Focus on Incentives, Not Budget Scores

In his excellent book, The Armchair Economist, Steven Landsburg writes that most of economics can be summarized in four words: “People respond to incentives.” The rest, he argues, is just commentary. Unfortunately, perverse incentives permeate our health care system, leading consumers and producers to generate mountains of inefficiency and waste, and relatively little innovation.

When Congress was debating the Affordable Care Act (ACA), the debate largely revolved around how many people would gain insurance and how the law would impact the deficit, not about how the law would address existing government policies and programs that largely created the perverse incentives. The ACA created several new perverse incentives, unfortunately.

ACA proponents, often pointing to Congressional Budget Office (CBO) estimates, argued that the law would reduce the number of uninsured people by about 30 million while producing federal budget savings. Five years later, low exchange enrollment suggests that far more people than projected will remain uninsured. New Mercatus research by James Capretta and Joe Antos shows that the projected federal budgetary savings result from provisions that are unlikely to be economically or politically sustainable.

Capretta and Antos show how four indexing provisions drive the projected budget savings. In this piece, I discuss three of these provisions: slowing Medicare payment increases; the growing reach of Medicare and investment taxes; and the growing reach of the Cadillac tax.

Slowing Medicare Payment Increases

The ACA reduces the annual increase in Medicare fee-for-service (FFS) payment rates with a new “productivity adjustment factor.” This factor bases increased Medicare rates on economy-wide productivity improvements rather than on the rising costs of medical goods and services, which have historically grown at a much higher rate.

Capretta and Antos highlight the concern that the lower adjustment factor, compounded year-after-year, will lead to such low provider payment rates that many doctors will stop treating Medicare beneficiaries. The federal government itself predicts that by 2040, half of all hospitals, 70 percent of skilled nursing facilities, and 90 percent of home health agencies will be losing money each year. These projections also show that Medicare payments will only be 40 percent of what private insurers pay in 75 years.

Medicare’s FFS schedule produces all the ill effects economists associate with price controls, including too many procedures when prices are set too high and too few procedures when prices are set too low. Moreover, markets adjust to what Medicare is reimbursing and what it is not, in a way that stifles innovation.

Although Capretta and Antos don’t mention it, past research shows that many providers adapt to lower Medicare payment rates by performing more services or by switching to more intensive, i.e. more costly, treatments in order to preserve their income. They do cite the Medicare “doc fix” experience as a reason that Congress will likely undo the ACA’s productivity adjustment change.

Nearly two decades ago, Congress created a different formula to ensure that Medicare’s rates to physicians did not increase faster than economic growth. Since 2002, Congress repeatedly enacted short term “fixes” that prevented the formulaic-driven cuts from taking effect. Earlier this year, Congress permanently repealed that formula.

In general, long-term projections unrealistically assume static conditions that fail to anticipate how innovation can transform markets. Even with large government interference and regulation in health care markets, it is impossible to know how innovation will change the health care system over the next several decades. For example, how will we know what hospitals will look like in 25 years, much less 75 years?

In a normal market, there is no need for government to mandate “productivity adjustments”; they just happen naturally as prices tend to decline and quality tends to improve with free market competition. Greater efficiency is unlikely to come from slowly trimming price controls. Real reform means working to eliminate the price controls that create the bad incentives and then allowing markets to work.

Growing Reach of Medicare and Investment Tax

The ACA contained a 0.9% add-on tax to the Medicare hospital insurance payroll tax and a new 3.8% tax on investment income for individuals earning above $200,000 and couples earning more than $250,000. The ACA did not index these thresholds to inflation, so each year as inflation and economic productivity push up nominal wages, more people will be subject to these taxes as the definition of “rich” gets lower and lower. Capretta and Antos argue this was intentional in order to “show ever-increasing revenue from the law, and consequently declining deficits.” The economic effects of these taxes, which Capretta and Antos estimate will affect 20% of households in 2033 and 40% of households in 2045, will reduce investment and work.

Of the indexing provisions, this one is probably the least vulnerable to repeal in the near term. In the absence of comprehensive tax reform, political supporters of the ACA can demagogue proposals to repeal it as a tax cut for the wealthy.

The Cadillac tax represented an attempt to remedy a longstanding problem with government health care policy – the federal tax exclusion for employer-sponsored insurance. Economists across the political spectrum tend to agree that the exclusion causes employers to offer overly expansive insurance, increases overall health care spending, and provides a disproportionate benefit to the wealthy.

The Cadillac tax, scheduled to take effect in 2018, imposes a 40% excise tax on policies valued above $10,200 for single coverage and $27,500 for family coverage. In order to amplify the effect of the tax over time, the ACA indexed these thresholds to general inflation rather than medical inflation. Capretta and Antos show that the average employer-sponsored family plan, increased annually by historic health spending growth, will exceed Cadillac tax thresholds in 2025.

A 40% excise tax represents a significant marginal rate, and many firms are already restructuring insurance plans, largely by increasing employee out-of-pocket payments, in order to avoid the tax. As employers make their plans less expansive, it is unlikely that the average plan will exceed the Cadillac tax thresholds until many years after 2025 assuming the Cadillac tax remains in place.

Politically, the Cadillac tax remains one of the most vulnerable parts of the law because of how much labor unions and big business oppose the tax. Replacing the Cadillac tax with a cap on the amount of coverage that can be excluded from taxes would be simpler and fairer, as several people have written about at this site.

Too Easy to Game a Favorable Score

During the drafting of the ACA, legislators changed and tweaked scores of provisions in order to hit desired enrollment and budgetary targets. A major contribution of the Capretta and Antos piece is to remind policymakers and the public that while these targets have some importance, they can be gamed.

Such “gaming” does not need to involve indexing. An example is the Community Living Assistance Services and Supports (CLASS) program in the ACA. CLASS created a new long-term care entitlement program and showed ten-year budgetary savings because the program collected premiums for five years before any benefits were paid. CBO’s March 2010 estimate attributed half of the ACA’s overall ten-year net deficit reduction to CLASS. Fortunately, Congress repealed CLASS before the government started enrolling people into another financially unsustainable entitlement program.

While politicians and policymakers will certainly continue to adjust parameters in legislation to produce favorable budget estimates, Capretta and Antos’ paper shows that we must consider the dynamic effect of policy changes and account for the plausibility of key assumptions underlying the estimates. More fundamentally, the manipulation of the scoring process by the authors of the ACA and the bad predictions by economic models of the ACA’s effect should lead policymakers to focus more on how new policies impact incentives and less on estimates produced by economic models.