Debt, Death and Business Cycles

There is a large literature on the impact of economic conditions on health outcomes. The original literature on the topic (Ruhm 2000, 2003, 2005a, 2005b; Dehejia and Lleras-Muney 2004) documented improvements in health outcomes and healthy behaviors when the U.S. economy worsens. This early work largely linked rising unemployment rates to health measured in many ways: death by cause, substance use, smoking, obesity, nutrition and exercise. Recent literature expands analysis of economic conditions influencing health by using finer measures of economic well-being including dips in the stock market (McInerney, Mellor and Nicholas, 2013; Cotti, Dunn and Tefft, 2015; Angrisani, Kadiyala and Lee, 2015) and changes in foreclosure rates (Currie and Tekin, 2015).

We examine the relationship between financial distress and mortality. Previous studies linking debt to health relied on either self-reported health and/or survey responses indicating financial difficulties (Drentea and Lavrakas, 2000; Lyons and Yilmazer, 2005; Grafova, 2007; Keese and Schmitz, 2010; Lau and Leung, 2011; Averett and Smith, 2014). Our work adds to this literature by the use of individual panel data that includes both observed explanatory and outcome measures and, due to the richness of our financial data, more detailed debt measures.

We exploit the individual panel aspect of the Federal Reserve’s Consumer Credit Panel/Equifax (CCP). The CCP is a nationally representative 5% random sample of U.S. consumers and their household members in the consumer credit data system. The quarterly panel follows individuals from 1999 to the present and contains credit balances and delinquencies for different categories of debt. In addition to age and location, the CCP includes an indicator of the quarter-of-death for individuals who died while part of the panel. This individual-level, objective measure of health can be linked directly to objective measures of the individual’s financial well-being including the Equifax credit risk score, total dollar value and share of debt in severe delinquency, bankruptcy, foreclosure, and credit utilization.

To examine the link between debt and mortality, we use an approach similar to that in Sullivan and von Wachter’s (2009) analysis of job displacement and mortality. The baseline analysis is a difference-in-differences model where treatment occurs in places hard hit by the mortgage crisis based on the 2008 mortgage default rate (MDR) in the zip code of an individual’s location in 2005. The main analysis is an event history derived from estimates of:

Death_izt = α + β₁Treat_i + γ_j(Treat_i X Q_t) + β₂Z_zt + Q_t + ε_izt

where Treat_i is a binary indicator for being in the treatment group where the MDR > the treatment threshold. Death_izt is the mortality flag for individual i in zipcode z at time t. Q_t is a quarterly fixed effect, and Z_zt is a vector of any zip code level controls. The event history plots the γ_j’s and the expected relationship is a flat area before the crisis, followed by increasing mortality. Alternatively, Treat_i is replaced with a continuous measure of exposure using the 2008 MDR for individual i’s 2005 zip code.