M Transportation Institute (TTI), 2012. Urban Mobility Report. Available online: http://d2dtl5nnlpfr0r.cloudfront.net/tti.tamu.edu/documents/mobility report 2012.pdf Updated February 4, 2013, Accessed April 2, 2013. TRB Committee for the Study of the Long Term Viability of Fuel Taxes for Transportation Finance, 2006 Special Report 2 85: The Fuel Tax and Alternatives for Transportation Funding. Transportation Research Board of the National Academies. Verhoef, E., Nijkamp, P., Rietveld, P., 1997. Tradable permits: their potential in the regulation of road transport externalities. Envir onment and Planning B 24 (4), 527 548. Viegas, J.M., 2001. Making urban road pricing acceptable and e ffective: searching for quality and equity in urban mobility. Transport Policy 8 (4), 289 294. Wang, X.L., Yang, H., Zhu D, Li C ., 2012, Tradable travel credits for congestion management with heterogeneous users Transportation Research. P art E, Logistics and Transportation Review v. 48, (2), 2012, p. 426 437 Whitty, J. M., & Imholt, B., 2005 Oregon's Mileage Fee Concept and Road User Fee Pilot Program: Report to the 73rd Oregon Legislative Assembly. Whitty, J., Svadlenak, J., & Capps, D., 2006 Public Involvement and Road User Charge Development: Oregon's Experience. Oregon Department of Transportation. Wikipedia, Fuel Tax in OECD Countries, 2010.png, 201 2 Available online: https://commons.wikimedia.org/wiki/File:Fuel_tax_in_OECD_countries,_2010..png Updated June 3, 2012, Accessed March 4,2013. Yang, H., Wang, X., 2011. Managing network mobility with tradable credits. Transportation Resear ch Part B 45 (3), 580 594.

PAGE 50

50 Zhang, L., McMullen, B. S., Valluri, D., & Nakahara, K., 2009 The short and long run impacts of vehicle mileage fee on income and spatial equity. Available online: http://wiki.umd.edu/lei/images/f/fe/Zhang2009b.pdf Updated August 26 200 9 Accessed April 4 2013.

PAGE 51

51 BIOGRAPHICAL SKETCH Patrick Amoah Bekoe was born in 1979 at Koforidua, Ghana. He attended the Kwame Nkrumah University of Science and Technology from September 1998 to June 2002 where he earned his Bachelor of Science degree in civil engineering in March 2003. From August 2002 to July 2003, he undertook his Ghana national service with Y AB Engineering Services, a private civil engineering firm located at Tema, Ghana. He was employed with the D epartment of Feeder R oads under the Ministry of Roads and Highway fr om October 2003 and awarded the Ministry of Transportation Fellowship to pursue of Florida in January 2008. He graduated with a Master of Engineering in civil engineering from the Unive rsity of Florida in August 2009 and continued subsequently to pursue a PhD in the same field after he was awarded the graduate school fellowship by the University of Florida In spring 2010, he was officially admitted to the D epartment of Economics to pursue a Master of Arts in economics; h is specialization was in transport ation economics. He graduated with a Master of Arts in economics and Doctor of Philosophy in civil engineering in spring 2013. Patrick Amoah Bekoe will be taking a faculty position at the Higher Colleges of Technology in the United Arab Emirates before returning to Ghana to work for the Ministry of Roads and Highway.