Economic Research on Direct-Purchase Health Insurance: New Models for Real Health Care Reform

Introduction

“Economists have realized that health policy has paid so little attention to direct-purchase health coverage that policy makers know very little about it.”

As U.S. governments grew, they continued passing laws to regulate the kind of health coverage people could purchase. By 1960, most people with coverage got it through their employer. Plans provided by hospital monopolies controlled the coverage market. The true cost of health care was hidden from covered individuals. Vast spending increases were the result. The introduction of Medicare and Medicaid in 1965 made the situation worse. Finkelstein found that Medicare increased real hospital expenditure by 23 percent between 1965 and 1970. Extrapolating from the Medicare estimates suggests that the spread of health coverage may “explain at least 40 percent of the rise in real per capita health spending” between 1950 and 1990.1

Only in the last two decades have economists reconsidered the effect that government regulations have had on the form health coverage takes, its cost, and how it encourages people to use health care. Government coverage programs obscure the real cost of health care and increase health spending by shifting the out-of-pocket cost of health care spending from the savings of individuals to the paychecks of working Americans.

Unlike government coverage or coverage purchased through employers, private coverage purchased directly by individuals encourages people to choose between health coverage and all other goods. It controls health spending by pricing individual risk, encourages substantial variation in plan design to accommodate differences in individual risk tolerance, and provides incentives for cost minimization.

All private health coverage arrangements must solve two problems. The first is to collect enough money to deliver enough health care to satisfy people who can stop paying if dissatisfied. The second is to control moral hazard, the tendency to use more care when other people are paying for it.

Early History of Health Insurance

By 1910, an estimated one third of all adult males in the United States belonged to fraternal societies. Like modern HMOs, some societies provided prepaid health plans that controlled the supply of medicine to member patients by refusing to provide care considered beyond the organizations’ means. They hired “lodge doctors” who worked for the plan and were less likely to side with patients in recommending more care than the plan wanted to deliver. The societies used a variety of tools to control moral hazard. These tools included home-visiting committees, waiting periods for benefits, and curfews on members receiving benefits. People not covered by the fraternal societies could choose from plans operated by trade and industrial unions, various governments, commercial insurers, and company-sponsored mutual benefit societies.

The Birth of the Blues

The Depression produced a dramatic change in health coverage. One year after the stock market crash of 1929 “average hospital receipts per person fell from $236.12 to $59.26, and average hospital deficits rose from 15.2 to 20.6 percent of disbursements.”2 To boost revenues, hospitals created community-wide hospital benefit plans. The American Hospital Association created Blue Cross to promote hospital benefit plans that conformed to guidelines that reduced interhospital price competition.

The hospital benefit plans offered guaranteed services in exchange for an annual premium. The Blue Cross monopolies were extraordinarily successful, particularly after the hospitals persuaded state officials to exempt them from insurance company capital requirements and taxes. By 1950, Blue Cross sold almost half of all hospital insurance. As economist John Goodman notes, Blue Cross and Blue Shield (the equivalent organization for physician services) so dominated the market that “they made it difficult for a commercial insurance company to adopt reimbursement procedures that differed fundamentally from those of Blue Cross and Blue Shield. If a commercial company attempted radical deviation, the medical community could threaten a boycott.”3 At the beginning of the 1990s, Goodman estimated that “net revenues (premiums minus benefit payments) on group policies [were] usually less than 5 percent of total premiums, [therefore] a 2 to 3 percent premium tax is equal to about 50 to 60 percent of net revenues.”4

Medicare and Medicaid expanded the Blue Cross coverage model to government coverage programs. It paid minimal attention to controlling moral hazard. It encouraged patients to use any hospital they saw fit. It reimbursed hospitals on a cost-plus basis. It emphasized one price for everyone in a given community, and it had little cost sharing and no waiting requirements. A structure that made sense when hospitals could do relatively little for patients ended up creating unprecedented increases in medical expenditures as medical technology improved.

The fraction of the U.S. population with health coverage stabilized in the 1980s. Since then, the percentage of Americans with health coverage has fluctuated in a narrow range between 84 and 87 percent. The favorable tax treatment bestowed on employer-based coverage after World War II encouraged most people to purchase coverage through their employers. In 2007, roughly 59 percent of Americans had employer-based coverage. Government plans, state-subsidized high-risk insurance pools for the uninsurable, Medicaid, Medicare, SCHIP, the Veterans Administration, and other forms of military health care, covered 28 percent of Americans. Just nine percent of Americans with health care coverage purchased their own coverage directly from insurers.5

After nearly 50 years in which employer-based coverage modeled after Blue Cross dominated, researchers in health policy often implicitly assumed that Blue Cross, Medicare, and Medicaid defined the form that health coverage should take. As costs exploded, researchers focused on cost control, ways to charge the same premium for all regardless of health risk, and ways to control costs in a system that bypassed patients with direct payment to providers. The tight controls on health care demand used by the guild systems and the fraternal societies were variously reborn in the managed care, evidence-based medicine, pay-for-performance, and “comparative-effectiveness” movements. Like the fraternal societies, some state governments forced insurers to issue policies to everyone at the same price, and focused on controlling the practitioners that people in state plans could see, denying certain kinds of care altogether, and rationing care by waiting.

The Market for Individual Insurance

Recent research has shown that such tight controls can do harm by reducing overall coverage and making medical care difficult to access. Economists have also realized that health policy has paid so little attention to direct-purchase health coverage that policy makers know very little about it. As late as 2002, Pauly and Nichols published an article entitled “The Nongroup Health Insurance Market: Short on Facts, Long on Opinions and Policy Disputes.”6 They wrote that “there is considerable disagreement among policymakers and policy analysts about how the nongroup insurance market works now” and that “factual disagreements about this market are true impediments to policy consensus.” In 2006, Marquis et al. noted that while a variety of proposals have been offered to overcome perceived weaknesses in the direct-purchase market, “until now, this market has been the subject of relatively little study. As a result, there is much disagreement about the importance of the advantages cited for this market, the extent of the barriers to participation, and the likely effects of policy proposals.”7

Direct-purchase coverage is owned by an individual. Premiums are based on the costs that an insurer expects a particular individual will generate in the future. People with low health risk, where risk is shorthand for expected future expenditure, pay less in premiums. Because premiums in the direct-purchase market are paid with after-tax dollars, individuals know exactly what their coverage costs. Because paying for health care with insurance coverage is more expensive than paying cash, people who buy direct-purchase coverage are more likely to insure against large costs, relying on out-of-pocket cash to cover routine expenses.

Because employer-related coverage is tax advantaged, employer plans generally follow the more expensive route of insuring even low-cost routine care. Employees “pay” the bulk of their employer-provided group health premium in the form of lower wages. Their artificially low out-of-pocket costs insulate them from the true cost of health care. This increases the demand for health care. According to Manning et al., data from the RAND Health Insurance Experiment show that the higher out-of-pocket payment characteristic of direct-purchase plans “reduces expenditures 31 percent relative to zero out-of-pocket price.”8

In 2008, the Kaiser Foundation Survey of Employer Coverage estimated that the average cost of employer-based group coverage was $4,704 for a single person and $12,680 for a family. On average, the out-of-pocket “premiums” that employers charged employees equaled about 16 percent of the total premium for single coverage, roughly $60 a month, and 27 percent of the total premium for family coverage, roughly $280 a month.9 According to the AHIP 2006-2007 individual market survey, average annual premiums for roughly 2.3 million direct-purchase individual policies sold to single adults, were $1,359 for people aged 18-24 and $5,090 for people aged 60-64.10

Direct-purchase markets pool risk extensively, charging high risk people far lower premiums than their health status might indicate. Most direct-purchase policies are renewable without additional underwriting. This means that as long as he pays the premium, an insured person can keep his policy no matter how sick he gets. Contrary to popular claims, state laws generally prohibit raising a sick individual’s premiums unless an insurer also raises the premiums of everyone else in his rating class.

Pauly and Herring report that direct-purchase insureds who had medical expenses about 4 times that of other people enjoyed premiums that were only 1.6 times as high. People who buy a policy and become ill have a strong incentive to continue paying. This may explain why age and sex were generally better predictors of direct-purchase premiums than chronic conditions.11 Marquis et al. concur, reporting that the individual direct-purchase market is “an important source of long-term coverage for many who purchase it” and that “there is substantial pooling” that “increases over time because people who become sick can continue coverage without new underwriting.”12

Adverse Selection Due to Community Rating

In states that are so wedded to the Blue Cross model that they have community rating laws requiring that premiums be the same for everyone in a given geographic region, premiums are higher for everyone. In a study of 26.8-million commercially insured people, Willey et al. found that just 0.8 percent of the population was severely ill. That group had average annual expenses of $29,273. Those with less-serious chronic disease spent $8,225 a year. Half of the population had expenses below $989 a year.13

Charging everyone the same price does lower the premium for high-risk people. This increases the probability that they will buy coverage. However, as more high-risk people enroll, premiums must increase for everyone in order to cover the plan’s higher costs. Some low-risk people respond by refusing to pay the artificially high prices for coverage caused by the one-price-for-all requirement. Insurance economists call this phenomenon “adverse selection.” Although economists usually use the term “adverse selection” to refer to the result of the individuals having more information than the insurance company, this is adverse selection due to regulation.

There are more low-risk people harmed by one-price requirements than high-risk people who benefit from them. Pauly and Herring find that regulations on insurance premiums raise the average premium considerably, by 12 to 15 percent, and that the number of people who drop coverage exceeds the number of higher-risk people who sign up. In all, “the overall proportion of eligible people in the individual market with insurance [falls] by 6.0-7.4 percent.” They note that policies that reduce administrative costs and create high-risk pools are likely to cover more people than government control of premiums. Avoiding price controls on premiums also “avoid[s] regulation-induced incentives to insurers to avoid money-losing high risks.” In short, allowing competitive markets to set the price for coverage will end up covering more people.

The Benefits of Variety

Results like this underline the fact that health insurance has value only because it allows those who buy it to pay for expensive medical care without forcing them to liquidate their assets or suffer big drops in current consumption. Some people are willing to pay a lot to have an insurance company bear the risk of large medical payments. Others are willing to pay very little. Cutler, Finkelstein, and McGarry find that individuals who have a higher risk tolerance, as revealed by engaging in risky behaviors such as smoking, drinking, and employment with high mortality rates, are less likely to purchase all kinds of insurance, including life insurance, private health insurance, annuities, long-term care insurance, and Medigap coverage.14 This suggests the variety of plans offered in the individual direct-purchase market, a variety not present in employer-based coverage or government plans, adds to individual welfare. Health coverage is not necessarily the most important thing in life, and as circumstances, finances, and risk tolerance change, people are better off if they can change their coverage to match their circumstances.

Assertions that people without coverage have no access to health care ignore the fact that people can and do access care by paying cash. Though the self-employed are less likely to have health coverage than comparable wage earners, Perry and Rosen show that they use health services at roughly the same rates as insured wage earners.15 Glied’s findings on the spending patterns of the uninsured suggest that, in addition to cash payment, even low-income “uninsured people have some de facto insurance coverage available to them through the uncompensated care system and Medicaid conditional coverage.”16 Overall, national estimates suggest that the uninsured pay for about half of their care and that government and private spending for uncompensated care equals roughly 3 percent of total personal health spending, about $150 for an average annual premium of $5,000. As the budgetary disasters in Massachusetts, Kentucky, and Tennessee have shown, government programs seeking to eliminate uncompensated care by requiring insurance for all typically cost more than the uncompensated care they seek to eliminate.17

Recent work also makes it clear that people adjust their coverage as their estimate of their individual risk changes, something that mandated coverage, controlled enrollment seasons, and government standardized policies discourage. Bundorf et al. find that higher-risk people are more likely to have coverage, even at incomes of less than 200 percent of the poverty level. Regardless of income, the probability that someone will have coverage rises 17 percent when his health risk changes from low to high.18
This increase in coverage is remarkable in view of the fact that low-income people are especially sensitive to premium cost. In Massachusetts, a $10 increase in monthly MassHealth premiums increased the odds of leaving the MassHealth program by five percent.19 Small increases in already low subsidized premiums have been associated with significant enrollment declines in government programs in Oregon, Vermont, Wisconsin, and Rhode Island. Gruber and Simon conclude that price-sensitive people substitute government insurance for private insurance in large numbers. For every 100 newly enrolled children in state children’s health insurance programs, they estimate, 60 had previously been privately insured.20

The substitution may also go the other way. George Borjas found that the 1996 law limiting Medicaid eligibility for immigrants increased the likelihood that affected households worked and, therefore, increased the probability of eligibility for employer-sponsored health insurance. Their health insurance coverage rates at least remained the same after their Medicaid eligibility was curtailed.21 Without government coverage programs, he wrote, “the targeted immigrants themselves would have taken actions to reduce the probability that they would be left without health insurance coverage.”

Although some people genuinely cannot pay even the small premiums charged by government programs, Bundorf and Pauly show that income based “affordability” calculations are not good predictors of whether or not a person chooses to sign up for coverage. For example, while 36 percent of people in families with incomes less than 200 percent of the federal poverty level are uninsured, 44 percent have private coverage. By some measures, coverage in 2000 was affordable for over 50 percent of the uninsured. 22

Conclusion

Eighty years after governments began seeking to determine the shape of U.S. health coverage, economists who study direct-purchase insurance have created a body of literature suggesting that market-oriented health care reform has far more promise than the decades-old plans that seek to impose the Blue Cross model on the entire population. In view of Medicare’s insolvency, reform plans calling for further expansions of the Blue Cross single-price guaranteed-issue model, both via Medicare/Medicaid expansions and so-called “public” policies designed by government and sold through government run connectors, are unlikely to succeed in providing more medical care for more people at current expenditure levels.

Direct-purchase of health coverage, along with guaranteed renewability, flexible premiums, and high-risk pools, is a promising approach to reducing health care costs, increasing the value received for health care expenditures, and making more healthcare available for all. Given that roughly the same percentage of people has had health coverage for the last 20 years despite numerous expansions of government coverage programs and a massive increase in illegal immigration, the direct-purchase results also suggest that expecting everyone to have coverage is not a sensible policy goal.

Another danger, one too infrequently considered in the drive to lavish government funds and attention on extending the 70-year-old Blue Cross model to basically healthy people, is that insuring everyone is an unattainable goal. No one can make an adult alcoholic or drug abuser sign up for anything. If government gets out of the way, basically healthy people should be able to make their own health financing arrangements. Furthermore, the focus on universal coverage diverts attention from the radical reforms necessary to save Medicare and Medicaid, two government programs based on the Blue Cross model that are dysfunctional, insolvent, and already failing the frail and vulnerable people who depend on them.