Thursday, December 12, 2013
By NATHAN TEFFT and CHAD COTTI
LEWISTON — According to the National Highway Traffic Safety Administration, from 2000 through 2009 there were more than 370,000 fatal U.S. automobile accidents, resulting in more than 410,000 deaths.
ABOUT THE AUTHORS
Nathan Tefft is assistant professor of economics at Bates College in Lewiston. Chad Cotti is assistant professor of economics at the University of Wisconsin, Oshkosh.
Yet, while the average number of fatal crashes had been relatively stable from 2000 to 2007 at about 38,000 per year, that number plummeted in 2008 and 2009 to about 32,400 per year, a historically huge decline.
Many news outlets have recently noted this remarkable change in the fatal crash rate.
But in our view, the news media brushed by the most important influence by far. To repeat a phrase famously used during Bill Clinton's 1992 presidential campaign, albeit with a different purpose: It's the economy, stupid.
Several studies have examined the relationship between economic indicators, such as the unemployment rate and personal income, and fatal accidents. An important series of papers during the 1990s and early 2000s found that there were fewer accidents during economic downturns.
In addition, less alcohol was consumed when the unemployment rate increased. This accumulation of research points to a consistent relationship between recessions and fatal automobile accidents.
In our own work, we have been able to expand on these earlier studies to explore the relationship between economic determinants and accidents using previously unavailable data.
These data not only allow us to specifically identify which accidents involved alcohol consumption, but we are also able to consider the most direct factor in the numbers of fatal accidents: the total number of miles driven on U.S. roads.
Before looking at the data, we expected that changes in the unemployment rate and personal income would be associated with changes in fatal accidents in two ways.
First, the quantity of miles driven would decline, because when individuals lose their job or work fewer hours they commute less and have less disposable income available to make purchases or travel for entertainment purposes.
Second, the composition of behaviors associated with driving would change with a fall in disposable income. For example, even if an individual drives to his local restaurant or bar for dinner or drinks, he may hold off on that last drink to save a little extra money.
Sure enough, our analysis revealed that from 2003 to 2009, states with higher unemployment rates and lower per- capita personal income saw fewer automobile accidents per capita. These two economic indicators also were associated with fewer total miles driven, which was in turn strongly associated with fewer accidents.
This led us to conclude that the recession perhaps unexpectedly led to the most straightforward way to reduce fatal accidents: less driving means fewer crashes.
In order to study the composition of driving behaviors, we looked at the relationship between the economic indicators and the accident rate between states in which miles driven were the same. We found that the unemployment rate was still associated with fewer accidents, but personal income no longer appeared to be related.
We then separately considered alcohol- and non-alcohol-related accidents and found that the unemployment rate relationship was present only for accidents involving alcohol.
Taken together, these results are suggestive that a rising unemployment rate leads to changes in behaviors associated with driving, such as consuming less alcohol when out.
What is most remarkable is that our findings are strong even though we accounted for unobserved characteristics of each state that remained unchanged from 2003 to 2009, as well as unobserved national- level changes in each year.
For example, our analysis nets out annual changes in gasoline prices and automobile safety technology that was consistent across the country.
And even if upgrades in automobile safety lagged in states with higher unemployment because there were fewer new car purchases, these differences would show up as a higher fatal accident rate, not lower.
Overall, our results suggest that the single most prominent reason for the recent dramatic decline in fatal crashes was simply the economy. This is a mixed blessing, however.
So, the unemployment rate, rather than personal income, was the more important factor in the decline in automobile accidents, indicating that a higher unemployment rate may have caused a drop in fatal accidents. However, this trend is likely to reverse as the economy improves.
At least we have the consolation that some lives were saved that would have otherwise been lost had the recession not occurred.
- Special to the Press Herald