Forecasting Transportation Revenue Sources: Survey of State Practices (2015)

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.

5 EARLY ADOPTION OF HIGHWAY USER FEE FINANCING Before 1920, most states’ funding for transportation systems, including highways and bridges, depended on allocations by state legislatures of general funds derived from property and income taxes (Martin 1923; Burnham 1961). But, as auto- mobile purchases rose at an unprecedented rate, states found themselves spending huge sums of money on roads and fall- ing far behind the perceived need for such facilities. For example, in 1922, 44% of the budget of California went to cover investments in transportation, including interest pay- ment on road bonds (Zettel 1946); yet drivers and interest groups such as automobile clubs demanded more investment. Gradually, the states came to believe that the majority of the responsibility for funding road construction and maintenance should devolve on road uses rather than general revenues because at the time, most citizens did not own vehicles; those who did drive had higher than average incomes. Truck traffic was increasing rapidly, requiring more durable pavements, and goods movement was derived from commercial ventures that legislators believed should be responsible for a fair share of the costs of roadways. Some states built toll roads, but tolls had to be collected manually, and the cost of collec- tion at some locations approached or exceeded the revenues obtained from the tolls (Brown 1999). In 1919, the state of Oregon instituted the first motor fuel tax on gasoline and diesel as a way of approximating revenue from tolls while incurring far lower costs of collection and administration (Burnham 1961). The first state excise taxes levied on motor fuels—a penny or two per gallon—were large in relation to the early sale price of gasoline, but accept- able to road users who widely acknowledged the importance of improving the quality and extent of the networks. Fuel taxes were collected at a relatively few wholesale distribu- tion points and the taxes were passed along at the gasoline pumps to the highway users, minimizing administrative costs and lowering rates of fraud, abuse, and evasion. As traffic grew, motor fuel excise taxes provided an increasingly lucra- tive and practical source of transportation revenue. By 1940, every state had a motor fuel tax, and to insure equity to those paying the fees, all but a few states dedicated or “earmarked” fuel tax revenues to transportation through state constitutional provisions or legislation. State trust funds became more com- mon, in the form of accounts reserved for surface transpor- tation expenditures into which fuel and other transportation user excise taxes were deposited. In 1932, the federal govern- ment added a national motor fuel tax, and in 1956, motor fuel taxes and the U.S. Highway Trust Fund (HTF) became the principal means by which the National Interstate and Defense Highways Act was funded (Brown 2001; Wachs 2003). Increases in both vehicle registrations and miles travelled per vehicle, coupled with modest increases in rates of taxation, ensured ample financing for one of the most extensive and heavily used public works systems in the world (Brown 1998; Brown et al. 1999). In addition to being lucrative, vehicle reg- istrations and travel grew steadily at stable and highly predict- able rates, meaning that the revenues produced by those taxes could accurately be predicted from year to year. Federal and state fuel taxes have been the principal source of revenue for transportation programs for nine decades. In 2004, for exam- ple, the AASHTO Center for Excellence in Transportation Project Finance showed that motor fuel taxes provided more than 82% of federal transportation revenue and slightly less than 37% of state revenue, which together yielded more than $62 billion in revenue for surface transportation programs. But even a decade ago, other revenue sources, including tolls, fares, vehicle use taxes, sales taxes, allocations of general gov- ernment revenues, property taxes, and other fees and charges, produced more funding in aggregate for surface transportation programs than did motor fuel taxes (Cambridge Systemat- ics et al. 2006). Today, more than $200 billion per year is invested in surface transportation by government agencies, and motor fuel taxes are declining precipitously as a share of the total (AASHTO Center for Excellence in Transportation Project Finance). THE CURRENT CRISIS IN SURFACE TRANSPORTATION FINANCE States and the federal government today face what is widely recognized as the most serious fiscal challenge in nearly a century. Fuel tax revenue has been declining in relation to road use as measured by VMT for more than 20 years. The federal motor fuel tax rate per gallon has not increased since 1993, causing the value of tax collections to steadily lose purchasing power. Although sales tax revenues automatically rise with inflation, and income taxes rise as people move into higher tax brackets, motor fuel taxes were historically set at a fixed amount per gallon for reasons that were more relevant than they are today. That the tax was based on the quantity of chapter two BACKGROUND: FISCAL CRISES AND CHALLENGES TO TRANSPORTATION REVENUE FORECASTING

6 fuel meant that taxes could be charged at the wholesale distri- bution points and did not depend upon the more variable sale price at the pumps at which consumers filled their tanks. The Institute on Taxation and Economic Policy (2014) recently reported that motor fuel tax revenues are on an unsustain- able course because Congress has refused to raise the fed- eral per-gallon tax rate, and many states have also been slow to raise per-gallon rates or change to another type of user charge. Purchasing power has plummeted as a result of infla- tion and rising construction costs. According to the institute, after accounting for growth in construction costs, the average state motor fuel tax rate has effectively fallen by 20% since it was last increased. States have seen the real value of fuel tax revenues drop by a nationwide total of $10 billion each year. Similarly, the fixed-rate federal gas tax has lost 33% of its purchasing power since it was last raised in 1993. In July 2014, U.S. Secretary of Transportation Anthony Foxx was joined by 11 former Treasury secretaries in issuing a state- ment calling upon Congress to enact a new long-term federal revenue strategy that would resolve the continuing precipi- tous fall in national HTF revenues and contributions to state transportation finance—although, tellingly, they did not offer a consensus recommendation for the most effective approach to accomplishing this objective (Hoitsma 2014). Inflation and failure to raise the motor fuel tax rate are not the only causes of the declining value of transportation revenue streams. There is also great uncertainty as to changes in travel among drivers in the United States. Light-duty vehicle VMT per licensed driver peaked in 2007 at 12,900 miles per year and decreased to 12,500 miles in 2012, and there is debate in the transportation community as to whether this trend will become more pronounced or will be reversed in the future as a result of the gradual economic recovery. The shift in VMT highlights the importance of travel behavior and its influence on energy consumption, and consequently on revenues associated with the taxes levied on motor vehicle fuels. Before the 2007 peak, travel behavior in the United States closely tracked economic growth. The Energy Information Administration (EIA) notes that: “Since 2007, trends in U.S. travel by [light-duty vehicles] have not followed the trends in economic indicators such as income and employment as closely as before. Although eco- nomic factors continue to influence travel demand, demo- graphic, technological, social, and environmental factors also have shown the potential to affect personal travel” (Hutchins 2014). This is very important, since this study indicated that in making projections of transportation revenue, many states rely on projections of future travel that are published by that agency. Although projections of factors that influence travel have always been subject to error, the growing uncertainty in socio-economic relationships influencing travel, illustrated in Figure 1, suggest that factors influencing future travel are becoming less certain and more subject to error than they have been in the past. Improved fuel economy is taking an additional toll, since gasoline consumption in the United States peaked in 2004 at 138,819 million gallons. Since 2008, national fuel consump- FIGURE 1 Economic indicators of travel changing over time (Source: Hutchins 2014).

7 tion has declined in absolute numbers every year, and in 2012 total national fuel consumption stood at 123,635 million gal- lons (Sivak 2014). The average fuel economy (window-sticker value) of new vehicles sold in the United States in July 2014 was 25.6 mpg—up 0.1 mpg from the value in June and up 5.5 mpg since October 2007 (University of Michigan Trans- portation Research Institute website). New and demand- ing Corporate Average Fuel Economy (CAFE) Standards will require achieving a new car fleet average of 54.5 mpg by 2025 in order to reduce dependency on imported energy, greenhouse gas emissions, and urban air pollution. Although consumers may be pleased that the fuel economy of vehicles is improving at a dramatic rate, this does cause difficulty rais- ing tax revenue for surface transportation programs. Electric vehicles, which use no petroleum-based motor fuels, are as yet a small portion of the vehicle fleet, but the promise of electric and hydrogen-powered autos suggests that at a future date liquid fuel consumption could drop precipitously. The biparti- san Simpson–Bowles deficit reduction commission called for increasing the federal fuel tax by 15 cents per gallon, nearly doubling the federal tax in order to stabilize the Highway Trust Fund and allow for programmatic growth (National Commis- sion on Fiscal Responsibility and Reform 2010). Congress has not taken action on this recommendation, and while some bills have recently been introduced that would raise the federal motor fuel tax, none has been enacted by either house. Under federal law, the Highway Trust Fund is not allowed to have a negative balance. When the balance becomes dan- gerously low, U.S.DOT takes administrative action to pre- vent the fund from heading into the red. When the highway account balance dips below $4 billion, U.S.DOT must either substantially delay payments to states or pay a reduced share of state costs. Similar measures take effect when the mass transit account balance dips below $1 billion. Gasoline tax revenues have fallen so drastically that since fiscal year 2008, Congress has transferred $54 billion in general fund revenues into the HTF to prevent insolvency. On July 1, 2014, Transportation Secretary Foxx sent letters to all state secretaries of transportation informing them that on August 1, 2014, scheduled and requested federal payments to states for transportation projects would be systematically reduced unless Congress acted prior to that date by transfer- ring additional general fund monies into the HTF (A. Foxx, Letter to State Departments of Transportation, July 1, 2014). Transfers authorized as part of the MAP-21 surface trans- portation bill were intended to keep the trust fund healthy through the end of FY 2014, but projections that the account would run out of money as early as August 2014 (Center for American Progress 2014) prompted Congress to pass a fund- ing bill before adjourning for its summer recess. An editorial in the Washington Post noted that: The measure provides $10.8 billion for infrastructure projects around the country, with more than half of the money supplied not by any real increase in revenue or reduction in spending but by an egregious budgetary gimmick known as “pension smoothing.” The law allows companies to put off otherwise mandatory contributions to their defined-benefit employee pension funds, which increases tax revenue for the Treasury, since those con- tributions would have been tax-deductible. Actually, smooth- ing increases tax revenue in the short run but decreases it later on, when companies have to make up for the missed payments (Washington Post, “The Post’s View,” August 19, 2014). In addition to the decline in revenues from what has tradi- tionally been the most reliable source of transportation fund- ing, uncertainty associated with future revenue is growing so quickly that it also impacts states’ abilities to forecast future rev- enues with accuracy. States’ receipt of federal funds for which they are eligible can no longer be taken for granted, since delays, reductions in funding, and “patches” using unprecedented sources of revenue are becoming more common; and infusions of general fund monies can come at times that are extremely difficult to predict. According to many experts and studies, the current condi- tion of the nation’s transportation system is poor and worsen- ing, and increasing the performance of the network will require increasing investments. It is currently estimated that 32% of the nation’s transportation network is in poor or mediocre condition, and that 24% or all bridges are in poor condition or are functionally obsolete (National Conference of State Legislators). Thus, as revenue becomes scarcer, the conse- quences become more severe. DEVOLVING THE BURDEN OF FINANCE TO THE STATES Faced with a depleted federal trust fund and poor prospects for new funding from a deficit-conscious Congress, states are seeking to increase revenues for transportation opera- tions, maintenance, and capital investment. In a recent issue of Innovation Briefs, C. Kenneth Orski reported that a sur- vey had revealed significant funding initiatives in 18 states. He reported that some states, including Maryland, Wyoming, Massachusetts, and Vermont, have raised gasoline taxes, while others, such as Pennsylvania, have shifted to a tax on fuel at the wholesale level. Still others (Arkansas, Virginia, etc.) have enacted dedicated sales taxes for transportation or floated toll revenue bonds (Ohio). Burdened with addressing growing revenue shortfalls in a context of increasing uncertainty and accelerating change in transportation finance, it becomes both more important and more difficult for states to forecast transportation revenues. It is more important because forecasts are needed to assess and modify new financing policies, and it is more difficult to do that in a dynamic environment. Revenue measures are increasingly politically challenging, and in many states sub- ject to voter approval. States are increasingly incurring debt in the short run, many issuing bonds to fund transportation

8 capital improvements and some borrowing from state infra- structure banks. These approaches to financing often require longer-term revenue increases, many of which have yet to be enacted. In addition, some states are devolving transporta- tion finance to counties and other units of local government. In California, for example, 20 counties that are home to 81% of the state’s population have enacted voter-approved sales taxes for transportation projects (California Self Help Counties Coalition). In the election of November 2013, approximately 80% of local and state transportation taxation ballot measures across the country were approved by voters (Center for Transportation Excellence 2013). It is difficult to forecast the outcomes of increasingly decentralized voter- approved transportation revenue measures. This study was motivated by these concerns, coupled with the fact that there is no other compendium of revenue forecasting practices in the states. EARLIER REVIEWS OF STATE FORECASTING PRACTICES With one exception, the review of the published literature revealed very few prior studies that specifically examined transportation revenue forecasts that had been done by states. In 1999, however, economist David Gillen assessed how well states forecast revenues from taxes and fees levied on high- way users and examined whether the models they employed in forecasting revenues appeared to be adequate. An academic economist, he was critical of the approaches that were in use at the end of the 20th century. Interestingly, the synthesis survey revealed that most of the approaches he found to be in use at that time continue to dominate state forecasting practices today, even though trends in the socio-demographic determi- nants of travel and the variables they suggest be used as pre- dictors of transportation revenue are more volatile or variable today than they were when he conducted his study. Gillen described three approaches to forecasting transporta- tion revenue. The simplest would be to develop a model that uses previous values of revenues in each category, perhaps giving more weight to recent values of variables than values from the more distant past. This approach simply matches a function to past data and extrapolates the values to create a forecast. The models used by most states to forecast travel and other variables affecting fuel tax revenues were described by Gillen as largely “accounting identities” or simple statistical relationships predicting one of the components of revenues; he criticized these as “simplistic and non-behavioral.” One com- mon feature of such models is their implicit assumption that the demands for travel, vehicles, and fuel are not especially respon- sive to changes in social, demographic and economic variables. A second approach would utilize some econometric time series techniques, such as the Box-Jenkins or Autoregressive Integrated Moving Average (ARIMA) model. Univariate Box- Jenkins models are extrapolation methods which employ past values to generate forecasts. When the absence of data or spec- ification errors make econometric models impractical, Box- Jenkins models are useful for time-series forecasting. The third approach, causal forecasting, develops an econo- metric model that explains the underlying causes or sources of variation in the factors that play roles in determining revenues from fuel taxes and registration fees. These utilize relevant demographic and economic variables in a set of “behav- ioral” equations to produce the forecast. Gillen considered this to be the richest approach, because once the model parameters are estimated, they can be used to develop fore- casts of the dependent variables; however, at the time of his study, this method was used far less frequently than the simpler approaches. Gillen proposed a modeling approach that could serve as the basis for state forecasts and could be useful as prices and sources or revenue become more dynamic. His own approach consists of a system of three equations: two rela- tionships (VMT and fleet fuel efficiency) and one account- ing identity (total fuel consumption) that would provide the requisite information to forecast fuel tax and registration fee and other fee revenues. The first two equations would be esti- mated through regression analysis, whereas the third would combine the results of the first two.