We develop a theory of rational self-medication. The idea is that forward-looking individuals, lacking access to better treatment options, attempt to manage the symptoms of mental and physical pain outside of formal medical care. They use substances that relieve symptoms in the short run but that may be harmful in the long run. For example, heavy drinking could alleviate current symptoms of depression but could also exacerbate future depression or lead to alcoholism. Rational self-medication suggests that, when presented with a safer, more effective treatment, individuals will substitute towards it. To investigate, we use forty years of longitudinal data from the Framingham Heart Study and leverage the exogenous introduction of selective serotonin reuptake inhibitors (SSRIs). We demonstrate an economically meaningful reduction in heavy alcohol consumption for men when SSRIs became available. Additionally, we show that addiction to alcohol inhibits substitution. Our results suggest a role for rational self-medication in understanding the origin of substance abuse. Furthermore, our work suggests that punitive policies targeting substance abuse may backfire, leading to substitution towards even more harmful substances to self-medicate. In contrast, policies promoting medical innovation that provide safer treatment options could obviate the need to self-medicate with dangerous or addictive substances.
We thank Jon Skinner, Tom Mroz, Matthew Harris, Melinda Pitts, Peter Savelyev, and seminar participants at the Atlanta Federal Reserve Bank, University of Tennessee, 2018 Workshop on the Economics of Risky Behavior, 2018 American Society of Health Economists Conference, 2018 European Health Economics Association Conference, IUPUI, Southeastern Health Economics Working Group, and UNC-Chapel Hill. The Framingham Offspring Study (FOS) is conducted and supported by the National Heart, Lung, and Blood Institute (NHLBI) in collaboration with the FOS Investigators. This manuscript was prepared using a limited access dataset obtained from the NHLBI and does not necessarily reflect the opinions or views of the FOS or the NHLBI. The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research.