At least three broad problems characterize U.S. health care and insurance: (1) high and rapidly growing costs, (2) large numbers of nonelderly people without insurance, and (3) enormous projected Medicare deficits and continued Medicaid cost growth. The health care reform debate and reform proposals have focused largely on expanding the number of people with health insurance. On November 7, 2009, the U.S. House of Representatives narrowly approved legislation to mandate that all individuals be covered by health insurance coupled with Medicaid expansion, premium subsidies for low-income persons, creation of a health insurance exchange (or exchanges) with strong restrictions on health insurance underwriting and pricing, and creation of a government-run health insurer to compete with private health plans. While the details differ, on November 21 the U.S. Senate voted 60-39 along straight party lines to approve for floor debate a bill with the same broad outlines.
Passage of health care legislation with these features would transform U.S. health insurance. Massachusetts is the only state with an individual health insurance mandate, enacted in 2006. The California Legislature rejected an individual mandate in 2008. Maine and Vermont programs offering subsidized health insurance without a mandate have attracted relatively few applicants. The Connecticut General Assembly overrode a veto by the state's governor to enact legislation in 2009 appointing a board to develop a public health insurance option to promote universal coverage, including low-income subsidies, to take effect by July 2012. The board is authorized "to evaluate implementation of an individual mandate." As part of its reforms, Massachusetts fines employers who fail to make reasonable contributions to employee health coverage. Hawaii has required employers to offer coverage to employees working at least 20 hours weekly since 1974. Subsequent employer mandates in Massachusetts, Oregon, Washington, and California were either repealed or never took effect.
Relatively few states have strict restrictions on health insurance underwriting and pricing of the type proposed in the Congress. Six states require guaranteed issue in the individual market (Kaiser Family Foundation, 2009). Ten states have a rate band system limiting permissible variation of rates based on health status. Five states have adjusted (modified) community rating laws that permit rates to vary in relation to factors such as age, location, and coverage, but not health status. New Jersey and New York have pure community rating, which requires an insurer to accept all applicants for a given type of coverage and location at the same rate. The small group health insurance market has more restrictions. In conjunction with federal law, all states require guaranteed issue. Thirty-five states have rating bands, 11 states have adjusted community rating, and New York has pure community rating. Three states and the District of Columbia have no rating restrictions. As an alternative to strict underwriting and rating restrictions, 34 states have a high-risk pool with guaranteed issue of basic coverage at subsidized (but still relatively high) rates, regardless of preexisting conditions.
Debate over the majority Democrats' proposals for expanding health insurance has been highly partisan. Democrats stress the importance of expanding coverage. Liberal and progressive members strongly favor a public insurer to compete with private insurers. Some favor a public plan as a significant step toward the ultimate goal of universal coverage under a single payer system. Congressional Republicans are nearly unanimous in their opposition to the Democrats' reform agenda, especially the creation of a public plan. They propose narrowly target reforms and market-oriented changes in health insurance markets and taxation to expand coverage while helping to control costs.
This article provides an overview of the U.S. health care debate and reform bills under consideration by the U.S. House and Senate, with a focus on proposals that deal directly with health insurance. It begins by briefly elaborating the main problems that confront U.S. health care and insurance: high and rising costs, a large uninsured population, and large projected deficits for Medicare. It then turns to the House and Senate bills, outlining the key provisions for expanding and regulating health insurance and CBO projections of the proposals' costs, funding, and impact on the number of people with health insurance. The next section considers the potential effects of the mandate that individuals have health insurance, premium subsidies, and proposed insurance market reforms. The proposed creation of a public health insurance plan and/or nonprofit cooperatives and provisions that would modify permissible grounds for health insurers to rescind coverage and repeal the limited antitrust exemption for health and medical liability insurance are then considered. The article concludes by contrasting the reform bills with market-oriented reforms and with brief perspective on future developments.
MOTIVATION FOR REFORM
Costs and Cost Growth
Figure 1 shows U.S. health expenditures as a percentage of gross domestic product (GDP) and annual growth rates in per capita health spending during 1962-2007. The percentage of GDP devoted to health care grew from under 6 percent to over 16 percent during that time. Real annual growth in per capita expenditures averaged 4.3 percent. Real per capita spending grew 6.2 percent annually during the 1960s, which included the creation of Medicare and Medicaid in 1965, and then 3.6 percent and 3.9 percent annually during the 1970s and 1980s, respectively. Real per capita spending growth slowed to 2.6 percent in the 1990s and has increased at 3.3 percent annually this decade.
Figure 2 shows per capita health expenditures in 2007 for OECD countries with available data, adjusted for U.S. purchasing power parity. The U.S. expenditure of $7,290 was 53 percent larger than that of the second highest country. Figure 3 plots compound annual growth rates in per capita health expenditures for Organisation for Economic Co-operation and Development (OECD) countries during 1997-2006 versus the countries' per capita expenditure in 1997. While the 6 percent (nominal) U.S. compound growth rate in per capita expenditures ranked 15th out of 25 countries, the U.S. growth rate is a clear outlier compared with trend.
Explanations of why the United States spends much more than other countries generally point to greater rates of technology adoption and diffusion and higher compensation for health care providers, along with the lesser role played by government in financing medical care. The consensus is that the U.S. system of government and private insurance has significantly increased expenditures and expenditure growth. Despite the large numbers of uninsured, the United States ranks well above average among OECD countries in the proportion of national health expenditures reimbursed by insurance (see Figure 4). It ranks first by a large margin in the proportion of spending reimbursed by private insurance.
The high average health expenditure in the United States is associated with high average health insurance premiums. The Kaiser/HRET survey of employer-sponsored health benefits reports an average premium (employer and employee combined) for family coverage in 2009 of $13,375, 131 percent greater than for 1999, with an average worker contribution of $3,515 (Kaiser/HRET, 2009). The average premium for single coverage in 2009 was $4,824, with the worker contributing an average of $779. Given greater average cost sharing and less generous benefits chosen, average individual health insurance market premiums in 2009 were much lower, despite higher expense loadings. According to an AHIP survey of 2.5 million policies, the average premium for single coverage in the individual market was $2,985, and the average premium for family coverage was $6,328 (AHTP, 2009). The average annual premium for individual (family) coverage ranged from $1,429 ($2,967) for 18- to 24-year-olds to $5,715 ($9,952) for 60- to 64-year-olds (see Figure 5).
The question of whether the higher cost of U.S. medical care produces significantly higher quality is much debated. U.S. infant mortality rates are high among developed countries. Americans do not have higher average life expectancies. The U.S. ranks highly on survival rates for certain cancers (Preston and Ho, 2009). It generally is characterized by greater innovation and more rapid diffusion of medical technology, new drugs, and biologies. Waiting times for noncritical surgeries are significantly lower in the United States than in many other countries. Health care expenditures and quality of care vary widely within the United States. A sizable literature, for example, documents large regional variations in Medicare spending and considers whether that variation is related to quality, as well as whether Medicare expenditures could be cut in high-cost regions without significantly reducing quality (see, e.g., Skinner et al., 2009; Cooper, 2009).
The high costs of health care and insurance influence many people to forego coverage. High premiums and the large number of uninsured have contributed to allegations that private insurance markets are substantially dysfunctional (see below). The most widely cited estimates of the uninsured population are based on the Current Population Survey (CPS). It is estimated from that source that approximately 46 million U.S. residents did not have health insurance in 2008, representing 17.4 percent of the nonelderly population. Compared with the insured nonelderly, the uninsured on average have significantly lower income and educational attainment, are less likely to be employed full-time, are more likely to be black and /or of Hispanic origin, are more likely to be young adults than middle aged, and are less likely to report being in excellent or very good health. Roughly a quarter of the uninsured were eligible for Medicaid but had not enrolled (Kaiser Family Foundation, 2009; see also NIHCM, 2008). Roughly 10 million lived in households where a member declined employer-sponsored coverage.
An estimated 38 million (20.4 percent) of the adult nonelderly population were uninsured. About 8 million were non-U.S. citizens. Estimates suggest that at least half of those persons are unauthorized immigrants (see NIHCM, 2008). Approximately 4 million had income above 400 percent of the federal poverty level9 in 2008 (Kaiser Family Foundation, 2009, Supplementary Data Tables, p. 3). The duration of time spent without insurance varies widely. The proportion of nonelderly uninsured has remained relatively steady since 1990, with a decrease in private insurance offset by an increase in public coverage (Cohen et al., 2009).
Uninsured rates vary widely across U.S. states in relation to income, age, race, ethnicity, and other socioeconomic and demographic factors. Figure 6 illustrates crossstate variation in uninsured rates during 2007-2008 (obtained from Kaiser Family Foundation, 2009) for the continental United States. It shows the average percentage of the adult nonelderly population without health insurance for quartiles of states ranked by the percentage uninsured, along with the within quartile averages of the percentage of the population with income below federal poverty level (FPL), the percentage of the state's population that was African-American (black), and the percentage of the population of Hispanic origin (Hispanic). States with the highest uninsured rates had considerably greater poverty and proportions of black and Hispanic residents than states with the lowest uninsured rates. Median household income is considerably lower in the states with high uninsured rates (not shown).
Again using data for the lower 48 states, Table 1 shows descriptive linear regressions (with no pretense of causal inference) of the percentages of the adult nonelderly population in 2007-2008 with employer-sponsored health insurance and no insurance as functions of state median household income, the proportion of the adult nonelderly population with public coverage (Medicaid and Medicare or military), and the proportion of the total population that is black or Hispanic. The employer coverage rate is strongly and positively related to median household income, and it is strongly and negatively related to the proportion of nonelderly adults with public coverage, and, especially, the proportion of the state's total population that is Hispanic. The uninsured rate is strongly and negatively related to median household income and public coverage, and it is positively related to the proportion black and the proportion Hispanic.
Estimates suggest that the uninsured pay about a third of the cost of their medical care and produced an estimated $56 billion in uncompensated care for providers in 2008, with government funding covering about 75 percent of the cost of uncompensated care and approximately $14 billion potentially being shifted to private health insurance (Hadley et al., 2008). n While causal inference is challenging given unobserved heterogeneity and related issues, the consensus is that lack of insurance negatively affects access to health care and health. The uninsured are entitled to hospital emergency/acute care to stabilize their conditions without regard to ability to pay, and many uninsured with low incomes obtain care from community health centers. But being uninsured on average is associated with a lower likelihood of having a usual source of medical care, less use of preventive medical care, greater likelihood of foregoing medical care due to cost, and, while the magnitude of the increase is debated, a greater likelihood of bankruptcy due to unpaid medical bills.
While the number of people that are uninsured in relation to preexisting conditions and loss of insurance after job loss and exhaustion of continuation of coverage benefits is not known, these sources of uninsurance and difficulty in affording health insurance are widely regarded as problematic. Figure 5 also shows individual health insurance denial rates by age group from AHIP (2009) survey data. The overall denial rate was 12.7 percent. The extent to which applicants denied coverage were able to obtain coverage from another insurer or source is not known. The AHIP survey also reports that 34 percent of offers were at higher than standard premium rates (36 percent of offers were below standard rates) and that 6 percent of offers included a waiver of coverage for one or more health conditions. While health insurance policy rescissions are unlikely to represent a significant source of uninsurance, health insurers' rescission practices have received scrutiny (see below). The possibility of being denied coverage, or having to pay a higher premium if disclosure is truthful, likely leads to more applications with misrepresentations or concealments and to higher rescission frequencies.
The Medicare/HealthCare Spending Deficit
Large projected Medicare deficits and continued Medicaid cost growth represent a third major problem confronting U.S. health care. The funding of Medicare in particular poses major challenges from real cost increases per enrollee and aging of the population. The Medicare Trustees (2009) estimated the present value of the projected Medicare deficit over the next 75 years at $38 trillion as of year-end 2008 (using their intermediate economic assumptions about real interest rates, general inflation, Medicare spending growth, GDP growth, and population growth). That figure is equivalent to about 2.6 times 2008 U.S. GDP, or about $250,000 per adult aged 16-64. While much of projected deficit reflects forecasts beyond 2020, the hospital insurance trust fund is projected to exhaust in 2017 under the status quo.
Of the $38 trillion projected present-value deficit, $13.4 trillion is for projected shortfalls in payroll taxes versus expenditures for the Medicare hospital insurance program (Part A). The remaining $24.4 trillion is for projected future general revenue transfers to pay the federal government's share (about 75%) of projected Medicare spending for outpatient services and prescription drugs (Parts B and D). The federal government transferred $184 billion of general revenues to pay its share of Medicare spending for outpatient services ($147 billion) and prescription drugs ($37 billion) in 2008. That $184 billion and future increases commensurate with GDP growth might be viewed as already built into the federal budget, so that the $37.8 trillion figure overstates the effective deficit. If the $184 billion were to grow at the Trustees' projected growth rates for GDP, the present value of the required general revenue transfers for outpatient services and prescription drugs would be $13.5 trillion less than the $24.4 trillion included in the $37.8 trillion figure. The combined deficit for excess of GDP outpatient service and prescription drug spending growth and the hospital insurance program is $27 trillion, about 1.9 times 2008 GDP, or roughly $175,000 per adult aged 16-64. The unsustainability of Medicare spending has significantly influenced the debate over how to finance expanded health insurance for the nonelderly.
HOUSE AND SENATE REFORM PROPOSALS
The U.S. House of Representatives approved the Affordable Health Care for America Act on November 7, 2009 by a vote of 220-215 with one Republican voting in favor. On November 21, the Senate voted 60-39 with no Republican support to approve the Patient Protection and Affordable Care Act for floor debate. The bill reflects a number of changes to the one approved by the Senate Finance Committee in October, including a proposed public option. If the full Senate approves the bill, with or without amendments, the House and Senate conferees will negotiate final terms for a vote by both chambers.
Table 2 summarizes the major features of both bills, which are similar in many key respects. A number of the bill's main features are consistent with President Obama's campaign platform for health care reform. Notable differences include the proposed mandate for adults to be insured and the proposed public insurance plan. Both bills would establish a pool for offering coverage to buyers with preexisting conditions at subsidized premium rates as a transition mechanism until creation of the health insurance exchange (or exchanges) with premium subsidies and implementation of marketwide underwriting and rating restrictions.
Both bills would require most legal residents to have health insurance that meets minimum requirements specified by the government, beginning in 2013 in the House bill and 2014 in the Senate bill. Eligibility for the taxpayer-funded Medicaid program would be expanded to all persons with income below 150 percent of FPL in the House bill and 133 percent of FPL in the Senate bill. Substantial premium subsidies would be provided to non-Medicaid-eligible buyers with incomes up to 400 percent of FPL through a sliding threshold of premium caps as a percentage of income, and lower-income households would be able to purchase coverage with a higher estimated actuarial value and thus lower cost-sharing at the subsidized rates. Figure 7 illustrates the maximum premiums that a family of four would have to pay and the associated actuarial values of coverage. Apart from very small establishments, the House bill would require employers to offer health coverage and contribute much of the cost or pay a tax up to 8 percent of payroll. The Senate bill would require employers with 50 or more workers who fail to offer coverage to pay $750 per worker. Both bills would provide modest tax credits for very small businesses that provide coverage. The CBO projects that the House (Senate) bill would result by 2019 in coverage of 96 percent (94 percent) of nonelderly legal residents, compared with approximately 83 percent today.
Insurance Market Reforms
Both bills would dramatically alter insurance markets and regulation. The House bill would establish a new federal regulatory and oversight agency. The Senate bill would utilize the Department of Health and Human Services and leave most enforcement to the states. Subsidy-eligible and other persons not covered through employmentbased coverage, Medicare, or Medicaid would be able to buy coverage through a new health insurance exchange (or, in the Senate bill, state-level exchanges) patterned after reforms enacted in Massachusetts in 2006 (discussed further below). The government would mandate broad coverage of services and levels of cost sharing from which consumers could choose (with additional limits on cost sharing for lowincome buyers as noted above). Health insurers would have to accept all applicants regardless of health status, without excluding coverage for preexisting conditions. Premium rates would be allowed to vary by coverage, geographic region, and, within a restricted range, a person's age. The House bill would permit a 2-1 age range; the Senate bill would permit a 3-1 range. The Senate bill also would allow variation up to 1.5-1 for tobacco use. Both bills propose ex post risk adjustment among insurers to help equalize underwriting experience across insurers.
The House bill would repeal the antitrust exemption for the "business of insurance" for health insurance and for medical liability insurance, subject to a safe harbor clause governing projected loss development and certain other activities. The House bill would require all health insurers to achieve a irtinimum loss ratio of 85 percent and to refund premiums, if necessary, to achieve that minimum, subject to regulatory discretion to relax the criterion to avoid undue market disruption. Both bills would prohibit insurers from rescinding policies for material misrepresentations or concealment unless the insurer could prove fraud or intentional by the applicant.
Both bills would create a government-run health insurer--a public plan--to offer insurance through the exchange in competition with private insurers. The bills stipulate that the public plan would be self-sustaining and would negotiate rates with providers. The Senate bill would allow states to opt out of the public plan provisions. Both bills also would provide grants and loans for the creation of nonprofit health insurance cooperatives on a state or regional basis. The public plan and cooperative proposals are discussed further below.
Funding Coverage Expansion
According to CBO 10-year projections, the expansion of coverage is projected to cost $1,052 billion under the House bill and $848 billion under the Senate bill (Table 2). The CBO projects $781 billion in taxes and fees under the House bill, including $460 billion in tax surcharges on high-income taxpayers and $168 billion in individual and employer penalties for noncompliance. Projected Medicare spending would decline by close to $400 billion under the House bill, including $170 billion in reduced reimbursement to Medicare Advantage. The CBO projects that the Senate bill would generate $486 billion in revenues, including $54 billion in new taxes on high-income earners for Medicare Part A, $149 billion in excise taxes on high-cost health plans, $60 billion in taxes on health insurers, and $41 billion in taxes on brand name drug and medical device manufacturers. Projected Medicare spending would fall by $436 billion under the Senate bill, including $118 billion in cuts in Medicare Advantage. Overall, the CBO projects that the House and Senate bills would reduce the 10-year federal deficit by $138 billion and $130 billion, respectively.
The CBO's cost, revenue, and deficit projections depend on numerous assumptions and are subject to considerable uncertainty, as well as to pay-as-you-go accounting. The cost projections would be significantly higher if not for the delayed implementation of Medicaid expansion and premium subsidies. The House and Senate bills project $102 billion and $72 billion in deficit reduction, respectively, from net receipts from creation of a federal long-term care insurance program, without reflecting the new program's projected accrual of liabilities (Harrington 2009). The projections of Medicare savings assume that payment rates for many providers would be held below the rate of inflation and that a proposed independent advisory board for Medicare would be "fairly effective in reducing costs" (see Elmendorf, 2009, regarding the Senate bill). In his November 19 commentary on the Senate bill projections, CBO Director Douglas Elmendorf stated that extrapolations beyond 10 years indicate that Medicare spending growth will average 6 percent over the next two decades (2 percent real growth per beneficiary), compared with annual growth of 8 percent in the past two decades (4 percent real growth per beneficiary). He concluded (Elmendorf, 2009): "Whether such a reduction in the growth rate could be achieved through greater efficiencies in the delivery of health care or would reduce access to care or diminish the quality of care is uncertain."
THE INDIVIDUAL MANDATE, SUBSIDIES, AND RATING RESTRICTIONS
A centerpiece of the House and Senate bills is the mandate for individuals to have health insurance, along with expanded Medicaid eUgibiUty, premium subsidies, creation of an exchange (or exchanges), and restrictions on health insurance underwriting and rating. The basic structure of the proposals largely mimics the 2006 Massachusetts reforms, which included an individual mandate, Medicaid expansion, premium subsidies for people with incomes up to 300 percent of the FPL, merger of the small group and individual markets, and an annual fine of $295 per worker on employers who fail to make a "fair and reasonable" contribution toward workers' health coverage. The Massachusetts' reforms were followed by an estimated increase in the insured population from 90 percent to 97 percent. Surveys indicate that a significant majority of people are satisfied with the reforms, although half of those surveyed who were forced to buy insurance disapproved (Blendon et al., 2008). The costs of Medicaid expansions and premium subsidies have exceeded projections, in part because take up has exceeded expectations. A state-appointed commission has recommended that the state move toward universal managed care with capitation as a means to control costs (Steinbrook, 2009).
The welfare effects of an individual mandate with Medicaid expansion, premium subsidies, and health insurance underwriting and rating restrictions are extremely complex (CBO, 2008).  An individual mandate might have informational and behavioral effects on purchase decisions of the uninsured apart from financial incentives provided by subsidies and penalties. Expanding health insurance coverage with a mandate, Medicaid expansion, premium subsidies, and underwriting and rating restrictions involves explicit Medicaid costs and premium subsidies and implicit (off budget) premium subsidies to buyers who are able to obtain insurance at belowmarket rates due to underwriting and rating restrictions. The approach also involves implicit (off budget) taxes in the form of above-market premium rates for some buyers. The net benefits of the proposals would depend among other factors on the magnitude of consumption externalities (the value placed by people on knowing that others have coverage); on the amounts, types, and costs of increased utilization of medical care; and on labor market effects. Reductions in uncompensated care would reduce the net cost of subsidies.
Explicit premium subsidies and a mandate with sanctions for noncompliance will increase demand for coverage. Underwriting and rating restrictions will lower the supply price for older and/or less healthy buyers, while increasing the price for younger and/or healthier buyers. By increasing demand, a mandate reduces the total cost of explicit subsidies needed to achieve any given increase in the percentage of people with insurance, including the costs that arise from crowding out unsubsidized coverage due to imperfect targeting of subsidies. A mandate also increases the size of the implicit tax base to fund below-cost premiums for older and/or less healthy buyers. The greater the penalties for noncompliance, the lower will be the explicit cost of required subsidies. A "weak mandate" will require larger subsidies and/or result in fewer people being insured than a "strong mandate."
The effects of health insurance underwriting and rating restrictions on decisions to insure and average premium rates also will depend on the strength of the individual mandate and the magnitude of explicit premium subsidies. Guaranteed issue of coverage without preexisting condition exclusions, prohibition of premiums based on health status, and limits on age-related premium variation will generate some degree of adverse selection as some younger and healthier people face higher premiums and delay buying coverage until they need expensive care, increasing the average cost of coverage that is purchased. The effects could be large without either generous subsidies or a strong coverage mandate with sizable penalties for failure to comply.
The Senate bill includes relatively weak penalties for an adult's failure to buy coverage compared to the House proposal (and Massachusetts law). The fine would start at $95 in 2014, increase to $350 in 2015, $750 in 2016, and be indexed to the Consumer Price Index thereafter. People who faced premiums for rrdnimunri coverage that exceeded 8 percent of their income would be exempt. The approval of similar low penalties by the Senate Finance Committee in early October generated substantial pushback by private health insurers, who had previously agreed to support proposed insurance underwriting and rating restrictions provided they were coupled with a strong mandate. PricewaterhouseCoopers (PWC, 2009) released a report, sponsored by AHIP, estimating that the Senate Finance Committee bill's weak mandate, in conjunction with its underwriting and rating restrictions, could increase average premiums for individual coverage by 47 percent by 2016 compared with current law, including the effects of proposed new taxes on several health care sectors and possible increased cost shifting from Medicare to private plans. Gruber (2009) responded that the PWC study did not consider proposed premium subsidies and ignored CBO projections that the Senate Finance Committee bill would result in lower premiums for comparable coverage than under current law. In a subsequent study sponsored by the Blue Cross Blue Shield Association, Oliver Wyman (2009) projected that proposed insurance reforms coupled with a weak mandate would produce a 50 percent increase in average medical costs per insured 5 years after reforms took effect compared with current law. While the assumptions underlying the PWC and Oliver Wyman projections are not transparent and debatable, the CBO's cost projections, also based on opaque assumptions, do not consider the potential for adverse selection.
Apart from the possible adverse selection issue, and without regard to policies that could reduce health care cost growth, an individual mandate with premium subsidies would be expected to put upward pressure on total health care expenditures. Utilization of health care, on average, will increase for people who obtain coverage in response to the reforms. In addition, a mandate necessarily requires government prescription of the types and amounts of medical services that must be insured. The proposed minimuin permissible coverage packages include broader benefits and less cost sharing than some people currently obtain voluntarily. Various provider groups will press for inclusion of their services. In principle, significant rninimum benefits are needed to achieve the basic goal of expanding coverage. They also may be needed to reduce the ability of lower risk people who face higher than market rates from underwriting and rating restrictions from sorting into low-coverage groups to mitigate implicit taxes. Increased coverage will lead to some increase in moral hazard and "excessive" utilization, a widely acknowledged contributor to high health care costs. Costs also could increase due to higher prices for medical services until the supply of health care providers expands to meet increased demand for care.
An individual mandate also has implications for the locus and scope of decisions about specific types of medical care that will be reimbursed by insurance and thus the amount of such care that will be demanded and supplied, including any movement over time toward the adoption of formal cost-effectiveness criteria. Decisions about private insurance coverage currently hinge on medical appropriateness and necessity as determined by custom and practice; by contracts between employers, employees, and insurers; by preferences of individual insurance buyers; by Medicare national and local coverage decisions that influence private insurance coverage criteria; and by numerous state and federal laws and regulations. The proposed individual mandate and insurance market reforms could be accompanied, if not initially then ultimately, by coverage determinations by the Department of Health and Human Services or other federal agency, perhaps along the lines of coverage determinations for Medicare. The possibility of expanded government control over insurance reimbursement has generated significant controversy. The overall scope of such expansion would depend on whether control eventually extended to the employer-sponsored market and/or the reforms eventually produced significant reductions in employer-sponsored coverage.
The question arises as to the rationale for a federal mandate, subsidies, and insurance market reforms given the dearth of activity by the states, especially with respect to mandates. A common explanation is that states generally cannot afford the subsidies needed to make mandates feasible. Gruber (2008, p. 67), for example, concludes that "states cannot meaningfully innovate in this area without a massive injection of federal funds." This statement begs the questions of why an insufficiency of state resources does not indicate that citizens, a significant majority of which are insured, are unwilling to pay the costs in higher taxes or reduced health services that mandates would require and how federal action could overcome that unwillingness to pay.
One argument for why a national health insurance mandate conceivably could be supported by voters who reject state-level mandates is that federal reform might help fund premium subsidies by substantially reducing Medicare and Medicaid cost growth. Another possibility is that state mandates are deterred by free-rider problems that a federal action might avoid. If a state were to enact the large, income-related subsidies needed to support a coverage mandate, it would tend to attract low-income people from other states, increasing its total cost of subsidies. The higher taxes needed from middle- and upper-income taxpayers would similarly encourage some outward migration, including small business owners and entrepreneurs, reducing the tax base for financing subsidies. Although any reduction in employment opportunities for low-wage workers would reduce inward migration to obtain subsidized insurance, the net result could still be a larger per capita burden on middle- and upper-income residents who remained.
Rate Restrictions and Incentives for Healthy Behavior
Unhealthy behavior is a major factor in obesity, diabetes, heart disease, and cancer. In principle, health insurance design can encourage healthy behavior through costsharing provisions and pricing (see, e.g., Bhattacharya and Sood, 2006, who consider insurance and obesity). Incentives for healthy behavior have traditionally been weak under employer-sponsored coverage, with little or no risk-related variation in workers' contributions to the cost of coverage. Turnover among employees and policy holders also reduces employers' and insurers' incentives to make long-term investments to promote healthy behavior. Regarding individual health insurance, basing initial premiums on factors such as weight and tobacco use provides some incentive for healthy behavior, but guaranteed renewability of individual health insurance at rates that do not reflect individual health and behavior dulls incentives for healthy behavior.
The House and Senate proposals recognize the potential benefits of providing financial incentives for healthy behavior in the employer-sponsored market. Existing regulation permits employers to vary employee contributions toward the cost of coverage by up to 20 percent to encourage healthy behavior under certain conditions (Mello and Rosenthal, 2008). Employers have been developing a variety of strategies to that end, including linking deductibles or premium contributions to tobacco use, weight control, blood pressure, and cholesterol levels. The bills would increase the permissible variation to 30 percent and provide regulatory discretion to permit variation up to 50 percent.
By guaranteeing issue of individual and small group health insurance without preexisting condition exclusions at rates that do not reflect health status, the House and Senate bills would make it illegal for individual coverage premium rates to reflect health-related behavior (except for smoking in the Senate bill). Benefit design and marketing of coverage also would be regulated in an attempt to keep insurers from trying to attract healthier people, and the proposals would authorize risk-adjustment mechanisms that would reallocate funds from insurers that experience lower medical costs to those with higher costs. If an insurer were to attract relatively more healthy people, or help keep more people healthy, it might have to forfeit some of any increase in profits to its competitors. While it is not clear how pricing incentives could be incorporated in the individual or small group markets if sufficient flexibility were permitted in an expanded individual health insurance market, the strict rating restrictions in the House and Senate proposals would likely deter potential innovation.
THE PUBLIC OPTION
Both the House and Senate bills include proposals to create a government-run health insurer to compete with private plans. President Obama and Speaker of the House Nancy Pelosi strongly support a public plan option as a means to promote competition, choice, and to "keep insurance companies honest." Some analysts argue that a public plan could improve competition and help lower costs by reducing profits, administrative expenses, and lowering reimbursement to providers (Hacker, 2008, 2009; Holahan and Blumberg, 2008; see also Nichols and Bertko, 2009). Pauly (2009) explains how a public option could provide choice to people who prefer dealing with government. Other observers stress that the case for a public plan is weak, that level competition would not be feasible, and that a public plan would inexorably crowd out private health insurance (e.g., Cannon, 2009; Francis, 2009b). The Senate Finance Committee bill proposed subsidies to create nonprofit cooperative health insurers (co-ops) at the state or regional level in lieu of a public plan. The House and Senate bills would subsidize creation of co-ops in addition to creating a public plan.
Competition and Market Structure
Private health insurance markets are characterized by high market concentration at the state level (Robinson, 2004; American Medical Association, 2007; GAO, 2009; see also Dafny, forthcoming). Concentration is much lower when measured at the national level. The extent and scope of economies of scale or other entry barriers at the state level other than some states' restrictive underwriting, rating, and coverage regulations is not clear. Effective entry and competition often depend on the ability to utilize relatively large provider networks and achieve sufficient scale to contract effectively with hospitals and physicians. In most states, insurers are able to contract with and utilize the services of large medical service organizations as an alternative or supplement to direct contracting. Consolidation in many private health insurance markets has coincided with increased consolidation of hospitals and hospital-provider networks, increasing insurers' ability to negotiate favorable rate with providers (and vice versa).
Over half of the employer-sponsored health insurance market is self-funded. Employers generally choose among insurers and numerous third-party adxninistrators for accessing provider networks and claims adrninistration. Reported insurance market concentration data generally do not reflect the self-funded market served by noninsurance third-party administrators. Those intermediaries and self-funding in general represent a significant source of competition for insurance companies in the employer-sponsored market except for small group coverage. Although often highly concentrated, buyers in the individual and small group markets have a choice among numerous insurers and plans in most states (except, for example, in New York, with pure community rating), including one or more nonprofit insurers.
Experience under the Federal Employees Health Benefits Program (FEHBP) indicates effective competition among health insurers without a public plan option. For decades federal employees and members of Congress have purchased their health insurance through this program, under which numerous private insurers compete for employees' business subject to oversight by the federal Office of Personnel Management. The FEHBP is generally acknowledged to work reasonably well, with high levels of employee satisfaction (Francis, 2009a). The provision of Medicare Part D coverage by private plans has been successful, with most seniors able to choose among numerous competing plans.
Profits and Administrative Expenses
Public plan supporters argue that health insurers' profits and adrninistrative expenses are excessive or even unnecessary, driving up the cost of coverage, and that a pubHe plan would achieve substantial savings on these dimensions. Administrative expenses are viewed as especially high for individual and small group coverage. Table 3 summarizes data on health insurers' profits, medical loss ratios, and administrative expense ratios from a variety of sources and time periods. Health insurers' profit margins (net income to revenues) typically average about 3 percent (less for nonprofits), medical loss ratios average roughly 85 percent (higher for nonprofits than for-profits), and administrative expense ratios average about 11-12 percent.
The aggregate margin for administrative expenses and profits in private plan premiums, including premium equivalents for self-funded plans, has averaged about 12 percent since the mid 1960s (with little or no trend). Sherlock (2009) reports administrative expense ratios of 11 percent and 16 percent for the individual and small group markets, respectively, in 2007 using data primarily from Blue Cross Blue Shield plans covering 36 million lives.
Insurers' administrative expenses include marketing, provider and medical management, account and member administration, general overhead, and state premium taxes (which average about 2 percent of premiums) (Sherlock, 2009; American Academy of Actuaries, 2009). Administrative expense ratios and medical loss ratios can vary widely across insurers in relation to (1) their mix of individual, small group, ASO, and Medicare/Medicaid related contracts; (2) how they account for ASO contract fees and expenses (including whether they are based on premium equivalents for those contracts); and (3) insurers' relative emphases on different types of managed care (Robinson, 1997).
Private insurers' administrative expense ratios are commonly compared with those of Medicare, which are about 1.5 percent of costs in the fee-for-service program (CBO, 2008). The low expense ratios for Medicare reflect a number of differences from private plans (Sherlock, 2009; American Academy of Actuaries, 2009), including:
(1) Per capita claim costs are much higher for Medicare, reducing administrative expenses as a proportion of total costs.
(2) Reported Medicare administrative costs usually exclude general overhead for the Center for Medicare and Medicaid Services.
(3) Enrollment and billing costs are reflected in Social Security Administration Accounts and not attributed to Medicare.
(4) Medicare does not negotiate with providers, engage in medical management, or spend much to reduce fraud and abuse.
(5) Medicare does not incur state premium taxes or incur regulatory compliance costs that affect insurance companies.
Private health plans have strong incentives to spend money to detect and prevent fraud and abuse if the expected savings exceed the expenditure. The resulting expenditures increase reported administrative costs. A public plan might not have comparable incentives. It commonly is argued that too little money is spent to combat Medicare fraud and abuse, with tens of billions of dollars lost annually.
Provider Reimbursement Under a Public Plan
A critical issue in the creation of a public plan is how reimbursement rates for health care providers would be determined. Private payers on average reimburse hospitals and physicians at significantly higher rates than Medicare (e.g., American Hospital Association, 2009, charts 4.6 and 4.7). The House and Senate bills' propose having the public plan negotiate rates with providers. That approach is inconsistent with many public plan supporters' goal of cutting costs by using Medicare reimbursement rates or Medicare rates plus a nominal percentage (Hacker, 2008). However, to the extent that Medicare reimbursements already entail significant cost shifting to private payers, an expansion of Medicare payment rates, with or without a modest markup, would further shift costs to and increase potential crowd-out of private plans, assuming that providers would accept or be required to accept such reimbursement rates. A strategy of linking public plan reimbursement to Medicare rates could threaten the financial stability of hospitals and physician practices that currently operate at low margins. Requiring a public plan to negotiate rates with voluntary participation from providers reduces this risk, although the risk remains that pressure for cost control would cause reimbursement and participation rules to tighten over time.
Is a Level Playing Field Feasible?
The market penetration of any public plan would depend on numerous factors concerning eligibility, pricing, and provider participation rules. A Lewin Group study prepared by Shiels and Haught (2009) estimated that an aggressive public plan reimbursing at Medicare rates would capture a large share of the overall market if open to employer plans. The CBO projects that the public plans proposed by the House and Senate would attract fewer than 5 milUon people by 2019, and that a public plan could have higher average premium rates as it could attract a less healthy population. It is not clear whether a tendency for a public plan to attract less healthy people would eventually be accompanied by increased risk adjustment to shift more costs to private plans.
Nichols and Bertko (2009) set forth criteria, shown in Table 4, for a public plan to compete equally with private plans. If legislation creating a proposed public plan reflected those criteria, it still could be difficult to ensure their implementation. For example, legislative language that public plan premiums include a contingency margin might not ensure self-sustaining premium rates in an environment of substantial pressure to make coverage affordable. Table 4 also shows two additional criteria for equal competition related to health insurers' capital and taxation.
Profits are needed to earn normal returns on capital that private insurers invest to back the sale of coverage and make promises to pay claims secure. The three largest publicly traded health insurers, UnitedHealth, Wellpoint, and Aetna, reported GAAP premium-to-capital ratios of 3.5, 2.9, and 3.3, respectively, at year-end 2008. The aggregate premium-to-capital ratio for the 13 largest publicly traded health insurers combined was 3.7 (A.M. Best, 2009b). The aggregate statutory accounting premiumto-capital (surplus) ratio for nonprofit Blue Cross Blue Shield plans was 3.1 (A.M. Best, 2009a). These ratios and underlying amounts of capital are associated with an "A" financial strength rating for the typical health insurer. Holding such capital may require a pretax margin in premiums of 2-3 percentage points (American Academy of Actuaries, 2009). If a public plan were required to hold a capital cushion as, for example, proposed in the Senate bill, and/or to maintain some form of premium stabilization reserve, it would not hold the amount of capital that a private insurer would need to achieve an A rating. It would hold less capital and ultimately be backed by taxpayers.
A public insurer also would not face the same premium and income taxes that private insurers face (including taxes on investment returns from holding capital, which increase the cost of holding capital noted above). Given that state premium taxes average about 2 percent of premiums, the total tax differential between a public and private plans could approximate 3-4 percent of premiums. As a result, a public plan could have a direct cost advantage related to capital and taxation of 5 percent or more of premiums.
The CBO projects that government-authorized, nonprofit co-ops would have little market penetration. The need for or role for co-ops is not transparent, given that nonprofit insurance companies already offer health insurance in many states and are dominant players in some states. Nonprofit insurers would be expected to expand and enter additional states if many new buyers who seek health insurance as a result of premium subsidies and /or the legal mandate to buy coverage prefer dealing with nonprofit insurers. Co-ops would not have any inherent advantage over private health insurers in establishing provider networks, negotiating with providers, and monitoring health care utilization and fraud.
The creation of government-authorized co-ops would create some risk of ongoing subsidies by taxpayers (if not by private health insurance buyers), of crowd-out of other plans, and of eventual conversion to a government-run plan if created as an alternative to a public plan. Like a proposed public plan, government-authorized co-ops would likely be backed implicitly if not explicitly by taxpayers. They would probably not have to hold the amounts of capital that private health insurers hold, and they would not have to pay income or premium taxes that private for-profit and nonprofit insurers must pay. There could be pressure for government-authorized co-ops to offer artificially low premium rates, with an attendant risk that they would experience persistent operating losses and require additional subsidies. Although co-ops would initially be required to negotiate their own reimbursement rates with providers, substantial pressure could arise over time for centralized negotiations. As would be true for a public plan, any ability of co-ops to undercut reimbursement would shift more costs to other payers, increasing crowd-out of other health plans.
POLICY RESCISSIONS AND THE ANTITRUST EXEMPTION
The House and Senate bills would override many states' laws regarding health insurance policy rescissions. The house bill would repeal the limited antitrust exemption for health insurance and medical liability insurance. An amendment to that effect could be proposed in the Senate.
Traditional practice, governed by common law, statute, and regulation, is for insurers to rely in underwriting and pricing on accurate disclosure by applicants without conducting a detailed investigation of medical history. Companies practice ex post auditing--conducting more detailed and costly reviews of a subset of applications following policy issue--sometimes when expensive treatment is sought soon after issue. This system lowers underwriting costs and premiums compared to more intensive upfront verification or to paying all claims regardless of the accuracy of disclosure. State laws permit rescission only on the basis of material information, that is, information that would have changed the insurer 's decision to offer coverage or the premium charged. Some states restrict insurers' rescission rights to instances where misrepresentation or concealment is directly related to the illness that produced the claim.
During the past few years health insurers' rescission practices have generated controversy, litigation, and new regulation in some states, and they have played a role in the health care reform debate. The House and Senate bills would prohibit rescission unless fraud (intent) could be established. The Subcommittee of Oversight and Investigations of the House Committee on Energy and Commerce held hearings on rescission practices during June and July, 2009. Congressional staffers' analysis of 116,000 pages of documents from three large health insurers identified a total of about 20,000 rescissions from several million policies issued by the insurers over a 5-year period (Committee on Energy and Commerce, 2009). Company representatives testified that less than one-half of 1 percent of policies were rescinded (less than 0.1 percent for one of the companies).
Congressional staffers highlighted 13 case studies of alleged abuse. Coverage was reinstated by the insurer in at least five of the cases. Five of the cases involved a rescission based on misrepresentation or concealment of a condition unrelated to medical claims for which reimbursement was sought. Two cases involved rescission of family coverage based on misrepresentation by the applicant; two involved agent misrepresentation. The practical effect of the House and Senate bills' requirement of proof of intent might be iniitimal given that the bills would guarantee issue of coverage without preexisting condition exclusions at rates that do not reflect health status. Otherwise, requiring proof of intent for insurers to rescind policies would be expected to increase underwriting costs, claim costs, and premiums, and it might increase denial rates.
Repeal of Antitrust Exemption
During October 2009 hearings by the Senate Judiciary Committee on possible repeal of the limited antitrust exemption for health and medical liability insurance, Senate Majority leader Harry Reid (2009) testified that "exempting health insurance companies has had a negative effect on the American people" and that "there is no reason why insurance companies should be allowed to form monopolies and dictate health choices."
The 1945 McCarran-Ferguson Act, which also codified state insurance regulation as national policy, exempts the "business of insurance" from federal antitrust law provided that the activities are (1) regulated by the states and (2) do not involve boycott, coercion, or intimidation. Until this year's health care debate, the long debate over the exemption's efficacy focused almost entirely on property-casualty insurance, including medical malpractice liability coverage, and specifically on the role of property-casualty insurance rating organizations, such as the Insurance Services Office and the National Council on Compensation Insurance. These organizations collect and analyze data on property-casualty insurers' loss costs, forecast loss development, and disseminate projections of future loss costs for hundreds of rate classes in different states. Depending on specific state law, property-casualty insurers can incorporate the forecasts in their ratemaking. In principle, this system helps produce more accurate property-casualty rates, thus improving financial stability, and it reduces entry barriers that otherwise would confront small insurers or insurers entering new markets. Cooperative production and distribution of loss development and future loss cost projections, as opposed to simply sharing historical data, would be unlikely to withstand antitrust scrutiny.
Despite allegations of large health insurers engaging in abusive monopolistic practices while enjoying protection from antitrust laws, there is no evidence that the antitrust exemption has contributed to higher health insurance costs, premiums, or profits, or, as implied by Senator Reid, of "health insurance monopolies . . . making health care decisions for patients. " In contrast to many property-casualty insurers, health insurers do not cooperate in estimation of medical claim loss development or projection of future claim costs. There is no evidence that the exemption has contributed to higher market concentration. It does not prevent review and challenge of mergers of health insurers by the Department of Justice. Mergers and acquisitions of health insurers are also subject to approval by state regulators. Repealing the antitrust exemption would not significantly increase competition, and it would not make health insurance coverage less expensive or more available. Repealing the exemption for medical liability insurance would not lower its cost or prevent future malpractice insurance crises, such as those that occurred in the mid-1970s, mid-1980s, and earlier this decade. The unintended consequences could include increased costs, reduced rate accuracy, and less competition in already fragile malpractice insurance markets.
The broad reforms in the House and Senate bills would transform U.S. health insurance. Significant expansion in health insurance coverage would be achieved through an individual mandate and by expanded eligibility for Medicaid, by substantial explicit and implicit premium subsidies, and by federal government prescription of individual and small group health insurance benefits, coverage, underwriting, and rating. These changes would improve access to and affordability of health insurance and health care for millions of residents, with significant costs to taxpayers and other insurance buyers, and uncertain long-run effects on the supply of medical care.
Enactment of proposed reforms would demonstrate that U.S. elections can have fundamental, long-run consequences. The House and Senate bills represent a clear contrast to market-oriented proposals for expanding coverage and helping to control costs, such as those included in Senator McCain's presidential campaign platform, through targeted insurance market reforms in conjunction with increased incentives for consumers to play a greater role in decisions regarding health and health insurance. Those alternatives include: (1) using tax credits for health insurance to help lowincome persons afford coverage, to equalize the tax treatment of employer-sponsored and individual coverage, and to reduce the tax subsidy for high-cost employersponsored plans; (2) expanding Health Savings Accounts, thus encouraging more consumers to assume greater financial responsibility for decisions regarding their health and medical care; (3) permitting people to buy insurance across state lines by authorizing health insurers that designate a "primary" state for regulatory oversight of underwriting, pricing, and coverage terms to sell insurance nationwide according to the rules of the primary state;36 (4) providing subsidies to state-based high-risk pools offering coverage, without regard to preexisting conditions, at subsidized premium rates that are high enough to discourage people from waiting to buy coverage until they need expensive care; and (5) providing additional, narrowly targeted subsidies to improve access to care for persons with very low incomes who do not currently qualify for Medicaid.
The longer-run effects of the broad changes proposed in the House and Senate bills on health insurance and health care will depend to a large extent on whether employersponsored coverage remains dominant, at least for large employee groups, with plan design and benefit determination governed largely by competition and private contracting. Under one scenario, a significant majority of the nonelderly population will continue for many years to receive coverage on that basis. An alternative scenario might see cost control pressures lead to the extension of government authority over plan design, financing, and reimbursable expenses throughout the market, and /or a steady reduction in employer-sponsored coverage and concomitant increase in coverage obtained through heavily regulated exchanges and /or a public plan.
Regardless of whether reforms based on the House and Senate bills are enacted, the cost of medical care and insurance will remain on the national policy agenda for the foreseeable future. Projected reductions in Medicare spending under the House and Senate bills would largely help pay for expanded coverage for the uninsured. Health care spending will very likely be back on the congressional agenda within a few years, especially if the costs of expanded coverage exceed projections. In particular, and despite projected spending reductions in the House and Senate bills, there is a good chance that the implicit Medicare debt will have to be renegotiated through additional spending cuts, tax increases, enrollee premium increases, and/or fundamental redesign of the program.
Scott Harrington is an adjunct scholar at AEI.