An earlier version of this article was published as “Follow the Money: Money Matters in Health Care, Just Like Everywhere Else,” American Journal of Law and Medicine, Vol. 36 (2010).

If all were well with the health care system, those observations would be of no real significance. No one would be much interested in the observation that auto mechanics, plumbers, actors, bicycle messengers, and newspaper reporters also respond to economic incentives — and that misaligned incentives can have adverse consequences. Yet, to say the least, all is not well with American health care. Whether the subject is the quality of care that insured and uninsured Americans receive, the cost of coverage and of receiving care, Medicare’s fiscal projections, the burdens Medicaid imposes on the states, the cost of pharmaceuticals, the availability of primary care physicians, the wide variation in cost and treatment patterns, the continued viability of employment‐​based health insurance, or the dysfunctions of the medical malpractice system, it is clear there is no shortage of problems with the U.S. health care system.

What all of those problems have in common is that some (and, more often than not, most) of the blame is properly attributable to misaligned economic incentives. Instead of trying to address that problem, past efforts have focused on a “collective search for villains” (to borrow David Goldhill’s line from his 2009 Atlantic article on American health care) with the specific identities of the villains varying depending on the political and philosophical commitments of the searchers. Those efforts have been time‐​consuming and have created steady work for legislators, lobbyists, lawyers, law professors, and policy wonks, but they have had about the same effect as the witch trials that swept Europe from 1400 to1600 and Salem, Massachusetts in 1692: deeply satisfying for those who perceive they are doing “God’s work,” intensely unpleasant (and sometimes lethal) for the targets, but providing little actual improvement in the state of the world. Stated more positively, unless and until we alter the core incentives created by our existing payment system, we will get more of what we already have — a dysfunctional non‐​system that delivers uncoordinated care of widely varying quality at high cost.

Money Matters: A Brief Compendium

It is difficult to overstate the extent to which economic incentives explain the structure, performance, and pathologies of the American health care system. I focus on three examples, all of which present variations on a common theme.

Fee‐​for‐​service | The American health care system primarily relies on an encounter‐​based, quality‐​insensitive fee‐​for‐​service system of compensation. In general, health care providers can lawfully bill for their efforts only when they physically interact with a patient or interpret a test that resulted from direct physical interaction with the patient. Each such interaction generates a bill, with the amount billed varying greatly depending on the nature of the service/​interaction. However, payment does not vary based on the quality of the service or on its medical necessity. There are also almost no constraints on the volume of the services that may be provided as long as a licensed health care provider deems them necessary.

The consequences of this compensation strategy are quite predictable: we have a system that aggressively delivers massive quantities of health care services in a highly fragmented non‐​system, but pays little attention to whether the services in question actually contribute to health. Worse still, there is usually no “business case” for improving matters: delivering higher‐​quality care and/​or keeping one’s patients healthier can actually make a provider financially worse off.

Consider the experiences of two different providers whose efforts to improve the quality and cost‐​effectiveness of the services they were providing ran head‐​first into these incentives. In 1998, Duke University Medical Center implemented a disease management program focusing on congestive heart failure, a major source of morbidity and mortality in the elderly population. The Duke Heart Failure Program emphasized a range of tactics (including aggressive use of medications and biweekly phone calls by nurse‐​practitioners) designed to keep patients with congestive heart failure healthy and out of the hospital. The program was extremely successful; the rate of hospitalization plummeted as the health of patients with congestive heart failure improved. Unfortunately, between the “extra” costs of the program and the “lost” revenue from hospitalizations that no longer occurred, Duke was financially punished for making its patients healthier. Duke eventually discontinued the program.

A few years later, the same dynamic snared Intermountain Health Care in Utah. New York Times economics writer David Leonhardt told the story in a November 3, 2009 column:

When Intermountain standardized lung care for premature babies, it not only cut the number who went on a ventilator by more than 75 percent; it also reduced costs by hundreds of thousands of dollars a year. Perversely, Intermountain’s revenues were reduced by even more. Altogether, Intermountain lost $329,000. Thanks to the fee‐​for‐​service system, the hospital had been making money off substandard care. And by improving care — by reducing the number of babies on ventilators — it lost money. As James tartly said, “We got screwed pretty badly on that.”

These problems are pervasive; as I summarized matters in a recent chapter in a book on health care fragmentation:

In health care, we get what we pay for — and what we pay for is the provision of specific services — virtually irrespective of whether they are provided efficiently, or even needed. Because payment is conditioned on the laying of hands (or eyes) upon a patient, time spent coordinating care doesn’t create a billing opportunity. When we don’t pay for something, it generally doesn’t get done. Similarly, providing integrated care doesn’t pay better than fragmented care — and in some instances, it pays worse. The results are entirely predictable — and until the incentives created by the payment system are modified, we will continue to get what we’ve already got: a fragmented non‐​system for delivering care of highly variable quality at high cost.

Stated differently, because we pay health care providers for what they do, not for what they accomplish, they have little or no direct financial incentive to improve quality and prevent errors. The baby steps we have taken toward payment reform (including bundled payments, tiering, and payment for performance (typically referred to as “P4P”) have been ineffective because they have not altered the core incentives created by our encounter‐​based, quality‐​insensitive fee‐​for‐​service compensation system. Unless and until we fix the core incentives, most attempts to improve quality and/​or value will result in providers experiencing internalized costs and externalized benefits — a combination that is extraordinarily unlikely to lead to an optimal level of quality.

Paying too little, or too much? | Paying the “right” amount for goods and services is very important. Paying “too little” results in one set of problems; paying “too much” results in a different set of problems.

What happens when we pay too little? Ask anyone covered by Medicaid — whose payment rates have historically been well below Medicare and private insurers — how easy it is for them to find a regular source of care other than the local hospital’s emergency department. A similar dynamic helps explain the dearth of generalists and cognitive‐​based specialists; the aggressive lobbying and litigation campaign that the American College of Cardiology has been waging over proposed changes in Medicare reimbursement; and the difficulties that some Medicare beneficiaries are experiencing in securing access — including the recent announcement by the American Medical Association that there are 22 “patient access hot spots” where access to care for Medicare patients “is already at risk.” As these examples reflect, the consequences of paying too little are the same as those of setting a price cap below the market‐​clearing price: the quality of the good or service degrades or it disappears entirely.

What about when we pay too much? Overpayment is obviously wasteful, but it is also skews the allocation of resources within the health care system. Overpayment also encourages inefficient unbundling and re‐​bundling of the delivery system as providers maneuver to capture the “excess” revenue. It is no accident that we have seen the emergence of physician‐​owned cardiac and orthopedic specialty hospitals, but no similar physician‐​owned hospitals for the treatment of trauma, burn care, or AIDS.

As with Goldilocks and the three bears, the key in purchasing health care is paying the “just right” amount. The bad news is that an administered pricing system (like Medicare) has great difficulty in doing that. Innovation creates one set of problems: what is the correct price for new treatments and improvements in existing treatments? But that is only the beginning of the complications. How much should prices vary by region? How much should prices vary by the type of provider delivering the service and/​or the location at which the service is provided? Should higher‐​value and/​or higher‐​cost services result in higher payment? Should lower‐​value services result in lower payment? Should payment rates incorporate an assessment of the social role of the institution (e.g., academic medical centers, safety‐​net institutions, etc.)? These problems are compounded by the pricing feedback loop (and the lack thereof): providers who are “underpaid” relative to the market price will show up and protest loudly, while those who are “overpaid” relative to the market price will never volunteer that fact — and will manufacture arguments explaining why they deserve every penny. These problems are further compounded by politics: legislators will lobby for special deals for providers in their districts and frame payment formulas to deliver more cash to their states. Even if payment reform is somehow implemented, these dynamics mean any victories are likely to prove temporary.

To be sure, no system of payment is perfect and we should not indulge in the nirvana fallacy in assessing the performance of administered pricing systems. If the alternative is a fragmented buy‐​side with limited ability to resist the cost‐​increasing demands of concentrated provider interests, then interest in rate regulation is likely to attract some enthusiasm from those convinced that they are right and Hayek is wrong. But the basic point remains: “compared to what” is the question that should be asked about all institutional arrangements and proposals to reform the same.

Subsidizing employer‐​based insurance | Roughly 160 million Americans obtain health insurance through their place of employment or that of a family member. Employer contributions to the cost of coverage are not treated as taxable income to the employee. Employees can also pay for their direct contributions to coverage with pre‐​tax dollars and self‐​employed individuals can deduct health insurance premiums to the extent that they do not exceed earned income.

The result is that employees who obtain employment‐​based insurance (and self‐​employed individuals who qualify to deduct their premiums) can purchase coverage with pre‐​tax dollars, while those who obtain insurance through other channels must purchase it with after‐​tax dollars. This subsidy has been estimated to exceed $200 billion in foregone tax revenue per year, with the precise value of the subsidy to any given taxpayer varying by income level and the cost of the coverage in question. (Because our tax system is progressive, the exclusion of employer contributions from income means the subsidy is worth more to people who make more — a peculiar design choice, to say the least.)

To summarize, the tax code creates a substantial financial incentive for the purchase of health insurance through one’s place of employment and creates a further incentive for employees (particularly high‐​income employees) to prefer richer benefits than they otherwise would. The result is an inefficiently high level of health care coverage for those who receive employment‐​based health insurance, even though almost everyone agrees we are already spending too much on health care. Worse still, the tax subsidy discriminates against those unable to obtain employment‐​based coverage (by making them use after‐​tax dollars in a market where a majority of other purchasers are using pre‐​tax dollars). Finally, it provides larger subsidies to those who need it the least.

Bill Simon, who was secretary of treasury under Presidents Nixon and Ford, memorably observed that our tax system should look like it was “designed [that way] on purpose, based on a clear and consistent set of principles.” It is hard to make the case that the current subsidy for employment‐​based coverage meets that standard — let alone the far‐​lower standard that it makes any sense whatsoever.

What About Reform?

Reform is supposed to make things better. Does the new Patient Protection and Affordable Care Act (PPACA) effectively address the problems identified above? The short answer is that although the legislation does have a few provisions attempting to address these problems, Congress and the Obama administration had bigger fish to fry. The legislation accordingly focuses on broadening access by means of insurance reform and not on changing the incentives driving health care treatment and overall spending. Indeed, on numerous occasions, Congress and the administration pulled their punches in addressing those problems — usually by trading stricter reforms in those areas for coverage provisions that they valued more.

More concretely, PPACA authorizes the creation of a Center for Medicare and Medicaid Innovation, along with a range of pilot programs and demonstration projects to address some of the dysfunctions created by Medicare’s payment system. These initiatives are promising, but significantly underpowered. In fairness, no one knows for sure which of these initiatives will actually work, but the limitations placed on them make it less likely that the proposed payment reform will have any effect whatsoever. History also suggests that Congress will likely cripple or kill “effective” initiatives (i.e., those that reduce payments to health care providers) and expand ineffective ones (i.e., those that result in increased payments to health care providers). Although PPACA authorizes expansion of effective pilot projects without further congressional approval, the administration is unlikely to do so if it will create push‐​back from Congress.

A similar fate is likely to meet the Independent Payment Advisory Board (IPAB) created by PPACA. IPAB is a 15‐​member board charged with presenting Congress and the president with proposals to reduce “excess cost growth” in Medicare. If cost targets set in PPACA are exceeded, IPAB must propose specific savings that will take effect unless the president vetoes them or a super‐​majority of Congress votes them down. However, nothing prevents Congress from allowing IPAB’s recommendations to take effect and then reversing them with a simple majority vote. PPACA also places significant limitations on the scope of IPAB’s recommendations; it may not target hospitals and hospices for reductions until 2020, and it is prohibited from making proposals that ration care, raise taxes or Medicare Part B premiums, or change Medicare benefit, eligibility, or cost‐​sharing standards.

IPAB is based on a statute that created a commission for determining which military bases should be closed. That model worked reasonably well at insulating base‐​closing decisions from political interference, but it is far from clear that the model will work in making national medical spending decisions, with all their attendant uncertainties and difficulties. As with the payment reforms outlined above, Congress is likely to neuter or eliminate IPAB if it runs afoul of political priorities.

Finally, there is already an independent agency that advises Congress on Medicare policy: the Medicare Payment Advisory Commission, better known as “Medpac.” On a bi‐​annual basis, Medpac issues reports that include numerous recommendations on the same subjects that IPAB will deal with. Just as regularly, most of those recommendations are ignored. Why should we expect IPAB’s recommendations to be more effective than Medpac’s? How likely is it that IPAB can make recommendations that will result in large‐​enough savings to “bend the cost curve,” but do so without creating a political backlash that will cripple or destroy the entire enterprise of reducing “excess cost growth”?

What of the tax subsidy for employment‐​based health insurance? The original House bill contained nothing on the subject, while the Senate bill imposed a 40 percent excise tax on “Cadillac plans” — i.e., plans whose cost exceeded a specified amount ($8,000 for individuals and $23,000 for families). The level was not indexed for inflation, so it would affect a growing percentage of the population over time.

Unions bitterly opposed the Senate proposal. Facing an ultimatum from a core constituency, the White House cut a deal, increasing the threshold for application of the tax and eliminating the tax for five years for benefits obtained through collective bargaining agreement. Those modifications dramatically reduced the likely cost‐​containment effects of the excise tax and created profound horizontal inequity while simultaneously removing an important source of the funding for health reform. Of course, it was simply a coincidence that the compromise provided a very substantial financial subsidy for an important Democratic constituency, as well as an incentive for unionization at a time when private‐​sector union membership had been declining for decades.

When it became clear that this solution was not acceptable, a revised deal was struck. PPACA imposes a tax on all plans beginning in 2018 if their cost exceeds $10,200 for single coverage and $27,500 for family coverage (with higher figures for those in high‐​risk professions). Those figures are also indexed for inflation. Although this structure eliminated the horizontal inequity of the earlier deal, there is a serious question whether a future Congress will allow even these modest back‐​loaded tax provisions to go into effect. So much for “reform” — particularly when the first‐​best solution was to eliminate the subsidy entirely.

The “pillars” of health reform | A useful lens for pulling the implications of these disparate strands together is to examine PPACA in light of the “four pillars of health reform” identified by a group of prominent economists in a letter to President Obama dated November 17, 2009. The four pillars are budget neutrality in the first decade and deficit reduction thereafter, an excise tax on high‐​cost health plans, an independent Medicare commission that would propose cost savings (i.e., IPAB), and delivery‐​system reform. In their letter, the economists hailed the Senate Finance Committee bill for including all of those elements — and the letter was widely seen as a boost to the administration’s reform proposal. However, in a subsequent letter dated December 7, 2009, most of the same economists and some new ones criticized the final Senate bill (which became the foundation for PPACA) for surrendering substantial ground on two of the four pillars (IPAB and delivery‐​system reform).

What of the other two pillars? Leave aside budget neutrality; the Congressional Budget Office is required to score the legislation as written and use a 10‐​year budgetary window, no matter how implausible the included provisions. Not surprisingly, health reform has provided numerous opportunities for proponents to reverse‐​engineer their way to whatever CBO score they want. The results of this process are wholly unreliable as a prediction of actual spending — “fantasy in, fantasy out,” according to the former head of the CBO. Indeed, subsequent reports have made it clear that PPACA’s spending projections (which were the basis for the claim of budgetary neutrality) were implausible on their face. Finally, PPACA’s neutering of the original proposal to tax “Cadillac plans” dramatically undermined one of the few incentives for cost‐​containment in the bill. Thus, judging by the pillars of reform emphasized by the economists in their initial letter, PPACA is deeply flawed, since the provisions that were supposed to make reform affordable were stripped out during the process of enacting the bill.

Lessons from Massachusetts

What, if anything, might we learn about PPACA’s prospects from Massachusetts’ 2006 health reform? Like the federal health reform bills, the Massachusetts legislation focused on access, leaving cost containment and quality for another day. This “dessert first, spinach later — we hope” approach was quite deliberate and was based on the theory that it would be easier to address the cost and quality problems once there was universal coverage. Proponents had no actual evidence to support this theory, but they nonetheless argued that there would be greater urgency in developing and implementing cost control once everyone was covered.

Not surprisingly, when it came time to actually eat the spinach, there was considerable push‐​back. The first step was the creation of a blue ribbon state commission to recommend payment changes that would contain costs and help ensure the delivery of efficient, high‐​quality care. The standard for success had been set so low that one proponent hailed the fact that the commission had been created (before it had actually accomplished anything) as proof that reform was working. When the commission issued its recommendations, it proposed greater use of “global payments” — or capitation, as it was known when it was first tried (and proven to be extremely unpopular) in the 1980s and 1990s. Health care providers promptly condemned the proposals.

Several months later, Massachusetts attorney general Martha Coakley issued a report indicating that provider market leverage was an important driver of increased health care spending. Under “Implications of These Findings for Cost Containment,” the report stated as follows:

One threshold question is whether we can expect the existing health care market in Massachusetts to successfully contain health care costs. To date, the answer is an unequivocal “no.” The market players — whether insurers, providers, or the businesses and consumers who pay for health insurance — have not effectively controlled costs in recent years. If we accept that our health care system can be improved by better aligning payment incentives and controlling cost growth, then we must begin to shift how we purchase health care to align payments with “value,” measured by those factors the health care market should justly reward, such as better quality.

Despite those findings, Massachusetts governor Deval Patrick concluded that insurers were the problem and he proposed caps on premium increases, along with emergency regulations requiring prospective approval of rate increases. The Massachusetts Department of Insurance (MDOI) promptly rejected most of the rate increases that had been submitted by insurers.

Insurers responded by refusing to quote rates for new coverage and filed suit. The MDOI attempted to force insurers to continue to quote rates. While the litigation was pending, emails were produced from the deputy commissioner of the MDOI observing that the premium caps had no actuarial support and were a “train wreck” that could lead to “catastrophic consequences including irreversible damage to our nonprofit health system.” The Massachusetts insurance commissioner responded by arguing that the deputy commissioner was responsible for solvency — not rate approvals — and that the premium caps would not lead to insolvency in the near term. Neither observation was relevant to the issue at hand — whether the rate increases were justified or not.

In the intervening months, a state appeals board reversed some of the decisions by the MDOI and Governor Patrick vowed to appeal. Of course, it was purely a coincidence that Patrick was locked in a battle for re‐​election, and he used the issue to bludgeon his opponent (who had run one of the major nonprofit health insurers in Massachusetts). As of right now, an uneasy truce seems to prevail — but it is hard to see how price controls are likely to lead anywhere, let alone anywhere good.

PPACA does not include explicit premium caps, but there is one provision that is likely to have similar consequences if enforced as written. PPACA requires insurers to spend at least 80 percent (small‐​group and individual policies) or 85 percent (large‐​group policies) of premiums on medical care and health care quality improvement activities. These “medical loss ratio” (MLR) thresholds go into effect in January of 2011 and are quite popular among providers (who believe most health care spending should go to them) and progressives (who believe that insurers should spend less on executive salaries and other overhead, and nothing on the distribution of profits to shareholders, since health care should be a nonprofit business). Insurers that fail to satisfy the requisite MLR threshold are required to pay rebates to their customers.

What effect will these MLR thresholds have on the market for coverage? Legislators clearly hoped that insurers would reduce their overhead so as to avoid paying rebates to their customers. That outcome is likely only for insurers with MLRs that are very close to the cut‐​off. Some insurers will probably try to reclassify administrative expenditures as “medical care and health care quality improvement activities,” but proposed regulations have already closed many of the obvious loopholes. A more long‐​term strategy is to push administrative responsibilities down to providers, since the resulting payments will then be counted as attributable to “medical care.” But the most likely strategy for insurers that simply cannot satisfy the MLR threshold is to be acquired by a larger insurer (as long as there are sufficient economies of scale) or drop out of the market entirely.

Consider what has happened since PPACA was enacted. During the fall of 2010, it became clear that one limited‐​benefit coverage option popular among many low‐​wage workers and college students (the “mini‐​med plan”) did not satisfy another provision in PPACA setting minimum annual coverage limits. Fast‐​food giant McDonald’s submitted a memorandum to federal officials stating it would be “economically prohibitive” to continue offering coverage to its hourly workers if PPACA was enforced as written. Other companies offering mini‐​med plans and other plans with coverage limits below the amount specified in PPACA made the same point. Faced with a large number of entities threatening to drop coverage, HHS issued waivers to more than 110 employers, insurers, and unions who collectively provide coverage to almost 1.2 million workers.

There are obvious rule‐​of‐​law problems with allowing HHS to pick and choose who should receive a waiver and for how long. However, these events prefigured how HHS would handle the roll‐​out of the MLR thresholds, where the problem was far broader than mini‐​med plans. In late‐​November of 2010, HHS issued interim final regulations that effectively exempted mini‐​med plans from the MLR thresholds for 2011 (by doubling their actual MLR before assessing compliance) and required insurers to provide information so that HHS can revisit the issue in 2012. The regulations also provide a process for states to obtain an exemption from the MLR thresholds if they can show that enforcing the thresholds will lead to market disruption. Georgia, Iowa, Maine, and South Carolina have already requested such waivers.

By proactively waiving the MLR thresholds for an entire sector of the coverage market and providing a pathway for states to obtain waivers, HHS avoided (or at least deferred) more bad press about the impact of PPACA on those who already have coverage. At the press conference announcing these regulations, the press release trumpeted that “the new rules will protect up to 74.8 million insured Americans” — but the director of HHS’s Office of Consumer Information and Insurance found it necessary to state twice that “no one is going to lose their coverage” because of PPACA.

These dynamics demonstrate that MLRs mirror the impact of price controls: irrelevant when they are set too high, and eliminating the supply or downgrading the quality of the price‐​controlled good when they are set too low. MLRs might also lead to further consolidation of the insurance market — even though reform proponents insisted that market concentration was part of the justification for PPACA. Finally, the correlation between high MLRs (i.e., 80–85 percent) and high quality of care has never been demonstrated, and the logic of that claim is far from self‐​evident.

To be sure, middle‐​men have never been popular in the United States, and insurance companies are especially unpopular middle‐​men. However, popularity is not the issue. It is far from obvious how price controls that will (best case) have almost no effect, or (worst case) result in fewer coverage options, more uninsured individuals, and a more consolidated insurance market should be viewed as an improvement on the status quo.

Conclusion

When incentives are misaligned, we should not be surprised that the results are not what we wanted — let alone what we hoped for. Stated differently, rewarding “A” and expecting “B” is a recipe for disaster. Yet, American health policy has long been based on exactly that approach.

Of course, this problem is not unique to health care. For almost a century, much of the market for legal services has relied on hourly billing to measure the value of its services. Unfortunately, “when you pay for hours, you get hours” — not all of which are cost‐​justified. But at least we do not allow people to buy legal services with pre‐​tax dollars, nor does government pay for anywhere near the same share of legal services as it does for health care. If we changed those factors, the market for legal services would quickly become almost as dysfunctional as the market for health care.

More broadly, problems will predictably result from the failure to understand how much incentives do matter and how malleable the background circumstances turn out to be, given such incentives. Most of the time, one can “live and learn” from such mistakes — but sometimes the consequences are irreversible. Consider a case study from the 1940s, in Java, as recounted by Bill Bryson:

The distinguished Dutch paleontologist G.H.R. von Koenigswald “had found … [a] group of early humans known as the Solo People from their site of discovery on the Solo River at Ngandong. Koeigswald’s discoveries might have been more impressive still but for a tactical error that was realized too late. He had offered locals ten cents for every piece of hominid bone they could come up with, then discovered to his horror that they had been enthusiastically smashing large pieces into small ones to maximize their income.

To summarize, money matters — meaning that if you get the incentives right, most of the big problems will take care of themselves, leaving a far smaller and more tractable set of problems to be addressed through regulation, litigation, and benign neglect. But if you do not get the incentives right, no amount of speeches, op‐​eds, law review articles, whining and hectoring, moral preening, regulatory oversight, legislation, lawsuits, or lectures about fairness and justice can take their place. Reformers should accordingly focus on getting the incentives right — and legislation that does not address the underlying incentive problem is not, in fact, “reform,” no matter what else it may accomplish.

Equally importantly, clever institutional design can mitigate the effect of politics, but politics never goes away. Congress and state legislatures will not permanently alienate their ability to deliver subsidies to favored groups and the temptations of rent seeking and rent selling are such that any mitigation is likely to be quickly eroded. Reform of the health care system is unlikely to prove an exception to this rule.