On highways, the 85th-percentile rule makes some sense. The evidence seems to suggest that raising speed limits slightly increases fatality rates, but not accid

On highways, the 85th-percentile rule makes some sense. The evidence seems to suggest that raising speed limits slightly increases fatality rates, but not accident rates, and doesn't have much influence on driver speed. At the other extreme, on neighborhood streets in residential areas, the rationale for the 85th percentile rule (reducing the speed differential between cars traveling in the same direction) is absurd: these speed differentials are much lower and much less important than other speed differentials. If a car driving down my street at 30mph has trouble dealing with a car going the same direction at 20mph, it will also have trouble with the kids playing basketball in the street, or the car headed the opposite direction without room to pass, or the car backing out of a driveway, or the bike going 15mph. Conflicts between drivers headed the same direction are a non-issue on these types of streets. And there are streets in the middle. A major commercial street in my town is four very narrow lanes (the right lane is actually less wide than a city bus, so when a bus is traveling the same direction you must enter the oncoming lane if you try to pass), with very small blocks, crosswalks at every intersection, tons of pedestrians and cyclists, tons of cars turning left across traffic, etc. Should this be set at the 85th percentile speed? Although it's more debatable, again, I don't think that the speed differential between cars traveling the same direction is likely to be a major contributor to fatal accidents here--on a busy urban street like this the danger is much more likely to arise from the speed differential between pedestrians and cars, a car turning left and an oncoming car, cyclists and cars, etc.