Venue: The Fuqua School of Business, Duke University, 1 Towerview Drive, Durham, NC 27708-0120
Presentation
Health Savings Accounts, High-Deductible Policies, and the Uninsured: Simulating the Effects of HSA Tax Policy Using a Utility-Maximization Framework
The Bush administration has offered a number of proposals to make health savings accounts (HSAs) and high-deductible health insurance policies more affordable and attractive for the uninsured. For example, current proposals include: 1) an income tax deduction for the premium on a HSA-qualified insurance policy, 2) an income tax credit for HSA-qualified policies purchased outside the employment-based group market, 3) an increase in the allowable HSA contribution amount.
While these proposals would likely lower the cost for HSA-qualified insurance, it is unclear whether such proposals would actually decrease the number of uninsured. The policies are designed to eliminate the tax advantage currently available to employment-based insurance. But eliminating that differential could adversely affect the employment-based market, causing firms to offer less attractive policies to employees, resulting in lower take-up rates, or to drop coverage altogether (Glied and Remler, 2005; Hoffman and Tolbert, 2006). The net effect could be an increase in the number of uninsured.
Our paper examines the impact of HSAs and the various tax proposals on the uninsured, including the effect of how adopting the tax deduction and credit policies for non-group insurance would affect the employment-based group market. The foundation for our work is a recently developed simulation model based on utility-maximizing, representative agents (Cardon and Showalter, 2007).
One advantage of our framework over previous simulation models is that it allows for the empirical calibration of 'deep' structural parameters of consumers' utility functions, and the risks associated with uncertain medical expenditures. This in turn allows us to determine consumers' valuation of products for which little or no data exists?e.g. HSA contributions and high-deductible insurance. Our methodology can evaluate such products because it accounts for both prices and consumers' underlying risk preferences.
We focus on three cases of interest: 1) the currently uninsured who do not have access to group coverage, 2) the currently uninsured who have access to group coverage, but choose to be uninsured, 3) the currently insured in group coverage. For each case, we segment the group in various sub-categories based on demographic variables such as income and age-specific health expenditures. For each sub-group, we calibrate our model to match the observable characteristics from the MEPS. Then we model the policy changes and assess the value of the change to the consumers and their likelihood of changing from the status quo. Of particular interest is the incentive for the currently insured in group coverage to switch to the non-group market, which could destabilize the employment-based market. We model the pooling within the group market and assess the likelihood of existing pooling arrangements being made unstable by the policy changes. Preliminary estimates suggest that HSAs result in a modest reduction in the percent uninsured for both the group and non-group markets.