(Archived document, may contain errors)

193

June 14,1994

REMRMUG HEALTH INSURANCE: ANALYZJNG OBJECT[ONS TO TME NICKLES-STEARNS BELL

INTRODUCTION

A graduate student in theoretical physics once produced a paper showing that it is aerodynamically impossible for bees to fly. "This is an excellent and completely persuasive paper," said his professor. "There is only one little problem-bees do fly. There is an equivalent of this salutary tale in today's health care reform debate. It is the theoretical argument advanced by some scholars that because of a phenomenon known as "adverse selection," Americans cannot be permitted to choose the health insur- ance benefits they want, or change plans. Adverse selection occurs when sicker people flock to some insurance plans, raising their costs far above their premium revenue. Ac- cording to the argument, if Americans are allowed to choose what plans and benefits they want in a market with any limits at all on the insurer's right to adjust premiums, adverse selection will cause the insurance market to collapse. This argument leads many scholars and lawmakers to the conclusion that Americans must be forced to buy a standardized benefits package and denied the right to choose the particular benefits they want. This one-size-fits-all benefits package is a central feature of "managed competition," which is the basis of health reform measures offered by Presi- dent Bill Clinton, Representative Jim Cooper (D-TN), Senator John Chafee (R-RI), and others. I Some conservative scholars raise precisely the same adverse selection objection as the proponents of managed competition. They maintain that adverse selection would destroy any insurance market in which families switch to the plans with benefits they want if there were any restrictions on the right of insurance companies to include all risk factors when quoting premiums. 2 Nothing short of pure risk rating is stable, according to this the- ory. The solution to adverse selection advocated by these scholars is for insurance compa- nies to charge sicker families much higher premiums. But this, of course, would lead to many high-risk families being priced out of the mar- ket-an unacceptable result in today's political climate-so these scholars are forced to modify their position. The problem for this group of conservative scholars is that their modifications lead toward a "solution" that the vast majority of other conservatives, and indeed -most other Americans, would find unacceptable. First, they would create an unstable dual insurance market which is part unregulated and part heavily regulated. These scholars would allow insurance companies to charge new enrollees the market premium, taking full account of the family's health status and all other risk factors. Current enrollees wishing to renew their policies, however, would enjoy premium increases held, by law, at the average for the insurer's portfolio of busi- ness. This dual market would be unstable. Insurers would try to dump costly policyhold- ers while telling young Americans with a family history of illnesses that they must pay substantially higher premiums than their peers if they want a plan. Second, they would create a new, subsidized health program for high-risk Americans run by federal and state governments. These scholars would create a new "risk pool" pro- grain for sicker Americans facing unduly high insurance costs. The dilemma is that either this new program would be an entitlement for all income groups meeting certain criteria or many Americans still would not be able to buy insurance because they would not be in the pool. Further, it would mean either general revenue funding by the federal govern- ment, or-more likely-a potentially huge new unfunded mandate on the states. Accord- ing to estimates used by some advocates of the risk pool idea, this program could become available for as many as 10 million Americans. 3 In addition, they call for a new "health safety net" program to pay for the essential expenses of uninsured working people once they have spent their above-poverty income and used up their assets. Notwithstanding the dangers implicit in such solutions, there is one other small prob- lem with the theoretical argument about adverse selection: In the real world, a health sys- tem exists which does allow Americans to switch to plans they prefer, does require insur- ers to ignore health status and other factors when setting premiums, and is not rendered unworkable by adverse selection. It is called the Federal Employees Health Benefits Pro- gram (FEHBP). It covers Members of Congress and their stafL Administration officials, and over nine million federal workers and retirees. It has proved stable and effective for 34 years. 5 Like the bee, despite the theory it actually flies. An improved version of the FEHBP forms the basis of the Consumer Choice Health Security Act (S. 1743, H.R.3698), introduced in the Senate by Don Nickles (R-OK) and in the House by Cliff Steams (R-FL). 6

THE EXAGGERATED PROBLEM OF ADVERSE SELECTION The Nickles-Steams bill has attracted strong criticism from advocates of managed com- petition, as well as from a group of conservatives, in part because it contains a provision that would allow Americans to switch health plans easily if they were dissatisfied with a plan's level of service or wished to obtain a plan with different benefits. Under the bill, insurers would not be permitted to take health risk (that is, pre-existing conditions or fam- ily medical history) into account when setting premiums. They could set rates only based on age, sex, and geography (with discounts available for group purchases and healthy life- style), a rating system known as "limited underwriting." Other features of the bill, includ- ing tax reforms, would ease the after-tax cost of insurance and out-of-pocket medical costs for millions of Americans. Thus, under the Nickles-Steams bill, if the wife in a family developed breast cancer, the family could switch to a plan with more extensive or better quality cancer treatment without a premium "penalty" for the wife's health status. Similarly, a young couple who conceive a child and then discover the fetus has a severe abnormality would know that they could obtain insurance coverage for their future child at reasonable cost. They would not face the horrifying choice between an abortion or full risk-based insurance that might take most of their available income. Thus, unlike today, or under estimated risk rating, a higher-risk family could always count on obtaining the coverage they need under Nickles-Steams without their medical condition making the price prohibitive. Managed competition advocates argue that this freedom of choice in a competitive in- surance market is inherently unstable. According to Stanford Professor Alain Enthoven, one of the architects of managed competition, there are two related and fatal problems. One is adverse selection, in which sicker people would gravitate to the plans offering the best coverage, thereby driving up their costs and making them uneconomic. Mean- while, insurers would try to "cherry pick" as they market their plans-that is, avoid high- cost individuals. The result, he says, would be an unstable market. The second pattern, which Enthoven argues would compound this problem, is"market segmentation." This means insurers would offer a combination of benefits designed to appeal only to a dis- creet, profitable share of the market. Enthoven raised these objections against the Nickles legislation, as well as the FEHBP, in recent congressional hearings. Said Enthoven: One problem is adverse selection, the good risks and the bad risks. The other problem is what is called segmenting the market. It is something we teach in business school. That is, if you do not want to compete you find a little niche where nobody else is and you get in there and make your product and you jack up your price. And if we as public policy people are trying to make the market competitive, then we want to combat market segmentation.... Plan A has wonderful vision care and no podiatry; Plan B has wonderful podiatry and no vision care. And that is anti-competitive because -those people who have bad eyes and good feet, they do not say "Am I going to join Plan A or Plan B based on price..." They just say, "This is the plan that covers my problems of bad feet and good eyes," or vice versa. 7 Enthoven maintains that allowing such competition would lead to what he graphically calls a "death spiral" of adverse selection within the insurance industry. This is the basis of his objection to the Nickles-Steams Consumer Choice Act. Like other supporters of managed competition, he proposed to deal with adverse selection and market segmenta- tion by limiting consumer choice to a range of plans with precisely the same services (all with eye care and podiatry-or none with these services), with these plans competing solely on price and quality. Some conservative scholars raise exactly the same objection. Echoing Enthoven, they maintain that if insurance companies cannot charge enrollees a premium fully in line with the anticipated cost of covering the individual, the market is unstable. 8 They say that such an unstable situation would occur if plans are "community rated" (that is, required to charge exactly the same price to all in an area, irrespective of any risk factor) or even if plans were prohibited only from including health factors in premium-setting, such as in the Nickles-Steams bill. Their argument is that relatively sicker people would cost the insurer more than the pre- mium it could charge in a competitive market if medical condition must be ignored in set- ting price. A competitive plan could not charge a healthy person significantly more in pre- miums than he would expect to use in services, or the person would switch plans or drop insurance. So to avoid heavy losses, there would be a remorseless incentive for compa- nies to try to avoid loss-making sick customers, drive down the quality of care for such customers on their books, and seek out profitable healthy customers. Meanwhile, sicker families would seek out the plans with the best services, causing these better plans to be swamped with costly enrollees and either to fail in the market or to cut quality and serv- ices in order to drive away people who need expensive care. The argument is identical to Enthoven's and paints a bleak picture of insurance market chaos.

THEORY MEETS REALITY The nightmare scenario laid out by Enthoven and others has a plausible logic to it, and patterns of adverse selection do exist in health insurance, lending some credibility to the "death spirar' theory. But, ironically, when Enthoven was testifying before Members of Congress and staff, he was talking to a group of people who are in a system which, ac- cording to the Stanford professor, cannot exist because of adverse selection and market segmentation. But it does, and it is stable and works well. It is the Federal Employees Health Benefits Program (FEHBP). Moreover, the Nickles-Steams legislation criticized by Enthoven, which is based on the FEHBP, makes crucial changes in the pricing of in- surance which would reduce significantly the degree of adverse selection that does exist in the FEHBP. It also would widen the choice of insurance plans and benefits available to ordinary Americans and enable all working families to obtain coverage that is right for them. The Federal Employee Health Benefits Program in theory should not exist, because it is an open invitation to the instability described by Enthoven and others. Under the FEHBP, the premiums charged to federal workers by competing plans must be commu- nity-rated. Thus a 19-year-old, perfectly healthy congressional staffer pays exactly the same premium for, say, his Blue Cross plan as a 60-year-old federal retiree with chronic heart disease who chooses the same plan. Once a year, each worker or retiree can change plans without restriction, and the new plan takes effect without any waiting period. More- over, plans may offer different medical benefits, and federal workers are free to pick the plans with the benefits that best meet their medical needs in a particular year. It is difficult to imagine an arrangement that should be more prone to the destructive pressures of adverse selection and market segmentation. According to theory, the FEHBP should have collapsed in ruins. But it has been serving federal workers well for 34 years. To be sure, adverse selection and segmentation do occur and do cause headaches for some insurers, and reforrns to make the system function better are long overdue. 9 In par- ticular, the requirement that the young and the old must be charged the same premiums poses some problems for high-benefit plans and should be changed. At the very least, FEHBP premiums should be related to age. But even with the perverse effects of today's community-rated premiums, private insurers consider this degree of adverse selection merely a price of doing business in the highly competitive FEHBP market, not the guar- antee of failure the theoreticians claim. Both buyers and sellers of FEHBP policies are quite satisfied. Enrollees, in fact, are so satisfied with the FEHBP that they have been lobbying hard to be kept out of virtually every "reforrif' plan proposed in Congress for the last several years, including the Clin- ton bill. Rather than scrambling to leave the FEHBP because of adverse selection, private insurers serving the system also have been lobbying hard to keep the program out of the 10 - Clinton-proposed system for as long as possible. Rejecting the lifeline of reform is curi- ous behavior indeed for insurers supposedly locked in a desperate "death spiral."

HOW NICKLES-STEARNS CORRECTS THE FEHBP'S PROBLEMS Adverse selection and market segmentation are merely irritants in the FEHBP insur- ance market, not fatal flaws. But the FEHBP performs less well than it should because of its community-rating principles. Community rating introduces an unnecessary degree of distortion and inefficiency by preventing insurers from pricing their policies in line with the risk they must accept. This is why the sponsors of the Nickles-Steams bill included important insurance reforms to reduce these problems in their proposal based on the con- sumer choice principles of the FEHBP. Thus, while the Nickles-Steams bill includes the right of families to pick a plan with the package of benefits that is right for them and al- lows them to switch plans if their needs change (as the FEHBP does), it includes crucial reforms that depart from the premium-setting principles in today's FEHBP. Thus, the market created by the Nickles-Steams bill would be even more stable, and much more ef- ficient, than the FEHBP is today. There are three key changes in the way insurance would be sold under Nickles-Steams, compared with the FEHBP. 1) Premiums would reflect most risk factors. The FEHBP is community-rated, and plans must accept all applicants, meaning every- one must be charged the same in a particular area for the same coverage whatever their in- surance risk. Under Nickles-Steams insurers would be able to adjust premium prices ac- cording to age, sex, and geography. I I Health status, however, still could not be consid- ered. But health insurers report that the risk factors allowable under Nickles-Steams un- derwriting provisions actually account for as much as 70 percent of the cost of serving different groups of enrollees, so insurers would be far less subject to adverse selection problems than they are under the FEHBP. Moreover, most Americans with employer- sponsored insurance have community-rated premiums within their workplace; that is, the young and healthy and the old and sick pay the same premiums if they work for the same firm. So this Nickles-Steams reform would shift today's whole market much closer to risk-based premiums. 2) A waiting period could be applied. Under the Nickles-Steams bill, enrollees would have to be covered for pre-existing con- ditions when they first bought insurance or changed plans. But insurers would be permit- ted to impose a waiting period before the coverage for these conditions takes effect if the person was purchasing his first insurance policy, or was buying a plan after a spell with- out insurance, or was upgrading his coverage. This is the normal practice today in em- ployer-provided insurance and other private plans and is considered quite sufficient by most insurers to protect them from this aspect of adverse selection. A waiting period is designed to reduce the potential danger to an insurer of families switching plans for a few months to get the best care for a particular ailment and then switching back to a less comprehensive plan. Since the enrollee is paying apremium but not receiving full coverage during the waiting period, it is a form of "entry fee," or like paying points when changing mortgage companies. By imposing an extra cost to obtain coverage or change plans, it discourages families from switching to "game" insurers. On the other hand, while a waiting period places some limits on the buyer of a policy, the policyholder knows he or she will have insurance coverage by a certain date. The FEHBP does not impose any waiting period, although in practice a form of waiting period of as much as a year can apply since federal employees can switch plans only at the end of each calendar year. Insurance companies report that this form of adverse selection is reduced to quite man- ageable proportions if a waiting period of between three months and a year (depending on the illness) is imposed. The Nickles-Steams bill links such waiting periods to its lim- ited underwriting requirements. So if a child in an uninsured young family is, say, diag- nosed with a chronic heart problem, that family could obtain coverage-even though they would have to wait some time before the insurance became fully effective. Today that family probably would be unable to obtain a policy to cover the heart condition, ex- cept perhaps at a prohibitive price, unless they gained access to a community-raised em- ployer-sponsored plan. 3) An insurance "pass-back" system would be established. The third protection against adverse selection in the Nickles-Steams bill is a pass-back provision. This is a mechanism designed to allow insurers to spread among themselves the risk posed when individuals with pre-existing conditions decide to change plans. It works like this. Say an individual diagnosed with cancer decides to switch coverage from insurer A to insurer B because B has better coverage for that condition. Insurer B must ac- cept him and offer him a premium based on age, sex, and geography but not adjusted for his higher health risk. Insurer B can impose a waiting period, as discussed above, which gives the insurer some protection against losses; but with the pass-back provision, Insurer B also can remit part of the premium to insurer A and require insurer A to cover costs as- sociated with treating the pre-existing condition. This gives additional protection against adverse selection to insurer B. It also, incidently, reduces the incentive on insurer A to try to "dump" the costly individual on another insurance company. These provisions would make the insurance market even more stable under Nickles- Steams than under the FEHBP. It would do so by eliminating community rating and other features of the FEHBP program that lead to an unnecessary degree of adverse selec- tion while assuring wide choice and flexibility for families seeking coverage. Why Good Plans Won't Fail Under the Nickles-Steams system, families would have an incentive to seek plans with the best services for their particular needs (as they do under the FEHBP). Thus, a plan boasting the best cardiologists in the region would tend to attract a disproportionate share of individuals with heart trouble. But there is nothing unstable in this, any more than fixed-price "all-you-can-eat" restau- rants are unstable because they attract people with big appetites. Plans emphasizing par- ticular services or high quality care would raise prices to a stable, market-clearing level to reflect the higher demand for those services. Thus, even though health risk could not be included as a risk factor in setting particular premiums, the pren-dum would still reflect the anticipated increased risk of enrollees as a whole. The insurance principle of risk spreading would still operate because both patient and provider would rarely be able to predict the final cost of any particular case. In the real world, all-you-can-eat restaurants are not caught in a death spiral of adverse selection. They price themselves to avoid that, and not all their customers are gluttons. In the real world, health insurance companies also would price themselves sensibly and would not attract only people who consume all the services available. The reasoning of the "death spiral" theory should apply to today's mortgage market, in which borrowers simply can refinance if rates fall. Adverse selection theorists should con- clude that today's mortgage industry is inherently unstable. The reason? Borrowers can switch to lower-rate 30-year fixed-rate mortgages whenever rates fall. So adverse selec- tion in today's fiercely competitive market in theory should lead to mortgage companies ending up with only two types of borrowers: those who have recently refinanced to re- duce their payments, and whose payments are close to the mortgage company's cost of money, and those who obtained a lower rate at some earlier time which is now below the current cost to the mortgage company of providing a new mortgage. The result: little profit on the first group and potentially huge losses on the second. Adverse selection the- ory thus shows that a competitive mortgage market cannot exist if borrowers can refi- nance. But just as bees fly, a mortgage industry actually exists. The real world is different from the theoretical world. A stable mortgage market exists because there are certain costs to boffowers from changing mortgage companies (points plus paperwork), and this discourages switching to achieve small payment reductions. Similarly, there are costs to policyholders from changing health plans to get better coverage (waiting periods, prob- able premium increases, and paperwork) that would maintain a stable market under the Nickles-Steams reforms.

THE DANGEROUS "SOLUTIONS" TO ADVERSE SELECTION Those who still maintain, despite the evidence, that adverse selection would render lim- ited underwriting approaches unstable generally support one of three solutions. The se- vere problems associated with two of thern-namely the single-payer system (which eliminates private insurance) and the managed competition approach (which outlaws a choice of benefits and retains community rating) - - have been discussed before in Heri- tage Foundation publications and elsewhere. 12 The other proposed solution, recommended by some free market economists, also turns out be very unattractive and has some very surprising features. While the specifics of this solution are unclear, since no version has been laid out with any detail, the general shape of the approach can be summarized as follows.

1) The health market would be more efficient, and medical providers would better serve the patient, if Americans paid directly for more of their health care and relied less on insurance.

Most conservative economists, including proponents of the Nickles-Steams bill, sup- port this feature of reform. 13 These economists point out that the tax code and other ele- ments of today's system encourage inefficient overinsurance. Legislative proposals to en- courage greater direct payment include tax-free medical savings (or "Medisave") ac- counts and neutrality in the tax treatment of insurance premiums and out-of-pocket medi- cal costs. Both of these elements are contained in the Nickles-Steams bill, as well as in other pieces of legislation. Some bills would restrict medisave contributions well below those contained in the Nickles-Steams bill, or limit the use of medisave funds only to the deductibles or copay- ments in an insurance plan. Under Nickles-Steams, medisave funds could be used for any legitimate medical purpose. 14 Moreover, unlike most other bills incorporating medisave accounts, the Nickles- Steams legislation permits the same tax relief for out-of-pocket medical spending (whether or not the payments are made out of a medisave account) as for insurance. To- day, the tax code effectively limits tax relief to insurance benefits for virtually all Ameri- cans. Thus, by making the tax code neutral between cash payments and insurance, the Nickles-Steams bill gives a stronger tax encouragement to all families to pay for more of their health care directly, not just if the family has the foresight or savings available to open a niedisave account. Curiously, the strongest conservative critics of the Nickles- Steams bill would not give tax relief to families who pay directly for care rather than file an insurance claim unless that family first had opened a medisave account. This restric- tion would encourage overinsurance among moderate and lower-income Americans, who are less likely to open medisave accounts. 2) Fully risk-rated insurance premiums would be encouraged. With an understandable hostility to premium regulation of any kind, some free-market economists reject all limits on underwriting. They maintain that the only defensible and stable insurance market is one in which insurance companies can charge a market pre- mium based on the actual risk to the insurer posed by the applicant, such that the pre- mium collected would be very close to the estimated additional cost and risk of serving that enrollee. In advocating full risk-rating, supporters ask such rhetorical questions as "Why should a heavy smoker, or a person who has contracted AIDS because of his lifestyle, enjoy the same premium as someone who maintains a healthy lifestyle?" There is a lot to be said for that view, and the Nickles-Steams bill does allow premium discounts to promote healthy behavior.

But the advocates of strict risk-based insurance go much further than these lifestyle cases. They would remove restrictions on insurance companies charging premiums ad- justed for any risk factor, including family medical history. Thus, under the system they advocate, a lawmaker voting for risk-based insurance would have to tell a constituent with cancer contracted through no fault of his own, or a mother who contracted AIDS be- cause of a blood transfusion after childbirth, or even an individual with a family history of cancer that if they wanted to buy insurance they would face significantly higher premi- ums than individuals not posing such a risk. Similarly, lawmakers would have to tell a young constituent with a chronic heart condition that paying a lot more for health insur- ance than her peers is just part of life; and they would have to wam all their constituents that as health insurance companies refine their techniques for calculating risk, such as gene analysis, even seemingly healthy people in the future could face high premiums- or be classified as uninsurable-under a pure risk-based system. This is how insurance broadly works in other areas. In automobile insurance, for in- stance, a driver with a bad driving history may well face prohibitively expensive insur- ance premiums or be turned down entirely, although he can at least avoid that cost by giv- ing up his car and taking the bus. But Americans are very reluctant to accept the notion of full risk-rated premiums if that means, as it does, that many ordinary, hard-working families with impeccable lifestyles would risk ruin because they could not afford medical insurance. The fear of many middle-class families today that at some time they will not be able to purchase affordable individual insurance helps fuel the political pressure for the Clinton Plan. Recognizing the problem posed by full-risk rating to high-risk families on modest in- comes, its advocates generally propose two modifications. Unfortunately, these are not solutions; they simply lead to more problems. Modification #1: Under the first exception to full risk-rating, policyholders would be given the legal right to renew their existing policies each year at a regulated price, but new enrollees in a plan would be charged the full risk-rated premium. In an unregulated market, when a health insurer discovers in a policyholder a new, heightened risk-perhaps a checkup revealed cancer or diabetes-that insurer has a strong incentive to raise the premium substantially or to drop coverage and usually does so. The same is true if a change in the health status of an employer-based group increases the insurer's risk. Many individuals and small businesses have experienced drastic in- crease in premiums or the loss of coverage because they or a member of their community- rated employment-based insurance group are discovered to have a potentially costly dis- ease. Americans understandably do not like this one-sided system of risk in which only the policyholder is vulnerable over the long haul. They want it changed so that families have greater long-term security in their health insurance. It is this anxiety among Americans in the real world that led supporters of the Nickles-Steams bill to support limits on the un- derwriting (risk calculation) used to set premiums. Under the Nickles-Steams bill, insurance companies would be obliged by law to renew coverage and to disregard health status in setting premiums. Other bills and proposals contain similar provisions. Because of these limited underwriting provisions, Nickles-Steams would still allow a sick family to obtain and renew coverage and would allow a higher-risk family to switch to another plan if they wished to do so. Even if their health deteriorated, the family would not be denied affordable coverage; nor would they have to stay with the same carrier if the insurer tried to cut levels of service to keep making a profit. While some conservatives attack this limited underwriting, they are strangely willing to accept tight price regulation and other requirements on insurance companies in pricing their existing business. But they insist on not applying the same regulation to the pricing of new policies. To defend this acceptance of price controls, they make the seemingly rea- sonable argument that it is better to regulate part of the insurance business than all of it. But by making this argument, rather than arguing for light underwriting limits on all health insurance, they overlook the destructive incentives for insurers in a mixed, regu- lated/unregulated market. Moreover, they must face the likelihood of very heavy regula- tion of the existing business segment of the insurance market. Without realizing it, these critics of the Nickles-Steams bill in practice would leave families virtually at the mercy of the health insurer. Consider a system in which insurers faced premium restrictions on renewals but could charge the market, risk-related rate for new enrollees. Unusually sick existing policyholders would mean a loss for the company; but if the same individual or family applied as a new policyholder, the insurer could charge the higher, profitable market premium. The insurer thus would have every incen- tive to reduce his higher-risk "old" business subject to regulation and replace it with un- regulated "new" business. Given the practices of insurers in today's markets, there could be little doubt they would use every trick and business strategy available to encourage ex- isting high-cost policyholders to drop coverage or pay a much higher rate. Because of the powerful financial incentive on insurers to do this, the federal and state governments would be pressured by citizens into introducing strong regulations to prevent "unfair" business practices. The insurer also would be vulnerable to adverse selection by policyholders. Policyhold- ers whose health record turned out to be better than the average would have the incentive to switch to another company as new enrollees and "lock in" at a lower premium. Insur- ers thus would tend to stick with their higher-cost business while always liable to lose their profitable policyholders. By contrast, if all business is subject to mild underwriting limits in premium setting, as is the case in the Nickles-Steams bill, there is little incentive to chum business in this way because old business and new business would be charged the same premium. 15 There are many other problems inherent in the dual regulated/unregulated market idea. Among them: V Welcome to "insurance lock." Under the perverse incentives in a regulated/unregu- lated market, many policyholders with deteriorating health would find themselves hav- ing to put up with the worst care the insurer could get away with as that insurer tried to keep their costs below the regulated premium or get the policyholders to drop coverage. The sicker family probably could not change policies if they had topay an unregu- lated risk-rated premium with the new carrier, so they would have to put up with the bad service. In other words, the family would experience what might be called "insur- ance lock," in that it could not leave an insurer giving poor service because of the pro- hibitive cost of moving to another carrier. Families also could face a severe financial blow or be forced to drop their insurance if they had to move to another area of the country where they would have to apply for insurance as new applicants at fully risk- based premium rates. This "insurance lock" would be very similar to today's "job lock" phenomenon, in which many sick or older workers feel unable to change jobs because they might not be able to qualify for health insurance in another firm or if they became self-em- ployed. Job lock is one of the features of today's system with which reform is sup- posed to deal, but replacing job lock with insurance lock is not much of a reform. Watch out for tier-rating. In this dual regulated/unregulated market, policyholders staying with a company would be particularly vulnerable to a phenomenon known as "tier rating," which is common in today's small group market. This practice involves splitting a group by offering healthy members a discount for switching to another pol- icy, leaving the sicker policyholders in the original group. Then the price of the policy can be raised substantially while still charging no individual a renewal price higher than the average for the group. Eventually the sicker individuals wishing to renew either are forced out off the insurer's books or end up paying close to the unregulated, risk-based rate despite their supposed right to renew at a regulated price. Make sure you overinsure - and don't pick the wrong plan. This mixture of regu- lated and unregulated market puts enormous pressure on responsible young families to overinsure so that they are covered in advance for every costly eventuality in later life. If families are not protected by limited underwriting, they would be vulnerable to any unforeseen drastic change in their health status because they could find the risk-rated cost of future insurance prohibitive. Moreover, since only renewals of a policy are subject to premium limits in the dual market idea, a young family had better choose wisely when buying its first plan. A family which develops a severe medical problem, say the birth of a child with a costly handicap, may never be able to switch to a new plan providing better care for the child. They will be bound to one insurer for life, so they literally cannot afford to make a mistake when choosing their first plan. Expect sticker shock. In the regulated/unregulated market, even young healthy fami- lies will encounter a few catches as they try to insure themselves against a change in their health condition that might not occur for many years. One is that if an insurer is required to renew policies at a controlled price that cannot be adjusted each year for changes in health status, it faces a higher potential risk when signing up a healthy young person, since an unusually costly sickness during any year could mean heavy losses for the company. This will mean significantly higher premiums for young peo- ple than today, making insurance less attractive to the young-and inviting financial disaster in the future. The reason for this higher cost for the extra risk associated with multi-year health in- surance is the same as the reason multi-year fixed-premium life insurance is much more expensive than yearly-renewable insurance. It is also why the interbst rate on a 30-year mortgage (where the mortgage company risks future interest rate hikes) is much more than an adjustable rate mortgage (where the buyer takes on that risk). Another catch is that controlling the premiums of renewed health insurance means in- surers would be far more careful about calculating the future risk posed by an apph- cant. Many young individuals in perfect health thus are.likely to face high premiums because their family health history indicates they may be at higher risk in later life. V Expect many more abortions. Consider a young, healthy couple with a modest in- come who opt for very basic health insurance or no insurance, given the cost of cover- age. Say the couple conceives a child but early tests reveal a severe handicap that would require permanent and costly services once the child is bom. Under the Nickles- Steams bill, that family knows that it could at least purchase additional coverage at rates that would not reflect the unborn child's unusually high future medical costs. But if the family had to buy coverage on the open, risk-based market as a new enrollee, the cost almost certainly would be beyond its means once the insurer received medical records indicating the condition of the fetus. The family faces ruin. The incentive to have an abortion in such circumstances could be overwhelming. It should also be noted that, in a sense, advocates of pure risk rating envision no "re- demption" for someone who makes an unwise decision and then runs out of luck. The young person who takes a chance and goes without insurance only to encounter a severe medical problem that forces him to seek coverage faces a much higher premium for life. Under Nickles-Steams, that family would have to pay a premium while waiting for cover- age to take effect (their "penalty" for the unwise decision), but then would pay the same premiums as their neighbors of the same age with the same coverage. These inherent problems associated with mixing regulated and unregulated businesses in the same market are not unique to health care. Perverse incentives leading to a deterio- ration of service in the regulated segment have been experienced in many markets. When housing units are rent controlled for an existing tenant but uncontrolled for new tenan- cies, as they have been in many U.S. and European markets, landlords simply cut mainte- nance and use other devices to "persuade" existing tenants to leave. And in the Medicare program, when in-hospital services became subject to price controls in the 1980s, hospi- tals immediately began to reclassify procedures as physician services or "out-patienf ' pro- cedures so that they fell into the unregulated sector. With existing health insurance cover- age subject to price controls but new coverage uncontrolled, insurers very quickly would begin to specify coverage and diagnoses with precision and to insert fine print requiring unregulated "new" coverage premiums for a host of ailments that policyholders thought were already covered. Thus, trying to have a mixture of regulated and unregulated prices in the same health insurance market would lead to an unstable insurance market with many disturbing side effects. In this dual market, there would be deteriorating service in the regulated market while complete deregulation in the other segment of the market would lead to a situation in which many families effectively were unable to obtain coverage. The natural response of government when faced with an avalanche of complaints about "gaming" by insurers in the dual market would be to place tighter and tighter regulations on insurance compa- nies. That is why limited underwriting for all insurance business, which avoids these prob- lems, turns out to be the best practical option in the real world.

Recognizing that in this dual market there would be a significant number *of uninsured people who want coverage but cannot afford it, some conservative scholars want to cre- ate a potentially huge new government program for high-risk people. Modification #2: Under the second exception to full-risk rating, the federal and/or state governments would create a subsidized insurance risk pool, and possibly other new government programs,.for families facing abnormally high insurance costs. Faced with the problem of how to cover high-risk individuals while rejecting limited underwriting in the private sector, some conservative scholars have taken the surprising step of arguing for the creation of a new, subsidized program for those unable to afford at least basic catastrophic insurance in the open, unregulated market. There are various ver- sions of this risk pool proposal, but the idea is as follows. Say an individual or family with a poor medical history attempts to buy insurance as a first-time purchaser, or switches from another policy, but is unable to purchase a policy costing less than a cer- tain premium level-or some proportion of their income. In this case, the family would be permitted to join an insurance risk-pool program operated by the federal government and/or the states. The government would seek bids from private insurers to run the pro- grain. Each enrollee would be charged a means-tested premium based on age or income, with the difference in cost subsidized by the taxpayer. In addition, a new "health safety nef' program might be created to pay for essential medical expenses of uninsured working Americans once they had spent down their in- come to poverty levels, and used up their assets, to pay for medical care.

THE FOLLY OF A NEW HEALTH PROGRAM FOR THE SICK The dangers of such a proposed program are legion. Although proponents claim that the program would be very small, painful experience suggests that it would not remain that way for long. There are many reasons to be pessimistic. First, such a national program could play into the hands of liberals whose ultimate aim is to create a single-payer national system for all Americans. With conservative support to establish a new program for certain categories of working-age Americans, there would be constant pressure from congressional liberals to ease the eligibility criteria for the pro- gram. Liberals no doubt would be aided by officials of the programs who would con- stantly discover new, unmet needs to be served. Moreover, families who did not quite meet the eligibility standards would testify that they faced onerous burdens in the private market. The result would be a bidding war in Congress between liberals and conservatives. Liberals would be offering hard-pressed families a lower eligibility threshold for subsidized government health insurance; conser- vatives would try to tell these families that the government could not afford to help them. Experience shows that liberals win those battles. Second, the insurance industry would have enormous incentives to dump high-risk in- dividuals onto the program by quoting high rates to anyone who might pose a high cost to the insurer in later years if they renewed. Proponents of the risk pool idea claim that very few people would fall into the risk pool in this or other ways, but they present no hard, defensible data to support that assertion. In any case, they overlook several things in making that assumption. One is that the uninsurable population in existing state risk pools is kept artificially low because the number of slots for entrants typically is controlled and because millions of higher-risk in- dividuals enjoy low premiums in company-sponsored plans which are community-rated within the firm, meaning that healthy employees heavily subsidize sicker employees. As this community rating unraveled in the less-regulated market proposed by advocates of the risk pool "solution," more and more individuals and families would find insurance un- affordable.

Most health economists advocating the risk pool idea are loath to describe it in any de- tail, and they present no information regarding its likely cost to taxpayers. Peter Ferrara of the Dallas-based National Center for Policy Analysis apparently envisions a risk pool as either a new unfunded mandate on the states or a federal subsidy available only when the high-risk family has been reduced to poverty by its medical costs: [S]tate insurance pools can back up [other people] who do not buy coverage and become uninsurable. These pools can charge high premiums up to some reasonable limit, with the state subsidizing remaining costs. Alternatively, or in addition, a means-tested program can pay for essential health care once above-poverty income and assets are used up. 16 In other words, states can figure out how to pay for the problem, and middle-class Americans can simply go broke and qualify for a new version of Medicaid. Third, as medical science improves and the ability of insurers to use genetic and other factors to calculate lifelong risk steadily improves, more and more individuals today clas- sified as healthy would find themselves re-classified by insurers as high-risk and discover their only realistic option was the government program. 17 Thus, unless insurance compa- nies were limited in their right to use steadily improving risk indicators through limited underwriting requirements-which are anathema to conservative proponents of risk pools -the government program would grow. In addition to these problems, supporters of the risk pool need to contemplate other likely results of the proposal. Among them: V A federal-state risk pool probably would lead to price controls and tightly-man- aged care. With a new entitlement program in place and subsidized by the taxpayer, painful experience shows that government costs will grow. There can be little doubt that in an effort to hold down the cost of the new program and so limit the level of sub- sidy, Congress would be inclined to introduce the same system of price controls and other regulations it has used to try to hold down the cost of the Medicare program. Thus, creation of a federal-state risk pool would become the Trojan horse for even more regulation of the medical profession. In addition, the new federal-state risk pool, like existing risk pools and government programs, can be counted on to incorporate tightly managed care in an effort to control physicians and patients. V A federaktate risk pool holds the potential of huge unfunded mandates on states. States have learned from the Medicaid program that Congress is very inclined to enact new requirements in federal-state programs but disinclined to pay for those new mandates. The proposed federal-state risk pool is much like a Medicaid program for high-risk individuals. State taxpayers would be wise to hold on to their wallets. Twenty-seven states currently operate risk pools enrolling 100,000 individuals. But to- day, many states place strict limits on the number of individuals who can be in these pools, as well as limits on spending by the pools. Illinois, for instance, places a cap on the number of people in its pool, and there is a waiting list. Many states also place a lifetime cap on the benefits enrollees can receive, leaving them uninsured again once the benefits run out. Despite this, most states have great difficulty keeping costs from exploding. Costs in the Oregon pool, for example, jumped from $3 million in 1989- 1991 to $17 million in 1991-1993 and are projected at $42 million for 1993-1995. 18 With a new federal entitlement on the states, it would be much harder, if not impossi- ble, for states to control access to, and the cost of, their pools. It is interesting to note also that states have taken steps to hide the true cost of their risk pools by shifting some costs to other citizens who buy insurance or medical care. States typically place assessments on insurance companies or apply a "service charge" to all patients entering hospitals. This hidden financing is important because one of the aims of the risk pool idea is said to be to make the subsidy explicit, instead of passing on the cost of high-risk people to other insured individuals. V A federal-state risk pool must be accompanied by an explicit benefits package. If a high-risk program is set up to provide catastrophic coverage with a specific de- ductible or stop loss, the government must specify the services that can be counted to- wards that deductible (and the services covered by the risk pool insurer). It must also define the private-sector catastrophic package the applicant must find too costly. Sim- ply specifying a deductible is not enough, since the deductible must be in a policy to cover something specific. Thus, conservative proponents of the risk pool idea must ac- cept the idea of a standard benefits package-something they usually claim is anath- ema to them.

CONCLUSION The health care reform debate, unfortunately, does not lend itself to neat theoretical so- lutions. There is no ideal way to reform the health care system-only second-best solu- tions. The aim of policy thus must be to find practical means to deal with the problems that worry Americans in ways that also are most likely to increase personal freedom, pro- vide greater scope to consumer choice and competitive markets, and limit government in- trusion.

The Nickles-Steams legislation accomplishes this. It provides Americans .'with greater security and freedom by giving them effective control over their choice of insurance cov- erage and the right to switch plans when they are dissatisfied with their current coverage. In this way, it does not lock sicker Americans into plans that serve them badly. It achieves this by creating an insurance market with individual premiums that actually are more in line with individuals' risk than the community-rated employer-sponsored groups which typify the insurance market today. The objections to this framework fall into two categories, both of which are erroneous. The first is the claim that such a market simply cannot exist because of adverse selection. This is refuted by the very existence of the Federal Employee Health Benefits Program, which is far more prone to adverse selection than the system that would be created by Nickles-Steams and yet is quite stable. The second is the claim that a better-functioning market, more satisfactory to Ameri- cans, could be created by introducing full-risk rating into health insurance and then deal- ing with the occasional problems that arise in such a market. But such a market would be technically and politically unstable. Worse still, for conservatives and indeed for most Americans, it could lead to a potentially huge new government-sponsored health insur- ance program and a deterioration of insurance services for millions of Americans.

Stuart M. Butler Vice President and Director of Domestic and Economic Policy Studies