Working Paper: NBER ID: w15689
Authors: Kristian Bolin; Daniel Hedblom; Anna Lindgren; Bjorn Lindgren
Abstract: Several past studies have found health risk to be negatively correlated with the probability of voluntary health insurance. This is contrary to what one would expect from standard textbook models of adverse selection and moral hazard. The two most common explanations to the counter-intuitive result are either (1) that risk-aversion is correlated with health -- i.e. that healthier individuals are also more risk-averse -- or (2) that insurers are able to discriminate among customers based on observable health-risk characteristics. We revisited these arguments, using data from the Survey of Health, Ageing and Retirement in Europe (SHARE). Self-assessed health served as an indicator of risk: better health, lower risk. We did, indeed, observe a negative correlation between risk and insurance but found no evidence of heterogeneous risk-preferences as an explanation to our finding.
Keywords: voluntary health insurance; asymmetric information; adverse selection; Europe
JEL Codes: D82; I1
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| healthier individuals (I12) | risk aversion (D81) |
| health risk (I12) | probability of holding voluntary health insurance (G52) |
| self-assessed health (I14) | probability of holding voluntary health insurance (G52) |
| observable health characteristics (I14) | probability of holding voluntary health insurance (G52) |
| insurers discriminate based on observable characteristics (J78) | probability of holding voluntary health insurance (G52) |
| low-risk individuals (G22) | attractive contracts from insurers (G52) |