Identifying and Quantifying Rates of State Motor Fuel Tax Evasion (2008)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

146 A P P E N D I X D Report: Addanki, S., Cohen, Y., and Dunbar, F. January 1987. Gasoline Tax Evasion. National Economic Research Associates. Washington, D.C. This study applies two approaches to estimating evasion of the federal gasoline tax: a comparison of nationwide gasoline consumption to gallons on which federal excise taxes are collected and a regression analysis trending Internal Revenue Service (IRS) taxed gallons to Federal Highway Administration (FHWA) and Energy Information Administration (EIA) data. The first approach to estimating evasion compared estimates of nationwide gasoline consumption from 1979-1986 to estimates of gallonage on which excise taxes were collected. The findings of this analysis suggested that there was little evasion from 1979-1982; however, following a tax rate increase in 1983, there was much greater evasion from 1984-1986. During the 1979-1982 time period, there was an estimated 1.8 billion gallon shortfall – which could have been a direct result of exemptions to fuel taxes – but the shortfall grew to over 7.1 billion gallons during the 1984-1986 time period. The 5.3 billion gallon increase in the shortfall correlates to $480 million in annual forgone revenue due to evasion. The second approach trended IRS data relating to gallons taxed to sets of consumption measures for the period 1974-1982, and then used the trend line to predict taxed gallons from 1984-1986. The predicted amount of fuel consumed was then compared to the amount of gallons taxed. The difference between the predicted values and actual collections during the 1984-1986 time period was 5.6 billion gallons per year, resulting in an annual evasion estimate of $510 million. A separate model was developed for application in New York State. The model estimated that unreported gasoline sales ranged from between 14.5 and 20.9 percent of reported sales. That range of evasion amounts to $168.4 and $254.5 million in annual lost state and local tax revenue due to evasion. Report: Auerbach, Alan. July 1999. On the Performance and Use of Government Revenue Forecasts. University of California, Berkeley and NBER, http://www.kerosun.com/irstax.htm This paper analyzes and evaluates the performance of government revenue forecasting models, specifically focusing on the forecasts performed by the Congressional Budget Office (CBO) and Annotated Bibliography

147 Report: Battelle Memorial Institute (BMI) and National Renewable Energy Laboratory (NREL). July 2001. “Vehicle Technology Assessment for MPG Impact and Forecast Highway Revenue Forecasting Model (HRFM),” Fuel Module. Columbus, Ohio. This study analyses the potential effects on fuel economy resulting from recent EPA emissions’ requirements, new vehicle technologies likely to be implemented in the next 20 years and pressure to reduce U.S. consumption of foreign petroleum. This information is required to update the fuel economy inputs to the Highway Revenue Forecasting Model. The four vehicle types classified and explored by this report are passenger cars, light trucks and SUVs, medium- duty trucks, and heavy-duty trucks. Fuel economy for each vehicle class is estimated for near- term, mid-term and long-term time frames, subject to certain assumptions. Report: Balducci, P. July, 2004. Current Rates of Evasion in the Areas of Federal Motor Fuel and Other Highway Use Taxes: Task 7 Report. Prepared by the Battelle Memorial Institute for the Federal Highway Administration. Portland, OR. This report identifies and provides an analysis of data sources currently available and relevant to motor fuel tax evasion. A general overview for each source of data is given as well as its source, collection frequency, method of collection, limitations and availability. This data is classified into three specific types: motor fuel volumes, tax collections and travel data. Data sources reviewed include Federal Highway Statistics, the Petroleum Supply and Reporting System, the Waybill Sample, the American Petroleum Institute’s Weekly Statistical Bulletin, Waterborne Commerce of the United States, the Commodity Flow Survey, the Statistics of Income Bulletin and the Transportation Energy Data Book. Report: Baluch S. January 1996. Revenue Enhancement through Increased Motor Fuel Tax Enforcement. Washington, D.C. This paper assesses the revenue productivity of motor fuel tax enforcement legislation and motor fuel tax compliance programs. The report notes that in 1994, FHWA estimated federal and state fuel tax evasion at $3 billion annually. Further, the paper presents the findings of an analysis conducted by FHWA to determine diesel fuel tax revenue growth attributable to increased compliance efforts. The paper attributes $600 - $700 million in revenue growth experienced in 1994 to improved compliance from enforcement efforts. Report: Battelle Memorial Institute and National Renewable Energy Laboratory. July 2001. “Vehicle Technology Assessment for MPG Impact and Forecast Highway Revenue Forecasting Model (HRFM),” Fuel Module. Columbus, Ohio. This report documents the update of the fuel module of the Highway Revenue Forecasting Model (HRFM). The HRFM was developed in 1981 as a means for the federal government to develop both short and long-term estimates of federal fuel tax collections. The model has been updated three times since 1981. The principle objectives of the HRFM model update were to assess the the Office of Management and Budget (OMB). Performance evaluation is based on the bias and efficiency of the forecasting models as well as how accurate CBO and OMB estimates are compared to private revenue forecasts. Auerbach concluded that the performance of government revenue forecasts doesn’t differ substantially from private forecasts. However, government forecasts were found to fail various statistical tests of efficiency.

148 The study constructed forecasts of future fuel economy growth by vehicle type. Passenger car fuel economy was forecasted to grow by 1.8 percent annually during the near term (1999-2005), 0.5 percent annually during the mid-term (2005-2010), and 1.5 percent annually during the long- term (2010-2020). Light truck and sport utility vehicle MPG was forecast to grow by 0.5, 2.0, and 1.5 percent annually during the near, mid and long-term, respectively. Medium-duty truck MPG was forecast to grow by 1.5, 0.5, and 1.0 percent annually in the near-, mid-, and long- term, respectively. Finally, heavy-duty truck MPG was forecast to grow by 0.75 percent annually in the near-term, 0.5 percent annually in the mid-term, and 0.5 percent annually in the long-term. Testimony: Brimacombe, Joseph, R. July 17, 2003. Statement of Joseph Brimacombe, Deputy Director, Compliance Policy, Small Business and Self Employed Operating Division, Internal Revenue Service. Testimony before the Full Committee of the House Committee on Ways and Means. Washington, D.C. This testimony presents motor fuel excise tax compliance issues and trends as perceived by the IRS, as well as strategies used to address compliance issues. Among the major evasion issues noted was the misuse of dyed diesel, boot legging, smuggling and cocktailing. The removal of tax exempt aviation fuel from the rack was also listed to be a major issue. Untaxed aviation isn’t dyed and therefore incentives exist to use it for taxable aviation purposes and for on-road diesel trucks resulting in lost revenue from federal and state aviation and diesel taxes. This testimony also discusses IRS strategies and efforts currently underway to curb fuel tax evasion. The Excise Summary Terminal Activity Reporting System (ExSTARS) has been a main focus of IRS efforts towards reducing evasion forms such as misuse of dyed fuel. Fuel finger printing is another technology that can address smuggling, bootlegging and cocktailing. A chemical figure print can be taken at a retail station and can be compared to the finger print taken at the terminal rack where the fuel came from. Another technology that the IRS is currently developing with the aid of the United States Departments of Energy’s Pacific Northwest National Laboratory (PNNL) is an Acoustical Identification Device (AID) that can be used to identify the contents of containers. The AID device can be used to identify smuggling at borders and would be much more efficient than the stop and sample method currently being employed. Finally, the IRS has developed a registration process for taxpayers that engage in tax free transactions of motor fuel. Report: Center for Balanced Public Policy (CBPP). March, 2004a. “Motor Fuel Distribution System.” Draft Report Prepared by the Center for Balanced Public Policy for the Federal Highway Administration. Washington, D.C. This report provides a detailed description of the distribution process for gasoline, diesel, aviation fuels, and other non-petroleum fuels. Physical features in the motor fuel distribution process such as refineries, terminals, bulk plants, barges and pipelines are described in general and specifically for the US. This is also done for non-petroleum products like biodiesel. Further, certain features of the nonphysical parts of the distribution system are described, such as exchange agreements (fuel transfers on paper) between oil companies and 637 registrations potential impact to fuel economy resulting from market penetration of alternative fuels, pressure to reduce dependence on foreign petroleum consumption and new emission control requirements required by the U.S. Environmental Protection Administration (EPA).

(registration for entities using fuel for tax exempt purposes). Finally, the report includes a brief summary of federal tax law related to point of taxation for gasoline, diesel, and aviation fuel. Report: Center for Balanced Public Policy (CBPP). March, 2004b. “Current Rates of Evasion in the Areas of Federal Motor Fuel and Other Highway Use Taxes: Motor Fuel Excise Tax Evasion Schemes.” Draft Report Prepared by the Center for Balanced Public Policy for the Federal Highway Administration. Washington, D.C. This report presents a comprehensive analysis of known motor fuel tax evasion schemes and other undocumented but possible evasion schemes. The CBPP report outlines 8 broad evasion methods: refunds and credits schemes, non-filing schemes, removals from bulk systems, imports and exports, daisy chains, dyed diesel and kerosene, false labeling and blending schemes. Within these categories, 35 actual or potential evasion techniques are described in depth. Each technique description includes affected fuels, distribution sector involved, factors contributing to the scheme and actual documented cases. The study notes that while categorization of techniques is useful in order to identify elements of an evasion scheme when it is encountered, it is not always the case that evasion schemes fit into neat categorical groups. In reality, evasion practices are usually very complex and involve elements of several evasion methods. Report: Council of State Governments and Council of Governors’ Policy Advisors (CSG&CGPA). 1996. Road Fund Tax Evasion: A State Perspective. Lexington, Kentucky. This report presents the findings of a joint project undertaken by the Council of State Governments (CSG) and the Council of Governors’ Policy Advisors (CGPA). The study classifies and describes evasion techniques, including failure to file information, filing of false information, filing of false exemptions, and failure to pay assessed taxes. The report documents the findings of three separate analyses of state fuel tax evasion conducted for this study. The three approaches used in this study to estimate evasion were a literature review and examination of prior research relating to tax evasion, a survey of state tax administrators, and the development of a statistical model. From the literature review, it was estimated that state fuel tax evasion was approximately $1.5 billion annually. The survey-based approach resulted in a $1.2 billion evasion estimate. Finally, the statistical model estimated $952 million in annual losses due to evasion. The basis of the model was to estimate anticipated motor fuels consumption, and then to compare the estimate to actual taxed gallons. Any disparity between the estimated consumption and tax gallons was assumed to be due to evasion. Variables used in the motor fuels demand model were: income/wealth, population characteristics, price variables, geographic dispersion variables and other variables. Report: Davis, Stacy C. October 2000. Transportation Energy Data Book, Edition 20. U.S. Department of Energy, Center for Transportation Analysis, Oak Ridge National Laboratory, ORNL 6959, Oak Ridge, Tennessee. The Transportation Energy Data Book is an annual compilation of statistics relating to transportation activity and factors that influence transportation energy use. The annual Transportation Energy Data Book focuses on petroleum production, energy use, transportation- related environmental externalities, vehicle data and non-highway modes. ORNL compiles data 149

150 from several sources, including: Federal Highway Administration, Environmental Protection Agency, Energy Information Administration, National Personal Transportation Survey, Bureau of Economic Analysis, Bureau of the Census, R.L. Polk Company and Eno Transportation Foundation. The Transportation Energy Data Book is compiled by the Oak Ridge National Laboratory under contract with the United States Department of Energy. Report: Davis, S., Hu, P., and Schmoyer, R. 1998. Registrations and Vehicle Miles of Light- Duty Vehicles, 1985-1995. ORNL-6936. Oak Ridge National Laboratory. Oak Ridge, Tennessee. This was a follow-up study to one performed by Davis et al. in 1994. The previous study estimated the amount of motor fuel used for off-highway recreation at the state level by vehicle type. In this study, recreational fuel was defined as federally taxed gasoline, gasohol, diesel fuel or special fuel used in recreational motorized vehicles on recreational trails of backcountry terrain. The project assisted FHWA in the determination of how the amounts transferred to the Trails Trust Fund could be apportioned equitably to individual states. This study reevaluated the model developed from the 1994 study. More current and accurate data sources were used to produce updated estimates for FHWA. Report: Denison, D. and Edger, Robert J. III. April 2000. “Tax Evasion from a Policy Perspective: The Case of the Motor Fuels Tax”. Public Administration Review Vol. 60, No. 2. This article examines fuel tax evasion from a policy perspective. Evasion techniques and federally implemented measures to fight fuel evasion are broadly described. Measures to increase fuel tax compliance by state legislators are examined in 16 southern region states. Four policy instruments are specifically explored: tax collection points, penalties and punishments, liability for fuel tax, and visibility and enforcement. There is a wide variety of collection points for the 16 southern states examined at the time of this article. Many states collect taxes at the terminal rack. Others license, bond and collect taxes from wholesalers. There also exists a wide assortment of penalties for fuel tax evasion between the southern states. Some states consider fuel tax evasion to be a felony while others deem it as a misdemeanor. It was noted that there is significant debate as to the effectiveness of penalties and punishments for deterring tax evasion. Most of the southern states hold corporate officers accountable for the tax activities of the entity. Finally, the majority of the 16 states examined for this study enforce dyed fuel restrictions and impose penalties for misuses of dyed fuel. Report: Denison, D. and Hackbart, M. July 1996. The Motor Fuel Tax Evasion Issue in Kentucky KTC-96-6. Kentucky Transportation Center. Lexington, Kentucky. This study was conducted in cooperation with the broader CSG/CGPA study of state fuel tax evasion. The report presents an overview of the Kentucky highway tax structure, details fuel tax evasion methods, and documents efforts aimed at reducing fuel excise tax evasion in Kentucky. The report uses the data and methods developed in the CSG/CGPA study to estimate fuel tax evasion in Kentucky. These methods include a survey of state fuel tax administrators and an

econometric analysis. Based on these methods, this study estimates that evasion of Kentucky fuel taxes costs the state highway fund approximately $26-$34 million annually. The report also makes several recommendations for improving motor fuel tax compliance including: participating in regional task forces, implementing an 11-point federal plan for reducing evasion, assessing the marginal costs of additional field auditors, modifying the state motor fuel tax administration model to bring it in line with the federal model, educating the public, conducting a study to precisely determine the level of fuel tax evasion in Kentucky and assessing evasion of other highway user taxes (e.g., registration fees, weight-distance taxes and other highway user fees). Report: Eger, Robert J. 2002. Wisconsin’s Off-Road Fuel-Tax Collection Process: A Midwestern Comparative Analysis and Assessment. Final Report SPR 0092-02-08 Prepared for the Wisconsin Department of Transportation. Madison, Wisconsin. This report examines the potential for evasion resulting from Wisconsin’s fuel tax exemptions for various off road uses. To address this issue, the report is organized into three sections. First, a statistical analysis of tax-exempt fuel consumption and refunds is conducted for Midwestern states. Second, an analysis is done comparing Wisconsin’s motor fuel tax law to other states in the region. Last, the study provides a series of policy options to improve the enforcement of Wisconsin’s motor fuel tax law. The statistical analysis involves comparing Wisconsin’s off-highway gasoline rebates to other Midwestern states. Using monthly panel data from 1994 to 2000, the study first compares Wisconsin’s gasoline refunds to its border states: Illinois, Iowa, Michigan and Minnesota. A second analysis is conducted combining the same data for the same time period into annual data. These analyses segregate refunds into three categories: agricultural refunds, industrial refunds and refunds for gasoline used for marine purposes. Controlling for a number of factors including average acres of farms, fuel tax, and industrial value, this study estimates that Wisconsin’s agricultural use of gasoline, as inferred by amount of refunds, is 1,600,000 gallons higher per month and 16,000,000 higher per year compared to other Midwestern states. This analysis indicates that evasion from agricultural gasoline is likely occurring in Wisconsin. To the extent that refund fraud is what explains the incidence of much greater refunds for agricultural use of gasoline, the resulting losses in revenue for the state of Wisconsin are upwards of $425,000 monthly and $4,000,000 annually. The opposite is the case for the analysis of industrial gasoline refunds. Based on the findings of this report, Wisconsin has approximately 900,000 gallons lower in monthly instruction gasoline refunds than average Midwestern states. For marine use of gasoline, Wisconsin is average with respect to its border states. Based on an analysis of Wisconsin’s tax law regarding motor fuel tax exemptions, a series of policy options are proposed in this report to improve Wisconsin’s enforcement of motor fuel taxes. First, the study found that exempt purchasers of gasoline file paper work with the supplier and the supplier was responsible for paying the tax and relate the exemption information to the state. The study recommends that statutes be amended to presume that motor fuel is used for highway vehicles so that refund applications get filed directly with the state. The deduction for tax free use of fuel would be removed and suppliers would be required to collect and remit tax. 151

152 The purchaser would now be responsible for filing the tax refund. Second, it was recommended that the amendment should be accompanied by a permit process for claimants to provide substantiation of exempt usage of fuel. Third, the study recommends that the state of Wisconsin gather additional information about claimants to facilitate audits, collections and enforcement processes. Lastly, the study recommended that criminal fuel tax evasion be penalized as a felony and fines should be enhanced for repeat violators. Report: Eger, Robert J. III. and Hackbart, Merle. July 2001. State Road Fund Revenue Collection Processes: Differences and Opportunities of Improved Efficiency. KTC-01-17/SPR- 99-192-1F Kentucky Transportation Center. Lexington, Kentucky. This report presents an appraisal of the effects of enforcement and audit practices on fuel tax assessments and makes recommendations to improve the efficiency of these processes in Kentucky. To assess the affect of administration processes on assessments, three models are estimated using regression analysis. One linear model is estimated for assessments per million VMT for trucks in FY 1997. Two models are estimated for FY 1997, one model in linear form and another in log-log form. The independent variables used for these models include the number of auditors in each state, the state excise tax rate, income per capita, miles of state owned urban road, miles of state owned rural road, amount of federal tax awarded to each state, and whether the state collects taxes through the revenue department. Since the authors were interested specifically in Kentucky and the region around Kentucky, a dummy variable was added to these models to include Kentucky and all states bordering Kentucky. Cross-sectional data used to model tax assessments in this study was attained from a survey of tax administrators for each state. The number of auditors in a state was found to be statistically significant for all the estimated models. The log-log model specifies auditors as elastic with respect to assessments. That is, a 10 percent increase in the number of auditors corresponds to 19.58 percent increase in assessments. Federal tax apportioned to each state is found to be statistically significant in both linear models. For all models, there are no statistically significant differences between Kentucky region states and other U.S. states. Likewise, there is no evidence from any of the models that assessments are affected by whether a state collects tax revenue from a revenue department or any other state agency. This study recommends six policy prescriptions to make the fuel tax administration process more efficient. Some of these recommendations do not result from the statistical analysis of this study specifically but rather are the result of general knowledge of the tax evasion problem as it relates to the circumstances facing Kentucky in particular. First, the authors recommend uniformity between Kentucky and its border states. Second, it is recommended that Kentucky consider increasing their registration fees because of opportunities for evasion due to higher fees in other states. The third recommendation is to consolidate Kentucky’s two audit groups into one since the statistical models showed no correlation between what agency performs audits and actual assessments. The consolidation may be more efficient than handling fuel tax compliance split between two separate agencies. Fourth, it is recommended that Kentucky increase its audit staff since the statistical analysis showed significant positive correlations between auditors and assessments. Fifth, the authors recommend that Kentucky evaluate multiple year estimates of

assessments to enable a time-trend analysis to analyze the effects of administration and audits on assessments. Finally, the authors recommend substantiation of their results through further model development and better data. Report: Federal Highway Administration, U.S. Department of Transportation. August 1997. 1997 Federal Highway Cost Allocation Study. Washington, D.C. The federal Highway Cost Allocation (HCAS) study assesses the equity and efficiency of the federal highway tax structure by comparing the costs that each vehicle class imposes on the road system to the share of revenue deposited in the Federal Highway Trust Fund (HTF) attributed to each user class. Travel, revenue and engineering data are used to calculate the cost responsibility of each highway user class. The travel data generated for the 1997 Federal HCAS were used to estimate total tax revenue (including fuel tax revenue) for each vehicle class. The base period of the study was 1993 to 1995. The Federal HCAS calculated equity ratios for each vehicle class by comparing cost responsibility to total revenue attributed to each class of vehicle. Equity ratios of less than 1.0 demonstrate that costs imposed on the road system by a class of vehicle exceeds revenues attributable to that class of vehicle. The equity ratio for passenger vehicles was estimated at 1.0, whereas the equity ratios for single-unit and combination trucks are 0.8 and 0.9, respectively. The study found that the equity ratio for single-unit trucks weighing in excess of 50,000 pounds was 0.5 while the equity ratio for combination trucks weighing in excess of 80,000 pounds was 0.6. The Federal HCAS also presents an overview of the Highway Revenue Forecasting Model (HRFM). FHWA developed HRFM in 1981 as a means for the federal government to obtain both short and long-term estimates of federal fuel tax collections. The model has been updated three times since 1981. The HRFM was used in the 1997 Federal HCAS to attribute federal highway user revenues to 20 vehicle classes and 30 weight categories. Within the HRFM, fuel tax revenues for individual vehicle classes are based on: vehicle fuel efficiency (i.e., miles per gallon or MPG), vehicle miles of travel (VMT) and the operating weight of the vehicle. The MPG estimates are developed for each vehicle class based on a typical engine type, transmission and vehicle performance characteristics, operating weight groups, and vehicle fuel types. Key input data for the MPG estimates come from the Vehicle Inventory and Use Survey [(VIUS; prior to 1997, VIUS was named the Truck Inventory and Use Survey (TIUS)], several publications by the EPA for new vehicle fuel efficiencies, and information from the United States Department of Energy, American Trucking Association, the United States Department of Transportation (U.S. DOT), as well as other private sources (e.g., Polk Company data). Given the number of different sources, engineering judgment and adjustments have been necessary at times to resolve inconsistencies among data and to weight the relative importance of each factor to the MPG. The HRFM estimates the gallons of fuel consumed by dividing VMT by MPG for each vehicle class and operating weight category, and the revenues are then a function of tax rate and gallons of fuel consumed. Report: FHWA, Federal Highway Administration, U.S. Department of Transportation. December 20, 1999b. Motor Fuel Tax Evasion Summary. Washington, D.C. http://www.fhwa.dot.gov/policy/summ.htm 153

This is a brief summary of fuel tax evasion and efforts to curb fuel tax evasion. It includes estimates of federal fuel tax evasion, major federal legislation relating to fuel taxes. and estimates of the benefit of state and federal enforcement practices. Report: FHWA, Federal Highway Administration, U.S. Department of Transportation. December 17, 1999c. Revenue Impact of Diesel Fuel Dyeing. Washington, D.C. http://www.fhwa.dot.gov/policy/fueldye.htm This is a brief summary of the federal diesel fuel dyeing program. It includes estimates of revenue losses resulting from evasion of diesel taxes and the revenue impact of, and how sates have responded to, the federal diesel dyeing program. Report: Federal Highway Administration, U.S. Department of Transportation. Office of Highway Information Management. August 1998b. TEA-21 and Estimation of Highway Trust Fund Tax Receipts Attributable to Highway Users in Each State. Washington, D.C. This report provides an overview of the process used to attribute federal fuel tax revenues to states. It also documents several budget provisions set forth in the Transportation Equity Act for the 21st Century (TEA-21) that will increase the need for improved accuracy in the allocation process. The report notes that during the attribution process, FHWA attempts to allocate fuel tax revenues to states based on where the fuel is consumed rather than where federal fuel taxes are collected. Because federal fuel taxes are collected at the terminal rack, fuel tax returns are generally filed by oil companies. FHWA estimates of highway use of gasoline are based on reports submitted by state fuel tax agencies. The factor used to attribute fuel tax revenues to states is the ratio of highway use of gasoline within each state to the highway use of gasoline in all states. The accuracy of the apportionment estimates has varied on a state-by-state basis. Total highway account excise tax receipt forecasted in TEA-21 lagged below actual receipts by 4.2 percent in 1998, 3.5 percent in 1999 and 7.5 percent in 2000. However, actual receipts of highway excise taxes in 2001 exceed TEA-21 forecasts by 6.0 percent. Report: FHWA, Federal Highway Administration, U.S. Department of Transportation. August 1997. Federal Highway Cost Allocation Study Final Report. Federal Highway Administration. Washington, D.C http://www.ota.fhwa.dot.gov/hcas/final/four.htm To evaluate highway-related costs attributable to various types of vehicles, FHWA performs periodic highway cost allocation studies. The primary purpose of these studies is to evaluate the equity of federal highway user fees by examining which user fees cover highway cost responsibility for different vehicle classes. Those paying more than their share of highway costs are, for all intents and purposes, subsidizing the operations of others. To discern how fair federal highway fees are, equity ratios are calculated for each vehicle class by comparing total revenue for each vehicle class to the costs each vehicle class impose on the highway infrastructure. An equity ratio of 1.0 means that a particular vehicle class is exactly covering its share of the cost 154

responsibility. The most recent analysis found that the equity ratio for combination trucks weighing less than 50,000 lbs. was 1.4 while the equity ratio for combination trucks weighing more than 100,000 is 0.4. Report: Federal Highway Administration, U.S. Department of Transportation. June 1992. Fuel Tax Evasion: The Joint Federal/State Motor Fuel Tax Compliance Project. Washington, D.C. This study provides an overview of the fuel distribution system and federal excise tax structure. Further, the study estimates total federal gasoline tax evasion at between 3 and 7 percent ($466.1-$1,087.5 million) and diesel tax evasion at between 15 and 25 percent ($645.2-$1,075.3 million). These estimates are based on a review of previous studies, congressional testimony, and IRS auditing records. The study also reviews and analyzes several tax evasion schemes and provides information relating to specific examples of tax evasion, as detected through auditing efforts. Finally, it identifies several techniques to close current avenues for evasion, thus reducing losses to the federal government. Survey: Federal Highway Administration, U.S. Department of Transportation. Various Years (a). National Personal Travel Survey. Federal Highway Administration. Washington, D.C. The Nationwide Personal Transportation Survey (NPTS) consists of a periodic survey of household-level data on demographics, motor vehicle ownership and vehicle travel. The NPTS was conducted in 1969, 1977, 1983, 1990 and 1995. The survey is produced by the United States Department of Transportation. The 1995 NPTS was based on three types of survey methods: Vehicle-Based Estimates – Vehicle miles of travel for household motor vehicles were generated based on owner estimates of travel and annual odometer readings. Average travel per vehicle was multiplied by the number of household vehicles to generate annual VMT. Driver-Based Estimates – Respondents were asked to estimate the annual number of miles they travel in all vehicles, including commercial during a 12-month period. Because this estimate includes commercial vehicle miles of travel, the driver-based estimate significantly exceeds vehicle-based estimates. Trip-Based Estimates – Trip diaries were used to estimate travel for all purposes, including commercial driving. Respondents were asked to itemize their trips in diaries during the previous day. Further, respondents were asked to note any trips in excess of 75 miles during the two weeks proceeding completion of the survey. For each respondent, VMT and trip information is matched with demographic and ownership data to analyze trends in vehicle ownership and usage for groups varying by age, gender and economic standing. As noted in Appendix B, the divergence between the 1995 NPTS VMT estimates and the 1995 FHWA VMT estimate for passenger vehicles ranges from 3.5 percent (personal estimates) to 0.6 percent (odometer readings). 155

Report: Federal Highway Administration, United States Department of Transportation. 1981- 2000. Highway Statistics. Federal Highway Administration, Office of Policy Information. Washington, D.C. FHWA annual highway statistics presents annual estimates of vehicle miles of travel (VMT), vehicle fuel economy, lane-mileage, state and federal highway revenues and expenditures and a number of other transportation-related indicators. FHWA presents VMT estimates for six vehicle classes (automobiles, motorcycles, light trucks, single-unit heavy trucks, combination trucks and buses) on an annual basis. FHWA also presents annual VMT estimates for each state and each functional class road system, as defined in the Highway Performance Monitoring System (HPMS) data system. HPMS VMT estimates are based on traffic counts performed by states using roadside traffic monitoring devices (e.g., pneumatic tubes, inductive loops and manual counts). Average annual daily traffic (AADT) is reported by states for each section of Interstate, National Highway System (NHS) and other principal arterial. VMT for each road segment is the product of AADT and centerline miles. That is, to estimate annual travel along a 50-mile segment of roadway with eight interchanges, a state could place one vehicle recording device on each of the seven stretches of roadway between each pair of interchanges and count the number of vehicles crossing those devices each day throughout the year. The number of vehicles counted at each roadway segment would, in turn, be multiplied by the corresponding centerline miles for each segment, such that if the distance between the first two interchanges was 10 miles, a one-year vehicle count of 100,000 cars would be multiplied by 10 to produce a VMT estimate of 1,000,000 for that first highway segment. For minor arterial, rural major collectors and urban collectors, VMT estimates are based on sample AADT counts, centerline miles, and expansion factors based on seasonal travel fluctuations for each roadway segment. AADT and travel reported by states are edited by HPMS software to remove unusual and erroneous data. FHWA also consults with states to smooth data and remove invalid values. FHWA requires states to perform counts on the entire HPMS system every three years. Where no data are collected during a single year, permanent vehicle counting station data and historical trend analysis are used to produce VMT estimates. Fuel economy data are presented in Highway Statistics on an annual basis. National fuel consumption is constructed from state fuel tax records. Total fuel consumption, and thus, miles per gallon (MPG) data for individual vehicle classes are derived by FHWA from states reports, the Vehicle Inventory and Use Survey (VIUS) (USCB 2002b) and other independent sources of data. Meeting Minutes: FHWA. Federal Highway Administration, United States Department of Transportation. Various Years. Joint Federal/State Motor Fuel Tax Compliance Project Steering Committee Meeting Minutes. Washington, D.C. 156

These are minutes of the annually held Joint Federal/State Motor Fuel Tax Compliance Project Steering Committee Meetings. These minutes generally contain a summary of reports made by individual states, regional motor fuel tax task forces, IRS, FHWA and other visiting organizations or individuals. Reports typically include discussion of any new legislation, tax compliance efforts, new studies and adoption of fuel tracking systems or electronic reporting systems. Report: Francis, Brian. 2000. Gasoline Excise Taxes, 1933-2000. Statistics of Income Bulletin. Washington, D.C. This report gives a history of federal gasoline excise taxes. It contains discussion of the highway trust fund, tax rates, legislative changes relating to gasoline taxes and historical federal gasoline tax collections. Presentation: FTA, Federation of Tax Administrators, ExSTARS/ExTOLE Presentation. August, 8 2004b. http://www.taxadmin.org/fta/meet/01am_pres/andersgen.pdf This presentation gives a brief overview of the ExSTARS and ExTOLE subcomponents of ExFIRS. Specifically, these systems are discussed in terms of their main functions, stage of development, usefulness to states and major benefits. Report: FTA, Federation of Tax Administrators. ExFIRS Background. August, 8 2004c. http://www.taxadmin.org/fta/mf/exfirs_back.html This document briefly summarizes the ExFIRS system. The report discusses ExFIRS historical development, purpose, benefits, subsystems and limitations. Report: Federation of Tax Administrators. September 2003. FTA Motor Fuel Tax Section Uniformity Project. Washington, D.C. This report provides documentation of ongoing efforts of the FTA Motor Fuel Tax Section Uniformity Committee. The Uniformity Committee adopted an 11-point plan in an effort to make administration of fuel taxes more efficient and more consistent between states, to improve information exchange and to encourage cooperative efforts between states to reduce evasion. The major points in this plan include: uniform definitions for imports and exports, federal identification codes that distinguish entities for reporting and information exchange, total accountability of fuel by licensing of all resellers and requiring third party reporting on the movement of fuel, uniform electronic reporting systems and trainings for auditors and investigators. There is a subcommittee for each point of the 11-point plan. This report documents the purpose and progress of each subcommittee. Items that document the subcommittees’ advancement included in this report are uniform definitions, tax administration forms and schedules and sub- schedules. Also included in the report is a model legislation checklist for states seeking to change their administrative procedures to curb fuel tax evasion. 157

Report: Festin, S. May 1996. Summary of National and Regional Travel Trends: 1970-1995. Federal Highway Administration. Washington, D.C. This report analyzes trends in VMT data in the United States and five regional areas during the 1970 – 1995 timeframe. Travel data are summarized by time increments (annual, monthly, weekly and daily), and the distribution of travel is compared among days, months and years. The study found that nationwide VMT has grown by an average of approximately 3 percent annually during the study timeframe, and that urban travel comprises 60 percent of total travel. The report also documents the phenomenon known as peak spreading, where peak travel times expand during morning and evening periods of high traffic. Model: Fleming, D. August 2001. Personal Communication with David Fleming of the Maryland Department of Transportation on the State of Maryland’s Motor Fuel Tax Revenue Estimation Methodology. Maryland has designed a model to estimate revenue generated by the state’s motor fuel tax. The model uses regression analysis to estimate future annual gallons of motor fuel purchased. The Maryland Department of Transportation (MDOT) uses data prepared by the economic forecasting firms DRI/WEFA and Economy.com to support its model. The model uses a two- variable, Real GDP and the Implicit Price Deflator for Gas and Oil, equation to forecast motor fuel tax revenue. Since 1990, the forecast has, on average, been accurate within 0.9 percent of actual receipts. The annual variance did not exceed 2.5 percent during the 1990 – 2001 timeframe. Report: General Accounting Office. June 2000. Highway Funding: Problems with Highway Trust Fund Information Can Affect State Highway Funds. Washington, D.C. This report describes the relationship between highway user tax receipts and funds available for federal highway funding programs. It reviews Treasury’s process for allocating tax receipts and FHWA’s process for attributing collections to individual states. It assesses the appropriateness of the mechanisms and assumptions used in the collection and application of the data used to distribute federal-aid funds to states. Finally, it provides recommendations to improve the accuracy of federal funding distributions to states. The report also provides an overview of the process for attributing revenue to states. The Treasury does not provide data on receipts at the state level to FHWA. To disaggregate data and distribute funds to states, FHWA relies on travel and fleet fuel efficiency data provided by individual states to estimate state-level contributions to the Highway Account (HA) of the federal HTF through what is known as the “attribution process”. The Treasury’s process for allocating tax receipts to the HA is analyzed within this report and problems with accuracy and timeliness of the data used in the process are identified. The report notes that a recent analysis conducted by the IRS recommends that the Treasury not require individual taxpayers to provide detailed information at the time they make deposits due to compliance costs. The IRS study also recommended: a) offering incentives to encourage 158

taxpayers to provide more detailed data at the time of deposit, and b) reviewing the issue in several years to determine if technological and data collection methods have advanced in a manner that makes the collection of more detailed data less burdensome from a compliance standpoint and, thus, more feasible. The report contends that the reliability of the data and methods used in the FHWA attribution process are questionable. The reliability, accuracy, and consistency of the data submitted to FHWA by the states is not verified through an independent review or audit, and the responsibility for implementing and verifying the attribution process rests with only two individuals at FHWA. The report also provides several recommendations to the Secretary of the Treasury for improving the accuracy and reliability of the data used in the federal fuel tax allocation process. Report: U.S. General Accounting Office. June 1997. Highway Funding: The Federal Highway Administration’s Funding Apportionment Mode l. (GAO/RCED-97-159). Washington, D.C. This report presents the findings of a technical review of the highway funding apportionment model designed for FHWA to perform formula allocation procedures during the development of TEA-21. The findings of the study conclude that the model captures the structure of the federal highway funding allocation process and is adaptable for use in evaluating new and proposed apportionment formulas. The report also notes that the data used for the model are questionable and are not properly certified by FHWA. The report makes recommendations for improving the model and accuracy of the data used in the model. FHWA offices charged with developing data for use in the allocation model are required to certify that the data are correct. In some cases, proposed apportionment formulas require the use of data not presently used in the federal highway allocation process. In these instances, the data may not be certified or properly scrutinized. Report: U.S. General Accounting Office. January 1996. Tax Administration: Diesel Fuel Excise Tax Change. (GAO/GGD-96-53) Washington, D.C. This document reports on the effects of the federal diesel fuel dying program established by the Omnibus Budget Reconciliation Act (OBRA) of 1993. It further discusses whether prominent concerns by stakeholders were addressed by the IRS regulations implemented in that program. The report states that diesel excise tax collections increased by $1.2 billion, or 22.5 percent, in the first calendar year after the diesel fuel dying law took effect. This increase does not include the increase in revenue due to the tax rate increase. It also reports that the Treasury Department estimates that increased diesel tax compliance resulting from the dying program is estimated to be $600 to $700 million. The report notes that even though tax collections had increased from the diesel fuel dying program, many opportunities still exist to cheat the system (i.e. refund fraud). The report remarks that the IRS at that time had only addressed stakeholder concerns about dying requirements by using red as the dye color and declining to use colorless markers at that 159

time in response to several stakeholder complaints that such a system would be too burdensome. The study also points to a number of unaddressed concerns. One unaddressed concern was the fact that kerosene hadn’t been dealt with in the OBRA legislation and kerosene could still be used in blending evasion schemes. Further, many stakeholders felt that the concentration of dye was too high and would adversely affect transportation and holding equipment. Recreational boaters were concerned that they would not have easy access to dyed fuel because boating marinas commonly only had one storage tank that would be used for undyed fuel for commercial uses and the costs of adding another for dyed fuel are too great. Report: U.S. General Accounting Office. May 1992. Status of Efforts to Curb Motor Fuel Tax Evasion, Report to Congressional Requesters. Washington, D.C. This report notes previous estimates of evasion, and focuses on the $1 billion total federal gasoline and diesel tax evasion estimate often cited in the literature. The report also provides an overview of the fuel distribution system in the United States and describes methods for evading federal fuel taxes. The inability of the federal government to identify and quantify total fuel tax evasion is discussed and an Appendix is dedicated to examining the numerous problems inherent with estimating evasion. The report notes the many anti-evasion techniques employed by the federal government to date, including moving the point of tax collection to the terminal rack and dyeing non-taxable fuel. The effectiveness of IRS compliance efforts is analyzed. Also, state initiatives with potential for application at the federal level are reviewed. Finally, the impact of compliance efforts on industry is also examined. Report: Gillen D. 2000. Estimating Revenue from User Charges, Taxes, and Fees: Identifying Information Requirements. Resource Paper, Transportation Economics Research Committee. Washington, D.C. This paper provides a comprehensive overview of highway finance systems at the state level and addresses the role of revenue forecasting. The report also reviews and assesses several methods used by states to forecast fuel tax revenues. The author suggests that analysts need better data and methods to accurately measure vehicle miles traveled, vehicle fuel efficiency and fuel consumption. Better analytical tools are needed, particularly those that can anticipate behavioral responses to changes in fuel prices and other relevant factors. The report contends that fuel tax forecasts are rarely accurate beyond a three to five year timeframe. The report noted that the level of expected revenue is directly tied to the form of a state’s fuel tax. Thus, fuel sales taxes are volatile, waxing and waning with fuel prices. Fixed per-gallon tax rates are less volatile but are unresponsive to inflation. The paper also reports that there are three approaches to consider when estimating fuel tax revenue: a) the simple model that is based on historical data, b) a model based on an econometric time-series approach such as the ARIMA model, or c) a causal forecasting model using relevant economic and demographic variables. 160

There are several states that forecast revenue based on historical data. The report points out that the use of a simple model implicitly assumes that demand for travel and fuel is not linked to economic and demographic variables. Report: Gittings, G. and Narayan, B. 1996. Federal Highway Revenue Estimation: Cost Allocation Perspective.” Transportation Research Record 1558. Washington, D.C. This report presents a revenue estimation methodology used to attribute federal highway revenue to individual passenger and heavy vehicle classes. It presents and analyzes variables that influence future growth in federal transportation revenue. The paper also presents both short and long-term recommendations for improving Federal revenue estimating procedures. An aggregate demand model was used to forecast transportation revenue based on three components: demand for a fleet of vehicles, demand for new vehicles and demand for VMT. New personal vehicle-sales forecasts were constructed using the following variables: GDP, unemployment, annual fixed capital costs of owning and purchasing a car, annual operation costs, and a variable on the relative burden of housing expenditures. In addition, a dummy variable was used to deal with market interruptions caused by labor stoppage in the motor vehicle industry. Truck commercial-vehicle sales were forecast as a function of: intercity truck-freight ton-miles, unemployment, annual ownership costs, real price of diesel fuel and two dummy variables used to account for legislative changes affecting the motor carrier industry and the deregulation of the trucking industry. Fleet size was forecast based on the fleet size of the previous year plus new vehicle sales minus scrappage rates. VMT was forecast as a function of fleet size, annual vehicle operating costs and a dummy variable used to account for fuel shortages. Report: Hallquist, Theresa E. December 1999. “A Comparison of Selected EIA-782 Data with Other Data Sources.” Petroleum Marketing Monthly. Energy Information Administration, Washington, D.C. This report presents an overview of fuel price and volume data collected by the Petroleum Division (PD) of the EIA using the EIA-782 survey from. The report compares EIA-782 price data with price data published by the BLS, and it compares EIA-782 fuel volumes data with data published in EIA’s Petroleum Supply Annual (PSA) and FHWA’s Annual Highway Statistics. The report notes that there is significant historical divergence between EIA and BLS gasoline price estimates because: the EIA survey does not include taxes but BLS price estimates do, EIA surveys target producers and distributors of gasoline while the BLS targets retail outlets, EIA prices are volume-weighted while BLS prices represent a simple average of monthly prices of varying grades of gasoline, and BLS prices represent a point-in-time estimate while EIA prices are weighted based on total monthly sales. 161

The report also notes that there are large and irreconcilable historical differences between gasoline consumption estimates produced by EIA and data published by FHWA, such that FHWA estimates of taxed gallonage actually exceed the EIA estimates of gasoline supplied to the transportation sector, as shown in the PSA. Further, EIA estimates of sales volumes, derived from form EIA-782, significantly exceed the amount of fuel supplied to the system prior to 1994. These obvious discrepancies are the direct result of the divergence between data collection techniques, where the data are collected, errors corrected over time (e.g., double-counting and missing some reporters) and treatment of blending fuels. Note that as data collection techniques have been improved and procedural errors have been detected and eradicated, the discrepancy between EIA and FHWA estimates has declined in recent years. Report: Hagquist, Ron and Dawn Doyle 1999. Fuel Tax Hijacking: How State Governments Are Responding. Government Finance Review. This report discusses fuel tax evasion, providing the history of how the issue became acknowledged, real evasion case descriptions, state actions and the results of those actions and likely future actions. Report : Harper, R. December 2000. Comparisons of Independent Petroleum Supply Statistics. Petroleum Supply Monthly, Energy Information Administration. Washington, D.C. This report compares data relating to crude oil production, crude oil imports, motor gasoline supplied and distillate fuel supplied for numerous sources. The article identifies major discrepancies between data sets and analyzes causes of data variations. The report also provides an overview of the data compiled by the PD of the EIA. The PD compiles and published data regarding the supply and distribution of petroleum in the United States. The data collected by the PD in aggregate comprise the Petroleum Supply Reporting System (PSRS). The PSRS is based on a series of surveys collected from the petroleum industry, state agencies and the Minerals Management Services (MMS) of the U.S. Department of the Interior. PSRS data are published in the Petroleum Supply Annual (PSA), Petroleum Supply Monthly and the Weekly Petroleum Status Report. EIA also uses market surveys to publish data on motor gasoline and distillate fuel supplied in the Petroleum Marketing Annual. Respondent data are collected from refiners, gas plant operators, importers and resellers or retailers. The American Petroleum Institute (API) publishes data relating to crude oil production, imports, motor gasoline supplied and distillate fuel oil supplied in the United States. Crude oil production data are based on information provided by state government agencies. Import data do not include crude oil imported by the Strategic Petroleum Reserve. Total gasoline supplied is based on an assessment of production plus imports (adjusted based on net stock change) minus exports. Import and export data are based on historical industry information provided by importers and operators of pipelines, bulk terminals and refineries. Estimates of distillate fuel supplied are based on distillate fuel oil delivered from primary storage facilities. Distillate fuel estimates do not include kerosene. 162

Additional sources of data included in the analysis are: crude oil production estimates published in the Oil and Gas Journal, data on United States oil and gas reserves published by EIA’s Reserves and Production Division, crude oil import data compiled by the United States Census Bureau and the Fuel Oil and Kerosene Sales Report published by the EIA. The report notes that there are large and irreconcilable historical differences between gasoline consumption estimates produced by EIA and data published by FHWA, such that FHWA estimates of taxed gallonage actually exceed the EIA estimates of gasoline supplied to the transportation sector shown in the PSA. Further, EIA estimates of sales volumes, derived from form EIA-782, significantly exceed the amount of fuel supplied to the system prior to 1994. These obvious discrepancies are the direct result of the divergence between data collection techniques, where the data are collected, errors corrected over time (e.g., double-counting and missing some reporters) and treatment of blending fuels. Note that as data collection techniques have been improved and procedural errors have been detected and eradicated, the discrepancy between EIA and FHWA estimates has declined in recent years Report: Henry, Eric 2002. “Excise Taxes and the Airport and Airway Trust Fund, 1970-2002.” SOI Bulletin, Winter 2003-2004. Internal Revenue Service. Washington, D.C. This report provides a history of aviation fuel excise taxes and the evolution of the aviation trust fund concept. The excise tax on aviation gasoline began with a 1 cent per gallon tax enacted by the Revenue Act of 1932. Today, the aviation tax rate is 21.9 cents per gallon. In 1970, the Airport and Airway Development Act created the Aviation Trust Fund, which was terminated in 1980 to be replaced with the Airport Improvement Fund of 1982. These fund pots have traditionally been used to fund airport improvements such as airport operations, safety measures and noise abatement projects. Today, the aviation improvement program is funded through the Airport and Airway Trust Fund. Report: Hu, Patricia, D. Trumble, and Lu, A. 1994. Fuel Used for Off-Highway Recreation, ORNL-6794. Oak Ridge National Laboratory. Oak Ridge, Tennessee. This Oak Ridge National Laboratory (ORNL) study estimates the amount of motor fuel used for off-highway recreation at the state level by vehicle types. In this study, recreational fuel was defined as federally taxed gasoline, gasohol, diesel fuel or special fuel used in recreational motorized vehicles on recreational trails of backcountry terrain. The project assisted FHWA in the determination of how the amounts transferred to the Trails Trust Fund could be apportioned equitably to individual states. Report: Hwang, Ho-Ling, Lorena F. Truett and Stacy C. Davis. 2003a. The Federal Highway Administration Gasohol Consumption Estimation Model. Oak Ridge National Laboratory ORNL/TM-2003/210. This report discusses ORNL’s review of the FHWA gasohol consumption model and the development of a new gasohol consumption model. The regression-based gasohol estimation model reviewed had been in use for several years prior to 1994, but based on an analytical assessment of that model and an extensive review of potential data sets, ORNL developed a rule- 163

based model. The new model uses data from the Internal Revenue Service, Energy Information Administration, Environmental Protection Agency, Department of Energy, ORNL, and FHWA. The model basically consists of three parts: (1) development of a controlled total of national gasohol usage, (2) determination of reliable state gasohol consumption data, and (3) estimation of gasohol usage for all other states. Report: Hwang, Ho-Ling, Truett, Lorena F., and Davis, Stacy C.. 2003b. A Study of the Discrepancy Between Federal and State Measurements of On-Highway Motor Fuel Consumption. Oak Ridge National Laboratory ORNL/TM-2003/171. This report assesses the discrepancy between state and federal estimates of motor fuel consumption. The Treasury Department collects highway taxes and puts them into the HTF. However, there is no direct connection between taxes collected and gallons of highway fuel used, which leads to a discrepancy between these totals. This study was conducted to determine the magnitude of the discrepancy between the Treasury Department’s estimated total fuel usage based on highway revenue funds and the total fuel usage used in apportioning HTF funds to states. Using data from Highway Statistics for 1991 through 2001, the analysis found that the overall discrepancy between these totals is relatively small, within 5 percent. Further, the discrepancy varies from year to year and varies among gasoline, gasohol and special fuels. Potential explanations for these discrepancies include issues on data, gallon measurement, tax measurement, HTF receipts and timing. For instance, evasion can result in a divergence between fuel use and taxes collected. Further, data anomalies such as deferment of tax payments from one fiscal year to the next can skew fuel tax data. Differences between state data reporting and collection processes will also impact fuel use data. Tax receipt data as conveyed through HTF can be impacted by refunds, credits and transfers. Lastly, a discrepancy can also be caused by timing issues such as calendar year vs. fiscal year. Report: Hwang, Ho-Ling. November 2000. Draft Technical Memorandum on Literature Review of Existing Methods and Models on Revenue Estimation. Working Paper. Oak Ridge National Laboratory. This report provides an overview of revenue estimation methods and models. It provides a summary of the general approach to revenue estimation and detailed descriptions of several different federal level forecasting models. Some forecasting models included in this report are the FHWA Highway Revenue Forecasting Model, Department of Energy’s National Energy Modeling System and IRS’s Excise Files Information Retrieval System (ExFIRS). Revenue estimation processes and methods discussed in this report include the Joint Committee on Taxation, Congressional Budget Office and the Office of Tax Analysis. Report: IRS, Internal Revenue Service. 2001. Criminal Investigation (CI) Annual Report. http://www.irs.gov/irs/article/0,,id=107541,00.html This report summarizes federal efforts to curb motor fuel tax evasion, gives examples of recent and specific evasion cases and statistics on initiated investigations and convictions by the IRS. 164

Report: Internal Revenue Service. 1993-2002. Statistics of Income Bulletin. Washington, D.C. The Statistics of Income (SOI) Bulletins document historical IRS excise tax collections. The SOI Bulletin presents fuel excise tax data stratified according to 25 fuel types – gasoline, gasoline floor stock, gasohol 5.7 percent, gasohol 7.7 percent, gasohol 10.0 percent, gasohol floor stock, gasoline for gasohol 5.7 percent, gasoline for gasohol 7.7 percent, gasoline for gasohol 10.0 percent, gasoline for gasohol floor stock, dyed diesel fuel used in trains, dyed diesel fuel used in trains floor stock, dyed diesel fuel for intercity or local buses, special motor fuels, special motor fuels floor stock, compressed natural gas, alcohol fuels, fuels used commercially on inland waterways, aviation gasoline, diesel fuel, diesel floor stocks, aviation fuel for non- commercial use, aviation fuel floor stock, aviation fuel for commercial use and kerosene. Since 1993, the IRS has published an SOI Bulletin in the spring and fall of each year. Prior to 1993, IRS tax collections were published in the Internal Revenue Report of Excise Taxes. Report: IRS, Internal Revenue Service. 1981-1993. Internal Revenue Report of Excise Taxes. Washington, D.C. Prior to 1993, the IRS prepared quarterly reports of excise tax revenue in these Internal Revenue Reports on Excise Taxes. Motor fuel excise taxes are reported by type of fuel. Excise taxes have subsequently been reported in the Statistics of Income Bulletins. User Guide: IRS, Internal Revenue Service. 1981-1993. Motor Fuel Excise Tax, EDI Guide. Washington, D.C. http://www.irs.gov/pub/irs-pdf/p3536.pdf This document provides general requirements, specifications and procedures for filing electronic forms for fuel terminal operators, carriers and transmitters. It also contains required electronic data interchange (EDI) record and file formats. Report: Jack Faucett Associates. February 1995. Fuel Efficiency, Alternative Fuels, and Highway Trust Fund Revenues. Bethesda, Maryland. This report assesses the future impact of alternative fuels on federal HTF revenue under several scenarios. The scenarios developed for this study are based on input from the ORNL Alternative Motor Fuel Use (AMFU) Model. The report quantifies the diminishing effects of increased fuel efficiency and market penetration of alternative fuels on federal HTF receipts. The report describes in detail the AMFU model and documents estimates of the long-term impact of alternative fuels on federal HTF receipts. Report: Joint Committee of Taxation (JCT). March 1998. Chairman’s Amendment Relating to Extension of Highway Trust Fund Excise Taxes and Related Trust Fund Provisions (Revenue Title to H.R. 2400). Washington, D.C. This report presents a detailed overview of roadway taxation in the United States. Specifically, it focuses on the implementation of highway excise taxes, motor fuel excise tax rates and motor fuel tax exemptions. 165

Report: Joint Committee of Taxation. January 1995. Written Testimony of The Staff of The Joint Committee on Taxation Regarding The Revenue Estimating Process, JCX-1-95. Washington, D.C. The Joint Committee of Taxation (JCT), established under the Revenue Act of 1926, is the official congressional scorekeeper in determining the revenue effects (i.e., budgetary implications) of any proposed tax changes. This report provides a good description of JCT’s revenue estimating methodology. At the JCT, a variety of econometric models are utilized to estimate the revenue impact of changes in tax laws relating to many issues. In some cases, such as the motor fuel excise tax, the information needed to calculate the revenue effects of a proposal may not be available from tax return data or may be available only for a limited number of potentially affected taxpayers. In these instances, the JCT staff must look beyond the Statistics of Income (SOI) data and construct a model that relies on alternative sources of data from other federal agencies. Report: Jorgenson, D. 1998. Growth, Vol. 1. Econometric General Equilibrium Modeling, MIT Press. Cambridge, Massachusetts. This book is the first of two volumes dedicated to the modeling of economic growth. The book presents the concept of an intertemporal price system, where demands and supplies for products and factors of production are balanced at various points in time. In this approach, a forward- looking feature (e.g., linking prices of assets to the present value of future capital services) is combined with a backward linkage between capital services, investment and capital stock. In doing so, the combined forward and backward-looking model captures the long-run dynamics of economic growth. The book includes chapters on aggregate consumer expenditures on energy, two-stage consumer demand for energy, linear growth models, the neo-classical model of development of a dual economy and econometric studies of energy policy and economic growth. Report: KPMG Consulting, Inc. December 2001. Motor Fuel Excise Tax Revenue Leakage Analysis. Prepared for Center for Balanced Public Policy. Washington, D.C. This report compares EIA estimates of domestic jet fuel supply with domestic fuel consumption, as reported by air carriers to the Federal Aviation Administration (FAA). Evasion estimates in this study are based on the assumption that the disparity between EIA supply data and FAA domestic consumption data are due to evasion. The study also considers the possibility that the jet fuel overage represents the use of jet fuel used as an additive or replacement to diesel fuel for taxable highway operations. Based on this set of assumptions, the study estimates a range of evasion resulting from the illegal use of jet fuel of $1.7 billion to $9.2 billion during the 2002 – 2011 timeframe. The range is due to the varying tax rates for fuel potentially displaced by non- taxable jet fuel (e.g., jet fuel for commercial use or diesel fuel for highway consumption). Report: Lazzari, Salvatore. 1997. The Tax Treatment of Alternative Transportation Fuels. CRS Report for Congress. National Council for Science and the Environment. Washington, D.C. 166

This report reviews federal tax treatment of alternative motor fuels in comparison with traditional petroleum highway motor fuels such as gasoline and diesel. The report discusses certain purposes of motor fuel excise taxes: revenue generation for highways, budget deficit reduction and energy policy concerns. Tax rates for several types of fuel are discussed including naphtha, benzene, benzol, casinghead gasoline, natural gasoline, liquefied petroleum gas (propane), liquefied natural gas, gasohol, ethanol blends, dieselhol, compressed natural gas and other blended fuels. It was noted that electricity is not considered to be a highway fuel in current tax code and therefore does not carry a highway tax. Report: Mingo, R., Chastain, R., Mingo, J., and Cummings, S. March 1996. Diesel Fuel Fee Non-Compliance Study. Report for the Oregon Department of Transportation. Salem, Oregon. This report documents the findings of a study conducted for the Oregon Department of Transportation. The study used an econometric model to estimate diesel fuel fee non-compliance rates for all 50 states. The model was specifically designed to estimate evasion rates at varying diesel tax rates in Oregon. Oregon is currently the only state in the nation that does not impose a tax on diesel fuel consumption by trucks weighing in excess of 26,000 pounds. The report also describes several methods for evading the federal and state fuel taxes and documents efforts to combat evasion. The report contends that federal and state diesel taxes experience high evasion rates. Evasion methods are continually evolving in response to changes in the point of collection and methods for tax enforcement. This study estimates the national average state diesel fuel tax non- compliance rate at 21 percent. The 21 percent evasion rate was calculated by comparing FHWA estimates of fee-based gallons to an amount calculated based on VMT and average fuel efficiency data contained in the 1992 Truck Inventory and Use Survey (TIUS). The findings of this study suggest that the level of evasion of state diesel taxes is most dependent upon: (1) the location of the state and its geographic proximity to coast lines and borders, (2) diesel fuel tax rates in neighboring states, (3) level of truck ownership and usage in states and (4) other truck tax rates within each state. Testimony: Mitstifer, G. May 1992. Shortfalls in Highway Trust Fund Collections: Hearing Before the Subcommittee on Investigations and Oversight. p. 238. Washington, D.C. In this testimony, diesel tax evasion estimates constructed by the National Association of Truck Stop Operators (NATSO) were presented to the United States Congress. The NATSO evasion finding was generated by comparing estimates of diesel fuel sold by truck stop operators across the nation to gallons taxed by the IRS. The comparison yielded a $3 billion estimate of federal and state diesel tax evasion. Report: New Jersey Commission of Investigation. 1990. Motor Fuel Tax Evasion: Hearing Before the State of New Jersey Commission of Investigation. Trenton, New Jersey. This report provides an overview of testimony provided by industry leaders and federal, state and local enforcement agency representatives. Testimony summarized in this report notes the 167

problem and financial implications of diesel fuel tax evasion in New Jersey, the methods used to evade taxes, weaknesses in statutes for combating evasion, possible solutions to address the problem of diesel tax evasion and recommendations for statutory change to combat evasion and reduce tax losses. Testimony provided by a representative of the New Jersey Division of Taxation indicates that the daisy chain method and the mislabeling of home heating oil are the most common forms of diesel tax evasion. The Division representative also noted that total fuel tax evasion in New Jersey represents an estimated $40 million annual loss in revenue to the state. Report: Oum, T., Waters, W., and Young, J. January 1990. A Survey of Recent Estimates of Price Elasticities of Demand for Transport. Policy, Planning and Research Working Paper, Infrastructure and Urban Development Department, The World Bank. Washington, D.C. This paper presents an in-depth literature review inclusive of 70 recent journal articles focusing on price elasticities of demand for transport. Data are presented for both passenger and freight transport, and are used to construct estimates of both own-price and mode choice elasticities. The economic principles underlying elasticity estimates are reviewed in detail, including the analysis of compensated, uncompensated, price, cross-price and mode choice elasticities. The paper addresses the relationship between each concept of elasticity. Data obtained from the literature review are used to present an estimated range for the elasticity of demand for transport, and are used to present a most likely estimate. The paper explores and attempts to qualitatively account for the variability of elasticity estimates between studies. The findings of the literature review suggest that transportation represents a derived form of demand. Further, the elasticity of demand for transport is inelastic (as prices rise demand is reduced by less than a proportional amount in percentage terms). Exceptions are evident for discretionary travel and certain forms of freight shipments due to intermodal competition. Testimony: Peters, Mary E. July 12, 2002. Statement of Mary E. Peters, Administrator Federal Highway Administration. Before the Committee on Finance United States Senate, Hearing on Schemes, Scams and Cons: Fuel Tax Fraud. Washington, D.C. This testimony contains a summary of the fuel tax evasion problem including descriptions of specific evasion schemes. It also describes the nature and impact of major federal legislation and efforts to curtail evasion. Report: Pickrell, D. and Schimek, P. May 1999. “Trends in Motor Vehicle Ownership and Use: Evidence from the Nationwide Personal Transportation Survey.” Journal of Transportation and Statistics. Washington, D.C. The Nationwide Personal Transportation Survey (NPTS) consists of a periodic survey of household-level data on demographics, motor vehicle ownership and vehicle travel. The NPTS was conducted in 1969, 1977, 1983, 1990 and 1995. The survey is produced by the US DOT’s Volpe Center in Cambridge, Massachusetts. Surveys were conducted using random digit dialing. 168

The 1995 NPTS was based on three types of survey methods: vehicle-based estimates, driver- based estimates and trip-based estimates. For each respondent, VMT and trip information is matched with demographic and ownership data to analyze trends in vehicle ownership and usage for respondent groups segregated by age, gender and socioeconomic factors. NPTS respondent data suggest the following: single occupant vehicle (SOV) trips grew from 60 percent of total trips in 1977 to 68 percent in 1995; the household vehicle fleet has continued to age, and that ownership of sport utility vehicle vehicles and vans grew dramatically between 1990 and 1995; and the aging of the fleet may have been caused by the recession of the early 1990’s. User Guide: RAILINC. August 2003. User Guide for the 2002 Surface Transportation Board Carload Waybill Sample. Cary, NC. This is a user guide for working with and interpreting carload Waybill data. The Surface Transportation Board conducts annual surveys of rail shipments for the U.S. This guide includes a summary of Waybill processing and record layouts. The data itself contains information such as date, individual waybill number, commodity code, billed weight, transit charges, origin and destination and reporting railroad. Report: Raven, Ronald. 1999. Deliver Us From Evil: Governmental Responses to Reports of Fuel Tax Evasion. Washington, D.C. This report reviews fuel tax evasion literature, including analysis of public testimony and government-sponsored studies. The paper presents alternative administrative models for fuel tax systems and documents how model adjustments (e.g., moving the point of collection up the distribution chain) have historically affected tax collections. The report documents relevant court cases, legislation and administrative program adjustments for the period 1981 to 1999. Report: Reno, A. and Stowers, J. March 1995. Alternatives to the Motor Fuel Tax for Financing Surface Transportation Improvements. NCHRP Report 377, TRB. National Cooperative Highway Research Program. Washington, D.C. This report contains recommendations for evaluating alternatives to the motor fuel tax. It provides an overview of numerous fuel tax forecasts and includes an analysis of the variables that affect tax collections (e.g., VMT and fuel economy). The report also notes that fuel efficiency changes can alter the per-mile revenue generated by fixed per-gallon motor fuel taxes. The report presents forecasts relating to on-road fuel economy, vehicle utilization (VMT/vehicle), and VMT and fuel consumption. Forecasts cited within this study include those performed by the Argonne National Laboratory, Data Resources, Inc., Gas Research Institute, Energy and Environmental Analysis, ORNL and the EIA. Report: Reno, Arlee T. October 1990. Measures to Curtail State Fuel Tax Evasion. Prepared for NCHRP, National Cooperative Highway Research Program. Washington, D.C. 169

This report synthesizes fuel tax evasion issues, discussing tax dodging methods and various fuel tax compliance methods and efforts. Fuel tax evasion schemes are discussed in four major categories: failures to file, the filing of incorrect information, filing false exemptions and failures to pay assessed taxes. Methods and practices of curtailing evasion are changes in point of taxation, screening, licensing, permitting, bonding, better information and reporting, uniform motor carrier fuel tax reporting, fuel purchase invoice requirements, audit efforts, cooperation among state agencies, criminal investigations and diesel fuel dying. These methods are both explained in detail and are described in terms of state experience in implementing them. Report: Revenue Canada. September 1996. 1996 Report of the Auditor General of Canada: Chapter 18, Excise Duties and Taxes on Selected Commodities. Ottawa, Canada. This report documents the findings of the Canadian Auditor General’s audit of Canada’s excise tax programs. The stated objective of the audit was to determine whether appropriate and sufficient controls, systems and practices had been used to maximize excise tax collections. This report analyzes three areas of risk to revenue collections: evasion realized through smuggling, illegal fuel production and failure by licensed producers to fully pay excise tax liability. The audit assessed the adequacy of Canada’s approach to finding and eradicating tax evasion. The report also highlights the revenue impact of auditing efforts. The report noted that the Canadian Department of Finance has compared estimates derived from Statistics Canada on motor fuel sales volumes with gallons taxed by Revenue Canada. Based on this analysis, revenue loss is estimated at $55 million, or 1.5 percent of the $3.8 billion in motor fuel taxes collected in Canada in 1994-95. The audit assigned relatively low confidence to the Department of Finance estimate because the data used by Statistics Canada were obtained directly from fuel producers. Because the data provided by fuel producers are likely the basis for assessing tax liability, the comparison of Statistics Canada and Revenue Canada data is thought to be more a reflection of the quality of record keeping on the part of Revenue Canada rather than an indication of the amount of evasion. Report: Runde, Al. 2003. Motor Vehicle Fuel and Alternative Fuel Tax. Prepared for Wisconsin Legislative Fiscal Bureau. Informational Paper 40. Madison, Wisconsin. This report outlines the history and current characteristics of the motor fuel tax in Wisconsin. This history includes a discussion of legislatively increased tax rates on various fuels and how tax rates have been indexed to account for inflation. Other characteristics of Wisconsin’s motor fuel tax administration system described are fuel tax exemptions, Wisconsin’s participation in IFTA, fuel tax refunds and a floor tax imposed on holders of fuel when fuel tax rates are altered. Report: Sinha, K. and Varma, A. 1997. “Long-Term Highway Revenue Forecasting for Indiana.” Transportation Research Record 1576. Washington, D.C. This report documents a revenue-forecasting model developed for estimating transportation tax revenue in Indiana. The report documents the methodology underlying the model and notes that the model was designed with an emphasis on the accuracy and availability of data needed to input into the model, simplicity of model inputs and ease of use and responsiveness to changing 170

energy, environmental, financial, legislative, socioeconomic and technology factors. The paper presents a summary of the highway revenue analysis, which includes a long-range forecast of all transportation revenue sources in Indiana. Report: U.S. Army Corps of Engineers. June 2002. Waterborne Commerce of the United States, Calendar Year 2002, Part. 5, National Summaries. New Orleans, LA. Waterborne Commerce of the United States is published annually providing statistics on both foreign and domestic waterborne commerce traversing U.S. waterways. It provides data relating to the transport of commodities at the ports and harbors located along the waterways and canals of the U.S. and its territories. Waterborne transport of petroleum products – e.g., crude petroleum, gasoline, kersosene, distillate fuel oil, residual fuel oil and liquid natural gas are examined in terms of foreign/domestic short tons transported, foreign/domestic short tons-miles and domestic barge transport in short tons. Report: U.S. Census Bureau. October 1999. US Vehicle Inventory and Use Survey - VIUS – 1997. Washington D.C. Other 1992, 1987, 1982, are found under Truck Inventory and Use Survey. The Vehicle Inventory and Use Survey (VIUS) compiles data from a sample survey of 150,000 commercial and private trucks registered in the United States. VIUS was known as the Truck Inventory and Use Survey (TIUS) prior to 1997. VIUS data were first collected in 1963. Since 1967, VIUS data have been collected once every five years, with the most recent survey being conducted in 1997. VIUS data do include state-level VMT and fuel efficiency data for trucks. The VIUS survey does not include passenger vehicles. Report: U.S. Congressional Research Service. March 29, 2000. The Federal Excise Tax on Gasoline and The Highway Trust Fund: A Short History, CRS Issue Brief for Congress, http://www.cnie.org/nle/trans-24.htm (CRS 2000). This report gives a history of federal gasoline excise tax revenues for fiscal years 1933 through 1993. It also provides a brief description of where these revenues have traditionally gone. Testimony: U.S. Congress. January 9, 1995a. Written Testimony of The Staff of The Joint Committee on Taxation Regarding The Revenue Estimating Process, JCX-1-95. This testimony provides a description of the Joint Committee on Taxation (JCT) methodology for estimating revenue effects from changes in tax legislation. Discussions of past revenue estimates, issues of importance when modifying revenue estimation methodology and approaches for improving the revenue estimation process. Report: USDOE, U.S. Department of Energy. February 2001.The Transportation Sector Model of The National Energy Modeling System, Model Documentation Report, DOE/EIA-M070 (2000), Energy Information Administration. Washington, D.C. 171

This report describes the National Energy Modeling System (NEMS) Transportation Model (TRAN), specifically focusing on its development, objectives and approach. The report also discusses model assumptions and other factors related to the structure of the model. TRAN consists of several partially independent models that pertain to various features of the transportation sector. The transportation model provides mid-term forecasts of transportation fuel demand by type, including gasoline, distillate, jet fuel and alternative fuels. The model also provides forecasts of passenger travel demand by automobile, air, or mass transit; estimates of the efficiency with which that demand is met; projections of vehicle stocks and the penetration of new technologies; and estimates of the demand for freight transport which are linked to forecasts of industrial output. Currently, NEMS forecasts extend to the year 2020 and use a base year of 1990. Energy consumption within several types (i.e. aircraft, marine, rail, light duty vehicles and truck freight) of transport are considered separately in NEMS forecasts. This approach is used to assess the impact of public policy and technological advances on particular transportation modes. Report: USDOE, U.S. Department of Energy. 1997. 1994 Residential Transportation Energy Consumption Survey. Energy Information Administration, DOE/EIA-0464(94), Washington, D.C. Other years are also available. The RTECS is based on a national multistage probability sample survey conducted on personal household vehicles. The first RTECS was conducted in 1983. Subsequent surveys were conducted in 1985, 1988, 1991 and 1994. No surveys have been conducted since 1994, and the US DOE Energy Information Administration (EIA) is presently investigating new methods to construct data formerly generated through the RTECS. Data on household, vehicle stock and fuel consumption is obtained through an initial personal interview. A subsequent telephone interview is conducted the following years, and is used to collect data on vehicle stock, turnover, new purchases and VMT. A third interview is conducted the following year. The RTECS is designed to construct VMT estimates for personal vehicles based on odometer readings. Vehicle characteristic information is obtained directly from the respondent. Prior to 1994, the RTECS based fuel consumption estimates on fuel consumption logs provided by the respondents. Due to budgetary constraints, however, RTECS fuel consumption estimates in 1994 were based on vehicle fuel efficiency data collected from the United State Environmental Protection Agency, adjusted for on-road degradation. Report: United States Department of Transportation. 1995. Revenue Impact of Diesel Fuel Dyeing. Washington, D.C. This report documents the estimated fiscal impact of dyeing diesel fuel and moving the point of diesel fuel tax collection to the highest point in the distribution chain. The Omnibus Budget Reconciliation Act (OBRA) of 1993 moved the point of collection from the distributor to the terminal level (effective January 1, 1994) and required the dyeing of all tax-exempt diesel prior to removal from the terminal. Treasury estimated that tax receipts available to the federal HTF grew by more than $1.2 billion in fiscal year (FY) 1994 after adjusting for the tax increase enacted on October 1, 1993. After adjusting for increased refunds stemming from moving the 172

point of collection to a point higher in the fuel distribution chain and growth resulting from increased travel and motor fuel consumption, the Treasury Department further estimated that $600-$700 million of the total growth in diesel fuel tax collections experienced in FY 1994 could be attributed to reduced tax evasion resulting directly from the enforcement provisions of OBRA 1993. This report also documents a spill-over compliance benefit realized by states as a result of federal diesel dyeing requirements and enforcement activities. In 1994, states reported an average increase in diesel tax collections of 6.9 percent, an amount equal to almost twice the expected growth in diesel fuel use. The report concludes that if half the growth experienced by states were attributable to increased compliance; state transportation programs would yield an additional $150 million due to reduced evasion. The report also notes that each dollar spent on auditing and enforcement activity yields an average of $10 to $20 in additional highway user revenue. Report: Varma, Amy and Kumares, Sinha. 1997. Long-Term Highway Revenue Forecasting for Indiana. Financial, Economic, and Social Topics in Transportation, TRR No. 1576, TRB. This paper presents a summary of the highway revenue analysis and highway revenue forecasting methodology developed for the state of Indiana. Indiana’s long-term highway forecasting model was developed with an emphasis on ease of data input and responsiveness to changes in various variables, including socioeconomic and technological factors. Concentration was placed on forecasting vehicle travel and vehicle registration. This paper includes an extensive list of references regarding federal and state forecasting models. Report: Virginia Department of Motor Vehicles. November 2000. Commonwealth Transportation Fund: Preliminary Revenue Estimates. Richmond, Virginia. This report presents revenue estimates for transportation revenues during the 2001 to 2006 timeframe, and documents the econometric model designed by the Wharton Economic Associates (WEFA) for the Forecasting and Analysis Office (FAO) of the Virginia Department of Motor Vehicles (DMV). The econometric model is maintained by the FAO. The report notes that gasoline and diesel revenues are modeled differently to reflect the markets in which these fuels are sold. Specifically, diesel consumption is more closely tied to economic growth, while gasoline consumption is more tied to personal income and fuel prices. In the Virginia Revenue Forecasting Model, gasoline is estimated as a function of real personal income in Virginia, a three-period moving average of personal expenditures on gasoline, expenditures on oil products relative to total consumer expenditures and three quarterly dummy variables. Taxable diesel gallons are estimated as a function of chain-weighted gross domestic product (GDP) and a three- period moving average of the refined petroleum producer price index. Report: Washington Interagency Revenue Task Force. February 2000. Transportation Economic and Revenue Forecasts. Olympia, Washington. This report provides an overview of a six-year forecast of transportation revenues in Washington State. Further, the report documents the regression equations used to forecast motor fuel excise 173

tax revenues in Washington State. Taxable fuel gallonage is estimated with an econometric model comprised of two equations. There are separate equations for forecasting gasoline and diesel taxes. Gasoline tax revenue is estimated as a function of Washington state real personal income, population, inflation, gasoline prices and average passenger car fuel efficiency. Diesel tax revenue is estimated as a function of Washington State real personal income and production activity in the lumber and wood products industry. The report notes that during the last 20 years, gasoline tax forecasts have been quite accurate, generally within 2 percent of actual collections. Diesel tax revenue forecasts have not been as accurate with the divergence between actual collections and forecast amounts reaching as high as 8 percent. Because diesel tax revenue represents only 15 percent of total fuel tax revenue in Washington State, the overall forecast has historically been within 3 percent of actual collections on an annual basis. Report: Washington Legislative Transportation Committee. 1996. Fuel Tax Evasion in Washington State. Olympia, Washington. This report documents the findings of a legislative task force established in 1995 to study the issue of fuel tax evasion in Washington State. The task force included representatives from the Washington State Patrol, the Department of Licensing, the Department of Revenue, the Washington State Department of Transportation, the Internal Revenue Service and representatives from private businesses. The objective of the task force was to estimate the magnitude of evasion and the methods used to evade fuel taxes. Further, the task force identified measures to counteract the methods of evasion and reduce tax losses to the State of Washington. The task force identified nine fuel tax evasion methods: cross-border smuggling, daisy chains, cocktailing/blending, fraudulent use of exempt licenses, fraudulent tax returns, fraudulent refund claims, failure to file, using dyed diesel in highway vehicles, and fraudulent licenses. The task force used three methods to identify and quantify evasion: literature review, a three-day border interdiction effort and random audits. Based on their findings, the task force estimated that Washington could recover between $15 and $30 million annually through enhanced enforcement of motor fuel taxes. Methods identified to reduce tax evasion included: change the point of taxation from the distributor to the rack, implement a dyed diesel fuel program, implement and encourage information sharing between enforcement agencies, develop a national evader database, require total fee accountability from origin to end user, encourage electronic reporting, foster working relationships with British Columbia and neighbor states, create an evasion investigation unit, institute a distributor bonding requirement, and educate public and prosecutors. Report: Weimar, M. R. et al. August 2002. Economic Indicators of Federal Motor Fuel Excise Tax Collections. Prepared by Pacific Northwest National Laboratory and Oak Ridge National Laboratory for the U.S. Internal Revenue Service. This report documents the development of a structural and statistical model that predicts gasoline excise taxes, discerns trends in tax collection and detects possible historical under-collection of motor fuel taxes based on macroeconomic variables. This model represents the first installment 174

of a larger model that would predict excise tax revenues for all motor fuels. Three stage least squares is incorporated to estimate fuel consumption for various vehicle classes. Further, supply and demand equations for VMT for various vehicle classes are estimated using economic variables, including non-farm employment and the price of gasoline. Fuel efficiency is also integrated into this model based on Corporate Average Fuel Economy (CAFE) and lagged endogenous variables. Quarterly data from 1981 forward obtained from federal, state and private sources were used to estimate this model. The structural approach of this model differs from other revenue estimation models in that most other models estimate revenue directly from VMT and this model first estimates consumption and then estimates revenue through tax rates and accounting procedures. Further, this model is designed to detect changes in levels of evasion as well as project fuel tax revenue. While it does not directly detect evasion, it can detect unexpected movements in collections which can be inferred as unexpected changes in the level of evasion. The model predicted higher collections than what was actually collected by the IRS prior to 1988. After the point in taxation was moved to the rack in 1993, the model results are on par with IRS collections. This is evidence that there was a systematic problem of under-collections prior to the change in the point of taxation. Report: Weinblatt, Herbert et al. 1998. Alternative Approaches to the Taxation of Heavy Vehicles. Prepared for NCHRP, National Cooperative Highway Research Program. Washington D.C. Transportation Research Board. This study reviews alternative state tax systems, specifically for heavy vehicles, and develops six criteria by which to compare taxation alternatives used to finance surface transportation. To assess existing state highway tax systems, previous studies on the equity of state and federal systems were reviewed and a survey was conducted of state agencies responsible for tax administration. Equity ratios that show the extent to which each vehicle class pays their share of the cost responsibility are presented. Further, tax systems in ten foreign countries were reviewed and major characteristics of those systems were presented to identify taxes and administration procedures that might be of interest for the US. Technologies that have potential for decreasing administrative and compliance costs were also reviewed. The criteria developed by this study to evaluate alternative tax systems are adequacy, administrative efficiency, equity, economic efficiency, evasion and avoidance and feasibility. Given these criteria, the authors found that there was no unambiguously superior taxation system. Rather, the choice between systems involves tradeoffs between the criteria. For instance, one important trade-off mentioned was between administrative efficiency and evasion. Enforcement can decrease evasion but at a cost to both the public and private sectors. Report: Wisconsin Department of Transportation. July 2001. Wisconsin Department of Transportation Revenue Forecasting Model Documentation. Madison, Wisconsin. This report documents the revenue forecasting model used by the Wisconsin Department of Transportation (WDOT) to estimate future vehicle registration and motor fuel tax revenues. The 175

revenue forecasting model is based on regression analysis, which relies on past behavior to predict future behavior. The gasoline consumption model is based on the assumption that fleet composition, real income and real fuel prices affect the demand for travel and fuel consumption, where: auto registrants = autos registered - autos scrapped + new auto registrations; and VMT = f (real disposable income, real gas price adjusted for fuel efficiency, dummy variables for abnormal years such as oil embargo years). 176