NEITHER MOLECULES, NOR BUGS. Many years ago, a colleague wrote a paper about the distribution of traffic speeds in which he suggested that drivers sorted themselves into the fast and slow lanes, rather than reacting in ways that could be simulated with fluid dynamics. Now comes Tom Vanderbilt's Traffic: Why We Drive the Way We Do (and What It Says about us.) The author also operates a traffic weblog that was the recipient of an Instalanche in mid-August.
There's a lot in the book that's routine at Cold Spring Shops, meaning a short Book Review N^o. 39 tonight. I'll just take it idea by idea. Mr Vanderbilt starts with "Why Does the Other Lane Always Seem Faster?" The answer is, when all lanes are equally congested, it doesn't. Queuing, however, is a behavior for which the social norms, or governing institutions if you insist, aren't well defined. Thus McDonald's can get away with its chaotic system in which you might be served next (or not) while banks (notorious for sending dial-up customers through voice-mail obstacle courses) and the post office (enough said) use a single queue for the next available agent. On the road, the norm is even less well-defined, leading to the single greatest source of anger on the roads, the lane closure. Most people immediately get into the lane that will remain open, which slows it down, and a few optimists sail ahead in the lane that will close, in order to barge just before the barricade. That behavior exacerbates the congestion at the choke point, and it breeds enough ill-will that truckers and some operators of other large vehicles will position themselves in such a way as to block the lane that will close behind them.

Mr Vanderbilt's analysis of the situation is straightforward enough (he suggests staying in the lane that will close and participating in a zipper merge at the barricade). He neglects an economic incentive that might help, namely the Principle of Indifference at the Margin. (The bargers expect more than enough time saving from going to the choke point to compensate from any antagonism when it comes to barging. If both lanes are equally unattractive, there's no point in barging. His suggestion, if followed by all drivers, would take away that incentive. I belabor this because he's going to be led astray later for not understanding the principle.)

He then explains "Why You're Not as Good a Driver as You Think You Are" (the task is generally not challenging, which leads people into difficult situations) and "How Our Eyes and Minds Betray Us on the Road" (our reactions being selected over millennia for walking speeds). Then come the bugs. "Why Ants Don't Get into Traffic Jams ..." It's complicated. In a swarm of locusts, the bugs toward the rear might be nibbling on each other because the front of the swarm has eaten all the food. The problems, again, are lack of social norms, here reinforced by lack of methods for drivers to communicate information-rich messages to other drivers. It's easy to signal irritation (the analog to biting another locust in the leg) but not to signal other information about road conditions. Next comes "Why Women Cause More Congestion Than Men." It's prosperity, overscheduled kids, and subdivisions without sidewalks, not the second income or the second car alone. Next, "Why More Roads Lead to More Traffic." Obvious. But here's where Mr Vanderbilt could use a refresher on the Indifference Principle, He tells a parable of a slow, narrow street that guarantees you a 1 hour trip, and a (wider? he doesn't specify) road that, when not crowded, gets you there in half an hour. He proposes that a traffic cop direct traffic onto the narrow road or onto the wide road in such a way that some people will get the faster trip. That sounds like a good way to breed contempt for the law, a funny thing to be advocating given a later correlation he draws between corruption and bad traffic, as well as a logically silly idea. In equilibrium, travel times will be equal, or the narrow road might get you there faster but with more traffic lights or wear and tear on your car. He also misinterpreted an interesting peak-shifting phenomenon in which a higher toll for a high-occupancy toll lane led more drivers to use it, because they inferred a severely congested substitute. There's definitely room for more research on high-occupancy tolling, a policy that emerged without any formal economic models to help the road commissioners implement it. Perhaps that's for the better.

A chapter with the provocative title "When Dangerous Roads Are Safer" surveys some of the traffic-taming methods we've reviewed previously. One of these days, the road commissioners will discover that simpler is better, and perhaps they'll be more inclined to synchronize traffic lights when there are fewer of them.

(Cross-posted to 50 Book Challenge.)