The Great Recession and Population Health: The Case of Work-Related Traffic Fatalities

Research has shown that one of the largest effects of economic expansions and downturns is on motor vehicle fatalities. Traffic safety tends to display a pro-cyclical pattern such that economic downturns significantly reduce traffic fatalities (Ruhm, 2000; Cotti and Tefft, 2011). Ruhm (2000) reports that the mortality elasticity is over five times larger for traffic fatalities than for overall mortality. Cotti and Tefft (2011) find that fatal crashes involving alcohol are more responsive to business cycles than are overall crashes. In this paper, we examine the effects of business cycles on traffic fatalities in the US with a special focus on work-related crashes. Specifically, we consider the macroeconomic fluctuations around the Great Recession in the late 2000s.

For our empirical analysis, we compile quarterly data for all 50 states from the 2004-2012 Fatality Analysis Reporting System (FARS). We identify work-related traffic fatalities in two alternative ways. FARS data includes, among other driver characteristics, whether an individual was at work or not. Using this classification, we construct the number of traffic fatalities in crashes that include at least one person who was at work at the time of the crash. Due to potential measurement issues, we also consider whether the crash involved at least one vehicle that is registered as business, company, or government vehicle. We estimate a series of conditional fixed-effects negative binomial and conditional fixed-effects Poisson models that include state, year, and quarterly fixed effects. Our models also account for other time-varying factors such as education, proportion of young drivers, proportion of light trucks, vehicle miles traveled per licensed driver, and a variety of policy measures including speed limits, seat belt laws, and handheld and texting bans.

We find that a one percentage-point increase in the state unemployment rate during this period decreases the number of traffic fatalities by 3.8-4.0 percent. However, work-related fatalities are particularly responsive to economic downturns as a one percentage-point increase in the unemployment rate leads to a 5.1-6.0 percent decrease in work-related fatalities compared to 3.7-3.9 percent decrease for non-work-related fatalities. These findings are rather robust across various estimation techniques and also to using alternative definitions of work-related fatalities. Our results provide insights into the potential mechanisms underlying the pro-cyclicality of motor vehicle fatalities. Specifically, the increase in traffic deaths during prosperous times is not only due to an increase in risky behaviors such as DUI, but also directly related to the increased economic activity possibly through higher traffic volumes, increased freight movement, higher proportion of larger vehicles, and more stressful driving conditions.