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.

59 4.1 Introduction Quality and availability of data are important factors when considering competing methodologies to be used for an evalu- ation of motor fuel excise tax evasion. Ultimately, the under- lying data set forms the foundation upon which an assessment is performed. Any methodology chosen to estimate motor fuel tax evasion at the state level will need to be supported by con- sistent and reliable data. By understanding the coverage, limi- tations, and collection methodology of various data sources, an appropriate methodology can be selected and discrepancies in the data can be realized, thus opening the door to possible cor- rective measures. In turn, reasonable assurance can be had that study conclusions are not erroneously based on data anomalies. Data needed for an analysis of state-level motor fuel tax eva- sion fall into the following broad categories: transportation data; motor fuel data; economic and population data; and tax administration, enforcement, and collections data. Transpor- tation data captures the number of vehicles by class traveling on state highways and roads, documents the number of vehi- cle miles of travel occurring by vehicle type, and examines the impact and nature of this travel as it relates to factors such as fuel consumption and motor fuel economy. Data relating to the fuel distribution system maps the movement of motor fuel within and between states through different points of the sup- ply chain (e.g., ports of entry, refinery, terminal, and retail). Economic data is a necessary component for an econometric- based motor fuel tax evasion analysis and is comprised of data such as population, gross state product, and employment. Tax administration, enforcement, and collections data comprise information such as historical state tax rates, motor fuel tax collection, and enforcement/audit budgets. 4.2 Overview of Data Categories This section presents a brief overview and analysis of several categories of data (transportation data, motor fuel data, tax collections and administration, and economic and population data) relevant to motor fuel excise tax EOE estimation. Fur- ther, it identifies and characterizes data elements falling within each category and, where appropriate, examines and compares various sources of data. 4.2.1 Transportation Data Data tracking the number and travel of vehicles in a state can indicate how much fuel is being consumed within that state. The Vehicle Inventory and Use Survey (VIUS), formerly known as the Truck Inventory and Use Survey (TIUS), pro- vides data on the number of private and commercial trucks operating, as well as the number of miles traveled by these vehicles within each state. This data is accessible through the U.S. Census Bureau. The primary limitation of this data is that it only captures travel by large transport trucks, vans, minivans, pickup trucks, and sport utility vehicles, only cap- tures a fraction of the travel occurring in each state by missing other vehicles types (e.g., passenger cars, motor homes, and motorcycles). Further, the VIUS survey is conducted only once every five years. Vehicle registrations and travel data also are available by means of the administration of IFTA and the International Registration Plan (IRP). IRP is a reciprocity agreement that allows motor carriers to pay state registration fees in a one–stop process based on the percent of total miles their fleet travels in each state. IFTA is a similar reciprocity agreement that enables motor carriers to pay their fuel taxes to all states they travel in while filing in just one state. As is the case with VIUS, these sources of VMT only capture a segment of the motoring pub- lic (i.e., those that are traveled by heavy trucks in this case). A more complete view of VMT data also is presented in FHWA’s Highway Statistics. States, required to submit highway-use data to the FHWA as part of the Highway Per- formance Monitoring System (HPMS), collect VMT on a con- tinual basis for multiple vehicle classes. The annual Federal Highway Statistics publishes VMT data for six vehicle classes: C H A P T E R 4 Data Availability

60 automobiles, motorcycles, light trucks, single-unit heavy trucks, combination trucks, and buses. HPMS VMT data are based on traffic counts performed by states using roadside traffic mon- itoring devices (e.g., pneumatic tubes, inductive loops, and manual counts). Shortcomings in FHWA VMT data include possible deficiency of traffic data on less-used road systems, the occasional malfunctioning of traffic counting devices and the lack of reliable estimates of the margin of error for this data. Even though the margin of error has not been reliably estimated for the nationwide HPMS system, there is evidence that VMT estimates produced by the states are reliable at the national level. Table 4-1 compares 1995 Federal Highway Sta- tistics VMT estimates to those generated through surveys completed for the National Personal Travel Survey (NPTS) (Pickrell and Schimek, 1998). FHWA estimates were within 0.6 percent (odometer readings) to 3.5 percent (personal esti- mates) of those published in the 1995 NPTS. 4.2.2 Motor Fuel Data Motor fuel volumes supplied, exported, imported, distrib- uted, sold, and consumed are published in a number of data sources that could be used to support a motor fuel excise tax evasion analysis (e.g., API, EIA, Bureau of Transportation Sta- tistics, FHWA). These estimates are presented at the state- level, and may prove useful in a state motor fuel excise tax evasion study. Historically, the volumes estimated by these various sources have proven to be in conflict with one another due to discrepancies between data collection methods and errors corrected over time (e.g., double counting, missing some reporters, inaccurate treatment of blending fuels). The U.S. Department of Energy (DOE)’s Petroleum Supply and Reporting System failed to adequately address the consump- tion of ethanol, methanol, methyl tertiary–butyl ether, and other blending stocks prior to 1993. In recent years, however, these errors largely have been addressed though discrepancies persist between data collection methods. Motor fuel consumption data, when used in combination with VMT estimates, can generate motor fuel economy esti- mates. National and state motor fuel consumption is gener- ated by state tax records, however, and, to the extent that evasion is reflected in the gallons not identified by state tax agencies, motor fuel economy estimates would be subject to error. There could be a form of circular referencing from an analytical standpoint if motor fuel economy estimates from motor fuel tax records are used within a model to examine motor fuel excise tax evasion. In addition to the sources of gross estimates of motor fuel gallons supplied, distributed, and consumed outlined above, there also are more detailed fuel tracking systems that have been deployed at the state and federal levels. At the federal level, the IRS is developing ExFIRS. ExFIRS has several component systems, not all focused directly on tracking motor fuel. The two components that do relate to fuel tracking are the ExSTARS and the Excise Classification Information System (ExCIS). Motor fuel tracking systems are also in operation within a number of states. As of 2004, 13 state fuel tracking systems were in operation (Anders-Robb, 2004; FHWA, 2003a; FHWA, 2001; FHWA, 1999a). In general, these federal and state systems provide sales and stocks information for various fuel types. One primary purpose of these systems is to collect and organize fuel tax collections data and data relating to the posi- tion of fuel volumes throughout the supply chain, allowing these two types of data to be compared. Another goal of these systems is to compare records of shipments to ensure account- ability and identify discrepancies between companies engag- ing in motor fuel transactions. By comparing tax collections and reported fuel volumes from various entities in the fuel dis- tribution process, a determination can be made as to whether all responsible for remitting fuel taxes are actually paying them. Further, comparing the sale and distribution of motor fuel reduces the disappearance of volumes from the distribu- tion system. Figure 4-1 depicts a system of reporting that pro- vides information on fuel position that can be input into a fuel tracking system. Limitations of ExSTARS and ExCIS are that these systems are still in the process of development, so current data may not be complete or reflect high-quality. Data from ExSTARS and ExCIS also may contain only information for the limited time that they have been in operation, thus limiting their use for time trend analysis. Further, ExSTARS identifies motor fuel volumes as they leave the terminal rack but does not capture downstream sales and transport of motor fuels. Finally, a frac- tion of the motor fuel transactions is still submitted in paper form and, for these transactions, only summary information is entered into the system. A shortcoming of data provided by state fuel tracking systems is that data may not be comprehensive. Figure 4-1 Source Universe Type of Data Trillion VMT 1995 NPTS Personal vehicles Reported by respondent 2.149 1995 NPTS Personal vehicles Odometer reading 2.215 1995 NPTS Travel period & day Trip diary 2.181 1995 Highway Statistics Light duty vehicles State traffic counts 2.228 Source: Pickrell and Schimek, 1998. Table 4-1. VMT estimates, 1995 NPTS and 1995 highway statistics.

61 presents a very thorough system of reporting to account for the movement of fuel at every step in the process. States may not have this level of fuel accountability and may only require reports from a few entities along the distribution process. Further, states will vary on which fuels they track and how they track them. For instance, some states may track kerosene while others do not. Also, while states are moving toward implementing the uniform report schedules recommended by the FTA uniformity project, many states may still require different sets of information from the fuel industry. Finally, only a handful of states have implemented fuel tracking systems, thus limiting the use of this type of data to certain states. 4.2.3 Tax Administration, Enforcement, and Collections Historically federal motor fuel excise tax collections were obtained from the IRS’ SOI reports. Within the SOI reports, motor fuel excise tax collection data are shown for 22 motor fuel types: gasoline, gasoline floor stock, diesel, diesel floor stock, kerosene, aviation gasoline, noncommercial aviation fuel other than gasoline, noncommercial aviation fuel other than gasoline floor stock, commercial aviation fuel, gasoline for gasohol 5.7 percent, gasoline for gasohol 7.7 percent, gaso- line for gasohol 10.0 percent, gasoline for gasohol floor stock, gasohol 5.7 percent, gasohol 7.7 percent, gasohol 10 percent, gasohol floor stock, special fuels, special fuel floor stock, dyed diesel used for certain intercity buses, dyed diesel used for trains, and dyed diesel used for trains floor stock. State tax collections data are held at the taxpayer level; however, the data made available for examining motor fuel excise tax evasion will represent aggregated collection totals by motor fuel type. Taxpayer-level information is treated as confidential. Motor-fuel volumes taxed are generally reported at a summary level, with detailed schedules identi- fying the date, volume, and supplier or customer for detailed transactions. Data are reported monthly or quarterly based on state tax code. The motor fuel excise tax collections data are typically housed within state departments of transporta- tion or revenue. State audit and enforcement data are of principal impor- tance when performing EOE estimation. On-road dyed fuel inspection data often are extensive at the state level, includ- ing thousands of records obtained annually. These records will typically include only summary information for stops that do not result in a violation (e.g., vehicle type, inspection result). When a violation is discovered, additional vehicle and company characteristics are identified. Audit data are generally available through at least four forms of audit: refund audits, IFTA audits, desk audits of tax- payers, and field audits of taxpayers. In some cases, data also may be available from other forms of audit, including retail audits. When conducting analysis of EOE, it is important to Figure 4-1. Reporting throughout the fuel distribution system. Source: FTA, 2004b.

obtain the following general types of information from audit records: • Business characteristics (e.g., years in operation, annual revenue, number of employees), • Operational characteristics (e.g., type of operation audited, states in which the company is licensed to operated), • Type of audit (field or office audit), • Date audit is performed, • Trigger for the audit (e.g., third party tip, random sampling, flagged return), • Auditor information (e.g., years in service, detection rate, rank), and • Assessment by type of violation found. Chapter 5 contains more detailed information about audit and enforcement data that could be used to estimate motor fuel tax EOE. Certain aspects of a state’s motor fuel tax administration processes may be important to address in an evasion analysis. These characteristics include state fuel tax rates, motor fuel tax collections, enforcement/audit budgets, auditing assessments, and collections generated through audits. For instance, one would expect that, if tax rates increase, a state will have more evasion because the incentive to cheat is higher (all other vari- ables remaining the same). Likewise, if an auditing budget increases, a state may experience less evasion because more evasion activities would be uncovered and possible evasion would be deterred (all else being equal). There has been no sin- gle source identified for state motor fuel excise tax collections, revenue, and auditing budgets. In most cases, this information should be available by request of motor fuel tax sections of state revenue or transportation departments. 4.2.4 Economic and Population Data Macroeconomic and population data may be used to sup- port an econometric examination of motor fuel excise tax evasion. In the econometric approach, mathematical models are set up to describe economic and structural relationships (such as the relationship between fuel tax evasion and fuel tax rates) and statistical techniques are applied to test hypotheses about those relationships and to measure the relative strength of the influences of certain variables on a dependent vari- able. For example, motor fuel prices and economic activity correlate with vehicle miles of travel and motor fuel consump- tion, with prices negatively correlating and economic activ- ity and income positively correlating. To the extent that an economy is expanding but diesel consumption, as measured by state tax records, is stagnant or in decline, evasion may be present. Economic variables must be controlled for when using the econometric method as they will be the primary driving force of trends in state fuel consumption and tax collections. Eco- nomic and population variables that would suit this purpose are gross state product, population, income, motor fuel prices, and employment. Gross state product for every U.S. state is reported annu- ally by the Bureau of Economic Analysis (BEA) as part of their annual Survey of Current Business. State population data are collected and reported by the U.S. Census Bureau as part of the Census of population and housing survey con- ducted every 10 years, with annual estimates generated based on updated data. State employment data are collected and reported monthly by the Bureau of Labor Statistic’s (BLS) Current Employment Statistics survey. This survey is the largest monthly survey of businesses within the United States. 62