We’ve just gotten our first look at the new US Department of Transportation performance measurement rule for transportation systems. The rule (nearly three years in gestation, since the passage of the MAP-21 Act) is USDOT’s attempt to establish performance measures to guide investment and operation of the nation’s urban transportation system. One of the criticisms—fair, in our view—of the nation’s transportation system is that there are few, if any, quantitative standards against which performance can be measured, and against which the merits and results of alternative policies and investments can be judged. The new DOT standards aim to do just that. In the next few weeks, we’ll all have an opportunity to weigh in on whether these standards help address this problem.

Standards like these may seem like technocratic trivia. But we’ve routinely witnessed obscure and seemingly innocuous rules of thumb about things like road width or parking requirements end up profoundly shaping our cities—and often in ways we don’t like. Getting these performance measures right can help push transportation investment in a direction that supports more successful cities. Getting them wrong runs the risk of repeating past mistakes. The proposed rules are voluminous, running to more than 400 pages in all—and incredibly detailed. So it will take some time to dig into them. But at first glance, we have some reactions. We’ll take a closer look in the days ahead, and revise and extend our comments as we (and others) go through the minutiae here.

Two key words: “Excessive” and “Expectations”

Initially, we’re focusing on three standards that the USDOT is proposing apply to metropolitan areas of a million or more population. The USDOT has also proposed separate standards for travel time reliability and freight that we’ll look at in a future post.

Briefly, the standards are outlined in the following table and further analyzed below. They deal with total hours of excessive delay, and the share of the Interstate system or national highway system where travel times don’t exceed 150 percent of a locally established “expectation.”

Performance Measures for Large Metropolitan Areas

Implementing these performance measures will require a heavy reliance on technology and a substantial investment. USDOT believes that it will be able to use speed data collected from vehicle telemetry, including cell phones to determine performance for individual street segments. While the technology is promising—it does have some limitations—an analysis of the performance of HOT lanes was flawed because it couldn’t distinguish between vehicles traveling in the tolled and free lanes. USDOT estimates that complying with the data collection and reporting requirements of this rule will cost states and local governments $165 to 224 million over the next decade.

Congestion: “Excessive hours of delay”

The core measure of whether a metropolitan area is making progress in addressing its congestion problem is what USDOT calls “annual hours of excessive delay per capita.”  This congestion measure essentially sets a baseline of 35 mph for freeways and 15 pmh for other roads.  If cars are measured to be traveling more slowly than these speeds, the additional travel time is counted as delay.  The measure calls for all delay hours to be summed and then divided by the number of persons living in the urbanized portion of a metropolitan area.

The proposed measure is, in some senses, an improvement over other measures (like the Texas Transportation Institute’s Travel Time Index) that compute delay based on free flow traffic speeds (which in many cases exceed the posted speed limit). But despite its more realistic baseline, this measure suffers from a number of problems:

—This is all about vehicle delay, not personal delay. So a bus with 40 or 50 passengers has its vehicle delay weighted the same amount according to this metric as a single occupancy vehicle.

—This ignores the value of shorter trips. As long as you are traveling faster than 15 miles per hour or 35 on freeways, no matter how long your trip is, the system is deemed to be performing well.

Interstate and National Highway System performance: “Peak hour travel times meet expectations”

If “meets expectations” sounds a bit squishy for a federal planning standard, it’s because it is.

Under the rule, state DOTs or Metropolitan Planning Organizations (MPOs) would establish “expectations” for how long (or how fast) trips would take on each segment of a metropolitan area’s major freeways and highways. Segments which experienced peak hour travel times that were 50 percent more than these “expected” travel times would be deemed to be congested. The metric would track the share of a region’s highway segments that didn’t experience this level of congestion-related delay.

The pivotal policy question here is what are “expected” travel times. Here the USDOT simply punts: It’s up to state DOTs or metropolitan planning organizations to make this call. As the DOT says: “Under this proposed approach, FHWA does not plan to approve or judge the Desired Peak Period Travel time levels or the policies that will lead to the establishment of these levels.”

In effect, this means that performance metrics are likely to vary widely from place to place. These performance measures beg the essential question of what constitutes a reasonable expectation of travel times. As we’ve pointed out, it’s a regular occurrence in daily life that Americans have come to tolerate very different levels of delay for the same service at different times, for example, when they order their morning coffee, as we documented in the Cappuccino Congestion Index. In our first reading, it’s also not clear whether states and MPOs can adjust expectations over time. It’s an interesting question: Should we adjust our expectations as conditions change, or once established, should “expected” travel times be an unchanging baseline against which performance is measured?

The “expectations” terminology begs a larger question as well: congestion reduction measures are seldom free. It costs money to expand capacity, improve transit, or implement other measures that might reduce travel times at the peak hour. The big question is whether the value commuters attach to such potential travel time savings come anywhere close to being commensurate with the cost of achieving expectations. Because USDOT offers no guidance or guidelines as to what might constitute reasonable expectations for travel times, and because they’re unmoored from any standard of cost effectiveness, this performance standard is likely to be of limited usefulness.

In addition, the “peak hour travel times meet expectations” measure is, of course, a variant of the classic travel time index that we—and others—have long critiqued. One of its chief problems is the denominator. In this case, the denominator is the size of a region’s highway system (in DOT parlance “segment length”.) The indicator is the percent of Interstate (or NHS) roadways that aren’t congested. So—at least in theory—if a region expands its “segment length” by building new, under-utilized highway capacity, it can improve the ratio of uncongested to congested roads—thus improving its performance. Conversely, a metropolitan area that doesn’t increase the segment length of its Interstate (or NHS system) in the face of increasing travel seems likely to see a decline in its rated performance. As a result, this measure seems to impart a strong “build, baby, build” bias to the indicators.

DOT whiffed on greenhouse gases

Despite some hopes that the White House and environmentalists had prevailed on the USDOT to tackle transportation’s contribution to climate change as part of these performance measures, there’s nothing with any teeth here. Instead—in a 425 page proposed rule—there are just six pages (p. 101-106) addressing greenhouse gas emissions that read like a bad book report and a “dog-ate-my-homework” excuse for doing nothing now. Instead, DOT offers up a broad set of questions asking others for advice on how they might do something, in some future rulemaking, to address climate change.

Three ideas for what DOT might have done

1. Make VMT per capita a core measure. Vehicle Miles Traveled (VMT) per capita is strongly correlated with important transportation system outcomes. It’s correlated with total system costs, costs to households, greenhouse gas emissions, crashes, injuries and fatalities.
2. Shift from excess travel time to total travel time. A total travel time measure, which recognizes the value of shorter trips, even when they occur at somewhat lower speeds better recognizes the economic and environmental value of more compact development patterns. Implement a “total travel time” measure that computes total travel time per resident, and gives equal weight to measures that reduce the distance of trips and the need for travel, especially at the peak hour, when it will have the greatest effects on congestion.
3. Establish a separate methodology for transit delays. How much additional time do transit riders incur from transit systems that don’t have average running speeds of some reference number (like DOTs 35 MPH for freeways and 15 MPH) for roads, or locally established expectations. The amount of this delay could easily be calculated from transit system operating records and ridership counts.

Our objective in writing about these standards is to encourage others to take a close look, and help provide a robust discussion of this important policy. We invite your comments and corrections—and we’ll update and add to this post as we learn more about the rule. Stay tuned.