Foreclosures and Health

Great paper by Janet Currie and Edral Tekin on the housing crisis and its impacts on health outcomes:

We investigate the relationship between foreclosure activity and the health of residents using zip code level longitudinal data. We focus on Arizona, California, Florida, and New Jersey, four states that have been among the hardest hit by the foreclosure crisis. We combine foreclosure data for 2005 to 2009 from RealtyTrac with data on emergency room visits and hospital discharges. Our zip code level quarterly data allow us to control for many potential confounding factors through the inclusion of fixed effects for each zip code as well as for each combination of county, quarter, and year. We find that an increase in the number of foreclosures is associated with increases in medical visits for mental health (anxiety and suicide attempts), for preventable conditions (such as hypertension), and for a broad array of physical complaints that are plausibly stress-related. They are not related to visits for cancer morbidity, which arguably should not respond as rapidly to stress. Foreclosures also have a zero or negative effect on elective procedures, as one might expect. Age specific results suggest that the foreclosure crisis is having its most harmful effects on individuals 20 to 49. We also find that larger effects for African-Americans and Hispanics than for whites, consistent with the perception that minorities have been particularly hard hit.

This study makes a few key contributions to the literature on economic vulnerability and health outcomes. First, it utilizes very detailed data that both enable precise estimates and a number of statistical techniques that can better address potential confounders. Second, the authors make great use of falsification tests. Basically, in theory you would expect stressful events to impact some diseases and not others (especially those that take time to evolve). The authors follow up this idea and actually demonstrate it in their data, which makes the results more believable.