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.