PS2-12
MICROSIMULATION METHODS FOR POLICY ANALYSIS: THE CASE OF THE CADILLAC TAX
Method: We developed a dynamic microsimulation model to simulate the effects of the CT on wages and health insurance benefits (i.e. premiums, contributions to health savings accounts, among others). We modeled the CT as a change in price of health insurance in terms of price and income effects, which are functions of their respective income elasticities. Our model generates two bivariate distributions of wages and health insurance benefits for individuals and families representative of the US population. The mean and variance of these distributions come from the Medical Expenditure Panel Survey (MEPS) and Kaiser Family Foundation Employer Health Benefits Survey. Drawing a sample of one million individuals and one million families from these distributions, we simulated the responses of each observation over time. We estimated the proportion of the population affected by the tax over time and separated them into two categories: those that will still purchase Cadillac plans and those that will cut back to the CT threshold. With these estimates, we calculated CT revenue (direct) and additional income tax revenue from diverted health benefits (indirect).
Result: In the first year of implementation, 6.5% of individual and 8.2% of family plans will be affected by the Cadillac tax; 99.8% and 99.7% of them will cut back to the CT threshold, respectively. The percentage of individuals and families affected by the CT will increase to 41% and 44%, respectively, by 2025. From the total revenue of the CT, 98% and 99% will be collected indirectly from individuals and families, respectively.
Conclusion: The proportion of individual and family plans affected by the Cadillac tax will increase steadily by 2025 reaching 41% and 44%, respectively. Most of the tax revenue will be collected indirectly through increased income tax revenues. Microsimulation models are useful tools to evaluate economic policies under different scenarios.