To be consistent with our previous work, in the new baseline simulations we assumed that individual HDHP policies had deductibles of $3,500 per year for single coverage and $7,000 for family coverage, and no cost-sharing (i.e. zero coinsurance) once the deductible is met. We also assumed that individuals enrolled in a single-coverage HDHP would contribute $1,000 per year on average toward their HSA, and those with family coverage would contribute $2,000 on average. However, we made a significant change from our previous report where we assumed that everyone with the same coverage type (single or family) would make the same HSA contribution. In the current simulations we relaxed that assumption by allowing HSA contributions to be roughly proportionate to enrollee age and income. Our rationale for this change is that high-income enrollees will find the tax advantages of larger HSA contributions more appealing, and that older enrollees with higher medical care use will want to make larger HSA contributions. Even though we allowed enrollee contributions to differ by age and income, we constrained the average HSA contribution to be $1,000 for single coverage and $2,000 for family coverage. This means that any change in HDHP take-up from our previous simulations would be due to a change in the composition of the contributions, not to a general increase in the contribution level.
The following table shows the matrix of 16 HSA contribution cells we used, where each cell represents a quartile of the age and income distribution of the eligible U.S. population:
Cell values for family-coverage HAS contributions were double those shown above, except in the highest income/oldest age cell, where the family contribution was $8,320. The population- weighted average of the cell values was $1,000 for single coverage and $2,000 for family coverage.
Next, we defined the initial premium variables that were used in the baseline simulations. Each premium had two components: the insurance premium and the tax adjusted HSA contribution as constructed above. Further description of each component of the premium is provided next.
- Premiums for individual HDHP policies were taken from the eHealthinsurance.com web site, which provides a monthly estimated premium cost based on county of location, age, family size, and health history. For single-coverage contracts, we assumed the "baseline" demographic category was a 40-year old, non-smoking male. For family contracts, we assumed the policyholder was a non-smoking, 40-year old married man with a spouse and two children under the age of 10. Premiums for other age groups were age-adjusted using information from a 2002 HIAA/eHealthinsurance.com survey of plans purchased in the individual market. Premiums were updated to 2006 using the medical care component of the Consumer Price Index from the Bureau of Labor Statistics web site.
- Premiums for "traditional" HMO and PPO health plans were taken from the linked 2001
MEPS data file, adjusted for coverage type, plan type, establishment size, and benefits.
Premiums for employer-sponsored HDHP were based on the four employers in our data base. Premium estimates for individual HMO and PPO plans used the adjustment for the smallest establishment size category, based on the assumption that this most closely approximates the loading charge for plan administrative costs in individual health insurance policies. Once we estimated the premiums we inflated them from 2001 to 2006 prices based on medical insurance price inflation during the period.
- We considered the HSA contribution to be similar to a premium because each individual chooses to make that contribution "up front" - just as he or she pays the monthly premium for insurance coverage. However, unlike the monthly premium for insurance coverage, the HSA contribution is made with pre-tax income up to a cap. Therefore, we converted the contribution into pre-tax dollars by multiplying times one minus the personal federal tax rate based on the policyholder's income. We could not adjust for state income tax subsidies because the linked MEPS survey does not identify the respondent’s state of residence.6 We capped the tax subsidy at $2,700 for individual coverage and $5,400 for family coverage, which are the limits specified by current law. The actual contribution was also in effect capped by our range of age and income related premium contributions.7 These caps are binding only for individuals in the highest income/oldest age cell. Consequently, the “premium” variable used in our baseline simulations equals the monthly insurance premium plus the tax-adjusted HSA contribution subject to the cap as described above.