Friday, October 14, 2011

Why salt is important to planners

The Texas Transportation Institute, out of Texas A&M, released its annual Urban Mobility Report and as usual, large urban metros are clogged with auto congestion. I am not going to get into the details, but know that methodologies used have been under increased scrutiny.

The top 10 cities come as no surprise: 1) Washington D.C., 2) Chicago, 3) Los Angeles, 4) Houston, 5) New York, 6) Baltimore, 7) San Francisco, 8) Denver, 9) Boston, 10) Minneapolis & 10) Dallas. Perhaps the bigger surprise is the preferences of the authors and its intended readers.

To make it simplistic, the authors assign an arbitrary amount ($16 per hour) to each drivers time lost in congestion. If a freeway is not level-of-service A, which is very hard to attain, then there is time lost to congestion.

The report makes the case that time lost to congestion is important, but time lost to commuting is not. For example, if a commuter (already a biased term) lives close to work, but spends all 10 minutes of the commute in congestion, they lose more time than the commuter who lives 40 minutes from work, but five of that is congested. So the latter commuter, which spends more time a day/week/year in traffic doesn't count as much in the congestion cost as the former, even though the former spends only a quarter of that time in the car overall.

Also, transit times are not considered at all into the equation. The certainty of knowing the train will arrive here at X time and get me there at Y time 95% + on schedule means nothing to the report. It does through a bone and says transit saves X amount of time in congestion by multiplying the riders of the system by the $16 an hour. I won't get into the flaws of this line of thinking, as I would hope they are somewhat obvious.

This report is often used as a reason for supporting more freeway lane construction. And again, this bias is skewed toward more freeways. If the latter commuter is ideal in the minds of the authors, then surely their recommendations will mirror that. On top of that, the recommendations ignore the induced traffic principle. I will devote a post to that sometime soon in the future, but just know, simplistically, it means travel is not a supply and demand function, but behavioral and people will travel based on capacity.

A lot of planners are coming around on the idea that more congestion is actually good. More people choose transit when the uncertainty of auto travel reaches a certain point. More people choose to live closer when travel times and costs are high. Others choose different routes, etc. Point is, most planners know more freeways means more low-density suburban development. I suppose the TTI does to.

That is why I say these things must be taken with a grain of salt, even if they support your point. In graduate school, I was working on a research paper for a statistical analysis class. I was working with numbers that were timed based. I ran the numbers at first by the year and didn't get what I wanted. So for fun I ran it by months and still didn't get what I wanted. Finally I tried the numbers by quarter years and it showed what I had hoped it would at the beginning. Needless to say, I switched to the quarter years.

This is why I do not put a lot of stock in non-peer-reviewed pieces, like the TTI's UMR. That saying that figures don't lie but liars figure can always be applied to public policy pieces.

No comments: