U.S. DOT: States overestimating VMT growth

By Chris McCahill

The U.S. DOT recently released its 2015 Conditions and Performance Report to Congress, describing the current state and future needs of the country’s roads, bridges, and other transportation infrastructure. The report hinges largely on estimates and forecasts of vehicle miles traveled (VMT) provided by the states through the Highway Performance Monitoring System (HPMS). There are issues with these raw numbers, however, so U.S. DOT has taken steps to correct them. As a result, the 20-year growth rate of 1.41 percent per year, resulting directly from HPMS, is adjusted down to 1.04 percent—the lowest forecasted growth rate going back fifty years.

In this newest C&P report, U.S. DOT recognizes that its past forecasts were too high, adding, “states have tended to underpredict future VMT during periods when actual VMT was growing rapidly and to overpredict at times when actual VMT growth was slowing or declining.” As the following figure from the report shows, the recent over-predictions were drastic.

State-provided long-term VMT forecasts compared with actual VMT (source: U.S. DOT)

State-provided long-term VMT forecasts compared with actual VMT (source: U.S. DOT)

One glaring reason the recent forecasts were so far off was the sudden drop in per capita VMT beginning in 2005. As the C&P report explains, “many states have been slow to adjust their models to incorporate emerging socioeconomic trends, as they wait to determine whether new data observations represent one-time phenomena or the start of new long-term trends.” But VMT stayed low for nearly a decade.

We saw a 2.3 percent increase in 2015, originally reported as 3.4 percent, as noted by the Frontier Group. Even with the strong economy and low gas prices, however, that growth seems to be happening more slowly than in the past, which will have important implications for predicting future trends.

Another reason the HPMS-based forecasts tend to be high is that they are typically made at the facility level then aggregated up without accounting for major changes in demographics or economic conditions. U.S. DOT’s national model accounts for factors like demographic characteristics, economic activity, employment, cost of driving, road miles, and transit service availability.

Tony Dutzik of the Frontier Group, who reports regularly on this topic, said via email that U.S. DOT is setting a good example for the states by correcting its forecasts. “As it turns out,” he explains, “state forecasts are still significantly higher than the national forecast, suggesting that public officials and citizens in those states need to approach what state DOTs are telling them about future traffic congestion and infrastructure investment needs with extra caution.”

Chris McCahill is a Senior Associate at SSTI.