The Cadillac Myth

Why the House-Senate conference should unload the tax on high-value health insurance.

Click here for a guide to following the health care reform story online.

A Cadillac

It’s widely agreed that health reform’s best remaining hope for “bending the cost curve” downward is the proposed 40 percent excise tax on “Cadillac” health plans in the Senate bill that passed early Dec. 24. These Cadillac plans are defined as any health insurance policy valued at more than $8,500 (for an individual) and $23,000 (for a family). A Dec. 22 Washington Post editorial praised the Cadillac tax as “an essential element of the package.” A letter to the White House signed by 23 distinguished economists said it is “critical” because it “offers the most promising approach to reducing private-sector health care costs.” Even the Center on Budget and Policy Priorities, a nonprofit that focuses relentlessly on how legislation affects low-income people, extended a (qualified) benediction. But the more I learn about the Cadillac tax, the more I’m convinced of its poor design. The tax will, over time, put the brakes on health insurance costs. But it will also put the brakes on health insurance benefits. Meanwhile, it probably won’t put sufficient brakes on doctor and hospital bills, which are the chief source of medical inflation.

The Cadillac tax was initially sold as a luxury tax on pampered investment bankers. Goldman Sachs, it was noted, lavished health insurance on its top executives worth $40,543, and neither the firm nor the employees paid a nickel’s worth of taxes on this benefit. President Obama warmed to the idea of “penalizing insurance companies who are offering super, gold-plated, Cadillac plans” because that wouldn’t “put additional burdens on middle class families.” This claim provoked a pungent backlash from labor unions, who pointed out that many union-negotiated health plans exceeded the Cadillac threshold. That backlash then provoked a counter-backlash demonizing Big Labor as a special interest group that feasts on sumptuous health benefits while rust belt industries wither and die.

Lost in the discussion were three compelling questions.

What makes Cadillac plans so expensive?  Excise taxes like the Cadillac tax differ from other taxes in that their principal purpose is not to raise revenue, but to change behavior. Excise taxes on gasoline, cigarettes, and alcohol, for instance, are aimed at reducing consumption of these goods. The Cadillac tax is aimed at reducing consumption of excess health benefits. Because the proposed excise tax on Cadillac health plans (40 percent) is higher than the top income tax bracket (35 percent), it’s assumed insurers will avoid paying it by lowering the value of their health insurance plans below the threshold. In so doing, insurers will cut out overly generous, perhaps even frivolous, health benefits.

The trouble with this logic is that overly generous benefits are almost never the reason why health insurance policies are unusually expensive. The January issue of Health Affairs has an interesting study by the nonprofit National Opinion Research Center (not available, alas, online) calculating the proportion of a family health insurance policy’s cost attributable to “benefit design.” More than half? Guess again. More than one-third? Nope. Try 3.7 percent. The Cadillac tax is targeted at altering a factor that accounts for less than 4 percent of a health plan’s overall cost. Add in the type of plan (Health Maintenance Organization, Preferred Provider Organization, whatever), arguably another variable in distinguishing a health plan that pampers its beneficiaries, and the zone of potential indulgence rises to a mere 6.1 percent.

The notion that greedy unions drive up the cost of health insurance took a serious beating in the Health Affairs study. Although 21 percent of firms with unionized workers had costlier plans (defined as exceeding not $23,000 but $13,350) a near-identical proportion of firms with unionized workers (17 percent) had less-costly plans. Far more significant factors were the type of industry and the geographical location. In the financial industry, about as many policies were above the threshold (7 percent) as below (8 percent). The industries tilted heavily toward costlier plans were, for some reason, “transportation/utilities/communications” (10 percent over, 5 percent under) and, ironically, health care (21 percent over, 9 percent under). Firms located in the Midwest and the West were likelier to be above the threshold than below it. Firms located in the Northeast and South were likelier to be below it than above it.

The Senate health reform bill raises the theshold for the Cadillac tax in certain instances based on occupation, geography, and age. Its determination of which occupations are likeliest to be dangerous, thereby raising health insurance premiums, makes prominent (and judging from the previous paragraph, appropriate) mention of people who “repair or install electrical or telecommunications lines,” but for the most part it seems weirdly random. Senate Majority Leader Harry Reid’s “manager’s amendment,” adopted Dec. 22, added longshoremen  to the earlier litany of police officers, fire fighters, paramedics, and emergency medical technicians, construction workers, miners, farmers, lumberjacks, and fishermen. The list still excludes most heavy industry, a glaring omission. A better-designed version of the tax would include more exemptions. But the more exemptions you add, however appropriate, the more pointless the tax starts to seem. “The problem with starting with that tax and trying to adjust for the real costs,” Rep. Joe Courtney, D-Conn., told the Washington Post’s Allan Sloan, “is that you end up with something that looks like a clown balloon.” The White House is arguing that the Cadillac tax will affect only 3 percent of health insurance premiums in 2013. That’s a selling point?

How would a Cadillac tax affect health benefits over time? Thereal kick in the Cadillac tax—and the reason it’s judged likely to lower premiums significantly over the long term—is that over time it will affect much more than 3 percent of all premiums. That’s because, although the threshold will be adjusted (“indexed”) upward over time, the adjustments won’t come anywhere near to keeping up with the pace of medical inflation. The rate of indexing will be the Consumer Price Index (i.e., the general rate of inflation) plus 1 percent. But it’s not unusual for the medical inflation rate to be more than twice the CPI. In 2002, for instance, the CPI was 2.4 percent. CPI plus one would therefore have been 3.4 percent. The medical inflation rate that year was 5 percent.

Which leads us to the third compelling question.

What impact would the Cadillac tax have on doctors and hospitals, the true drivers of health inflation?

It’s certainly true that the price and availability of insurance coverage has a significant effect on medical inflation. If insurers spend less, doctors and hospitals will spend less. But how much less? The prices that doctors and hospitals charge aren’t determined only by private insurance. They’re also determined by the availability of government money (which will increase substantially under health reform) and by what individuals feel compelled to pay out of their own pockets (which, in the case of serious illness, can be quite a lot). It’s therefore easy to imagine the Cadillac tax squeezing patients between ever-lowering benefits and still-rising prices for medical care.

I don’t expect these patients to suffer in silence. As more and more people see the value of their health insurance decline relative to doctor and hospital fees, they will likely complain to Congress. It’s easy to imagine so many people complaining that the Cadillac tax would be repealed.

But wait, you say. How can you square this with your support for a public option, which would also exert downward pressure on health insurance premiums, perhaps even as doctor and hospital fees rose? Easy. If private health insurers, caught between limits on what they could charge and what doctors and hospitals demanded to be paid, ended up manhandling their customers, then yes, the proposition that the public option existed to keep health insurers honest would be disproved. But the customers would have a means of escape: They could switch to the public option. Private insurers might suffer from their inability to compete. But that would be their problem, not mine. At any rate, we aren’t going to get a public option.

A far better alternative to the Cadillac tax would be the House bill’s 5.4 percent surtax on family incomes above $1 million and individual incomes above $500,000. Would it “bend the cost curve”? No. But it would do a better job paying for health reform, and it wouldn’t take an existing problem—the declining quantity of necessary medical care that premium dollars will purchase—and make it worse. The Senate should dump the Cadillac tax in conference. If it doesn’t, I’d lay decent odds it ends up doing so a few years down the road.

E-mail Timothy Noah at