Equity performance evaluation of two different pricing options: fuel tax per gallon and VMT fee
Currently, there has been extensive discussion on various pricing alternatives to the present tax per fuel gallon that is in effect. There has been great interest to identify additional resources that will increase the federal and state revenues allocated to the maintenance of the existing surface transportation network. The widely suggested policy measure of the vehicle-miles-traveled fee (VMT fee) is an alternative pricing option that has drawn great attention by researchers and policymakers, particularly regarding its equity performance among various social groups.
In this context, the objective of this thesis is two-fold. The primary objective is to identify which social sub-groups are mostly affected under the current fuel tax option and the alternative VMT fee option. To achieve this, the author collated information on socioeconomic-, geographic-, and vehicle-specific attributes at the household (HH) level from the original Household and Vehicle Files of the 2009 National Household Travel Survey. The identification of the social sub-groups is realized via a three-stage least squares (3SLS) model development at the national level of analysis, where the dependent variables in the model specification for each pricing option are the average vehicle fuel efficiency and the vehicle-miles traveled for the fuel tax and the VMT fee option respectively.
The second research objective of this thesis is to identify if the model specification at the national level may be applicable at a more localized level of analysis, for example for Iowa. The analysis of variance (ANOVA) and the asymptotic t-test of the individual coefficients are applied to test for differences in the two levels of analysis.
The results of the first part of the analysis show that particular social sub-groups, such as HHs located in rural areas, or HHs that are located in states with lower fuel taxation, operate vehicles of lower fuel efficiency at the HH level, thus consume more fuel for the same distance traveled, and therefore have higher fuel-related expenditures. Such groups are expected to shoulder greater of the fuel tax burden, as their relation to the average vehicle fuel efficiency at the HH level via the model specification is decreasing. On the other hand, based on the second national model specification, HHs such as those that own vehicles of higher fuel efficiency, or are located in rural areas, or have a higher average income generate more trips at an annual basis, thus have a higher VMT at the HH level.
Regarding the second part of the analysis, the differences between the national and local model lie in three different levels, suggesting that, despite the similarities, the development of a distinct local model is statistically supported.
* There are statistically significant differences in the VMT at the HH level, suggesting that VMT in Iowa exhibit different trend than the national average VMT.
* The two levels of analysis share quite a few common variables in the model specification, but the location-specific variables did not participate in the local model development.
* Based on the asymptotic t-test of equality of individual coefficients, there is a statistically significant difference detected in the coefficient estimates of all variables, suggesting that their effect magnitude on the average vehicle fuel efficiency and VMT at the HH level differs between the national and the local model.