Speed Kills?

Freakonomics cites a recent paper on road fatalities:

“According to a recent paper by Lee S. Friedman, Donald Hedeker, and Elihu D. Richter, the lifting of the federal 55 mph speed limit in 1995 was responsible for 12,545 deaths between 1995 and 2005. That’s about 45 percent more American fatalities than we have suffered in 9/11, Iraq and Afghanistan put together.”

The freakonomics post by Eric A. Morris is worth reading , as is the link to Robert Yowell’s more skeptical article here.

Friedman’s paper cited above uses the Fatality Analysis Reporting System (FARS) to gather data on fatal road deaths which they correct for traffic volume and density. They found a 3.2% increase in road fatalities after the National Maximum Speed Law was repealed, allowing states to increase the limit above 55mph (90kph). From this they work out their figure of 12,545 deaths.

But before the US nannies (and the New Zealand nannies) start looking at reducing the speed limit further on roads, there are several things wrong with their conclusions. While it is perfectly true that the faster your speed the more likely you are to have an accident and the worse the accident will be, the evidence is that speed limits, of themselves, produce little reduction in the road toll. When the US introduced the 55mph limit in 1974 in response to the 1973 Oil Crisis, there was an immediate drop in road fatalities. This drop was, unfortunately, quite transient. It was also unsurprising that reducing high speed crashes in 1974 would have a serious impact on fatalities. Few seat belts, no roll bars, crumple zones or airbags – it is amazing anyone survived a high speed crash then.

Fast forward to today and even the cheapest, low-end car is an order of magnitude safer than the best 1974 car. For a high-speed accident to be fatal today you have to be either driving a very old car, ignoring the seatbelt alarm or driving very fast indeed. Head on collisions would be the other major cause of fatalities but these would be as fatal at 90kph as they would at 120kph (the difference between a 180kph and a 240kph collision is immaterial). I have long held the opinion that New Zealand’s poor road statistics has far more to do with our relatively ancient fleet of cars than it has to do with the speed limit (and don’t get me started on the pitiful state of our highways).

Modern cars do have one feature that makes fatal accidents more likely. They are fast. A decade of improving engines means that, again, even a low-end car can travel far in excess of any speed limit. There have been three fatal accidents on the road in front of my house in the past 5 years. None of the cars were doing anywhere near the posted speed limit of 80kph. My lowly Hyundai Sonata has a clock that goes to 220. While I suspect it might not quite hit the end of the speedo, I have no doubt that I could take the car close to the top. I would then be driving at double to speed limit (note to cops: I have never, ever tried this!). It seems to be that this ability to greatly exceed the speed limits is more likely to be the cause of the small increase in fatal crashes. The car might feel stable at those speeds but one bump in the road and you are flying.

Friedman’s paper does contain one interesting statistic. The urban interstates that have retained their original 55mph limit have by far the greatest increase in fatal accidents (12.88% – not quote in the abstract), while the urban interstates with slightly increased speed experienced a slight drop. The authors attribute this to speed adaptation and spillover effects that occur when drivers, coming off high-speed roads, continue to drive faster than those already on the same road. Although this is a well-observed phenomenon, there is no reason that this should cause an increase in fatalities over and above the increase on non-urban interstates. Admittedly the traffic density would be greater but this has already been controlled for in the statistics. I suspect what is happening is that drivers coming off the faster roads are becoming more impatient and are taking more chances. This would explain the improvement at slightly higher speeds.

It is easy to draw simplistic conclusions about figures such as speed and fatalities. But the problem is considerably more complex than more speed = more fatalities. Politicians do not like the complex, which probably explains why they are so ready to listen to superficial explanations such as this.