Do mileage-based congestion fees hit low-income drivers harder?

By Chet Edelman
While there is mounting evidence that demand-based pricing—or congestion tolling—can more efficiently manage highway use, serious concerns continue to arise regarding the system’s disproportionate impacts on low-income drivers. However, a recent study by researchers at Purdue University has found that a less onerous tax alternative may exist—one that combines congestion tolling with mileage-based user fees or a VMT tax.
Known simply as a congestion tax, the policy consists of charging individual vehicles a per-mile fee only during peak congestion hours. Although several pilot VMT tax programs exist in states such as Oregon, California, and Minnesota, there is currently no congestion tax in widespread practice, making the results of the study especially useful when constructing policy.
Using data collected from the Oregon Household Activity Center, the researchers analyzed socioeconomic demographic data and weekday travel records from 14,000 households in Oregon. The goal was to determine the driving patterns of different socioeconomic groups during peak congestion hours and assess the equity impacts of a hypothetical 1.5 cent per-mile tax. Additionally, the researchers looked at the spatial distribution of the 14,000 households by classifying them along a five tier rural-to-urban scale to identify whether geographic inequities might exist.
After applying the congestion tax to the existing driving data, the researchers found the burden of the tax was almost exactly proportional to income. In other words, the congestion tax resulted in each household paying close to an equal fraction of their earnings. Using a metric known as the Suits Index, a measurement of tax incidence on a scale of 1 to -1, with a value of 1 indicating the burden of a tax falls entirely on high-income earners and a value of -1 indicating the burden falls entirely on low-income earners, a congestion tax was found to score between .0248 and .0369—close to a perfectly proportionate tax value of zero.
Simply put, wealthier households tend to drive more during peak congestion hours and therefore are likely to pay a slightly higher proportion of their income under a congestion tax.  In comparison, the Oregon state gasoline tax had a Suits Index value of -.30—a discrepancy which suggests the congestion tax could act as a more progressive tax alternative to existing transportation revenues.
Where the study did find the tax to be harmful, however, was with regard to rural households. A congestion tax was found to have a value between -.110 and -.088, meaning rural households are more likely to shoulder the burden of the tax. Considering rural households typically must commute farther distances than those in urban areas, these results do not come as a surprise. The authors suggest some of the revenues from the tax could be applied to rural transportation projects to help offset this inequity.
Since a congestion tax has yet to be implemented in the field, it is difficult to discern how effective it would be in altering vehicle usage. Furthermore, privacy concerns exist regarding a system in which vehicle mileage and usage are closely tracked. Even in lieu of these concerns, the less harmful nature of a congestion tax would negate much of the need to offset inequity. The authors note that administratively-challenging tasks such as redistributing tax revenue in some form to lower-income households could potentially be avoided altogether.
Although further inquiry is likely needed before a congestion tax can realistically be implemented, the results of this study suggest it poses a promising alternative to either congestion tolling or a flat VMT tax going forward.
Chet Edelman is a Project Assistant at SSTI.