Let’s back up. Every year, the Texas Transportation Institute releases a nationwide study of urban mobility in U.S. cities. Among the things they model is traffic congestion: the number of hours commuters spend stuck in traffic; the percentage of extra time it takes to make a trip during rush hour, as opposed to during other times of day; and other similar measures.
But the TTI methodology has changed—in some cases radically—over the years. In particular, they recently started to account for congestion-reduction strategies, such as metered on-ramps and rapid responses to highway accidents. Accidents can cause major traffic slowdowns, and clearing them quickly can significantly reduce traffic tie-ups.
The graph to the right—taken from the Seattle Times website—shows that Seattle-area congestion plummeted between 1999 and 2000. All of that decrease can be attributed to the change in methodology.
And this brings me to another point—for years, the TTI ranked Seattle as one of the most congested cities in the country, usually ranking in the top 5 on most measures of congestion. But I lived in Washington, DC nearly a decade ago, and I’ve spent a bunch of time in Boston. From my experience, Seattle’s congestion today doesn’t hold a candle to either of those cities. And, indeed, using the new methodology greater Seattle’s congestion is merely average—and certainly not one of the most congested cities in the country.
The TTI data certainly has its uses. But it’s simply wrong to interpret the data as suggesting that greater Seattle’s congestion has suddenly dropped from the top 5 in the nation to about the national average.
Changing the methodology used to study congestion does not change congestion itself. The reporters who wrote these stories should understand this.
UPDATE: OK, it looks as if Seattle Times already covered this story a year ago, when the Texas Transportation Institute first changed its methodology. But last year’s story interpreted the numbers (correctly) as a change in methodology:
The Seattle-Everett metropolitan area’s congestion ranked 12th among 75 cities. The region tied for fifth in last year’s edition of the report, easily the best-known, most-publicized continuing congestion study in the nation.
At least part of Seattle’s drop stems from changes in the way the institute measures traffic. This year, for the first time, researchers included in their calculations the effects of public transit, HOV lanes and such operational improvements as freeway-ramp meters.
That article gets it right—when the methodology changed, the amount of congestion estimated by the transportation model fell. That didn’t mean that congestion itself fell.
Despite its flaws, I still think the TTI study is valuable. Looking over the data a few years ago, I learned some important (though obvious) things from it:
- Big cities are more congested than smaller ones—which suggests that as cities grow, they get more congested
- A city’s overall population density doesn’t have much to do with the amount of time its residents spend stuck in traffic.
- In general, building more highways only slows congestion growth little, if at all. Cities that have been on road-building sprees (think Atlanta) still have bad traffic.
- Cleveland has done really well at minimizing congestion.
At one point I tried to figure out what Cleveland was doing so well, and here’s what I came up with: it held down its population growth (in fact, the city and inner suburbs shrank); it emptied out its urban core; and it sprinkled new development in the far-off exurbs. That’s one recipe for no congestion, but it’s not one I care to see replicated widely.