Thursday, January 05, 2006

The return on mobility investments

Bob Poole writes a newsletter I really like for the Reason Foundation called the Surface Transportation Newsletter (how's that for a compelling, clever, sexy newsletter name?). I think this recent article is really on to something big.
Congestion and Regional Economies

Back in issue #22 I noted that the much-cited Texas Transportation Institute annual congestion cost figures tell only part of the story. They estimate the annual cost of being stuck in traffic to individual motorists—in terms of lost time and wasted fuel. Even though that number is large ($63 billion in the most recent year analyzed), it's hardly the total cost of traffic congestion. In that issue I spotlighted a report from the National Cooperative Highway Research Program that did some pioneering work to estimate how much congestion costs businesses in urban areas, in terms of things like increased inventory costs and employer/employee skills mis-matching. (It's NCHRP Report 463, in case you missed that issue.)

Now I'm pleased to tell you about another effort to estimate the cost of congestion to an urban area's economy, this one from Europe. Remy Prud'homme and Chang-Woon Lee's November 1998 paper is "Size, Sprawl, Speed and the Efficiency of Cities." Their basic idea is to look at the effective size of an urban area's labor market—i.e., what fraction of the total job possibilities can a worker access in a reasonable commuting time? Although larger metro areas have more total jobs, the bigger the city (geographically), the harder it is for workers to get to them, especially if the transport system is inadequate. Thus, in a small metro area, the effective size of the labor market is nearly the entire market, but that may decline to as little as half in a large, congested one like Seoul.

In the paper, after a preliminary look at Korean cities, Prud'homme and Lee used data from 22 French cities to calculate their effective labor market sizes, from the standpoint of both employers and employees. Using data on labor productivity from each one, they find a robust relationship between the two—i.e., a larger effective labor market leads to (or at least is correlated with) higher productivity. Their average estimate is an elasticity of 0.18. In English, that means when the effective labor market size increases by 10%, productivity (and hence output) increases by 1.8%.

The effective size of the labor market is negatively affected by sprawl and positively affected by speed of travel, a relationship which they tested with the 22-city data. For the speed variable, they concluded that a 10% increase in average speed leads to a 15-18% increase in labor market size, other things being equal.

In a study on mobility options for Atlanta published last year, Wendell Cox and Alan Pisarski applied the Prud'homme and Lee analysis to look at the economic productivity differences among several alternative transportation scenarios over the 2005-2030 period. The "maintain congestion" scenario (which would invest enough to keep congestion, and hence travel speeds, from getting worse) would lead to a 2.4% increase in gross personal income in the Atlanta area, compared with the “present plan” scenario. The "50% congestion reduction" scenario would lead to a 3.5% increase in gross personal income and the "70% congestion reduction” scenario would produce a 4.5% gain in personal income. Translated into increased per-capita income in 2030, those three scenarios would lead to increases of $2,450, $3,560, and $4,620, respectively. That's not chicken feed.

The Prud'homme & Lee study does not compare the relative costs and benefits of reducing sprawl versus increasing travel speed. That's a worthy subject for further research. And they also do not assess the relative cost-effectiveness of investments in transit versus highways for increasing travel speed. In both of these areas, the results will surely differ from metro area to metro area. But my guess for most of America's low-density metro areas is that reducing sprawl would be very difficult and costly, and that expanding highway capacity will generally be more cost-effective than expanding rail transit. But we should all be open to letting the numbers tell us more about these issues.

I think you can assume that Atlanta scenario is extremely close to Houston's situation: big congestion improvements would lead to big productivity and income gains. Keep in mind that when he talks about increasing the "effective labor market," he doesn't just mean population growth, but population growth along with enough mobility to actually reach the jobs within a reasonable commute, which, of course, is substantially affected by the average speed.

Sprawl critics will jump on the finding that sprawl negatively affects the size of the effective labor market - which is as you would expect putting more distance between residents and jobs. But, of course, those people are making those choices for a reason, whether it's better home value, size, quality, neighborhoods, or schools. People weigh off these choices against their commute, their access to job opportunities, and ultimately their job-fit and productivity. This is one reason I support aggressive tollway expansion and congestion pricing, both to relieve congestion and increase average speeds (and therefore the accessible job market and productivity), but also so the costs of the housing tradeoff decision are more explicit (live closer vs. pay tolls).

I obviously don't have the hard data to back this up, but my subjective judgement based on everything I've seen about different cities is that Houston might be close to a "sweet spot" on the mobility-population tradeoff. Smaller cities have a lower effective population, and therefore lower productivity and lower incomes. Larger cities run into gridlock from inadequate freeway investments (LA) or long transit commutes (NY, Chicago), reducing mobility and therefore realistic job opportunities and productivity. I think other similar-sized, high-growth, sprawling-with-freeways, sunbelt cities might be in roughly the same "sweet spot": DFW, Atlanta, and Phoenix. One missing piece from this analysis is the effect limited mobility has on skyrocketing housing prices (can't add supply within a reasonable commute to the jobs), so even if you have a higher average productivity and income (as Chicago and NY do), you probably end up with less discretionary income after paying for housing and other higher costs of living.

I predict some urban economist will make a big name for himself by deeply figuring out these equations for American cities, based on data analysis and computer simulations, to the point where the cost-benefit analysis for mobility investments (freeways, surface street improvements, or transit) will be much more clear-cut - maybe to the point where taxes can be raised to fund desperately needed improvements because the return-on-investment will be so obviously compelling (kind of like a well-designed TIRZ). "We're going to raise your taxes a dollar to make mobility investments, but it will increase your income two dollars". Of course, self-funding toll roads can avoid the tax-raising problem, but surface street, freeway, and transit improvements can't.

Any ambitious economists in the audience?


At 1:17 PM, January 07, 2006, Anonymous Anonymous said...

I do occasionally teach economics at a community college, ambitious no.

I strongly agree with this sentiment - "This is one reason I support aggressive tollway expansion and congestion pricing"

I've been to a few other cities, notably DC, where some highways were designed for commuting specifically and were not lined with commercial/industrial development. Why don't we have more streets like Memorial and Allen Parkway that aren't clogged with shoppers instead of commuters?

At 6:41 PM, January 07, 2006, Blogger Tory Gattis said...

Here's why:

and here's a little more:


Post a Comment

<< Home