The Great Crash: Estimating the Effect of the Financial Crisis and Recession on Automobile Fatalities

In the wake of the 2008 financial crisis and recession, automobile fatalities fell by almost 25%, and the decrease in fatalities spanned all age groups and affected both alcohol and non-alcohol related fatalities. Although health economists have thoroughly evaluated the effect of public policies such as increases in alcohol taxes, minimum legal drinking age (MLDA), seatbelt laws, and texting bans on driving accidents, the reduction driving fatalities in the post 2008 era has not been comprehensively discussed in the literature. In this paper, we examine the extent to which changes in employment rates and other macroeconomic conditions such as the housing foreclosure rate and migration affected automobile fatalities.

We use detailed data on fatal traffic accidents in the United States between 2004 and 2016 from the Fatal Accident Reporting System (FARS). The FARS includes information on the date and location where the crash occurred, characteristics of the automobiles involved in the crash, the age and other demographic characteristics of the driver and other occupants, the blood alcohol content of the driver, and many other characteristics of each fatal traffic accident. We leverage the FARS data to examine many different characteristics of fatal accidents, including the make and model of cars, the days and times of crashes, alcohol and non alcohol-related crashes, and demographics of the drivers to evaluate the heterogeneous effects of macroeconomic conditions on fatal accidents among several sub-groups of population.

We aggregate the FARS data to the county and month level, which allows us to connect changes in automobile fatalities to changes in local macroeconomic conditions. To bolster the empirical connection between macroeconomic conditions and fatal accidents, we exploit plausibly exogenous variation in macroeconomic conditions, following Maheshri and Winston (2016). Specifically, we instrument for changes in a county’s macroeconomic conditions using changes in surrounding counties’ macroeconomic conditions.

Our work adds to the literature examining the relationship between macroeconomic conditions, begun by Ruhm (2000) and continued in several more recent papers (e.g. Ruhm, 2003, 2005, 2007, 2015, 2016; Gerdtham and Ruhm, 2006; and Ruhm and Black, 2002). We add to the literature examining the determinants of automobile fatalities (e.g. (Adams and Cotti, 2008; Adams et al., 2012; Carpenter, 2004; Chaloupka et al., 1993; Cheng, 2015; Cotti and Tefft, 2011; Lovenheim and Steefel, 2011).