The exclusion of employment-based health insurance from income and payroll taxes is thought to increase the generosity of insurance coverage and, in turn, increase the overutilization of low-value health care services. We examine this inefficiency of overinsurance by quantifying the change in expected utility across 4 benchmark plans varying in actuarial value (AV) and focus on the distribution of each of these estimates across different groups of people varying in health status. Specifically, we quantify the changes in health care spending due to moral hazard and the changes in uncertainty tied to risk aversion using data from the nationally representative sample of adults with employment-based coverage from the 2007-2016 Medical Expenditure Panel Survey, and produce estimates of expected utility for 24 groups of people based on their age, gender, and preexisting conditions. Our model suggests an average preferred AV of 78% without the tax exclusion, with 29.0% of the population preferring a 60% AV, 6.5% preferring a 70% AV, 18.1% preferring an 80% AV, and 46.4% preferring a 90% AV. When incorporating the distortionary effect of the employment-based tax exclusion, the preferred plan increases to an 83% AV for low-income people (with 71.0% of the population preferring a 90% AV) and an 84% AV for high-income people (with 76.0% of the population preferring a 90% AV). We estimate that policy changes to make subsidies independent of a plan's AV could result in increases in utility equal to about 2.7% of total health care spending, but with those net gains concentrated among the healthy.
