A Medicare Reform Proposal Everyone Can Love: Finding Common Ground among Medicare Reformers

Policy Reports | Health

No. 306
Saturday, December 01, 2007
by Andrew J. Rettenmaier and Thomas R. Saving

Estimated Effects of the Reform on Medicare Spending

Because the HIRA reform makes all future retirees more cost conscious than they would be under current Medicare, health care spending can be expected to decline relative to current projected spending.  In addition, because total cost sharing rises with lifetime income, the level of cost consciousness will rise with income.  Two methods of accounting for the effects of changing cost sharing are used to identify the cost paths depicted in Figures IIa and IIb.  These (Methods 1 and 2) are based on the RAND health care experiment. [See the sidebar, “Effects of Deductibles on Health Care Spending.”] The estimates are derived by using the experience of the medium worker as the representative consumer.

Three series are presented in each graph.  The Current Medicare series from the 2007 Medicare Trustees Report reflects total Medicare costs as a share of GDP.  These costs must be paid through payroll taxes, taxes on Social Security benefits, premiums from beneficiaries and general revenues.  The series for Reformed Medicare Spending includes both the effect of lower total spending on Medicare-covered services due to the incentives produced by the rising total deductibles, as well as the spending that will be funded by retirees' HIRAs. At first, the reform's effects are modest, but as an increasing number of retirees have annuity payments from their HIRAs, the effects on total spending grow.

Given that the HIRA reform requires contributions from workers equal to 4 percent of their earnings, these contributions are added to the Reformed Medicare Spending Series to arrive at the final series reflecting the total reform costs.  Comparing this series to the Current Medicare series shows the timing of the costs for the two alternatives.

Upper-Bound Estimate of Medicare Spending (Method 1). As seen in Figure IIa, until 2048 the reform is more expensive than continuing the current arrangement.  Thereafter the reform is less expensive.  In 2050, Reformed Medicare spending is 79 percent of Current Medicare and by 2080 it is 76 percent.  By 2080, the total reform costs, including contributions to HIRAs, are 91 percent of what Current Medicare would cost.

Lower-Bound Estimate of Medicare Spending (Method 2).  Figure IIb presents the estimates when Method 2 is used to identify the effects of the higher deductible on health care spending, the total costs of the reformed program are less than Current Medicare beginning in 2032.  Under Method 2 assumptions, reformed Medicare spending is 68 percent of Current Medicare in 2050 and by 2080 it drops to 66 percent.  Total costs for the reformed system are 88 percent of the current system's costs in 2050 and 81 percent by 2080.

“The RAND health experiment showed that modest deductibles reduced health care spending without affecting health outcomes.”

The two methods of accounting for the deductibles' effects provide upper- and lower-bound estimates of the cost of the reform due to demand-side effects only.  In either case, the reform is less expensive than the current arrangement in the long run. 19 If supply-side effects were included the change would be more dramatic.  As Amy Finkelstein has shown, the introduction of Medicare caused dramatic supply-side effects, such as the entry of new hospitals into the market and the adoption of new, more expensive practice styles. 20 Finkelstein estimates that the market-wide spread in third-party payments can explain half of the growth in per capita spending from 1950 to 1990, a much larger estimate than those previously made.  Clearly, greater cost sharing due to the HIRA annuities and base deductibles will not only affect the level of spending, but also the spending growth rate.

Read Article as PDF