Voice and exit in health care policy: what can we learn from the recent controversy over mandated birth control coverage?

Henderson, M. Todd

Birth control was briefly the center of the U.S. health care debate last year when President Obama proposed requiring insurance companies to provide generous benefits for birth control costs. Some religious institutions that oppose the use of certain types of birth control criticized the proposal on grounds of religious freedom. Others characterized the issue as one of respect and fair treatment for women, citing examples of bad outcomes that allegedly flowed to individuals denied insurance coverage for birth control. Name-calling and exaggerated claims of doom and gloom then gripped both sides. But there were larger lessons about health care policy and regulation that were lost in the firestorm about religion and feminism. If we learn the central lesson from this debate, we may be able to improve our regulatory outcomes.

Who Pays for Birth Control?

Putting aside the (large and important) issues of religious conscience and the First Amendment, the issue is a simple one: who should pay for birth control? At first blush, one might reasonably conclude that the individual consuming the birth control should pay for it. In a capitalist economy, individuals generally pay for the things they consume. If the average person wants cable television, the money for it must come from his pocket, not someone else's. To be sure, there are subsidies for the very poor, like food stamp programs, but even the worst off in our society generally bear most of the costs of recurring monthly expenses. This is especially true for rather trivial expenses like birth control (about $10 per month for a generic prescription). This is not to say that there is no one who cannot "afford" birth control; clearly there are such people. But many of the insureds who would be covered by the mandate would not be considered poor.

The big question is whether birth control should be paid for via health insurance or some other means. Insurance seems like an odd fit for a small, predictable recurring expense like birth control. Insurance is usually thought of as a risk-sharing mechanism for unforeseeable expenses. I insure my car against loss, but not the costs I incur to refill it with fuel each week.

Although we might expect most individuals to pay for their own birth control, most insurance companies cover birth control, and President Obama proposed that all be required to do so. Why? There are several possibilities. Importantly, each of them is simply a reflection of the aggregated desires of individuals in the common pool-that is, the insured. Insurance companies provide services their customers value in a way that tries to maximize the value of the business; governments try to do the same for the welfare of society as a whole.

Obviously, those who plan to use birth control would want an insurance company to subsidize the purchase price by making others who were not planning to use it pay some or all of the price. Everyone likes getting something for nothing. But whether they can get it depends on the number of people wanting to use birth control, the number of people not wanting to, and the ability of insurance companies to create viable businesses for just the latter group. For instance, if non-users are not a separable insurance pool and users are important customers, then a cross subsidy might arise. In addition, the non-users may not object to paying for users' birth control because the additional costs to them will be trivial, so long as the ratio of users to non-users is not too high. Most non-users (e.g., men) probably don't even know whether their insurance plan covers birth control; those who do may be married and therefore effectively a user from an economic standpoint.

[ILLUSTRATION OMITTED]

But there is a reason even non-users of birth control might be willing to pay for others' birth control: doing so may be cheaper than not doing so, given that the insurance will likely cover the medical costs of not using birth control. Babies are more expensive than their prophylaxes. Covering birth control may economize on total expected payments and therefore allows the insurance company to offer lower rates. This reason is about the costs of insurance for all insureds, not just expected users of birth control.

Making others pay for an individual's birth control on this ground must be based on an assumption that the individual will not pay for it herself out of her own funds. This is an economically irrational decision if one doesn't want a baby or would suffer negative health consequences as a result of not using birth control. But individuals might overweight a current expenditure and discount future ones (e.g., childrearing), and therefore consume less than the optimal amount of birth control. If others in the same insurance pool believe birth control will be under consumed if not covered by insurance, then they would agree to pay for someone else's birth control in order to lower their own premiums. This is true whether or not one expects to consume birth control in the future. And it should, as a purely economic matter, be independent of one's views about the propriety of birth control. But perhaps those who are opposed to birth control would be happy to pay for the costs of not using it, even at a high level, given that they put such a negative value on paying for birth control and such a high value on producing children.

For those with idiosyncratically strong preferences about the sexual matters of others, whether for religious or other reasons, they can find out whether various insurance plans cover birth control and then choose accordingly. If there are sufficiently large numbers of such people, they could form an actuarially significant pool such that the plan could exclude payment for birth control. And, in any event, the prices of insurance should reflect individuals' willingness to pay for the basket of services offered. If there are individuals who do not want to pay for birth control, the price they pay for insurance may be higher, since there may be more births or other medical costs that result. The price difference between two plans-one that offers birth control and one that does not-should reflect the value of this idiosyncratic preference. The market should price these choices. In a reasonably functioning market-that is, one with competition and information available to purchasers-those who do not want to cover birth control should bear the costs of that decision.

The reason for government involvement is to remedy any market failure. So far we have seen two possibilities. First, if individuals are too poor to afford birth control or are irrational (that is, not making smart economic decisions in the long run), they may under-consume birth control relative to their private and the social optimum. Second, if the market is not providing insurance alternatives such that preferences are priced, then individuals may find themselves either coerced into a plan that does not offer them what they want or those who have idiosyncratic preferences about others' sexual behaviors might not have to bear the full costs of those preferences.

In the first case, the solution is to lower the cost for individuals by having other people pay. If the problem is purely one of money, then this can be done easily by taxing rich people and giving money to poor people. Milton Friedman called this approach a "negative income tax," meaning direct payments to individuals to get them to at least a minimum income level. The idea would be to guarantee income, and then harness the power of individuals acting in markets to deliver the socially optimal level and type of goods and services. This assumes that with the money, poor people would make rational decisions-e.g., pay $10 per month for birth control to avoid having to drop out of school or leave work and bear the huge costs of (unwanted) child rearing. Of course, if the assumption is that poor people are irrational, then this approach might not work. In that case, a better approach might be to artificially lower the cost through an insurance man date to cover birth control. Note two things about this approach, however. First, it still assumes some rationality. A mandate that insurance companies offer birth control does not compel individuals to consume it. Second, unless tied to an income level, the result is not just a subsidy from non-users to poor users, but to all users. Ironically, some of the most vociferous proponents of the mandate were relatively wealthy women.

The second type of market failure could be the lack of competition or the ability of some people to "externalize" costs onto others. Pollution is the classic case of an externality. Ifa factory does not have to pay for the environmental damage it inflicts on a local stream, the factory will produce more than the socially optimal level. Effective environmental regulation is about forcing the factory to pay all of its costs and therefore optimize output. Without law-either regulation of outputs or lawsuits by injured parties-we would privilege those firms or individuals who are best at making other people pay, not those who are the most efficient producers.

This logic obtains for both users and non-users of birth control. If women do not use birth control but can make others in the insurance pool pay for their child birth and other associated costs of child rearing, then they will engage in an inefficient level of production, just like the polluting factory. On the other hand, non-users with idiosyncratic preferences-that is, those who discount these child-related costs because they highly value not subsidizing birth control, or having control over others' sexual practices, or simply having more babies--will only make sensible decisions if the costs of those preferences are fully borne by them.

The best approach to this problem is to create a robust competitive market where insurance companies offer a variety of baskets of services. If there is only a single insurance company in a particular location, all people desiring insurance have no choice but to accept the basket offered by that company. For some it will be optimal; for others it will not, and the losses suffered by that group in terms of satisfaction will be gained by others in the insurance pool or the insurance company. It is for this reason that some health care reform advocates have focused on the ability of insurance firms to sell across state lines and other reforms that would increase market competition.

Who Decides?

We've seen so far there may be a compelling case for either private or public cross-subsidization of birth control. But there is a big difference between the logic and practical effect of insurance company cross-subsidies and government-mandated ones. Governments and insurance companies both offer risk sharing pools. In theory, citizens pay taxes that subsidize others in order to make everyone better off, insured individuals pay premiums that improve the welfare of others in the insurance pool. Two things are notable: First, government mandates will, by definition, have the potential to reduce social welfare in that they tolerate less choice. Imagine there are 100 individuals buying insurance. Half the people prefer a plan that covers birth control and half do not, whether it is for moral, ethical, or economic reasons. If these two groups create a viable risk pool, we can imagine insurance policies tailored to deliver to them their preferences. A government mandate would necessarily make half the people (the ones who prefer no coverage) worse off, while not necessarily improving the welfare of the other half.

Of course, if there are a large number of individuals who desire coverage (whether or not they can pay) who are unable to sort into these two policies, then a mandate might make sense. This would only be true if the gains to the people who get the coverage are greater than the losses to the people who are harmed by providing it. (This could be the case, for example, if society puts a low value on individuals trying to control others' birth control choices.) The point is simply that government mandates tolerate less local variation, which leads to fewer choices and potential destruction of social welfare. The government could mandate that all cable companies show only PBS, and this would likely make some people better off, but it would make many more worse off. Less ambitiously, the government could simply require all cable companies to carry PBS regardless of demand. This cross-subsidy from non-watchers of PBS to watchers of PBS would be justified only if the PBS watchers would be otherwise unable to watch PBS and the gain from doing so exceeded the cost of the cross-subsidy.

The important takeaway here, however, is that any decision, by either a government or an insurance company, to create a cross-subsidy is based on an imperfect assessment of whether or not the transfer is in fact welfare-enhancing. There is a big difference along this dimension between governments and insurance companies. Governments make one-size-fits-all decisions based on the opinions of experts, while markets operate based on the tacit knowledge of hundreds of millions of individual actors. While command-and-control can make sense in some cases, what Fredrich Hayek called the "knowledge problem" will plague any attempts to answer complicated questions based on limited information available to experts. We will return to this issue in a moment.

Voice or Exit?

There is another problem that was revealed starkly by the birth control flap. That difference has to do with what economist Mbert O. Hirschman called the choice between "voice" and "exit." Individuals have these two choices to influence the world around them. First, individuals can exercise "voice," that is, some direct control over the goods or services they are offered. Voting, whether it is for representation or directly on an issue, is the classic manifestation of voice. Individuals-say, parents in a school district-can vote for individuals to represent their interests on the school board or they can vote directly on a particular spending or curriculum issue. This is representative versus direct democracy, and there are arguments for both in various contexts.

Second, individuals can "exit"; that is, express their preferences about a particular policy offered by a particular institution by choosing to leave that institution and have their desires satisfied elsewhere. Customer choice is the classic manifestation of exit. If one has a bad meal at a restaurant, one does not try to get the chef replaced or seek input into the recipes used; one simply goes to another restaurant. Although this seems a silly example, exit is the primary way we shape our lives. Voice is the exception and only used in cases in which exit is unavailable or very costly.

The interplay between voice and exit is crucial to understanding how the birth control controversy implicates larger issues in health policy. In other contexts, we see that voice and exit work as rough substitutes for each other. Where individuals have lots of exit options, then we do not expect them to exercise much voice. Individuals won't demand it and we won't observe it in practice. The restaurant example above shows this point, but it is an easy case. Consider instead a case in which the individuals exercising voice or exit are not just customers, but owners of the institution in question.

Corporate America is such a place. Holders of shares in large American corporations have very little power over how those firms operate, even though shareholders in essence own the company. Shareholders elect the board of directors, which governs the firm, but those elections are more like elections in North Korea than North Dakota. Board members are hand-picked by the chief executive officer and are very rarely replaced. Shareholders also don't have much, if any, say on corporate policies or decisions. Why is there so little voice for corporate owners?

The reason is that exit costs are so low. Ownership stakes in firms are readily transferable at extremely low cost. Shares of large companies are traded in highly liquid public markets, like the New York Stock Exchange, and shares can be sold online through discount brokers like E'Trade for a few dollars per trade. In addition, there is no market for any particular stock; there is just a market for the risk-return trade off offered by each stock, for which there is an infinite number of alternative combinations. Thus, stocks look like restaurants. The rational thing for a shareholder in a company that makes a bad decision to do is to sell the stock and buy another company instead. This can be done for less than $10, compared with the enormous expense and uncertainty involved in trying to change corporate policy.

Some companies, however, do not trade in liquid markets, and therefore selling shares is much more costly. Small, closely held firms are in this group, since ownership is usually concentrated in a family, and there are restrictions on who can sell, when, and how much. As such, we would expect, and do see, much more active roles in governance played by the shareholders in these firms. These shareholders demandvoice because the costs of exit are high.

The consequence for these corporations with illiquid shares and more shareholder voice is conflict. The corporate law casebooks are filled with disputes among rival owners, typically family members, squabbling intensely about the governance choices of small firms, While large, publicly traded firms rise and fall as stock is bought and sold based on individual investor sentiment, small, closely held firms frequently find themselves in court fighting over corporate policy.

Hirschman's logic can be extended to the political realm. One might ask, for instance, why voters seem to care more about elections for the federal government than for local governments. One reason is the relative costs of exit. Although it is expensive to move from one town or state to another, the costs of moving abroad are dramatically higher. The U.S. annual migration rate over the past 20 years has averaged about 15 percent, meaning about one out of seven Americans moves each year. But very few Americans emigrate to another country. Local laws affect property values, and as such are priced by the market. If Illinois raises taxes dramatically, house prices will fall as people move to Texas, where house prices should rise. Although not everyone will move based on such considerations, marginal consumers set prices, and enough bad laws will cause people to move. It is more difficult to avoid federal law, and therefore it is more important to get it right.

For governments, this feature is called federalism. The chief virtue of a federalist model in which sub-national states have much authority over their citizens is competition. In a famous dissenting opinion, U.S. Supreme Court Justice Louis Brandeis described it this way: "It is one of the happy incidents of the federal system that a single courageous State may, if its citizens choose, serve as a laboratory; and try novel social and economic experiments without risk to the rest of the country." If there is uncertainty about some policy, say whether or not government should compel insurance companies to create cross-subsidies for birth control, then allowing states to experiment with different policy approaches reveals information about the costs and benefits of each. This is simply a limited market for government in the spirit of Hirschman's duality. Massachusetts may compel birth control coverage while Maine may prefer to stay out of the business of telling insurance companies what services they must provide. Then we watch and see what happens. Later, Massachusetts, Maine, and all the other states may take information learned from this experiment and update their preferences about what is and what is not good policy.

In contrast, a federal policy is a single experiment that compels compliance by the entire nation. Article VI, clause 2 of the U.S. Constitution dedares federal law "the supreme law of the land," which means it trumps any state law that directly conflicts with or contradicts or impedes the purpose of federal law. While federal experimentation is possible, it is more costly to write rules governing the entire country and change happens more slowly. As the supreme lawmaker, however, the federal government has other options, such as the ability to coerce states into experimenting. In this role, the federal government could set a target say, increasing the percentage of individuals with health insurance and then require states to deliver that target in whatever way they see fit, by a certain date. As leverage, the Supreme Court permits Congress to use carrots and sticks unrelated to health care, such as the withholding of federal funds for highways, education, or the like.

But what can we expect to happen if the federal government deploys the Supremacy Clause of the Constitution to make one federal policy for a particular health care issue?

Two Approaches to Allocating Health Resources

There are two (and only two) ways to allocate all scarce resources: markets and fiat. Market allocations are made using the price mechanism. Buyers and sellers are matched at mutually beneficial terms by reducing their preferences to a single price at which they are willing to buy or sell. The chief virtue of price is that it encourages both producers and consumers of a product or service to reveal information about the value of the goods or services in question.

Fiat, on the other hand, is a mechanism that works based on hierarchy. Those higher up in a particular hierarchy make decisions about who will do what or receive what, and those allocations are enforced by the hierarchy. At the macroeconomic level, this is the approach of planned economies, such as the former Soviet Union, which had "experts" develop five-year plans for the allocation of all the resources in the entire Soviet economy. Gosplan, as it was known, was a state agency staffed by economists, business people, politicians, and scientists from various fields. They were charged with determining the optimal way in which everything from wheat to steel to shoes to health care would be produced and delivered.

One danger of this approach is the fact that, despite experts' confidence in their own views and the public's confidence in them, experts are often wrong. In his book Expert Political Judgment: How Good Is It? How Can We Know? psychologist Philip Tetlock shows how experts often get it wrong, sometimes spectacularly so. The book documents how various decision making heuristics and biases-e.g., the confirmation bias, the saliency or availability heuristic, and so forth-plague experts as much as the rest of us. Experts can be subtlety led to particular outcomes that confirm their hunches, which may not be correct or even supported by the data. One need only pick up the newspaper to see how scientists declare x one week and not-x the next. In the lay press, this is known as the everything-that-was-bad-for-you-is good-for-you problem. Butter is better than margarine, drinking (in moderation) is good for you, so too is chocolate, too much exercise can kill you, and on and on and on. More concretely, a recent news story described how scientists at Amgen failed to replicate 47 of $3 "landmark" cancer studies published "in top journals, fi'om reputable labs."

Errors can be expected, so the question is whether an expert-driven approach is more likely to uncover them and reform them vis-a-vis a diverse approach. Imagine 15 scientists, politicians, and other experts tasked with making all scientific decisions or judging which scientific discoveries to implement into policy. Although such a top-down approach might have certain virtues in overcoming the irrationality of crowds in some areas of science policy (perhaps global warming is such an issue, although perhaps not), this approach probably strikes us as absurd. And yet, as we'll see in a moment, it is precisely the approach to the allocation of health care resources taken by the Affordable Care Act.

Another problem with an expert-driven approach is that it can lead to complacency and dependency. The philosopher Immanuel Kant identified the problems that flow from individuals relying on the decisions of others in his essay "Answering the Question: What is Enlightenment?" Kant argued powerfully against paternalism and rule by expert, claiming that humankind could only be truly capable of achieving its full potential when freed to deploy individual human reason. He described humans, then under the tutelage of church and state paternalism, as like cattle, incapable of thinking for themselves, leading to a perpetual state of immaturity and therefore abuse.

This state of unthinking and blindness is manifest as well in leaders. A well-known chestnut to demonstrate this point is the story of the Soviet ambassador to London who, upon marveling at the abundance of a local bakery, asked to meet the man responsible for ensuring the bread supply in London. Such a question strikes capitalists as absurd, but it was perfectly sensible for someone living under Gosplan. To these people, asking whether government should provide a particular service, like producing, distributing, and selling bread, is completely natural because they cannot imagine it being produced, distributed, and sold without government. (Many in the United States today probably feel this about health care, where the government spends more than all but three other countries in the world.)

But experts should not be discounted entirely, despite the failure of every five-year plan and communism in general. Markets can be inefficient in some cases, and they do not always allocate resources in socially optimal ways. Environmental externalities, as discussed above, are the obvious example. So is research in basic science, which may, for a variety of reasons, be under produced by the market. Where market failures are identified, the question is, what is the best way of contending with the problem? In the environmental area, there are examples of success based on both a price-based approach to regulation and a command-and-control approach. To solve the problem of acid rain in the 1970s and 1980s, the Environmental Protection Agency developed a market for tradable emissions credits for sulfur dioxide that dramatically reduced emissions, resulting in huge social benefits. In contrast, there may be particular pollutants, such as arsenic or mercury, that experts judge to be harmful beyond a certain level, and which are best solved by simply banning the emission of them beyond that amount.

In health care policy, the choice between markets and experts is manifest in the political ideologies professed by different segments of the electorate. Market-oriented reform proposals focus on empowering individuals to make decisions (with cash subsidies for poorer individuals) that will hopefully lead society in the direction of the optimal and efficient allocation of resources. The Affordable Care Act, on the other hand, implements a top-down approach in which politically appointed experts will make decisions about who will be able to buy what health care goods and services. The Independent Payment Advisory Board will consist of 15 voting members appointed by the president and confirmed by the Senate. The board will be responsible for ensuring that the growth rate of Medicare spending does not grow faster than a target rate. It will achieve this by rationing care-that is, deciding what health care procedures, devices, and drugs will be available, and to whom.

This approach can theoretically achieve reductions in spending because the board could make an assessment of various health care alternatives and decide to exclude those procedures that are not cost effective or not supported by strong evidence. I'm sure readers can imagine circumstances in which such an approach could be successful and beneficial at improving health outcomes or reducing costs. But I'm also sure readers can imagine instances in which this approach could lead to worse outcomes. The question is whether we trust these 15 people, subject to some limited oversight, to get the answers right.

The expert model is not unprecedented, as it is used in other jurisdictions with some efficacy. In Germany, private health insurance is common, but an expert agency akin to our new 1S-member board deems certain procedures to be unnecessary and therefore not reimbursable. The Institute for Quality and Efficiency in Health Care (Institut fur Qualitat und Wirtschaftlichkeit im Gesundheitswesen, or IQWiG), which is similar to the National Institute for Health and Clinical Excellence in the United Kingdom, investigates drugs, medical devices, and all medical treatments to determine efficacy. If the treatment is deemed not cost effective, then another group-a committee representing doctors, nurses, insurance companies, and hospitals-decides whether to authorize reimbursement.

According to a senior minister in the German Health Ministry, "most patients and doctors usually accept IQWiG's recommendations." American health care policy expert Uwe Reinhardt, who saw the IQWiG in action and marveled at its effectiveness, told critics of the Independent Payment Advisory Board: "Go to Germany, study [the IQWiG], and you will find that this really works. ... It's civilized."

What We Can Learn from the Birth Control Debate

Can the rationing of U.S. health care spending be civilized? This brings us to last year's very public debate about birth control coverage. It was not civilized, as illustrated by radio commentator Rush Limbangh's insult of mandate proponent Sandra Fluke. The question is why such a firestorm would arise, when in fact almost every health plan covers birth control and the vast majority of Americans have no problem with the implicit cross subsidy.

While birth control or some other drug, device, or procedure might not be salient for most members of the common pool, be it insure& or taxpayers, it may become so when the issue becomes political. By political, I mean that the issue is decided by majority vote, crudely speaking, such that 50 percent plus one of the people set the rule for everyone. W'hen the president sought to make birth control a mandatory part of all insurance plans, this was a political decision regarding health care. This is not to disparage political decisions in general, but merely to point out this feature of them, that they bind those who disagree.

In contrast, apolitical or market decision-making involves individuals choosing what maximizes their own interest. This depends on there being choices that exist to satisfy individuals' preferences. Monopolies can undermine this result. It also depends on the choices of some individuals not harming others; if my decision to swing my fist in the air does not impact your nose, you have little ground to complain; but if it does, then certainly the law should speak loudly to set right the wrong and deter future acts of this sort.

Whether political or not, the common pool always votes. For governments, the voting is clear. For insurance companies, the choice is made individually by buyers of insurance, but their individual choices are aggregated at the firm level to produce a suite of services that will be offered. Looking at it another way, one can vote about what insurance plans should offer, or one can vote about which insurance to buy, thus expressing a preference for what insurance plans should offer. Again, if there are enough people who express a particular view-one way or the other about a particular service being part of the common pool, then it will be offered by an insurance plan. In this way, the competitive forces working on insurance companies result in a more continuous satisfaction of individual preferences than the dichotomous choice presented by a political calculation. There may not be an insurance plan to satisfy everyone's ideal basket of services and price, but there are more of these provided by the market than by political decision making.

With this as background, what we should take from the birth control debate is that while the German experience may suggest that, for that population, political decision making or expert decision making can be effective as a means of deciding who gets what medical care, such a policy is much more fraught in our country. A relatively simple, low cost, and widely accepted practice like birth control became a firestorm when individual choice and local variation were overridden on the grounds of improving social welfare. The airwaves and print media were filled with analysis, name-calling, and hyperbole. Kitchen tables, like my own, were filled with debate about how we should vote about the financing of other peoples' use of birth control. The reason for the intensity of the debate-the powerful expression of voice-was the fact that exit options were dramatically limited. This was a debate about a federal rule that would apply to everyone. Hirschman predicted intense expressions of voice, just as we saw.Just imagine what the debates will look like when the stakes become-as they inevitably will-whether expensive cancer thempies, surgeries, or other procedures will be paid for, or whether more controversial matters like abortion, gender reassignment, and the like will be paid for.

We saw this in a way during the health maintenance organization boom of the early 1990s. For a period, HMOs were quite aggressive in trying to ration care based on their assessments of the efficacy of various health treatments. So-called "managed care" did what it was supposed to do: the only decade since World War II that did not see health care costs increase more rapidly than general cost of living was the 1990s. But the system was untenable because political leaders could not stomach the stories of people harmed when denied care. In addition, choice was not robust among providers, and poor individuals were not given subsidies to purchase insurance. Without a flourishing market and with tragic (but perhaps rational) cases making the nightly news, the system became political and therefore died. Costs have therefore continued to skyrocket, as the system is not designed to deliver efficient care. The Independent Payment Advisory Board is the purported silver bullet, but it is likely doomed by politics, and probably unreliable regardless.

When we vote with our feet and our wallets, our preferences can be satisfied, so long as there is choice and we are not imposing costs on others through our choices. When, instead, matters are decided by experts or politicians, mistakes can be made and made in ways that necessarily are coercive. This coercion does not admit easy exit, as one can exit an insurance policy, especially if done at the federal level. The central lesson is that centralized power over complex matters risks making larger mistakes than decentralized power, admits less innovation, provides for less tailored satisfaction of preferences, and generates greater political conflict. Ironically, those risks may undermine the important work that government must do to improve the world we live in.

READINGS

* Exig Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States, by Albert O. Hirschman. Harvard University Press, 1970.

* Expert Political Judgment: How Good Is It? How Can We Know? by Philip Tetlock. Princeton University Press, 2005.

* "How Germany Is Reining in Health Care Costs: An Interview with Franz Knieps," McKinsey Quartet 3. February 2010.

* "The Independent Payment Advisory Board: PPACA's Anti Constitutional and Authoritarian Super-Legislature," by Diane Cohen and Michael Cannon. Cato Policy Anal3,sisNo. 700, June 14, 2012.

* "Valuing Laws as Local Amenities," by Anup Malani. Harvard Law Review, Vol. 121 (2008).

M. TODD HENDERSON iS professor of law and Aaron Director Teaching Scholar at the University of Chicago Law School. This article is taken from Henderson's paper "Voice and Exit in Health Care Policy," Pathology Case Reviews, Vol. 17, No. 4 (July/August 2012).