Deaths of Despair or Drug Problems?

The United States is in the midst of a fatal drug epidemic. This study uses data from the Multiple Cause of Death Files to examine the extent to which increases in county-level drug mortality rates from 1999-2015 are due to "deaths of despair", measured here by deterioration in medium-run economic conditions, or if they instead are more likely to reflect changes in the "drug environment" in ways that present differential risks to population subgroups. A primary finding is that counties experiencing relative economic decline did experience higher growth in drug mortality than those with more robust growth, but the relationship is weak and mostly explained by confounding factors. In the preferred estimates, changes in economic conditions account for less than one-tenth of the rise in drug and opioid-involved mortality rates. The contribution of economic factors is even less when accounting for plausible selection on unobservables, with even a small amount of remaining confounding factors being sufficient to entirely eliminate the relationship. These results suggest that the "deaths of despair" framing, while provocative, is unlikely to explain the main sources of the fatal drug epidemic and that efforts to improve economic conditions in distressed locations, while desirable for other reasons, are not likely to yield significant reductions in drug mortality. Conversely, the risk of drug deaths varies systematically over time across population subgroups in ways that are consistent with an important role for the public health environment related to the availability and cost of drugs. Put succinctly, the fatal overdose epidemic is likely to primarily reflect drug problems rather than deaths of despair.