< Previous | Contents | Next >

Attribution Paper

ATTRIBUTION PAPER

Ralph Erickson, FHWA

CAVEAT: One of the primary purposes of this paper is to assist the FHWA staff who manage the attribution process to understand all the issues under the topic. Several of the issues addressed are outside the box of traditional attribution practices. An understanding of what those issues are, and why they have been handled as they have, is essential knowledge for those who deal with attribution in great depth. This document does not stand as a description of the current attribution process, and is not advocating any changes to the current procedures.

Introduction

Since FY 1984, motor fuel data has been used in the apportionment of Federal-aid highway funds. State-reported highway use of motor fuel is used to attribute Highway Trust Fund tax payments to highway users and commercial vehicle contributions to highway users in each state.

Think of a pie as a simple analogy to illustrate the attribution procedure. Highway Trust Fund tax payments, as reported by the Department of Treasury, determine the size of the pie of each motor fuel type (gasoline, gasohol, special fuels -- meaning diesel and special fuels), and the gallons on-highway use of each motor fuel type consumed in a state determine that state's slice of the appropriate pie. As special cases, the Federal tire tax, truck tractor and trailer retail excise tax, and the heavy vehicle use tax revenues are also reported by Treasury, which determines the size of those pies, and each state's slice of those pies is based on gallons of special fuels.

Internal Revenue Service (IRS) data provides the tax payment amounts, but IRS does not track where or how the fuel is consumed. The Federal tax collection is too high up the distribution chain to identify ultimate place or type of use. Therefore, the attribution process uses state reported motor fuel tax revenue data to determine highway use of motor fuels. FHWA also uses statistical methods to estimate missing or not-reported state data.

Under the Transportation Equity Act for the 21st Century (Public Law 105-178) (TEA-21), motor fuel data are used in the apportionment of Federal Surface Transportation Program (STP) funds, National Highway System (NHS) funds, Interstate Maintenance (IM) funds, and the Minimum Guarantee. The following shows the use of these factors in FY 1999.

• Highway Trust Fund payments to the Highway Account are used as a 35 percent factor for distributing about $4.9 billion in FY 1999 STP funds.
• Diesel fuel used on highways is used as a 30 percent factor for distributing about $4.6 billion in FY 1999 NHS funds. Unlike the other bullets listed here, this factor uses gallons, not dollars, as the measurement unit.
• Commercial vehicle contributions to the Highway Account are used as a 33.3 percent factor for distributing about $3.6 billion in FY 1999 IM funds.
• The Minimum Guarantee, under which each state is guaranteed that its share of apportionments and highway priority projects will be at least 90.5 percent of its share of attributed contributions to the Highway Account of the Highway Trust Fund, and is estimated to be about $5.7 billion in FY 1999.

Since the Surface Transportation Assistance Act of 1982 (STAA), Congress has expressed desires to manage the cross-subsidy between states (the donor-donee problem). With certain states (called donee states) receiving more in highway Federal-aid than they contributed in payments to the Highway Trust Fund (HTF), pressure from the states who contribute more than they receive (donor states) has culminated in congressional mandates to limit the Federal cross subsidy. Some of the programs to implement this policy include:

• Minimum Allocation (STAA of 1982);
• 90 Percent of Payments (ISTEA);
• Donor State Bonus (ISTEA); and
• Minimum Guarantee (TEA-21).

Objective

Recognizing the increasing importance of accurate, timely reporting of motor fuel and related attribution data in determining state-funding shares, FHWA is reviewing the motor fuel data reporting system used to collect this information. As part of the review process, FHWA is evaluating the attribution process to determine the continued quality of its attribution methodology and to identify areas where improvements can be made.

While FHWA believes it is identifying and appropriately dealing with the vast majority of attribution issues, the primary purpose of this paper is to identify those issues that need to be better addressed, and possibly streamline the attribution process by eliminating issues that have been addressed, but do not warrant the continued effort involved in collecting and analyzing the data for attribution purposes.

Background

It is a surprise to many that the state-by-state contributions to the HTF are not available from the IRS. The Federal fuel taxes, which make up more than 80 percent of the HTF's receipts, are imposed when the fuel is first removed from bulk storage and the tax is paid by the seller. Thus, the typical Federal fuel taxpayer is an oil company. While paid initially by a company, the costs of these taxes become part of the purchase price of the products and are ultimately paid by the highway user. The heavy vehicle-use tax is the only Federal highway-user tax paid directly to the IRS by the vehicle owner, and even then, the return captures only the business address of the owner, not the state or states where the vehicle is operated. Using oil company tax data would only cause problems in attribution because the state in which the motor fuel tax is paid does not reflect where it will be shipped, stored, or - of particular interest in the attribution process - used.

As Federal tax records do not yield the desired information, FHWA estimates the HTF contributions from highway users by looking at tax revenue data in each state. The method for attributing Trust Fund receipts to each state has changed over time, with the last changes made in 1985. A data series, collected from the states by Federal government, on motor fuel use on and off the highway goes all the way back to 1919. FHWA first started making the state HTF payment estimates based on motor fuel data in the early 1970's, in response to general interest in donor-donee issues. With the passage of the Surface Transportation Assistance Act of 1982, the attribution of HTF receipts became a factor in calculating the 85-percent minimum allocation. FHWA solicited comments from the states on its attribution methodology, and modified the methodology to reflect the concerns of the states and the Congress that the attribution employ use-based factors. The resulting methodology, first used to attribute fiscal year 1984 Trust Fund receipts, was published in the Federal Register on June 21, 1985 . These procedures have been in use since that time.

The Department of the Treasury reports the tax receipts deposited in the HTF for each tax type. The net receipts, after refunds and transfers, are the contributions to the HTF that are attributed to the highway users in each state. The basic premise is that the Federal motor fuel tax receipts (by fuel type) to the Highway Account are received from each state in proportion to the highway use of that fuel type in each state compared to the total use in all states.

Commercial vehicle contributions -- the diesel fuel tax, truck tire tax, truck and trailer retail sales tax, and the heavy vehicle use tax -- are attributed to the states using highway use of diesel and special fuels (special fuels are very small amounts). This method is considered to be the best available proxy for truck use in each state, and was formalized in the Federal Register Notice of June 21, 1985 .

The attribution analysis is basically a process of determining a state's on-highway motor fuel use from state tax data by looking at any given set (in the mathematical sense of set) of motor fuel data, and determining if it is an on-highway use. Typically, state revenue departments have data on motor fuels that are exempted, refunded, or taxed at other rates. From this kind of data, FHWA may be able to identify the use of the fuel and therefore its place in attribution. In many cases, this data does not exist at the state level and FHWA estimates usage from other sources. When FHWA estimates and state's data are both available, a choice must be made about the relative quality of the data, and one or the other is selected.

Historic Attribution Approaches

Historically, there have been two basic methodologies or approaches which have been considered as attribution options. One approach is based on gallons of motor fuel, and the other is based on vehicles miles traveled (VMT). Truck registration data was once used to attribute truck taxes, but was found to inaccurately reflect truck on-highway usage. The June 21, 1985, FHWA Federal Register Notice last examined attribution policy, and determined that motor fuel should be the attribution measure. In the time context of the 1985 Notice, the emphasis was on attributing the newly imposed and increased truck taxes, and therefore the VMT method received significant consideration. While FHWA continues to pursue improving both motor fuel and VMT reporting, the VMT methodology does not provide the accuracy and veracity that the motor fuel methodology does. It is worthwhile, however, to briefly review both methods.

Advantages and disadvantages of these approaches for attribution purposes

• Accuracy of data:
  • VMT is an estimate and truck VMT is especially subject to uncertainty
  • State motor fuel data are capable of being double-checked thru audits
• Ease/cost of collection
  • Estimated data are cheaper to produce, especially if little data collection is used in the estimation process.
  • States are already collecting motor fuel tax data for their own purposes (though not always in the form FHWA requires).
  • Diesel fuel used in interstate commerce is taxed under International Fuel Tax Agreement (IFTA), which was a procedure not in place in 1985 when attribution policy was last addressed. IFTA is discussed further below.

Interstate motor-carrier fuel use is treated differently from other fuel categories. The intent is to tax major interstate fuel users (typically motor carriers), on the basis of the quantity of fuel used within the state rather than on the basis of fuel purchased in the state. While most of the motor fuel is diesel, gasoline and special fuels are reported. Almost all states and Canadian provinces now use IFTA provisions for taxing motor fuel used by interstate motor carriers in place of Interstate Motor Carrier (IMC) taxation. At least on the interstate level (leaving out intra-state diesel fuel usage, which is captured under normal state reporting,) state use of the IFTA taxation procedure addresses the place of use instead of place of purchase. A problem with IFTA data however, is the reporting lag. IFTA processing takes place on a quarterly basis and a state does not know its IFTA net tax revenue until five to six months after tax liability occurs. Current FHWA guidance asked for state data within sixty days of the end of the reporting month.

FHWA concludes that highway use based on motor fuel data and identified by state tax data or estimated by FHWA continues to be the better approach to attribution. This conclusion is based on three primary facts:

• The current reliance on motor fuel data by the Congress (initiated in the policy guidance of 1985, and confirmed in three subsequent reauthorization bills;
• The general agreement of the states and related organizations (such as the American Association of State Highway and Transportation Officials (AASHTO) that the attribution process an acceptable methodology; and
• The factual basis for motor fuel accounting (versus estimation of VMT).

A Closer Look at Current Legislation

The existing attribution process is not in precise agreement with the language of the TEA-21 legislation. On close inspection, the following discrepancies are revealed:

TEA-21 directs that the Surface Transportation Program (STP) be apportioned with "35 percent of the apportionment in the ratio that the estimated tax payments attributable to highway users in all states paid into the Highway Trust Fund (other than the Mass Transit Account) in the fiscal year for which data are available." Note that it is based on the state's share of tax payments by highway users, which should exclude Federally tax exempt public vehicles using either gas or diesel motor fuels, as these vehicles do not pay a Federal highway tax. Under current motor fuel reporting procedures, gallons of diesel fuel powering public vehicles should not be included, but in reality many states cannot report this use separately. In contrast, gallons of gasoline used to power public vehicle is also not Federally taxed, but it is either reported by the state or is estimated by FHWA models and is subtracted from total gasoline volume reported, and therefore is treated appropriately in attribution methodology.

TEA-21 directs that 33.3 percent of the Interstate Maintenance Program be apportioned "in the ratio that the of each state's annual contributions to the Highway Trust Fund (other than the Mass Transit Account) attributable to commercial vehicles." Note that commercial vehicles do not include public vehicles. Therefore, in Interstate Maintenance apportionments, gallons of motor fuel used in public vehicles should be excluded from attribution. As above, gallons of diesel fuel may not be reported properly, but gallons of gasoline are either reported or estimated.

TEA-21 directs that the National Highway System component should be apportioned as follows: "30 percent in the ratio that total diesel fuel used on highways in each state bears to the total diesel fuel used on highways in all states." Note that this language does include Federal, state, county and municipal diesel, and that it does not include LPG. Current motor fuel reporting procedures do not ask for public use of diesel (but it is included in some state's data), and attribution procedures therefore do not currently distribute public diesel correctly. Attribution procedures also currently include LPG fuels. Furthermore, notice that the apportionment depends on each state's share of gallons of diesel fuel directly, not Federal tax dollars as apportioned by each state's share of diesel motor fuel. This last point is taken into account in the current attribution process.

Internal FHWA discussion only: how important is it that FHWA follow the letter of the law in this situation, or how important is it that FHWA not change attribution procedures to maintain a continuous data series, or to avoid criticism from states hurt by the changes. Is there any reason that the information provided is more accurate under existing procedures than by procedures that could be developed to meet Congressional language.

Discussion of the Ideal Attribution

Ideally, the FHWA should be asking the following questions in connection with determining attribution of motor fuel:

• Is the unit of fuel in question consumed in the process of propelling a vehicle on a public highway system in the United States ?
• What type of fuel was taxed (based on Federal definitions of the fuel in question).
• If no Federal tax was paid on the unit of fuel used to propel a vehicle on a highway, for what purpose was the motor fuel consumed?
• In what geographical location (state) was the highway on which the unit of fuel was consumed.

The astute reader will immediately begin to ask questions. What about electric vehicles and motor fuels not measured in gallon units? What about motor fuels which are not currently paying any highway tax? How come some highway uses of motor fuel are not included for attribution? And what about the problem of fuel purchased tax-paid in one state, and then consumed on highways in another state? The purpose of this paper is to raise these issues, recognizing that some of them have not been answered in any definitive way.

Where Existing Attribution Procedures Do Not Meet the Ideal

The proceeding sections have discussed attribution methodologies. In the actual world, a perfect attribution procedure cannot and will not be achieved. However, serious effort can be made to identify the significant attribution shortcomings and to analyze the potential to achieve second-best solutions, or some form of satisfactory solution. Several principals should be considered:

• Equity among states, and (as a corollary issue) does data inaccuracy create biases between certain states. (Equity can be defined as treating entities with certain characteristics or parameters alike.)
• Accuracy of reported data.
• Accuracy of estimated data (either by states using unrelated, individual estimation methodologies, or by consistent, single-source estimation procedures run by each state or run by FHWA for all states)
• Size of the inaccuracy. Small inaccuracies may not be worth the cost to resolve. Careful analysis of the costs and benefits of the additional data collection is needed.
• Other

The above points should be considered when analyzing and discussing the following individual issues. The following discussion is grouped into two categories:

• Institutional - management or intergovernmental issues
• Methodological - data collection and analysis issues

Some of the reasons the ideal is un-achievable include:

Institutional

• Because of the need to combat tax evasion, Federal and state legislation is pushing tax payment up the motor fuel distribution chain to the wholesaler or distributor with the objective of having fewer taxpayers to track. And this is good tax policy. But for attribution purposes, direct highway user taxation at the retail, or use, level would provide the best data for attribution.
• What was meant precisely when Congress passed the legislative language on attribution, and how has this translated into actual practice? Section 157 of the STAA of 1982 defines the Minimum Allocation as: ". . . . shall not be less than 85 per centum of the percentage of estimated tax payments attributable to highway users in that state paid into the Highway Trust Fund, other than the Mass Transit Account, in the latest fiscal year in which data are available." The Federal Register Notice of 1985 describes attribution with gallons of highway motor fuel use as the attribution weighting factor: " . . . . . Federal motor fuel receipts to the Highway Account of the Highway Trust Fund are received from each state in proportion to the use of gallons of motor fuel in each state compared to the total highway use of gallons of motor fuel in all states." FHWA documents leading up to the Federal Register Notice describe how the decision was made to attribute based on highway use, and how this was coordinated with Congressional staff.
• Federal government definition issues: for example, the Interstate Maintenance Program is 33 percent based on commercial vehicle contributions to the highway account, while the National Highway System Program is 30 percent based on highway use of diesel. FHWA guidance asks states to provide private and commercial use of diesel fuel - this should not include motor fuel consumed by government diesel highway vehicles (which should be included in highway use of diesel.) However, many states cannot split out government uses from private and commercial diesel. The final result of all this is FHWA cannot accurately attribute either commercial vehicle use, or highway use of diesel.
• Cross border transfer of fuel in the tank connected to the vehicle's propulsion motor. This issue is largely a gasoline matter as diesel fuel taxes are more directed at place of use. States don't pursue cross-border taxation due to the way the Interstate Commerce Clause of the U.S. Constitution has been interpreted by the courts. Several states have legislative language that also directs them to ignore propulsion-tank transport of motor fuel.

Methodological

• Federal/state tax legislation asymmetry. As discussed earlier, the way motor fuel is taxed in each state is different from every other state and from the Federal motor fuel tax. A good example of this is gasohol: the Federal government defines three gasohol types (5.7 to 7.7 percent gasohol, 7.7 to less than 10 percent gasohol, and 10 percent or greater gasohol). None of the states report these three levels, only one state recognizes three types, and only a few recognize two types of gasohol. For this and other reasons, FHWA estimates gasohol for use in attribution. In general, for accurate attribution, each state's data must be analyzed and gallons of motor fuel adjusted to meet FHWA's criteria of on-highway motor fuel use. Adjusting state data to meet standard criteria creates equity among the states in the use of motor fuel data in attribution.
• The inability of state motor fuel tax reporting to capture data required to meet Federal needs. For example, state and local motor fuel is Federally tax exempt, but most states do not capture data that identifies these gallons. In these cases, it has been FHWA's policy to include these gallons for attribution since they were consumed on the highway, even though they did not pay the Federal tax.
• Attributing Federal non-gallon based taxes (the truck and trailer excise, tire excise, and heavy vehicle use taxes) using special fuels gallon data as a proxy.
• Refunds non-highway consumers are entitled to but do not apply for (or exemptions not taken) which will be (but shouldn't be) included as highway use.
• Reporting of the value of assessments (without penalty amounts added in) converted to gallons of motor fuel, captures motor fuel used in a state in the past, not the current reporting period.
• Several states collect fees from the sale of decals in lieu of per unit taxes for highway use of fuels not normally considered as highway fuels (alternative fuel decals are a prime example). While these fees are very poor measures of highway use of these types of fuels, the states that utilize non-distance based decals currently do not get any credit for these highway uses which do pay the Federal motor fuel tax.
• Using the measure of gallons of motor fuel for attribution creates a potential problem as several alternative fuels are not unitized in gallons. Most of these fuels can be treated uniformly by using mandated or scientifically known conversion rates, and therefore they must be adjusted before use in attribution. Furthermore, it may be impossible to separate these fuels for use in attribution at this time, since it is typical that both revenue and unit data are combined with the diesel fuel to create a special fuels category. Furthermore, the use of these fuels is so small that it may not be cost effective to pursue corrections unless the uses of these fuels becomes a larger portion of total highway uses.

How to Deal with Insufficient Data

Assuming FHWA chooses attribution on the basis of where the tax is paid and using motor fuel gallon data, then the question is how to deal with data not available or not of sufficient quality to use for attribution. The options include:

• FHWA estimation of the data using a single methodology;
• FHWA acceptance of state estimation of the data using one standard methodology;
• FHWA acceptance of state estimation by one of several pre-approved methodologies (sometimes called best practices);
• FHWA acceptance of as state's proposed estimation procedures;
• FHWA acceptance of a recommendation from a peer group of states that a state's estimation procedures are acceptable; or
• make a determination to not take any action, and therefore not to use in attribution.

The data on state's ability to track any of the following information came primarily from the Re-assessment survey completed May 1999. Other sources occasionally provided additional data.

Currently FHWA estimates:

• three grades of gasohol (about 32 states have good data on 10 percent gasohol);
• public uses of gasoline (Federal--which 18 states can track, state--which 10 can track, local--which seven can track, and public school--which five can track);
• off-highway uses of gasoline (generally, states do track these) including:
  • aviation;
  • marine;
  • agriculture;
  • construction;
  • industrial and commercial;
• off highway recreational use of gasoline (most states don't track this).

Currently we do not estimate (which means they are assumed to be zero unless a state reports data in accordance with FHWA's Guide to Reporting Highway Statistics):

• public uses of non-IFTA diesel (Federal--which 11 states can track, state--which nine can track, local--which seven can track, and public school--which six can track);
• neat (85 percent) alcohols: M85, E85 (three states can track these);
• other alternative fuels, in some unit other than gallons: liquefied petroleum gas (LPG)--18 states can track, liquefied natural gas (LNG)--eight states can track, compressed natural gas (CNG)--13 states can track, kerosene-five states can track;
• electric vehicles;
• travel by decal-registered alternative fuel vehicles (20 states have this form of tax).

Currently FHWA makes some determinations (recognizing our determination are sometimes arbitrary, and largely based on attempting to be consistent across all states):

• gasoline loss allowances -- percentage losses to be capped at 1 percent. The reason for capping this number at one percent is FHWA's belief that the fuel is not actually lost, but is consumed on the highways. States typically use this mechanism for reimbursing taxpayers for administrative costs of paying the tax, and the fuel is not lost. Actual losses, however, are considered to be real losses and are treated as non-highway gallons in attribution. (30 states have no allowances, 10 have a specific percent amount, and four report actual losses.)
• deletes from state-reported diesel data: refunds, exemptions, and other identifiable non-highway and public diesel uses. (Six states can identify private uses, seven can identify commercial uses, six can identify Federal uses, five can identify state uses, five can identify local uses, and three can identify school uses of diesel.)

Issues Under Non-estimation

Not estimating a number means that zero is the estimate. In fact, a bias is most likely introduced in these cases. How big a problem is this?

Conclusions

Initial thoughts on conclusions:

FHWA will probably continue to use gallons of motor fuel reported by states as the proxy for on-highway use, developing estimates where needed and including highway uses where states cannot separate out uses (like state and local gasoline and diesel, and public use of diesel.)

Under the re-assessment process FHWA has been conducting, the following general findings summarizes FHWA position at this point. (January, 2000)

FHWA believes the current attribution is sound:

• the principals on which attribution are based are sound;
• the procedures and processes are sound; and
• decisions need to be made on the need to meet Congressional language on the inclusion or exclusion of public vehicles by program, and the distinction between diesel and special fuels.

Improvements in data are needed:

• improved uniformity
• improved accuracy
• improved consistency

Improvements in documentation are needed:

• revised Guide
• attribution description
• IFTA reporting description
• FHWA models and analyses documentation.

Reporting Problem Areas

The two biggest barriers to better reporting for almost all states are:

• Public use of diesel, which is inadequate, in the sense that most states cannot identify public uses separately, and inconsistent, in that those state that can separate it are disadvantaged.
• IFTA, which needs better procedures for report timing, in that most states collect the data quarterly, and that it takes longer for states to administrate the IFTA procedures than the procedures IFTA replaced, and accuracy, while we are not sure why, many states data reveals falling revenues when economic indicators would support opposite conclusions.

Barriers to More Equitable Attribution - Gasoline/gasohol

• Treatment of gasoline flat % loss allowances
• State/Federal mismatch of gasohol definitions
• General improvement/refinement of the gasohol model
• Native American unreported gasoline

Barriers to More Equitable Attribution - Special Fuels

• Inability to report public use of diesel separately/Inclusion of public diesel in gross volume reported
• Native American unreported diesel
• Special fuels not adequately reported
• Unit conversions for special fuels ( energy equivalents and non-gallons units)
• Method of taxation (decals fees in lieu of per unit taxes)
• Significance (many special fuels are currently insignificant compared to total motor fuel volume)

Proposed Improvements

• Develop model to estimate public use of diesel
• IFTA reporting

Allow additional time before initial report

Permit alternative reporting procedures

• Eliminate one percent loss allowance cap
• Investigate separation of diesel/special fuels for attribution purposes
• Native American reporting

Tribal negotiation

State‑developed estimates

FHWA‑developed estimates

Improvements are also needed in the following areas:

• Revise FHWA Form 551M to four columns entitled gasoline, diesel, gasohol and other (other to be defined by the state.)
• Develop a standard to allow inclusion of highway use of alternative fueled vehicles that states currently tax via registration fees rather than per gallon taxes.
• Determine where states include data that they cannot separate and let those states that can separate add the gallons in. It is unfair to penalize states that are going by the Guide to Reporting Highway Statistics (Guide) when most cannot.
• Revise FHWA's guidance to cover these issues and to add a glossary.
• Continue to communicate the methods used in attribution to improve stakeholder understanding.
• Document the attribution process.
• Produce a Federal register notice to re-issue attribution policy, and to revise Guide as discussed above.