What does traffic really cost?

The news media is reporting a new study by AAA on the costs of auto crashes and on traffic congestion. The report is available here (1MB PDF).

Short answer: we worry too much about congestion, when crashes are the more expensive problem. To the left are two graphs based on the report's data. The report also contains a number of nice graphs.

The study compares the economic costs of crashes vs. the costs of congestion. Costs include delays, lost work, and so on. See the report for a complete discussion of the methodology, which looks to me to be thorough and balanced. This study isn't performing a trivial analysis, but I think the work will support some important conclusions:
  • The cost of traffic is dominated by crash costs. Except for a very few places that manage snarled but not terribly deadly traffic, lowering crash costs is a much more promising strategy than striving for lower congestion costs. This is perhaps most striking in Seattle, my home town. Congestion dominates the traffic discussion, even though congestion is not particularly bad. Conversely, Seattle's streets are among the most dangerous, pushing overall traffic costs near the top in the US. If Seattle had crash costs as low as San Francisco, our overall traffic costs would be the lowest of any major US city.
  • The smaller the place you live, the more traffic generally costs you. Not because of congestion of course, but because of crashes. Partly because of greater travel distances, the further you get from the city, the more crashes there are. (When looking at major cities, the study includes nearby suburbs. )
  • It's impossible to build enough roads for everybody to go at full-speed wherever they want whenever they want, but even if we could, it wouldn't have that much effect on the cost of traffic, because as vehicle miles rise, so do accidents and the costs associated with them.
  • These are just economic costs, not the human costs. The human cost of waiting in traffic isn't much. The human cost of being in a car crash is often extreme. Nor are the environmental costs considered. These are real economic opportunity costs that we pay every year.
  • Some places are doing much better than others. Miami and Phoenix are just categorically worse than the rest of the country for traffic. Conversely, places as diverse as New York, San Francisco, and Eugene are doing much better than everyone else. It's interesting that the places with low traffic costs are also places that are known as pedestrian- and bike-friendly places. This suggests that effective pedestrian and bike safety facilities are actually profitable: they can save more money in crashes than they cost to install and maintain.
Update: see Erica Barnett's Slog posting here on the AAA study.


Esoteric knowledge of the Blogosphere

"Actually, before you post to the internet, get someone who passed middle school English class to proof read your witting please. Its riddle with errors and makes my eyes bleed to read it."

Actual comment from here.

Sometimes the truth is right there in front of us. Sometimes its riddle with errors, but its there.


Joe Weizenbaum

Joseph Weizenbaum 1923 - 2008.

Joe Weizenbaum, pioneering and uncompromising computer scientist, passed away this last week. Weizenbaum wrote the computer program ELIZA, and was a relentless foe of oversimplifications of cognition and a defender of what it really means to think. Weizenbaum was a scientist and a thinker in the truest senses of the words.

I didn't really know Joe well, but we were acquaintances. One time, Joe told me the most wonderful story about his daughter. This is when Joe was at MIT and his daughter was about twelve. A colleague of his, an MIT mathematician, was over for dinner. (I forget who.)

Joe's daughter reached for last piece of bread. Joe said, "don’t take that." She asked why. Joe said, "because you don't take the last piece of bread." Joe's daughter sat silent for a minute.

"Then nobody can have any bread."

Silence around the table.

Finally, the mathematician couldn't restrain his curiosity as to how a twelve-year old would advance an induction proof. "Why?" he asked.

"Well," she said, "if you can't eat the last piece of bread, it isn't really the last piece...."

Joe's delight in this story was not just pride in his daughter, but the window it opened into one of his favorite subjects: the nature of human cognition. Joe will be missed. We'll miss the person, his utter lack of tolerance for convenient and comforting yet wrong explanations, and his unwillingness to keep his mouth politely shut in the face of bullshit.

More on Joe here and here. Terry Winograd's tribute to Joe on his winning the Norbert Wiener award is here.