Sponsored by:
Over the past two decades, technological and scientific advances have led to substantial improvement in the predictive power of genetic data, including for diseases and health conditions. This has led to concerns that further improvements in genetic prediction will create problems in health insurance markets due to adverse selection. Adverse selection could arise if, for example, potential insurance consumers who know they are genetically predisposed to develop certain diseases decide to buy more health insurance to protect themselves against the possible onset of these diseases, while insurers are unable to charge more to such consumers. This, in turn, could lead the market for such insurance products to unravel. We make two contributions toward understanding the likelihood of such adverse selection occurring in a future where the predictive power of genetic data approaches its theoretical maximum. First, we develop an econometric method to estimate key parameters needed to assess how much selection genetic prediction could create on insurance products. Second, we apply the method to data from the UK Biobank, a large dataset with genotypic and health-related data, and use economic models to estimate the risk of adverse selection for a suite of common medical conditions. Because of the still limited accuracy of genetic predictors for these conditions, there is little selection today. However, preliminary results suggest expected improvements in genetic prediction over the next few decades could greatly intensify the level of selection for most of the medical conditions. This, in turn, could lead to high levels of adverse selection and to the unraveling of health insurance markets.