MICHAEL CANNON: The problems with America’s health care delivery system, rather than being the result of market failure, arise because government protects low‐​quality providers by stifling competition between different ways of paying health care providers. Rather than allow a level playing field between different payment systems, government tips the scale toward fee‐​for‐​service payment. The federal Medicare program is the largest purchaser of medical services in the world, and it operates largely on a fee‐​for‐​service basis. The federal tax code encourages fee‐​forservice payment and discourages pre‐​payment. State level clinician and insurance licensing laws also discourage or place disproportionate burdens on pre‐​paid health plans. With all of these interventions, government creates a steady stream of revenue for low‐​quality providers and penalizes innovations, including electronic medical records, bar code scanners for prescription drugs that help prevent medical errors, and surgery checklists that would have the effect of preventing unnecessary services.

The problem is not fee‐​for‐​service. Feefor‐ service is as legitimate a way of paying doctors and hospitals as pre‐​payment or any sort of blended payment system. The problem is that we don’t have open competition between all payment systems so that we can realize the quality gains that each can provide.

In the face of these government failures, the response from Congress is to keep digging. The reforms currently making their way through Congress would further stifle quality‐​enhancing competition between different payment systems. For example, the Senate is considering a bill by Sen. Ted Kennedy that would create a new government program with a payment system modeled on Medicare’s. It would lock even more of the marketplace into fee‐​for‐​service payment, resulting in more of the same rather than the sorts of reforms that would improve the way we deliver care.

Improving the delivery system cannot be done by government dictate because the government’s decisions are going to be unduly influenced by those who want to protect the status quo. For example, the government is currently trying to improve health care delivery with rifle‐​shot reforms, like subsidizing health information technologies, comparative effectiveness research, coordinated care, error reduction, and pay‐​for‐​performance. These are all things that open competition between payment systems will promote, but when the government tries to promote them, those efforts are either blocked by incumbent providers or made more expensive than they need to be.

The key to reforming the delivery system is to let consumers’ choices and competition do the reforming. Let consumers control their health care dollars. Let them choose their own health plan. Let them buy insurance across state lines and let clinicians take their licenses across state lines, so that even pre‐​paid plans can compete on a level playing field and consumers can choose a Kaiser or Group Health Cooperative or a fee‐​for‐​service plan. It takes Medicare decades to make even minor changes to its payment systems. Consumers who control their health care dollars can change payment systems in a heartbeat. You get what you pay for, and there isn’t any better way to ensure that we’re going to be paying for quality than if we let consumers control the money.

DAVID HYMAN: I’m going to focus my remarks on the “three Ms” of a public plan: monopoly, monopsony, and maverick.

First, monopoly. Proponents of a public plan argue that the market for health insurance is monopolistic and that a public plan will make it more competitive. Both of those claims are deeply problematic as a matter of competition law. The assertion that the health insurance market is monopolistic is usually based on market concentration statistics computed by the American Medical Association on a state‐​by‐​state basis along with some claims about the number of mergers of health insurers.

There are a few problems with this approach. The first is that counting up the number of mergers doesn’t tell you anything useful. Even if you assume you’re dealing with the same product market, mergers across geographic markets don’t really raise antitrust issues. Mergers within geographic markets may raise antitrust issues, depending on the particulars, but the particulars are important.

Second, although states are a natural unit to look at, because they’re the basis on which we regulate health insurance, the marketplace for insurance doesn’t usually track state borders very well. Market concentration ratios for something that isn’t a market are just meaningless. They don’t help you understand what’s going on in the market. Third, concentration ratios or percentages of markets are just a screening tool to tell you “Gee, we shouldn’t worry about this market,” or, “Gee, maybe we should look closer at this market to try and figure out whether there’s in fact an antitrust problem.” For the past three decades, among antitrust enforcers from administrations in both parties, nobody’s thought that de‐​concentrating a market in the absence of an actual antitrust violation was a strategy that would go anywhere in court or that had much to recommend itself as a general policy.

None of that is to suggest that there aren’t problems with health insurance markets. Nor that some markets might not, in fact, be oligopolistic. But you can’t answer these kinds of questions in the abstract. You actually have to go look and try to figure out what’s happening. If we conclude that there’s actually a monopoly problem in certain markets, we have a way of dealing with it — called the antitrust laws. We file an antitrust suit, we go in to court, we prove our case — or not, if you look at the hospital merger record — and then we use the remedies provided, the principal one of which is structural. You break up the monopoly and restore competition to the market.

As far as I can tell, in the entire history of antitrust, no one has ever thought a plausible response to a monopoly was for the government to go into the business of providing monopolized services. The government is currently investigating Intel and Google, and it previously prosecuted Microsoft for antitrust violations, but I don’t know anyone who thought the correct remedy was for the government to go into the business of developing computer chips, web browsers, and search engines. If you want more competition in the market for health insurance, identify barriers to entry and get rid of them. Don’t assume that creating a situation where the government is both a competitor and regulator is going to accomplish your intended objectives.

The second M is monopsony. If a public plan can rely on Medicare’s purchasing power and pricing — and the Commonwealth Fund, as one of its three options for designing a public plan proposes explicitly that — it can probably underprice private insurance. Two observations. If proponents of a public plan are right that it can do that, then private insurers don’t have the degree of monopoly power in the market that proponents thought they did. And, the degree of monopoly power was the premise for wanting the public plan in the first place. Leaving all that aside, let me just remind people that monopoly and monopsony are bad things. Setting up a monopsony purchaser of healthcare services is just as bad as having a monopoly seller. You don’t want to do that if you can avoid it. Proponents seem to view monopsony purchasing power as a feature when it’s actually a bug.

Third is maverick. The claim is that a public plan will discipline the behavior of private plans, although it’s not quite clear how that will happen. The difficulty here is that if the public plan is subject to the same set of rules and taxes as a private plan and it can’t access government subsidies, it’s kind of hard to see why it is going to behave any differently than any other private plan. It’s important that we have the same sets of rules, not because we have any particular love for the private insurance market, but because the logic of competition is that the outcome reflects people’s actual preferences.

For example, if we subsidized hybrids and tax SUVs, or did the opposite, nobody would think that the resulting purchasing patterns would tell us anything useful about the actual demand for hybrids and SUVs. You need to treat them the same and then look at the outcome and say, “Oh, people really do want a fill‐​in‐​the‐​blank,” as opposed to, “I’ve given you a huge sum of money to buy this and I’m going to tax you heavily for buying that. Which one do you prefer?” Either we have equal treatment on a level playing field or we need to stop pretending this is about competition.

Would a public plan have lower administrative costs? Medicare appears to have lower overhead for a couple of reasons, most of which won’t apply to the public plan we’re talking about. Medicare has a monopoly on the over‐​65 population. It doesn’t incur marketing or advertising costs. Medicare doesn’t form networks and doesn’t do much to control utilization. It does not do as much as it should to control fraud. Presumably none of these will be true for the public plan. Medicare also relies on Social Security and the IRS to do some of its bookkeeping and to collect its premiums, which won’t be the case with a public plan. So it’s hard to see that the kind of magnitude of difference in a public plan will look nearly as big as it does when comparing Medicare to private plans.

The public plan may have some comparative advantage in overhead, but the advantage is not going to be that large. If we wanted to know how large that advantage might be, we ought to look at the overhead in selffunded state plans, which public plan proponents offered as their model. Then we can actually compare apples to apples.

Of course the government plan might not work as hard to avoid high‐​cost individuals, which means it will probably attract a sicker population, eliminating some of the purported cost advantages — unless you risk‐​adjust. The problem with risk‐​adjustment is that it is hard to do right. The challenge is in differentiating whether costs are lower because of favorable riskselection or because you’re delivering higher‐​quality care to a chronically ill population. f you get that wrong you mess up the incentives. And there’s no reason for thinking regulators will favor the “home team,” is there?

AARON YELOWITZ: Several arguments are used to convince people to support employer mandates. Let’s do a little bit of digging into each of them and try to see whether they hold water.

The first is that very few firms are affected. A quote from the California Medical Association in support of California’s “pay‐​or‐​play” mandate back in 2003 said that “Senate Bill 2 was actually a moderate and reasonable step that would affect less than 5 percent of California employers.” But there is a key difference between employers and the number of workers. There are very few large employers, but they employ a lot of people. In California, the top 5 percent of firms employ 61 percent of workers. There are very few Microsofts and Walmarts, which have tons of employees. So, even though it doesn’t affect many firms, a mandate certainly affects a lot of workers, and that will impact costs.

A second argument is that large firms already offer health insurance. Eric Schlosser, author of Fast Food Nation, in support of the pay‐​or‐​play mandate, said “Among employers with two hundred or more workers, 99 percent already provide health insurance. Among those with 50 to 150 workers, 94 percent do.” That leaves the impression that there is maybe one percent or five percent “bad, scoundrel firms” out there that are not doing what they are supposed to do, but that the overwhelming majority of firms wouldn’t be affected.

The problem is that offering health insurance to your employees and your employees taking up the health insurance are two different things. Most large firms are offering health insurance, but they don’t necessarily offer it to all their employees. If you’re part‐​time or seasonal, the odds of getting insurance are much lower. Offering it to full‐​time, full‐​year employees is different from offering it to part‐​time employees. For most firm sizes, around two‐​thirds of employees are taking insurance. Even in firms that are already offering health insurance, a mandate would increase their costs for that final one third of employees.

The final point that I want to make about employer mandates is that, in the California debate at least, there has often been a focus on how much it would cost to cover the uninsured. The mandates do not simply say “You must only cover the uninsured,” but rather, “You must provide ‘Cadillac coverage.’ ” Imagine that the University of Kentucky, which pays 50 percent of my premiums each month, was forced to pay 80 percent of the monthly premiums for my family plan. That is a new, significant cost to the university. There are lots of employees like that in California, lots of people who already have employer‐​sponsored insurance but whose expenses are now being raised. In fact, almost half of the cost of the legislation is for people who already have health insurance and who presumably have come to some kind of agreement with their employer on the right compensation package for them, in terms of wages and health insurance.

It takes some work to try to break out of those sound bites that are easy to say, such as, “All large employers are already offering health insurance to their employees and paying for a good share of it.” But when you look at the numbers carefully, a lot of the claims don’t stand up.

MICHAEL D. TANNER: An individual mandate is a unique and unprecedented violation of individual liberty and choice. But despite its intrusiveness, it is likely to be unenforceable in the long run. The idea that you are going to track down every undocumented alien, every homeless person, every mentally ill person, people who change jobs, people who move in and out of a state, and find out if they have insurance and then penalize them for failing to get it is unrealistic.

An individual mandate is also the first in a series of dominoes that would almost inevitably lead to greater government control of health care. If you are going to have a mandate for insurance, it will have to be heavily regulated and heavily subsidized. You have to define, for example, what insurance meets the mandate. Once you start down this road to mandating what this product that everyone has to buy will be, you create a special interest bonanza, as every interest group, provider, and disease constituency demands to be included in the product. As they are included, the costs rise both in terms of premiums and the subsidies necessary to keep this affordable for people. As the premiums and subsidies rise, the public demands cost controls, and you begin to put in premium caps or other forms of cost control containment.

The primary reason we’re told we need to have an individual mandate is to get people insured to deal with the problem of uncompensated care. But let’s keep it in perspective. The cost of uncompensated care is actually about 2.5 percent of total health care spending. It is a much more manageable problem than is commonly believed. We should also note that the imposition of mandates does not necessarily eliminate uncompensated care. We haven’t seen an elimination of uncompensated care in Massachusetts. In fact, the hospitals there say that they still need their subsidies for uncompensated care.

We are also told that we need to have an individual mandate in order to bring more young and healthy people into the pool, which will lower premiums for everyone. That’s true only in so far as you prohibit actuarial underwriting of insurance. If people are underwritten on the basis of their health, it doesn’t matter whether you have young and healthy people or old and sick people in the pool. Everybody’s premium is based on their own health.

If we want to bring young and healthy people into the pool, reducing the cost of health insurance by eliminating things like community rating — which drives up the cost of insurance — might help. You would think that if you want people to buy a product, creating legislation that drives up the cost of that product isn’t a good way to do it. Yet we do things that make insurance more expensive for young, healthy people to buy, and then we’re surprised when young, healthy people don’t buy insurance. New York State was a classic example. When they introduced community rating, some 500,000 people, mostly young and healthy, dropped their insurance because of the increase in premiums. There are ways we can do this better.