The Washington PostDemocracy Dies in Darkness

One curious effect of the recession may bolster the case for driverless cars

A self-driving Ford Fusion hybrid car is test driven Aug. 18 in Pittsburgh. (Jared Wickerham/AP)

Two researchers say they've discovered a curious relationship between deadly car crashes and the jobless rate — and the finding may have important implications for next-generation vehicle technologies like self-driving cars.

Fewer people die in car crashes, it turns out, when the economy is hurting. That's the report from Clifford Winston, a senior fellow at the Brookings Institution, and Vikram Maheshri, an assistant professor of economics at the University of Houston.

Writing in the Journal of Risk and Uncertainty, Winston and Maheshri claim that every 1 percent increase in unemployment is linked to a decrease in traffic fatalities by about 5,000 annual deaths. This seems like a momentous statistic by itself. But even more important is the reason behind it: It's because America's least-safe drivers stay off the roads during economic downturns, according to the researchers.

How do they know that? We'll get to the specifics of the study in a second. But the connection to driverless cars is clear: If partially or fully automated vehicles can help risky drivers drive better, we may be able to achieve the same reductions in traffic fatalities the researchers observed during the recession without, you know, actually having to go through a recession.

Winston and Maheshri's study probably wouldn't have been possible without some advanced technologies that are just becoming mainstream. They relied on data from State Farm, the insurance company, some of whose customers had opted in to being tracked (probably in hopes of receiving a discount on their premiums). This individual-level data give us valuable demographic and behavioral insights that we don't get from a national-level accounting of vehicle miles traveled, a fuzzy and imprecise metric.

Understanding which drivers are over 60 years old, own older cars that may not be as safe, have recently experienced a crash or who don't live with a family — all risk factors — allowed the researchers to compare those groups' driving behavior against statistically safer demographics. And what they found was that the riskier drivers drove less, as measured by their own number of vehicle miles traveled.

The State Farm data may be skewed by the fact that it comes from drivers who opted into the company's tracking program, and these people may already be safer than the norm to begin with. And it's also drawn from a single state, Ohio, which may not be nationally representative. The study also can't explain why the recession may have these effects: It's possible that middle-aged people respond to the recession by needing to work more, and therefore need to drive more to perform said work. Or that drivers who are statistically more likely to crash then drive less because they perceive a heightened state of risk during unstable situations (like recessions).

Still, the researchers say, the findings at least highlight how different groups might benefit from driver-assist technologies, and provide a relative sense of how massively driverless car technology could cut down on traffic deaths. Vehicle automation could effectively take risky drivers "off the road," substituting for the unemployment effects Winston and Maheshri looked at in their study. The technology could also enable those drivers to keep participating in economic lives they would have otherwise avoided. But that's a subject for another time.