New travel demand projections are due from U.S. DOT. Will they be accurate this time?

By Eric Sundquist
When the U.S. DOT’s most recent Conditions and Performance Report to Congress hit the streets in 2012,[1] it forecast that national vehicle-miles traveled would reach 3.3 trillion that year. A few months later we learned that their estimate was almost 11 percent too high.
That overestimate is the equivalent of adding travel from five average-sized states to the total. And the overestimate came in the year of the release, not year 20. This is troubling in a report that is widely regarded as a gauge of the “need” for funding new highway capacity.
Such a gross error made us wonder how good C&P estimates have been historically. With help from Tony Dutzik of the Frontier Group, who has mined VMT statistics for reports issued by the advocacy group U.S. PIRG, we reviewed the 20-year projections in the six C&P reports going back to 1999 and compared them to actual reported VMT.
Enough time has passed by now that 61 yearly projections[2] can be compared to the reported VMT. And in 61 cases out of 61, the C&P estimates were too high. For example, the 1999 C&P overshot 2012 reported VMT by more that 22 percent—almost 11 extra states’ worth of driving.
In fact, though the national VMT trend line began flattening in the 1990s and actually turned down in the 2000s, the slope of the C&P projections has remained nearly constant. (See Figure 1.)

Figure 1. U.S. VMT (in trillions) as tracked by FHWA's Travel Volume Trends (“Actual") and as projected by U.S. DOT's C&P reports (by year reports are dated).
Figure 1. U.S. VMT (in trillions) as tracked by FHWA’s Travel Volume Trends (“Actual”) and as projected by U.S. DOT’s C&P reports (by year reports are dated).

To be fair, while U.S. DOT publishes the C&P and its demand projections, the projections come from state and local traffic estimates done as part of the Highway Performance Monitoring System. The HPMS manual explains how future traffic estimates are obtained:

Future AADT [average annual daily traffic] should come from a technically supportable State procedure, Metropolitan Planning Organizations (MPOs) or other local sources. HPMS forecasts for urbanized areas should be consistent with those developed by the MPO at the functional system and urbanized area level.
This data may be available from travel demand models, State and local planning activities, socioeconomic forecasts, trends in motor vehicle and motor fuel data, projections of existing travel trends, and other types of statistical analyses.

In other words, the projections can be as simple as assuming some level of growth from a base number. So the assumptions are critical. If HPMS growth estimates that show up in recent C&Ps had assumed modern VMT growth trends would continue, they would have come much closer to predicting reality. But the rolled-up trend estimates show essentially the same slope year after year, indicating that agencies providing HPMS data generally have not updated their models and assumptions to account for current conditions, as if they expect the year to be 1980 forever.
And not only are these data being aggregated into the national report, but they also are being used in project selection and development around the country. High estimates of VMT have several negative implications, including 1) they imply a level of “needed” spending that is politically unachievable, 2) they can spur overbuilding on projects, draining resources from critical preservation and multimodal investments, and conversely 3) they can discourage construction of lower-cost, lower-throughput streets that improve livability and property values.
A few blocks from U.S. DOT, the Energy Department is noticing the new trend lines. Its just-released Annual Energy Outlook predicts 0.9 percent annual increases in light-vehicle VMT, a 50 percent reduction from the forecast growth in the 2006 report. Over time the difference becomes huge. The new 2030 estimate is 24 percent lower than the earlier forecast. The growth in truck VMT is also down, albeit less dramatically.
All six of the U.S. DOT’s most recent C&Ps have dramatically overestimated VMT. The 2012 edition is due soon. Will this be the year that shows transportation-planning assumptions are adapting to reality? Or will this new report again overstate the need for highway capacity?

[1] The most recent C&P report, released in 2012, was dated 2010 and based on data from 2008. In Figure 1 the trend line from the 2010 report begins in 2008. Other trend lines reflect similar patterns.
[2] The 1999 report provides projections for 16 years for which we now have actual data (1997-2012); the 2002 report provides projections for 13 years (2000-2012), and so on.
Eric Sundquist is Managing Director of SSTI.