Abstract
A growing number of states are pursuing strategies to combat declining fuel tax revenue and fund road construction and maintenance, including the use of sales taxes, income taxes, and tolls; raising fuel tax rates; and adopting road mileage user fees. We use data from a nationally representative survey to compare public acceptability of a mileage user fee with each of these alternative revenue mechanisms. We find that support for the revenue options varies from 13.4 percent for income taxes to 33.8 percent for tolls, with higher gasoline tax rates, mileage user fees, and sales taxes in the middle. The evidence also points to stronger intensity of opposition than intensity of support across all alternatives. Finally, we find that, conditional on opposition to the mileage user fee, public acceptability is highest for tolls, followed by higher fuel taxes, sales taxes, and income taxes. Policy implications are discussed.
There is no topic of more importance to the future of public administration than the quest for taxes that the public may tolerate. With some segments of public opinion hardening against taxation of any form, government officials are struggling to find adequate revenue to support even the most basic functions of government. One of those basic functions is road construction and repair. Although the gasoline tax was once seen as a reliable and dedicated finance source for highways, it has become outmoded by the increasing fuel efficiency of motor vehicles, the rising cost of construction, and the inability of politicians to muster consensus for periodic raises to the rates of gasoline taxation (Wachs 2007; Duncan and Graham 2013).
The annual tax revenue generated by the federal fuel tax is more than US$20 billion lower than the US$54 billion required each year to maintain highway performance at its current level (Kile 2011). The shortfall is even more dramatic when compared to the annual investment of US$220 billion required to maximize net benefits from spending on highways, bridges, and transit infrastructure (American Society of Civil Engineers 2011). Similar shortfalls exist at the state level. For example, future funding shortage is partly responsible for Indiana House Bill HB1104, which “… authorizes [the] Indiana Department of Transportation to hire a third party to study methods that could be used to pay for the state’s infrastructure.” 1
As the federal government relies more on general revenue to supplement funds from the federal gasoline tax, a growing number of states are pursuing alternative strategies to combat declining fuel tax revenue, including the use of sales taxes, income taxes, and tolls; raising fuel tax rates; and adopting road mileage user fees. Duncan et al. (2014a) argue that this strategy of relying on a mix of revenue sources is consistent with the public’s view of benefit-based taxation. Roughly a third of Americans support benefit-based taxation in the context of road financing, which suggests that relying on multiple revenue sources that vary in the extent to which they are consistent with the benefit principle of taxation might be a responsive political strategy (Duncan et al. 2014a).
However, there is not much systematic evidence on how the public feels about the specific financing options being pursued by governments. The objective of this study is to compare public acceptability of five alternative revenue mechanisms: a mileage user fee, higher fuel tax rates, greater reliance on tolls, higher retail sales tax rates, and higher income tax rates. 2 In particular, we ask the following questions: what is the level of support for each alternative revenue option and, conditional on opposition to the mileage user fee, which of the four alternatives is most acceptable to the public? Because of the renewed interest in the mileage user fee among policy analysts and government officials despite widespread public opposition, we argue that the latter question ought to be of particular interest to policy makers. We also assess whether support for alternatives varies by demographic, political, and driving behavior characteristics.
Most of the existing research on these revenue options includes only some of the options, not all of them. The most comprehensive effort may be by Agrawal and Nixon (2013), who compare a mileage user fee, a sales tax, and a fuel tax increase, at the federal level. They find that 19 percent of their respondents support a mileage user fee with a flat rate of US$0.01/mile. This is lower than the level of support for a new US$0.50 national sales tax (51 percent support) as well as a US$0.10 increase in the gasoline tax used to finance roads (67 percent support). 3 A similar synthesis of 110 surveys/polls was conducted by Zmud and Arce (2008), who find that a majority of existing surveys/polls report public support, rather than opposition, for tolling and road pricing initiatives. 4 Additionally, Zmud and Arce find that tolling/road pricing garners greater public support relative to tax-related initiatives that sought to increase the fuel taxes. 5
While the synthesis provided by Zmud and Arce (2008) is informative, it is difficult to compare results across surveys due to differences in the designs of the survey instruments, the modes of administration, and the national representativeness of sampled respondents. Furthermore, it is worth noting that Zmud and Arce found that the results of the surveys/polls were associated with factors of the survey design, including the wording of questions, and the sponsor of the research. An important advantage of our analysis, then, is the fact that all of the revenue options are presented to the same nationally representative sample of respondents and presented under the same context with the same mode of survey administration. This makes comparison of the public acceptability of the revenue options more valid than would be achieved by simply comparing results across the existing studies.
Our data on public opinion are drawn from the Indiana University School of Public and Environmental Affairs (IU-SPEA) Mileage User Fee Survey that was administered in the summer of 2013. The survey elicited public opinion about mileage user fees from 2,087 individuals who are weighted to be representative of the noninstitutionalized US adult population.
We find low levels of support for each of the revenue options as means of funding road construction and maintenance, with levels of support ranging from 13.4 percent for income taxes to 33.8 percent for tolls and with higher sales taxes (18.3 percent), mileage user fee (20.9 percent), and higher gasoline tax rate (28.8 percent) filling out the middle. The evidence also points to stronger intensity of opposition than intensity of support across all alternatives. Finally, we find that, conditional on opposition to the mileage user fee, public acceptability is lowest for personal income tax (12 percent) followed by sales tax (18 percent), gasoline tax (29 percent), and tolls (31 percent). Of the revenue measures considered, support for an increase in the gasoline tax rate is the one most influenced by individual characteristics such as household income, education, political affiliation, and knowledge of current federal gasoline tax rates.
The findings suggest that policy makers at the federal level and some state officials are pursuing policies that enjoy less public support than a mileage user fee. For example, the Federal Highway Trust Fund is currently being supplemented by transfers from the general fund, which is financed mainly by income taxes. Although there is a greater mix of strategies at the state level, some states rely heavily on sales taxes and others have considered raising income taxes to help finance transportation infrastructure. More importantly, our results suggest that the support for higher fuel tax rates may be artificially low due to a misimpression: approximately 70 percent of respondents believe the federal fuel tax rate is higher than it actually is and support for higher fuel tax rates is lower among these individuals. This misimpression is important because higher fuel tax rates are often seen as the most viable short-term solution to the current funding crisis, while a road mileage user fee is seen as a longer-term solution. The results also imply that greater reliance on tolls, where feasible, is one of the more viable options, at least from a public acceptability standpoint and may represent a transition to a more comprehensive road mileage user fee. This too, is an important observation because, unlike fuel tax rate increases, tolls can be used as part of both a short-term and long-term solution to the current funding crisis.
The remainder of the article is organized as follows: The next section describes the data. This is followed by a description of the findings and finally a discussion of policy implications.
Data
This section of the article provides a description of the survey. The primary objective of the survey was to elicit public opinion regarding policy proposals to replace fuel taxes with mileage user fees. Because the focus of this article is on public perception regarding alternative revenue sources, we only describe those features of the survey that allow us to answer those research questions. A more detailed description of the survey is provided in Duncan et al. (2014b).
The data are taken from an Internet-based public opinion survey that was designed and sponsored by the IU-SPEA and conducted in August and September 2013 by the GfK Group. Well-designed Internet surveys, when compared to surveys based on random-digit dial telephone methods, have been shown to be comparable or superior with regard to key criteria of concern (e.g., representativeness and response quality) to the survey research community (Yeager et al. 2011).
Additional individual socioeconomic, demographic, and political characteristics are provided by GfK as part of their standard set of variables available for panel members. These include age, sex, marital status, household income, employment status, education, region, political party affiliation, and political ideology. These variables are primarily used to evaluate the representativeness of our sample. However, we also explore whether these characteristics are related to level of support for alternative revenue sources.
The Generic Mileage User Fee, Alternative Revenue Sources, and Related Questions
We ask respondents for their opinion about what we call a “generic” mileage user fee, that is, one for which respondents are not provided specific information about how the mileage data would be collected or how the fees would be administered. Duncan et al. (2014c) compared attitudes toward a generic fee with attitudes toward specific fees based on odometer readings or various GPS systems and found that levels of support—while not identical—are similarly low. The specific question used for the mileage user fee, after a short definition of the concept, is: Would you support an effort to Some have suggested the following alternatives to raise the money needed for road maintenance, repairs, and construction. Please tell me whether you strongly agree, agree, disagree or strongly disagree with the following statements? (NOTE: Statements were presented in randomized order.) The gasoline tax rate should be increased. The gasoline tax should be replaced with a higher general retail sales tax rate. The gasoline tax should be replaced with a higher personal income tax rate. The gasoline tax should be replaced with more tolls on roads.
At the end of the survey, we ask a series of questions about participants’ driving-related characteristics, including number of vehicles owned, number of miles driven weekly, and use of interstate highways. We use these questions to explore whether the level of support/opposition to the alternatives is related to respondents’ driving behaviors. For example, people who drive a lot of miles may be less likely to support a mileage user fee. We also elicit respondents’ knowledge of the federal gasoline tax rate using the following question: If the average price of gasoline is US$3.50 per gallon, how much of that price would you guess goes to the federal gasoline tax? Even if you don’t really know, please give your best guess. Respondents selected from one of the following categories: less than US$0.10 per gallon, US$0.10 to US$0.24 per gallon, US$0.25 to US$0.49 per gallon, US$0.50 to US$0.74 per gallon, US$0.75 to US$0.99 per gallon, US$1.00 to US$1.24 per gallon, US$1.25 to US$1.49 per gallon, US$1.50 to US$1.74 per gallon, US$1.75 to US$2.00 per gallon, more than US$2.00 per gallon.
Representativeness of Sample
The web survey was fielded to a nationally representative sample of 3,325 US adults aged 18 years or older from GfK’s KnowledgePanel®. The KnowledgePanel is a probability-based online panel. Members are recruited using random-digit dialing and address-based sampling methods that include both households with and without Internet access, thus providing nearly complete coverage of the US population. 7 Probability-based Internet panels have advantages compared to random-digit dialing telephone surveys and other methods, including the potential for reduced measurement error, lower cost, and increased timeliness (Chang and Krosnick 2009; Yeager et al. 2011).
The survey was fielded from August to September 2013. Five reminders to encourage participation were sent to nonrespondents during this time period. We obtained 2,142 respondents for an American Association for Public Opinion Research completion rate of 64 percent among panel members. Approximately 2.5 percent of respondents (fifty-five respondents) are removed from the sample due to short survey completion time: less than six minutes. 8 Short response time is an indicator used by survey researchers to identify speeders who complete the survey without reading and carefully answering the items (Olson and Parkhurst 2013). This leaves us with a sample of 2,087 respondents.
Data are weighted to adjust for unequal probabilities of selection and to reflect Current Population Survey estimates of the US population on demographic characteristics, including age, sex, race/ethnicity, education, household income, home ownership status, Internet access, Census region, and metropolitan area status. Standard errors have been adjusted to account for weighting.
Findings
The objective of this study is to compare public support for replacing the gasoline tax with a mileage user fee to support for four revenue alternatives: higher fuel tax rates, greater reliance on tolls, higher retail sales tax rates, and higher income tax rates. We are also interested in examining the extent to which opponents of the mileage user fee support each of the four alternatives. This section of the article discusses our main findings. It is important to recognize that support for a given revenue instrument is conditional on how that instrument is framed in the survey.
Support for Revenue Mechanisms
Our measure of support for the mileage user fee is a dummy variable that is equal to 1 if the respondent supports an effort to replace the gasoline tax with a generic mileage user fee and zero otherwise. We define our measure of support for each alternative in two ways. The first is a dummy variable that is equal to 1 if the respondent agrees or strongly agrees with an alternative and the second is a count variable on a four-point Likert-type scale of strongly agree, agree, disagree, and strongly disagree. The latter variable allows us to measure the intensity of support/opposition for each alternative.
Table 1 summarizes the level and intensity of public support for the mileage user fee and the four alternatives. We find that there is some variation in public support for the alternative revenue sources for road financing, as the level of support ranges from 13 percent for income taxes to 34 percent for tolls, with sales taxes (18 percent), mileage user fee (21 percent), and higher gasoline tax rate (29 percent) filling out the middle. Although the levels of support shown here are statistically different from each other at the 1 percent level, they mostly point to very high levels of opposition to the main road-financing options currently being considered by policy makers around the country. Additionally, results reported in columns 2 to 5 of table 1 show that the intensity of opposition is stronger than the intensity of support across all alternatives. Most supporters tend to simply “agree” with the revenue options, while a larger share of opponents tends to “strongly disagree.”
Support for Alternatives.
Note: Reported are the percentage of respondents who support each alternative as well as the intensity of support/opposition for each alternative indicated in each row. Numbers have been rounded to the nearest whole number. Number of observations is 2,064.
To explore whether support for the four alternative revenue sources is associated with individual characteristics, we regress the binary measure of support for each of the alternatives on demographic, socioeconomic, political, and driving behavior characteristics of the respondents. 9 The linear probability models are estimated separately for the four alternatives and the results are presented in table 2. We find that support for the benefit principle as defined by UserFeePrinciple is positively associated with support for revenue sources that are linked to degree of road usage such as tolls and gas tax and these correlations are statistically different from zero at the 10 percent and 1 percent levels, respectively. 10 The correlation between socioeconomic characteristics and support for alternatives varies across alternatives. Support for the gasoline tax is higher in the South, and higher among men, whites, Democrats, high-income households, and the highly educated. We also find that respondents who indicated that the federal gas tax rate is more than US$1 per gallon are significantly less likely to support an increase in the gasoline tax rate relative to respondents who indicated that the gasoline tax rate is less than ten cents a gallon; the correct rate is 18.4 cents per gallon.
Correlates with Support for Alternatives.
Note: Reported are the estimated coefficients of a linear probability model where the dependent variable is equal to 1 if respondent supports the alternative indicated in column title and 0 otherwise. All models are estimated with sample weights and robust standard errors are in parentheses. HS = high school.
aUserFeePrinciple is a dummy variable that takes a value of 1 if individual i agrees that those who use the road should pay more than those who do not use the road or only those who use the road should pay, and 0 otherwise.
bParty affiliation is constructed from a seven-point Likert-type scale where the Moderate group includes individuals who “Lean Republican,” “Lean Democrat,” or are “Undecided/Independent/Other.”
*Significant at 10 percent.
**Significant at 5 percent.
***Significant at 1 percent.
Support for more tolls is largely uncorrelated with the political and socioeconomic variables. The notable exceptions are households with very high income and gender; support is higher in very high-income households and higher among women. Similarly, support for higher income tax rates is negatively correlated with age, but higher among political moderates and Democrats. On the other hand, support for higher retail sales tax rates is higher among men, moderates, Democrats, and respondents in the Midwest and South and is positively correlated with income. We find that an individual’s employment status, marital status, and driving characteristics such as number of miles driven per week, frequency of interstate usage, and number of vehicles owned are not associated with support for any of the four alternative revenue sources.
Support for Alternatives among Mileage User Fee Opponents
This section describes support for each of the alternative revenue sources conditional on opposing the mileage user fee. Our measure for conditional support is coded two ways. First, we create a dummy variable that is equal to 1 if the respondent opposes the mileage user fee and agrees or strongly agrees with the alternative and 0 if she or he opposes the mileage user fee and disagrees or strongly disagrees with the alternative. The second definition characterizes the intensity of conditional support and is equal to 1 if the respondent opposes the mileage user fee and agrees with the alternative, 2 if the respondent opposes the mileage user fee and strongly agrees with the alternative, and 0 if the respondent opposes the mileage user fee and disagrees or strongly disagrees with the alternative. Finally, we define respondents who disagree with the mileage user fee and all of the alternatives as an opposer.
Table 3 summarizes the level and intensity of conditional support for each alternative. We find that support for an increase in the gas tax and greater reliance on tolls, conditional on opposing the mileage user fee, is 29 percent and 31 percent, respectively. Corresponding figures for an increase in retail sales and personal income taxes are 18 percent and 12 percent, respectively. 11 These results indicate that there is considerable variation in support for each alternative among those who oppose a mileage user fee. We also find that intensity of support for alternatives among mileage user fee opponents is relatively modest, with roughly 1 to 4 percent of the respondents who oppose the generic mileage user fee “strongly supporting” any of the alternatives. The “opposers” (respondents who oppose the mileage user fee and all four alternatives) constitute approximately 34 percent of the sample (687 respondents). 12
Percent and Intensity of Support Conditional on Opposing the Mileage User Fee.
Note: Number of observations is 1,610.
aReported is the weighted percentage of respondents who support each alternative conditional on opposing the mileage user fee.
bWeak support measures the percentage of respondents who oppose the mileage user fee and agree with the respective alternatives.
cStrong support measures the percentage of respondents who disagree with the mileage user fee and strongly agree with the respective alternatives.
To explore whether support of any of the other four revenue choices among mileage user fee opponents is associated with individual characteristics, we regress our binary measure of conditional support on demographic, socioeconomic, political, and driver characteristics. The linear probability models are estimated separately for the four alternatives and the results are presented in table 4. 13
Conditional on opposing the mileage user fee, we find that respondents who agree with the benefit principle of taxation as defined by UserFeePrinciple are 14 (significant at 1 percent level) and 3 percent more likely to support an increase in gas tax and greater reliance on tolls, respectively and are 5.1 (significant at 5 percent level) and 2.5 percent less likely to support an increase in retail sales and personal income taxes, respectively. These results provide further evidence that the support for the benefit principle is positively associated with likelihood of supporting alternatives that are linked to degree of road usage.
The results in table 4 also show that socioeconomic characteristics are correlated with the likelihood of supporting an increase in gas tax conditional on opposing the mileage user fee. We find that white respondents are more likely to support an increase in the gasoline tax than nonwhite respondents. 14 There is also evidence that Democrats are more likely to support the gas tax relative to Republicans and that education and household income are positively correlated with the likelihood of supporting an increase in the gasoline tax. Respondents’ knowledge about the federal gasoline tax rate is also associated with the likelihood of support; respondents who believe the current federal tax rate is higher than it really is are less likely to support the gas tax.
Correlates with Support for Alternatives Conditional on Opposing the Mileage User Fee.
Note: Reported are the estimated coefficients of a linear probability model where the dependent variable is equal to 1 if the respondent opposes the mileage user fee but supports the alternative indicated in column title, and 0 if the respondent opposes both the mileage user fee and the alternative indicated in column title. All models are estimated with sample weights and robust standard errors are in parentheses. HS = high school.
aUserFeePrinciple is a dummy variable that takes a value of 1 if individual i agrees that those who use the road should pay more than those who do not use the road or only those who use the road should pay, and 0 otherwise.
bParty affiliation is constructed from a seven-point Likert-type scale where the Moderate group includes individuals who “Lean Republican”, “Lean Democrat,” or are “Undecided/Independent/Other.”
*Significant at 10 percent.
**Significant at 5 percent.
***Significant at 1 percent.
There is no evidence that the likelihood of supporting greater reliance on tolls, conditional on opposing the mileage user fee, is correlated with individual characteristics except for the respondent’s knowledge of the federal gas tax rate and political party affiliation; support is lower among those who believe the federal gasoline tax rate is higher than it actually is, and lower among Republicans (relative to Democrats). On the other hand, support for the retail sales tax, conditional on opposing the mileage user fee, is higher among respondents who disagree with benefit-based taxation; higher among men; higher in the Midwest, South, and West (relative to the Northeast); and higher among political moderates and Democrats (relative to Republicans). We also find higher levels of support for higher income taxes, conditional on opposing the mileage user fee; among Moderates and Democrats; and in the Midwest, South, and West (relative to the Northeast), as well as among younger respondents.
Conclusion and Discussion
We use data from the IU-SPEA Mileage User Fee Survey to assess public opinion about five revenue mechanisms aimed at addressing the inadequacy of the current fuel tax. We find low levels of support for each revenue measure, as the level of support ranges from 34 percent for tolls to 13 percent for income taxes, with higher gasoline tax rate (29 percent), mileage user fee (21 percent), and sales taxes (18 percent) filling out the middle. The evidence also points to stronger intensity of opposition than intensity of support across all revenue options. Finally, we find that people who oppose the mileage user fee are more likely to support tolls, an increase in the gasoline tax, higher sales tax rates, and higher personal income tax rates, in that order. Roughly a third of the respondents (the opposers) report opposition to all revenue financing alternatives.
We recognize that support for road-financing instruments depends on the framing of the instruments in the survey. Therefore, caution must be exercised when comparing our findings to the existing literature. Nonetheless, our findings are somewhat consistent with Zmud and Arce (2008) who report that tolls have greater public support than tax alternatives (including higher gasoline tax rates), and Agrawal and Nixon (2013, 2014) who find that a US$0.10 increase in gasoline tax is preferable to a flat rate US$0.01/mile mileage user fee. Interestingly, the level of support identified in Agrawal and Nixon (2013, 2014) for these two mechanisms is almost identical to what we find in our study, despite important differences in survey design.
Our results raise important policy implications that are worth considering, as policy makers aim to solve the road-financing problem facing the United States. First, we find that there is widespread misperception of the current fuel tax rates in the country and that this misperception has a nontrivial association with the extent to which individuals are willing to support increases in fuel tax rates and tolls. Opposition to an increase in the federal gasoline tax is higher among those who overestimate the current size of the federal gasoline tax. Therefore, it is possible that increasing the salience of the federal fuel tax rate—it is currently applied at the wholesale level and embedded in the price of gasoline at the pump—may actually increase public support for higher fuel tax rates, assuming that salience has an educational effect.
Second, we find that, conditional on opposing mileage user fees, respondents are more likely to support higher gas tax rates and tolls than sales taxes and income taxes. One possible interpretation of this finding is that, other things equal, people prefer revenue sources that satisfy the benefit principle of taxation. In other words, they opt for revenue sources that are consistent with the benefit principle even as they avoid the mileage user fee. This finding is interesting because the mileage user fee is arguably the strongest form of a benefit-based revenue source for road financing but remains objectionable for a number of reasons including implementation costs and perceived invasion of privacy. Therefore, we suggest that policy makers seeking short-term solutions to the current highway funding crisis give greater emphasis to revenue sources that are consistent with the benefit principle of taxation. This does not mean that policy makers should abandon all of the revenue instruments that they currently use. Instead, our results suggest that policy makers may want to change their revenue mix such that a greater share of revenues comes from those instruments that are more widely supported by the public.
For public administration scholars and practitioners, we urge an expansion of public opinion research to expose factual misunderstandings of the size of current tax rates (of all forms) and to measure the effectiveness of different informational/educational approaches aimed at curtailing misperception. Priority should be given to the misperception issue because increasing revenue from fuel taxes (by raising the fuel tax rate and/or indexing for inflation) is one of the easiest and cheapest short-term strategies for solving the road-financing problem. There are many contributors to the widespread public opposition to increasing revenues for basic public services, some that are very difficult to combat. In light of that complexity, priority should be given to tackling the simpler perceptual errors first. Understanding and reducing public misperceptions of the size of common tax rates (e.g., personal income, corporate income, sales and property tax rates) is a crucial starting point.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.