In December 2015, Congress delayed implementation of the so-called Cadillac tax, a 40 percent excise tax on high-cost employer health benefit plans. Now scheduled to be implemented in 2020, the Cadillac tax would effectively cap the current tax exclusion for employer health benefits. By excluding health benefits from taxable incomes, the current tax break creates incentives for employers to pay more of employees’ compensation via health benefits instead of taxable wages, possibly leading to overuse of health care services and driving up health costs.

Economists are generally keen to reduce or cap “tax expenditures,” like the current tax exclusion, to help prevent overuse of tax-favored items. We also tend to worry about the relative inefficiency of the health sector in particular, and the adverse impact of ever-higher health costs on federal and state budgets.

As a result, many prominent economists love the Cadillac tax. Former Treasury Secretary Lawrence Summers and Harvard economist Gregory Mankiw praised the Cadillac tax with bipartisan gusto in a recent op-ed. A long list of economists signed a letter urging Congress not to repeal it. Jason Furman, head of the Council of Economic Advisors, told The Boston Globe that the tax “is perhaps the single biggest leverage we have on health costs in the private sector” and issued a spirited and comprehensive rebuttal to the tax’s critics.

However, these economists may be the tax’s only supporters. Political leaders and presidential candidates on both sides of the aisle are in opposition. Labor unions and business groups fought to delay or repeal it in 2015 (Note 1). In response to concerns, the Administration has proposed a modification to the tax intended to reduce its burden in states with high health care costs.

There are two key practical issues with the tax: the indexing problem and the adaptation question. In our opinion, these are serious enough issues to warrant continued caution before implementing the Cadillac tax.

The Indexing Problem

The Cadillac tax was enacted as part of the Affordable Care Act (ACA), which set the tax’s thresholds at $10,200 for single coverage and $27,500 for family coverage in 2018 dollars. The thresholds are set to grow by the consumer price index (CPI) plus 1 percentage point in 2019 and 2020, and by the CPI thereafter. Any health premiums above those thresholds would be taxed at 40 percent. Although health plans administering employer benefits would pay the tax, the cost would be passed through to the health plan enrollees.

Long-term projections from the Congressional Budget Office (CBO) currently foresee CPI growth of 2.4 percent per year beginning in 2019. Thus, the tax thresholds in 2020, the first year of the tax’s implementation, would be about $10,900 for single coverage and $29,400 for families.

Although the tax would only affect the highest-cost employer health plans at first, most projections of growth in health insurance premiums range between about 4 and 7 percent per year. For example, CBO projects that premiums will grow by about 5 percent per year. Thus, over time more employer plans would be affected by the tax simply because its thresholds are indexed more slowly than plans’ likely costs.

Furthermore, the distribution of premiums, even within industries, turns out to be very wide. Therefore, looking at averages only tells part of the story about which industries will likely be hit first and hardest by the tax. Exhibit 1 shows the range of premiums in each major industry group. For example, single premiums in the service sector ranged from a low of $2,000 to a high of $18,000 in 2014.

Exhibit 1: Range of Single Premiums by Industry, 2014

jeff_lemieux-figure1_revised

Source: KFF/HRET 2014 Employer Benefits Survey, calculations by INFORUM.

Exhibit 2 shows our reckoning of the percentage of employees affected by the tax, based on our forward model of 2014 Kaiser Family Foundation (KFF) data. When we say “affected,” we mean employees whose health benefits plan would trigger the tax if benefits were not modified or premium costs reduced in some way.

For this illustration, we used a medium-fast growth projection of 6 percent, a percentage point or so faster than CBO’s baseline. But either way, with premiums expected to grow much faster than the indexing rate of the tax thresholds under either scenario, more and more firms and employees would be affected. For example, we estimate that over 70 percent of workers in the health care industry itself are in health plans that would be affected by the tax by 2030.

Exhibit 2: Percent of Employees Affected Under a Medium-Fast Premium Growth Scenario (6 percent), No Adaptation (Benefit Changes), and Annual CPI Growth of 2.4 percent

jeff_lemieux-figure2

Source: KFF/HRET 2014 Employer Benefits Survey, model calculations by INFORUM.

It is notable that some of the keenest analysts of employer benefit and cost data have issued warnings about the indexing problem. Jon Gabel, who developed the benchmark KFF employer health benefits survey, issued an early warning that the Cadillac tax could become a “Chevy” tax over time. Gary Claxton, who now oversees the employer survey, reinforced Gabel’s conclusion in a recent brief. Research from the Congressional Research Service using the Agency for Healthcare Research and Quality’s (AHRQ’s) Medical Expenditure Panel Survey’s insurance component came to similar conclusions.

The Adaptation Question: Higher Deductibles or Innovations in Payment Systems?

It is certain that employers will adapt to the Cadillac tax by modifying their health plan offerings in some way. More work is needed to better understand what employers will most likely do. Some believe the tax will cause employers to find creative new ways to reduce health costs (and therefore premiums), as Jason Furman and Matthew Fiedler argue in their recent op-ed.

We worry the tax will cause employers to continue to shift costs to beneficiaries through higher and higher deductibles. There are several reasons why further increases in deductibles wouldn’t necessarily be desirable.

First, we looked at the distribution of deductibles by premium level. Exhibit 3 shows the relationship between premiums and deductibles among manufacturing firms in 2014 KFF data. The inverted “S” curve shows single premiums ranked from the lowest on the left, to the highest on the right. Most manufacturers in 2014 had single premiums between $4,000 and $8,000. The blue dots represent the deductibles for each plan on the curve. Clearly, plans with the lowest costs tend to have higher deductibles, and plans with the highest costs tend to have lower deductibles. The relationship is indicated by the orange trend line. However, the pattern is not uniform. Quite a few high-cost plans already have high deductibles, and some low-cost plans have low deductibles.

Exhibit 3: Premiums vs. Deductibles, Manufacturing Industry, 2014

jeff_lemieux-figure3_revised

Source: KFF/HRET 2014 Employer Benefits Survey, calculations by INFORUM.

Note: Firm by firm results are arrayed on the horizontal axis from left to right by premium, from lowest to highest.

Gabel’s report using KFF data also found that only a small percentage of the wide variation in employer premiums was associated with benefit levels (such as deductibles) in the first place. Instead, variations in premiums were largely due to the health status of enrollees, regional variations, or other factors. The wide variation in premiums and the fact that high premiums are not always linked with high benefit levels means that it will likely be difficult for regulators to fine-tune the Cadillac Tax toward only firms with extensive or lavish benefits, as opposed to firms that may have ordinary benefits but a sicker-than-average pool of enrollees.

Second, deductibles and other forms of patient cost sharing have already jumped in recent years, and it’s not clear that high patient cost sharing only limits unnecessary or lavish care. A recent study followed a multi-year natural experiment where a large employer switched from a zero-deductible plan to a high-deductible plan. The reductions in care that resulted from the switch were across the board, among the sicker and the healthier, and among preventive and acute care services alike.

The employer in this particular study had high average worker incomes ($100,000+ median), and presumably a highly educated workforce. As such, it was striking that the deductible seemed to cause such a clumsy response. Other studies have had similar findings, and the conclusion that patient cost sharing is, at best, a blunt tool for restraining health costs dates back to the RAND Health Insurance Experiment.

As noted, Furman and Fiedler argue that, instead of causing employers to increase deductibles, the Cadillac tax will instead spark new payment arrangements designed to improve health system efficiency:

To realize the full potential for systemwide improvement, the private sector will need the right incentives to discover and deploy novel approaches to improving the health care system….. Unfortunately, the U.S. tax code has long undermined such incentives. Because employees pay income and payroll taxes on wages but not on compensation provided in the form of health care benefits, it is rational for employers to skew compensation packages away from wages and toward excessively costly and inefficient health benefits…..[Under the Cadillac tax] many employers will probably focus instead on encouraging more efficient care delivery, by deploying innovative payment models…. and finding creative ways to steer patients toward more efficient providers — investments that were often difficult to justify when the federal government was picking up much of the tab for inefficient care.

It’s an interesting argument, but we don’t think private insurers and health plans have been holding back on implementing innovative payment arrangements because of insufficient incentives. On the contrary, we have written previously about private sector approaches to improving care coordination and quality, and have worked with manufacturers implementing innovative health plans intended to control costs while fostering employee recruitment, health, and productivity.

In fact, we think some private health plans and their large employer clients were a bit ahead of Medicare in at least trying to spark innovative payment and reimbursement regimes with health care providers. What the private health plans lacked, however, was Medicare’s market clout to really make the new systems stick and bring them to scale.

Watchful Waiting

Even if Furman and Fiedler are right, and the Cadillac tax would indeed sharpen private plans’ incentive to implement payment systems that reward quality of care rather than volume, higher deductibles could still result.

When we saw the results of Inforum’s Cadillac tax model, we were struck not only by the huge variation in premiums across firms and within industries, but also by the escalating nature of the tax, with a deeper and deeper bite each year due to the indexing problem. Absent additional deferrals or delays in implementation or adjustments to the indexing, even the most innovative firms could struggle to find alternative ways to cut costs quickly enough to avoid the tax, and might have to default to continuously higher deductibles.

Overly high deductibles, in turn, could become clinically inefficient. That is ever-higher deductibles could lead to the delay or avoidance of appropriate and necessary health care for too many patients, especially those with low incomes. In those cases, the deductibles could cause reductions in health status sufficient to raise overall or long-term health costs, not lower them.

On the other hand, the Administration’s proposal to modify the tax thresholds in high cost states could relieve the pressure of the tax in places that would otherwise be hardest hit. However, that proposal has not been enacted in law, and we have not attempted to model its impact.

For several years, economists have expected a modest acceleration in health spending from its recent historically low growth. Although the volume of health care consumption has increased recently, and hospital employment started rising briskly in 2015, overall national health spending remains relatively restrained, at least compared with long-term averages.

Medical professionals sometimes use the term “watchful waiting” in clinical cases where potential interventions may be harmful or may turn out to be unneeded. Waiting for additional information on health costs and trends in employee health benefits before implementing the Cadillac tax would seem to make sense in this case.

Note 1

The National Association of Manufacturers (NAM) is one of the business groups opposed to the tax, and NAM funded Inforum’s modeling of the tax’s sectoral and economic impact.