Skip Navigation

INTRODUCTION

The Congressional Budget Office (CBO) and the Joint Committee on Taxation have prepared this analysis of House Majority Leader Richard Gephardt's health proposal. The analysis is based on the text of the proposal as printed in the Congressional Record on August 10 and on subsequent revisions specified by the Leader's staff. It comprises a review of the financial impact of the proposal and a brief assessment of its economic effects and factors that could affect its implementation.
 

FINANCIAL IMPACT OF THE PROPOSAL

Congressman Gephardt's proposal would assure universal health insurance coverage with a guaranteed package of benefits. People not eligible for the existing Medicare program (Hospital Insurance and Supplementary Medical Insurance, Medicare Parts A and B) would be required to enroll in a private health plan or in a new public program (Medicare Part C). Employers would be required to offer health insurance coverage to their workers and would generally be required to pay 80 percent of the premiums. Individuals would be required to pay that portion of the premium not covered by their employer, but low-income people would be eligible for federal subsidies.

Medicare Part C would replace the current Medicaid program for acute health care services. Benefits under the current Medicare program would be enhanced by adding some new benefits and expanding others. Spending for all parts of Medicare would be subject to stringent limits on growth.

The estimated federal budgetary effects of Congressman Gephardt's proposal are displayed in Table 1 at the end of this document. Tables 2 and 3 show its effects on the budgets of state and local governments and national health expenditures, respectively. In the process of extending health insurance coverage to the entire population, the proposal would significantly increase national health expenditures. The estimated changes in mandatory spending, revenues, and the discretionary spending limits, however, would not add to the federal budget deficit.

Coverage and Benefits

Congressman Gephardt's proposal would achieve universal health insurance coverage by requiring people to purchase health insurance for themselves and their families starting in 1999. People not enrolled in Medicare Part A could obtain coverage by enrolling in Medicare Part C or in a certified health plan offered by their employer or purchased individually. The mandate on individuals would be accompanied by a mandate on employers requiring them to offer coverage to employees and their dependents and to contribute at least 80 percent of the cost of that coverage.

Certified health plans and Medicare Part C would both offer a guaranteed national benefit package. That package would include the benefits currently covered under Medicare plus several enhancements, including unlimited hospital care without coinsurance, a prescription drug benefit, and a cap on out-of-pocket spending. The annual deductible amount would be $500 for an individual and $750 for a family (indexed after 1994), with a separate $500 deductible for prescription drugs.

The Congressional Research Service and CBO estimate that such a benefit package would initially be 3 percent more costly than the average benefit package of privately insured people today. This estimate assumes that the management of services for mental illness would hold the cost of the expanded mental health benefit to the levels projected under the current system. To cover benefits not included in the guaranteed benefit package, individuals and employers could purchase a standardized package of supplemental insurance; they could choose from up to 10 such packages.

Employers' Responsibilities

The proposal would impose different requirements on large and small firms. Within a firm, the requirements would vary with the characteristics of the workers and their families.

Beginning in 1997, large firms (those with more than 100 employees) would be required to offer qualified employees a choice of at least one managed care plan (if available) and one health plan with an unlimited choice of providers. Those firms would generally be required to pay at least 80 percent of the cost of the plan for each enrollee, with payment prorated for employees working less than 35 hours a week. Beginning in 1999, small firms would be required to offer their full-time workers Medicare Part C or a choice of private health plans, and would generally pay 80 percent of the premium. They could, however, be eligible for a temporary tax credit that would defray part of the cost.

Employers offering a choice of private health plans could, but would not have to, enroll certain part-time or seasonal workers in those plans. If they did not, they would be required to pay 80 percent of the Medicare Part C premium for qualified workers who earned more than $100 a month. The contribution would be prorated according to the number of hours worked.

Employers could also offer an alternative benefit package that combined a high-deductible version of the guaranteed benefit package and a tax-favored medical savings account. Employers would not be required to offer high-deductible plans to their workers, and workers covered under Medicare or receiving a premium subsidy would be ineligible for the option, even if their employer offered it.

Married couples with children and with both spouses working could choose to purchase health insurance through either spouse's employer, termed the enrolling employer. Married couples without children could designate one employer as the enrolling employer or could enroll separately as individuals. For each worker who obtained coverage through a spouse, the nonenrolling employer would be subject to a tax equal to 80 percent of the Medicare Part C premium for a single individual. After a phase-in period, that tax would ultimately be rebated to employers who paid premiums for two-parent families. (CBO also assumed that the credit would be available to employers of any married workers whose spouses did not work.) For the first few years, however, a significant portion of the revenue from the tax would be retained by the Treasury.

Changes in the Insurance Market

Congressman Gephardt's proposal would change the market for health insurance in several ways that would become fully effective in 1999, when the individual mandate came into force. In the meantime, various transitional rules would apply.

Health plans would be required to sell coverage to all eligible individuals and groups and provide for an annual period of open enrollment. They would be prohibited from excluding or limiting coverage on the basis of preexisting conditions and from imposing waiting periods for coverage. Plans other than group- and staff-model health maintenance organizations would be required to include in their network any health care provider that was willing to accept the plan's terms for participation. All plans would be required to contract with an extensive list of "essential community providers." National quality standards for health plans would be established, and each plan would be evaluated annually with respect to access to care, effectiveness and appropriateness of care, and consumers' experience and satisfaction. An individual or a health care provider would be able to bring legal action against a health plan for failure to comply with the terms of the plan or with federal or state law.

In effect, the market for health insurance would be divided into three sectors: Medicare Part C, a large-employer market, and a community-rated market. Individuals and small firms that did not participate in Medicare Part C would purchase insurance in the community-rating area in which they were located. Small employers and multiple-employer welfare associations would be prohibited from self-insuring. Associations meeting federal standards would be allowed to sell community-rated insurance plans to their members. Within the community-rated market, premiums could vary only by class of enrollment (single adult, one-parent family, or two-parent family). A risk-adjustment mechanism would be developed to even out risks among insurance plans in the community-rated market, but no adjustment of risks would be made among the three market sectors.

Small employers and eligible individuals could also obtain coverage through a new universal Federal Employees Health Benefits Program (universal FEHBP), which would contract with and offer a variety of health plans in each community-rated market. Plans offered through the universal FEHBP would charge enrollees the same premiums as they charged others in the community-rated market but could offer an administrative discount. Federal employees would be fully integrated into the universal FEHBP after a seven-year transition period, which could start no earlier than 2000. Initially, federal employees would remain in the existing FEHBP, but the benefits would be conformed to the guaranteed national benefit package and the government's average contribution would increase to reflect the provisions of the mandate on employers. In the fifth through seventh years of the transition period, federal employees could enroll in either the current program or the universal FEHBP.

States would have considerable flexibility to set up their own health reform programs as long as they assured universal coverage, provided the guaranteed national benefit package, and controlled costs. They could establish a single-payer system, voluntary or mandatory consumer purchasing cooperatives, or an all-payer system to reimburse health care providers.

Medicaid

Medicaid would no longer cover acute care services, except for emergency benefits for illegal aliens through 2001, but would continue to cover long-term care. States would be required to make maintenance-of-effort payments to the federal government based on the amount by which their Medicaid spending was reduced. The maintenance-of-effort amounts would be computed separately for Medicaid beneficiaries who received benefits from Supplemental Security Income (SSI) or Aid to Families with Dependent Children (AFDC) and for those who did not receive cash benefits. States would pay 100 percent of the full maintenance-of-effort amounts in 1999 through 2001, 96 percent in 2002 and 2003, and 86 percent thereafter.

Medicare Parts A and B

The existing Medicare program (Parts A and B) would be expanded by adding a prescription drug benefit and various preventive benefits, increasing coverage of mental health services, eliminating the lifetime limit on inpatient hospital days, and capping out-of-pocket expenditures starting in 2004. Savings would be achieved by imposing 20 percent coinsurance on home health services, reducing disproportionate share adjustments for hospitals, scaling back payments for the indirect costs of medical education, and making other, smaller changes. The rate of growth of Medicare spending would also be tightly limited, as described below.

Medicare Part C

Medicare Part C would begin in 1999. Net of subsidies for low-income families, the program would be financed largely by premiums paid by enrollees and their employers. For the first four years, however, premiums would be established under the assumption that 60 percent of the eligible population was enrolled in the program.¹ Also, disabled SSI recipients would be excluded in calculating the premium. General revenues would be used to make up the shortfall resulting from these two constraints on premiums.

Enrollment in Part C would be open only to people (and their families) who did not work full time, worked full time for a small employer that did not offer coverage under a private certified health plan, worked for a small employer and were eligible for a federal subsidy of their premium, or received benefits from SSI or AFDC. Alternatively, nonworking people, subsidized employees of small firms, and SSI or AFDC recipients could enroll in a certified health plan offered in the community-rated market.

Because no one would be required to enroll in Medicare Part C, estimating the number of people covered by the program, their use of health care services, and the required premiums is particularly difficult. The estimates assume that 80 percent of nonworking people and former Medicaid beneficiaries and 50 percent of people connected to small firms and part-time employees in large firms would ultimately enroll in the program. They also assume that the participation rate of small employers, some of whom might initially be reluctant to enter the new public program, would rise from 10 percent in 1999 to 50 percent in 2004.

Reimbursement of health care providers under Medicare Part C would follow the approaches currently used by Medicare Parts A and B. The estimates assume that Medicare payments would initially be 10 percent below the amounts that would be paid on behalf of Part C enrollees if they had private insurance. Under these assumptions, the estimated average premiums in 1994 for the three classes of enrollment are as follows:
 

With similar premiums inside and outside Medicare, as the estimates assume, private health insurance could continue to compete and coexist alongside Medicare Part C. Such a scenario would also require premiums inside and outside Part C to have similar rates of growth. If Part C became the insurer for disproportionate numbers of high-risk people, its premiums could soar and it could end up dealing with a smaller group of high-cost enrollees, much like the present Medicaid program. At the other extreme, if premiums for Medicare Part C were low and small employers wished to simplify their administrative costs for insurance by not offering private insurance, Part C could become dominant and drive private health insurance out of the small-group market. Neither of these alternative outcomes can be ruled out.

Cost Containment

The proposal would set target rates of growth for the Medicare program (Parts A, B, and C, together) and for the private sector. It would set Medicare's payment rates accordingly and would establish a standby system of cost containment for the private sector.

Medicare's cost controls would go into effect in 1996 for Parts A and B and in 1999 for Part C. The target for total Medicare spending per capita would increase by the rate of growth of gross domestic product (GDP) per capita plus 1.8 percentage points in 1996 and by lesser amounts thereafter. In 2000 and beyond, the target would increase by the five-year average rate of growth of GDP per capita. The per capita estimates would be allocated among 10 or more classes of health care services using complex procedures specified in the proposal The Secretary of Health and Human Services (HHS) would set reimbursement rates for providers, with the goal of meeting the targets.

Spending targets would also be established for the private sector. The per capita targets would be allocated by class of service, as in Medicare, and by state of residence, and the Secretary of HHS would determine maximum payment rates that corresponded to the targets. The maximum payment rates would be only advisory through 2000. Starting in 2001, however, they would become mandatory in states that exceeded their per capita spending target.

The Congressional Budget Office believes that expenditure limits enforced by rate setting could be reasonably but not totally effective in controlling Medicare spending. The Health Care Financing Administration collects most of the data necessary to set rates and track spending relative to the targeted amounts for Parts A and B. It also has considerable experience in setting payment rates and estimating the responses of providers. Nonetheless, the history of cost control efforts both in this country and abroad strongly suggests that setting payment rates is not sufficient for achieving full control over health expenditures.²

CBO's estimates assume that the limits on Medicare spending would ultimately prove to be 75 percent effective and that providers would shift one-fourth of the Medicare savings to private payers. Although the limits would apply jointly to Parts A, B, and C, initially they would probably be more successful in Parts A and B than in Part C, which would be new and untested. In Parts A and B, the 75 percent rate of effectiveness is assumed to apply from the start. In Part C, however, the maximum rate would be reached only gradually, as the quality of data improved and experience with the program grew. The estimates assume that the expenditure limits in Part C would be ineffective in 1999 and 2000, 25 percent effective in 2001 and 2002, and 50 percent effective in 2003 and 2004.

The limits on non-Medicare spending are more likely to be breached and to be less effective. The task of establishing a reporting system for national health expenditures as specified in the proposal would be formidable. States would be permitted to operate their own payment systems as long as the growth in health care spending did not exceed what it would have been under the maximum rates--a difficult calculation to make. The estimates assume that the limits on private health spending would be ineffective in 2001 through 2003 and 25 percent effective in 2004.

Low-Income Assistance

The proposal would offer three types of low-income assistance: premium subsidies, cost-sharing subsidies, and wraparound benefits.

Premium Subsidies. Low-income people would be eligible for federal subsidies to reduce their liability for health insurance premiums. Qualified Medicare benficiaries--those with income up to 120 percent of the poverty level--would be eligible for special subsidies for Part B premiums. (Currently, Medicaid pays those premiums for qualified Medicare beneficiaries.) Temporary subsidies would be provided to certain early retirees and to employers required to pay for the health benefits of retirees. Small firms with low average wages would be eligible for a tax credit.

A family with modified adjusted gross income (AGI) below a threshold (approximately equal to the federal poverty level) would receive a subsidy equal to its portion of the Medicare Part C premium. From 1999 through 2001, the subsidy would phase out between 100 percent and 200 percent of poverty. The upper limit of the phaseout range would increase to 220 percent of the poverty level in 2002 and to 240 percent in 2004 and thereafter. People participating in Part C would have their tax liability reduced. People participating in certified health plans would be given a premium certificate, or voucher, equal to the appropriate percentage of the premium for Part C or the certified health plan, whichever was lower.

In addition to receiving any regular premium subsidies for which they were eligible, certain early retirees would have their premium liability limited to a percentage of modified AGI. The provision would apply to people ages 55 to 64 in 1994 who did not work full time and had income below $30,000 for an individual and $40,000 for a couple. The cap would equal 7 percent of modified AGI in 1997 and 1998 and fall to 4 percent in 2001 and thereafter. As a result of this provision, the federal government could pay 50 percent or more of the costs of health insurance for some early retirees who would not otherwise have received subsidies. Moreover, similar retirees who were not members of the specified age cohort would receive no additional financial assistance from the government.

Employers who paid anything for retirees' health coverage in January 1994 would be required to make maintenance-of-effort payments and would be eligible for special subsidies. Employers subject to this requirement would have to pay 80 percent of the cost of coverage (or 80 percent of the Part C premium, if less) for all retirees ages 55 to 64 in 1994 and their dependents, regardless of the amount they previously paid. Such employers would be eligible for a federal subsidy, however, equal to 40 percent of the applicable Part C premium.

From 1999 through 2005, small firms with low average wages would be eligible for a tax credit to reduce their liability for the costs of health insurance. Employers with no more than 25 employees and an average wage of no more than $14,000 per full-time-equivalent employee would receive a credit of 50 percent in 1999 through 2003, 30 percent in 2004, 15 percent in 2005, and nothing thereafter. For employers with 26 through 50 employees, the credit would equal 37.5 percent in 1999 through 2003, 20 percent in 2004, and 10 percent in 2005. The credit would be reduced proportionately for small employers with an average wage between $14,000 and $26,000.

Cost-Sharing Subsidies and Wraparound Benefits. Cost-sharing subsidies would be available to qualified Medicare beneficiaries as under current policy. Qualified Medicare beneficiaries with income below 100 percent of the poverty level would receive assistance for paying deductibles and coinsurance under Parts A and B of Medicare.

Cost-sharing subsidies and wraparound benefits would also be provided for other people with income below the poverty level, children and pregnant women with income below twice the poverty level, and AFDC and SSI recipients. Those beneficiaries would be relieved of all cost-sharing requirements, and payments would be made to certified health plans based on the cost-sharing amounts for Medicare Part C. In addition, those beneficiaries would receive wraparound benefits--that is, benefits not included in the guaranteed benefit package. Among those benefits would be early and periodic screening, diagnostic, and treatment services for children and vision and hearing care for adults. CBO's estimates assume that children covered by this provision would receive benefits equivalent to those currently provided by Medicaid.

Other Spending and Revenues

The proposal would increase spending on various public health programs, establish a capped entitlement program to provide grants to states for long-term care, and provide for enrollment in certified health plans offered by the Department of Veterans Affairs and the Indian Health Service. Outlays for Social Security retirement benefits would increase slightly because the assurance of access to health insurance and the provision of subsidies to low-income retirees would encourage some workers ages 62 to 64 to retire earlier. The federal government would also incur additional costs to administer the provision of low-income assistance, Medicare Part C, universal FEHBP, and private-sector cost controls and to regulate the health care system.

The Budget Enforcement Act of 1990 divides spending into two categories for purposes of budgetary control. Spending provided in annual appropriation acts is termed discretionary spending and is subject to dollar limits on budget authority and outlays. Spending established by permanent law is called mandatory spending and, along with receipts, is subject to a pay-as-you-go requirement. Spending for the veterans health and Indian health programs is currently discretionary, but the proposal would make much of that spending mandatory. The proposal would reduce the discretionary spending limits in the Budget Enforcement Act to take account of these shifts in classification, but it would not raise the limits to allow for the increases in discretionary administrative costs.

The Joint Committee on Taxation has estimated the impact of the provisions of the proposal that would affect federal revenues, other than Medicare Part C premiums and payments by nonenrolling employers. Requiring employers to contribute to the cost of health insurance would significantly increase their spending on health. The estimate assumes that most of those additional costs would be passed on to workers in the form of lower cash wages. As a result, federal revenues would fall because the lost wages would no longer be subject to income and payroll taxation. The decline in revenues from this source would reach $58 billion in 2004. Revenue-raising provisions of the proposal include a 2 percent excise tax on private health insurance premiums, a phased increase of 45 cents a pack in the tax on cigarettes, the extension of Hospital Insurance taxation and coverage to all employees of state and local governments, and an increase in the Supplementary Medical Insurance premium for high-income people.

Budgetary Treatment of the Mandate

A mandate requiring that individuals purchase health insurance would be an unprecedented form of federal action. The government has never required individuals to purchase any good or service as a condition of lawful residence in the United States. Therefore, no budgetary precedents or concepts exist that would provide conclusive guidance about the appropriate budgetary treatment of such a mandate. Sound arguments can be made both for and against including in the federal budget all of the costs that individuals and firms would incur in complying with the mandate. It is only appropriate, therefore, for policymakers to resolve the issue through legislation.³ Pending resolution of the issue, Table 1 includes the outlays and receipts of Medicare Part C, which would clearly be a federal program, but does not include the outlays and receipts of certified health plans purchased in compliance with the mandate.
 

OTHER CONSIDERATIONS

Like many proposals to restructure the health care system, Congressman Gephardt's would require extensive changes in the current system of health insurance. It differs in fundamental ways, however, from most other proposals--including those that would impose a mandate on employers to provide coverage. Congressman Gephardt's proposal is unique in its inclusion of an option sponsored by the federal government as well as private-sector health insurance, and in the features of the regulatory structure for containing health care costs.

CBO's estimates assume that the proposal could be implemented within the scheduled time frame and would work basically as intended. Nonetheless, there is a significant chance that the substantial changes required by this and other systemic reform proposals could not be achieved as assumed. The following discussion summarizes the major areas of potential difficulty as well as some other possible consequences of the proposal.

Cost Containment Initiatives

The proposal adopts a complex regulatory approach to containing health care costs, with parallel but separate cost containment systems for the private sector and Medicare. Both systems would be critically dependent on reliable, timely data, many of which are not currently available. Although CBO believes that the system for containing Medicare's costs would have a better chance of success than the one for containing costs in the private sector, even the constraints on the Medicare system would be difficult to implement. Moreover, as designed, the cost containment systems could have a variety of unintended consequences and might raise concerns about both equity and efficiency.

The same underlying principles would guide the cost containment programs for both the private sector and Medicare, but the methods of implementation would differ. In both sectors, the federal government would attempt to control costs for 10 or more separate classes of service including inpatient hospital services, outpatient hospital and ambulatory facility services, physician and other professional services, home health and hospice care, rehabilitation services, diagnostic testing services, durable medical equipment and supplies, prescription drugs, nursing facility services, and mental health services. Bach class would be allocated a per capita spending target based on a share of an overall per capita spending target. The overall targets for Medicare A/B and the private sector would be based on actual per capita spending in those sectors in 1993, inflated by the estimated rate of growth in per capita spending in each sector until 1995.⁴ The targets for 1996 and subsequent years would be the 1995 estimates inflated by the rate of growth of GDP per capita plus specified percentages that would decline over time. The target for Medicare Part C would be based on an estimate of per capita spending in that program in 1999. By 2000, the per capita targets for Medicare and the private sector would all be increasing at the rate of growth of GDP per capita, which would be the permanent growth rate for the targets.

The Secretary of HHS would establish the classes of service and would also define the services to be included in each class. The classes could not subsequently be changed without Congressional approval. Because the classes would be allocated different portions of the overall per capita amounts and would be permitted to grow at different rates, health plans and providers would have a keen interest in the initial allocation of services to classes.

The share of per capita spending allocated to each class of service would depend on the initial share in the base year and a fixed trend factor, representing an assumed annual rate of growth for the class. In the private sector and Medicare Part C, the trend factor would be the average rate of growth in private per capita spending for the class for the five-year period ending in 1995 (modified, if necessary, for Part C). For Medicare A/B, by contrast, the trend factors have been written into the proposal and, for some classes of service, are likely to differ considerably from the 1990-1995 average rate of growth. Regardless of how the trend factors were determined, however, over time the classes of service with higher trend factors would increase their shares of total spending relative to classes with lower trend factors. Moreover, the relative growth rates of different services would probably diverge increasingly from unconstrained relative growth rates, affecting treatment patterns, changing the incentives for the development of new medical technologies, and limiting the ability of the health care system to adopt more efficient methods of health care delivery.

Implementing Cost Containment Mechanisms in the Private Sector. The provisions to contain health care costs in the private sector would be implemented at the state level. Accordingly, the Secretary of HHS would have to establish state-specific per capita spending targets, using adjustment factors for each state that reflected differences in prices and patterns of service use. The per capita targets would be based on patients' state of residence, regardless of where they received their care.

Beginning in 2000, if the Secretary of HHS determined that a state's actual per capita spending exceeded its target, maximum payment rates would be imposed on providers. Those rates would be estimates of the rates necessary to achieve the national per capita spending targets for each class of service. Because no measures would be taken until 2001, spending in some states might be significantly out of compliance by that time. (Maximum payment rates would be only advisory through 2000.)

The amount of data required to develop and implement this system would be enormous. Establishing the baseline allocation of spending among classes would require detailed information on the allocation of spending by health maintenance organizations (HMOs)--which is not generally available now-as well as claims data from other providers. Computing the state adjustment factors would require data on states' private per capita expenditures (for state residents) and patterns of service use, as well as demographic information. (These requirements would be in addition to the information on geographic variation in wages, prices, and other costs of medical practice that Medicare currently uses when setting reimbursement rates.) Moreover, information on states' per capita expenditures would be needed on an ongoing and timely basis to determine whether states were in compliance with the spending targets. To obtain the necessary information, the Secretary of HHS would have to rely on reporting from health plans and providers, and they would not have strong incentives to be either timely or accurate.

Another concern raised by this approach to cost containment is one of equity for providers. The tracking and regulation of spending would be based on patients' residence rather than providers' place of practice. Providers could, therefore, be penalized because of rapid growth in spending by their state's residents for health care received from providers in other states. This outcome could be a particular problem in the many states with multistate metropolitan areas.

Implementing Cost Containment Initiatives in Medicare. In contrast to the approach for the private sector, Medicare's cost containment provisions would be implemented at the national level. The Secretary of HHS would establish a per capita expenditure target for Medicare, which would be the weighted average of the targets for Medicare A/B and Medicare Part C. Per capita spending would be allocated to classes of service separately for A/B and C, and a weighted average would then be calculated for each class. Medicare's payment rates would be set so that the combined Medicare expenditures for each service would be consistent with the combined per capita allocation to each class.

This approach seems more likely to reduce the rate of growth of spending than the private-sector approach because Medicare spending would be easier for the government to track and because each class of service would be separately monitored and regulated on an ongoing basis. Moreover, rate setting would be prospective rather than retrospective; that is, rates would be set to hit spending targets rather than modified after the fact if the targets were exceeded. Nonetheless, the approach would still be extremely complicated and have extensive requirements for data.

Effects of the Cost Containment Provisions. The combined effects of the cost containment provisions for the private sector and Medicare are difficult to predict. If implemented as intended, the overall per capita spending amounts would be constrained to the same maximum growth rates after 2000, although the rates of growth of per capita spending for different classes of service in Medicare and the private sector would differ, as would the relative prices of services. Those differences might result in different patterns of service use in the private and public sectors. But if cost containment proved to be more successful in the Medicare program than in the private sector, providers would generally have incentives to allocate more resources to private-sector clients and fewer to Medicare beneficiaries. Moreover, if efforts to contain costs in the private sector proved ineffectual, providers might be able to recoup some of their reduced Medicare revenues by charging more to private-sector clients. (Such behavior would, however, result in a growing divergence between Medicare Part C and private premiums, causing more individuals and small employers to enroll in Part C.)

Setting Medicare's payment rates jointly for A/B and C would raise a variety of complex problems because--for each class of service--analysts would have to predict the behavioral responses of providers to price changes in two different markets. The rates of growth of spending in those markets would differ, as would the characteristics and patterns of service use among their respective enrollees. Although A/B and C would have separate targets for the allocation of per capita spending among the classes of service, those targets would probably not be met individually because prices would be set to meet the joint per capita spending target.

Market Structure. Adverse Selection, and Risk Adjustment

The issue of how effectively community-rated markets would function has been a concern in all the health care proposals that CBO has analyzed. If the average health status of enrollees varied significantly among health plans, plans with less healthy enrollees would have difficulty competing unless appropriate steps could be taken to compensate them for their higher-risk clientele. But the development and implementation of reliable risk-adjustment mechanisms is likely to remain an elusive goal, at least in the immediate future.

The problems of adverse selection could become more severe as people's health insurance choices expanded, giving them greater opportunities to self-select into groups according to their health status and preferences for health care. Depending on the proposal, those choices could be of four basic kinds:

• whether to obtain health insurance at all, a choice that arises only in proposals that do not mandate health coverage;
  
  
• what market institution or agent to use, a choice that arises in all proposals that do not mandate the use of a single purchasing organization;
  
  
• what benefit package to choose, a choice that arises in proposals that do not require a single standard benefit package; and
  
  
• what type of plan to choose--for example, a plan that allows an unlimited choice of providers versus an HMO.

Since Congressman Gephardt's proposal would require everyone to have insurance coverage, the decision whether to participate would not be an issue. But the other three types of choices would all arise.

Assessing possible responses to those choices is difficult, both because of the range of options that individuals and small employers would face and because of ambiguities in the bill. Not only could individuals and small employers choose between Part C and private insurance, but those selecting private insurance could purchase it directly from insurance companies, through the universal FEHBP, through a purchasing cooperative (if one was established), through association plans (if they were members of an association sponsoring a plan), or through state-sponsored enrollment sites. Some people enrolled in private-sector plans could also choose between standard and catastrophic coverage.

Although Medicare Part C is clearly intended to be outside the community-rated market, the status of some of the private-sector purchasing options is uncertain. Resolving their status is important because risk adjustment would occur only within the community-rated market. The proposal's apparent intent is that all individuals and small employers purchasing private health insurance should participate in the risk-adjustment process, and that is what CBO assumed in its estimates. This requirement would reduce the problems of adverse selection that would arise if some of the purchasing options were excluded. But the existence of multiple purchasing arrangements would complicate the process.

It is also unclear whether risk adjustment would occur across plan types within the community-rated market. If risk adjustment was implemented within but not between plan types, the option to choose catastrophic coverage would provide another avenue for adverse selection.

Because Medicare Part C would be the default insurer for anyone who did not have private coverage, that program might enroll an unfavorable risk pool. But that conclusion is by no means certain. Some of the uninsured who did not seek out coverage, and ended up in Part C by default, would be young and healthy. By contrast, some subsidized people with poor health would probably take advantage of the option to purchase insurance outside Part C. As discussed below, however, the lack of risk adjustment between the private sector and Part C means that the potential for unstable premiums and enrollment in either Part C or the community-rated pool is significant. That result could occur if either pool experienced serious problems with adverse selection at the expense of the other.

The Path of Premiums and the Sustainability of the System

CBO's estimates of Congressman Gephardt's proposal assume that Medicare Part C and a private, community-rated health insurance market could coexist over time with no risk adjustment between them. This assumption in turn embodies additional assumptions about the relationship between the premiums for Part C and those for private insurance.

Initial Premiums and Subsequent Growth. Whether individuals and small employers would have a meaningful choice between public and private-sector insurance coverage would depend on the relationship between Part C premiums and private premiums--initially and over time. A variety of scenarios is possible. For example, Part C premiums might start out higher than private premiums, in which case few people would choose to enroll except for those who were fully subsidized (and who therefore would not have to pay any of the difference in the premiums) and those who ended up in Part C by default because they did not actively seek out a private plan. In these circumstances, the program might never expand beyond a very limited base.

Alternatively, Part C premiums might start out lower than private premiums but grow more rapidly because, say, of adverse selection. In that case, Part C would lose enrollees over time as its competitive advantage evaporated. The higher rate of growth would mean that Part C premiums would eventually exceed private premiums, again resulting in an enrollee base composed primarily of people whose premiums were fully subsidized.

A third scenario--the one assumed by CBO--is that lower reimbursement rates and more effective cost containment would result in premiums for Part C starting out lower than private premiums and growing more slowly. Achieving the long-run rate of enrollment in Part C would take several years under this scenario. The proposal requires, however, that initial premiums be set on the basis of the program's ultimate enrollment. That assumption would probably result in premiums that reflected a population with lower average risk than the initial enrollees in Medicare Part C. Consequently, additional outlays would be necessary to make up the shortfall in the first few years.

A variation on the previous scenario that cannot be ruled out is that the gap in growth rates between Part C and private premiums might be much larger than CBO has assumed. In that case, Part C enrollment would increase more quickly than assumed here, and the Part C program could eventually become the dominant insurer for individuals and small firms and possibly their only choice in some markets.

Sources of Uncertainty. The federal outlays required to make up the shortfall in Part C premiums in the early years of the program would--in effect--constitute additional premium subsidies to Part C enrollees. If Part C enrollment (and the corresponding risk composition of enrollment) did not reach the levels assumed by CBO or did not do so as rapidly, the federal government might have to continue such subsidies indefinitely. Letting Part C premiums rise to their actuarially correct level would only exacerbate the situation, because such an increase would further slow (or reduce) growth in enrollment while converting outlays to cover premium shortfalls into direct subsidies for low-income families enrolled in Part C. In addition, higher Part C premiums would increase federal outlays for subsidies to people enrolled in private plans.

Considerable uncertainty also surrounds the assumption that Medicare Part C would constrain the health spending of its enrollees more effectively than would private insurance plans. Alternative assumptions about the relative effectiveness of the cost containment measures would have important implications for subsidy costs.

Depending on how it occurred, cost containment in the private sector that proved more successful than CBO assumed could either increase or decrease federal subsidies. Subsidies would increase if private-sector cost containment resulted from preferential risk selection, which caused more higher-risk people to enroll in Part C. As Part C premiums rose to reflect the changing mix of risks, a greater fraction of private premiums would be subsidized. Subsidies would decrease if more effective containment of private-sector costs resulted from reduced spending per capita (say, because of lower reimbursements to providers). At the same time, however, more effective containment of private-sector costs would reduce people's incentives to enroll in Part C.

Cost containment under Part C that was less effective than CBO assumed would have two effects. First, it would reduce people's incentives to enroll in Part C, thus increasing costs and subsidies for Part C enrollees. Second, it would increase subsidies to people in private plans even if Part C remained less expensive than those plans (because subsidy amounts would depend on the level of the Part C premium).

Insurance Costs for Moderate-Sized Firms

Under Congressman Gephardt's proposal, participation in the community-rated market would be restricted to individuals and employees of firms with 100 or fewer employees. Participation in Medicare Part C would be restricted to those groups plus AFDC and SSI beneficiaries, as well as part-time, seasonal, and temporary employees. Larger firms would have to self-insure or offer experience-rated coverage obtained from an insurance carrier.

Moderate-sized firms (those with, say, 101 to 200 employees) might face relatively high premiums under this structure, not only because they would be experience-rated but also because of the requirement that they offer their employees a choice of at least two plans. The same types of concerns have arisen with other proposals that have similar provisions. But the potential problems are particularly pronounced in Congressman Gephardt's proposal because a firm's size would be determined by the total number of employees--both full-time and part-time--rather than the number of full-time or full-time-equivalent employees. Enrollment in some employers' plans could therefore be extremely small because part-time employees, as well as employees in families with two workers, could obtain coverage elsewhere.

Small enrollments would result in high administrative costs. Also, one employee with a costly medical problem could raise a plan's premiums significantly. Some plans could end up with increasing premiums and shrinking enrollment as employees either switched to a cheaper plan offered by the firm or sought coverage elsewhere, if they had that option. In some cases, any plans offered by the firm could prove quite expensive.

Insurance Costs for Federal Employees

The proposal's FEHBP and universal FEHBP provisions could result in some federal employees paying more for health insurance, although the effects would vary in markets across the country. The intent of the provisions is to allow people who would be eligible to enroll in community-rated health plans to have access to the same choices as federal employees. Ultimately, federal employees would all be enrolled in universal FEHBP plans and would be charged the same premiums as everyone else enrolled in those plans, but the integration would take place over a period of seven years. At the end of that period, federal employees would no longer have the choice of enrolling in national plans as they do today; they could enroll only in universal FEHBP plans offered in their community-rating area.

The federal employees who might end up paying more under this structure are those who live in high-cost markets and who can currently obtain lower premiums by enrolling in national rather than local FEHBP plans. But some federal employees in relatively low-cost markets might find themselves better off, with a wider range of health insurance choices available at local community rates. Many other federal employees might initially be better off because the federal government would pay a higher percentage of their premium than at present, reflecting the requirements of the mandate on employers. Over time, however, wages or other fringe benefits would probably adjust to reflect the increased share of compensation going to health care.

Effects on HMOs

Most health care proposals would affect the market position of HMOs relative to fee-for-service plans in a variety of ways. Because Congressman Gephardt's proposal would build on the Medicare model to expand health insurance coverage, some analysts believe that it would promote fee-for-service medicine relative to managed care--since Medicare is still primarily a fee-for-service program. Extrapolating from Medicare's experience in the managed care market to Part C is risky, however, because Part C's enrollees would be so different from the current Medicare population. (The fact that only a small percentage of current Medicare beneficiaries enroll in HMOs may be more a reflection of the preferences of the elderly and disabled populations than an inherent feature of the program.) The proposal actually contains a variety of opportunities, incentives, and disincentives for people to enroll in managed care or indemnity plans. It also includes some provisions that might weaken HMOs' ability to contain their costs.

Part C enrollees could select an HMO, if HMOs chose to participate in the program. HMOs would be paid in essentially the same way as they are today by Medicare; that is, for each enrollee, HMOs would receive about 95 percent of the average per capita cost of comparable Part C enrollees in their community who were not enrolled in HMOs. The willingness of HMOs to participate under those conditions would depend on the relationship between their average costs for Part C beneficiaries and Medicare's payment (and would reflect their success in enrolling relatively healthy Part C beneficiaries).

Private-sector enrollees could select a health plan that offered an unlimited choice of providers (a UCP plan) or an HMO, if one was available. That provision would expand choice for all people whose employers currently offer only one plan. Since employers are increasingly shifting away from indemnity coverage to managed care, more people would probably continue to have access to UCP plans than would have in the absence of the requirement to provide a choice of plans.

The incentives for Part C and private-sector enrollees to join an HMO would differ considerably. Part C enrollees would pay the same "premium" through the tax system regardless of the type of plan in which they enrolled. Consequently, the only costs they would be concerned about would be deductibles, coinsurance, and copayments. Because the UCP standard option would have relatively high cost-sharing requirements, an HMO would probably be an attractive option to moderate-income families who were ineligible for cost-sharing subsidies and could not afford to purchase a cost-sharing supplement. Private-sector enrollees, by contrast, would base their cost comparisons on premiums as well as cost sharing--much as they do today. Employers offering private-sector plans would, however, be required to contribute at least 80 percent of the premium of the lowest-cost plan of each type, possibly reducing the incentives for their employees to select the lowest-cost type of plan.

The extensive subsidies for cost sharing in Congressman Gephardt's proposal would significantly reduce incentives to enroll in HMOs, regardless of whether the eligible populations were enrolled in Part C or the private sector. Unlike some other proposals with cost-sharing subsidies, this one would not generally require people who were eligible for subsidies to pay even nominal cost-sharing amounts, essentially providing them with first-dollar coverage--that is, coverage with no deductibles, coinsurance, or copayments. (The subsidies would be based on the applicable cost-sharing amounts under Medicare Part C.)

The proposal would also place several requirements on HMOs that would restrict their ability to control costs through tightly managed networks of providers. Some of those provisions were incorporated in other proposals. Examples include requiring HMOs to contract with an extensive range of so-called essential community providers and centers of excellence; prohibiting HMOs from requiring women to obtain referrals to obstetricians and gynecologists; and requiring all HMOs that use networks of providers to allow any licensed provider to participate on the same terms as other providers in the network. (Because of some qualifying language in the bill, it is unclear how the latter requirement would actually be interpreted.)

Responsibilities of the Federal and State Governments

Most proposals to restructure the health care system incorporate major additional administrative and regulatory functions that new or existing agencies or organizations would have to undertake. Questions arise, therefore, concerning the capabilities of government agencies to fulfill their responsibilities.

The federal government would play a larger role in the health care system under Congressman Gephardt's proposal than under most other recent health care proposals. The greater federal involvement would result from the proposal's extremely complex regulatory structure, the establishment and operation of Medicare Part C, and the proposed approach to implementing the system of subsidies. The states would also have important new tasks to perform, but because they would not be responsible for implementing the subsidy system, they would have fewer obligations than under some other proposals.

Since the enrollment process for Medicare Part C would be run primarily through the tax system, the Treasury would assume major new responsibilities under this proposal. It would have to track the tax obligations of individuals enrolled in Medicare Part C at some time during the year, taking into account the reduced tax obligations of those eligible for subsidies. In addition, it would have to track the Part C taxes owed by employers, including employers' tax obligations for their nonenrolling employees.

The Department of Health and Human Services would also have a greatly expanded role. It would be responsible for paying Part C claims, issuing and redeeming vouchers for premium subsidies for low-income people enrolling in private plans, administering cost-sharing subsidies for low-income people enrolled in Part C and private plans, and developing and implementing the cost containment initiatives for the private sector and Part C (Presumably some of those tasks could be contracted out to private-sector organizations, just as claims processing is handled under Medicare now.) In addition, the department would bear the primary responsibility for setting standards for certifying health plans' data systems and quality assurance mechanisms, developing and implementing a system for verifying enrollment, establishing a reporting system for national health expenditures, designing supplemental benefit packages, and developing model risk-adjustment mechanisms.

Similarly, the Department of Labor and the Office of Personnel Management would carry out important functions. The former would be responsible for certifying and monitoring self-insured health plans and operating a reserve fund to pay the claims of insolvent plans; the latter would design and implement the universal FEHBP and integrate it with FEHBP, which would be an extremely complicated undertaking.

Although states could develop single-payer or managed care systems, operate state reimbursement systems, or establish purchasing cooperatives, they would be under no obligation to do so. They would, however, have to assume a variety of responsibilities related to the effective functioning of the health insurance markets and quality assurance. States would, for example, certify health plans, provide uniform information for consumers on all insured health plans, provide enrollment assistance and establish enrollment sites, and set up guaranty funds to pay the claims of insolvent carriers. They would also monitor health plans' compliance with quality assurance requirements, assess patients' satisfaction, and publish annual reports on the performance of health plans.

The Effects on Health Spending by Employers

Congressman Gephardt's proposal would maintain the central role of employers in financing health care but would alter the distribution of costs among employers and workers. Total spending by employers would increase significantly because they would have to pay for insurance for workers who are currently uninsured. CBO estimates that in 2000, all employers together would pay over $110 billion more for health insurance under this proposal than if the current system continued unchanged. The increase in spending would be even larger in subsequent years, exceeding $170 billion a year by 2004.

Even though the proposal would raise the overall cost of health insurance for employers, it would have widely differing effects on individual firms and industries, raising costs in some cases and reducing them in others. Three factors account for most of the diversity. First, the requirement for all employers to contribute to health insurance would raise spending by firms that currently do not. Second, the requirement for small firms to participate in either Medicare Part C or the private, community-rated market would probably raise the insurance costs of small firms employing younger, healthier workers and lower them for small firms employing older, less healthy workers. Third, the temporary subsidies to small firms with low average wages would reduce their cost of insurance relative to the cost faced by larger firms or firms with higher average wages.

Who Ultimately Pays for Health Spending by Employers?

Although employers initially pay a large portion of the bill for health insurance, other people ultimately bear these costs. Workers may pay them in the form of lower wages, consumers in the form of higher prices, and shareholders through lower returns on their investments. But economic theory and empirical research both indicate that workers bear most of the cost of employers' premiums for health insurance. Thus, the significant increase in costs that Congressman Gephardt's proposal would produce compared with current policy would be largely passed on to workers in the form of lower wages.

This increase could be particularly burdensome for families with low income. For example, consider a family of one adult and two children, with income just below the poverty threshold in 2000. If the adult worked at a firm with more than 50 full-time employees, the firm would pay more than $5,400 on the worker's behalf for insurance (80 percent of CBO's estimated single-parent premium in that year); that amount would represent roughly 45 percent of the family's income. If the adult worked at a smaller firm with average wages below $26,000, the firm's payment would be reduced by up to $2,700 but would still represent more than 20 percent of the family's income. At those prices, the family might well have preferred not to buy insurance at all, especially if it could obtain publicly provided emergency care for serious health problems, as many people can today.

Effect on Job Opportunities of Certain Minimum-Wage Workers

Although most workers would bear the mandated insurance costs through lower wages, the cash wages of workers earning close to the minimum wage could not fall. As a result, the net cost of employing those workers would increase under this proposal, and fewer adult low-wage workers would be able to find jobs.

Under Congressman Gephardt's proposal, the cost of employing minimum-wage workers would increase significantly above the $4.57 per hour that employers currently pay to cover the federal minimum wage and their portion of the payroll tax. For example, in 2000, unsubsidized employers would have to pay a minimum of $6.25 per hour for a single worker and $8.40 per hour for an enrolling family worker. Moreover, the subsidies for small firms with low average wages would not greatly reduce those hourly costs on average because the subsidies would apply to only a limited group of minimum-wage workers, would be less generous than the subsidies in other health proposals (such as the Administration's), and would be phased out over time.

Some employers would respond to those higher costs by hiring fewer adult minimum-wage workers. Although estimates of such impacts are highly uncertain, the number of employed adult minimum-wage workers could fall by half a million once the economy had fully adjusted to those higher costs. (That estimate is relative to employment levels in the absence of the proposal and, in some respects, may be conservative.)

Those losses in jobs for adults might be partly offset by job gains among other workers. For example, firms might replace some of their low-wage adult workers with more highly skilled workers; or they might employ teenagers under 18 or full-time students under 24, who would be exempt from the mandate. Although economic theory does not suggest an unambiguous gain in jobs for teenagers or other workers and the empirical literature on the subject is not extensive, the few empirical studies that exist tend to confirm such a substitution.

Work Disincentives

The proposal would discourage certain low-income people from working more hours or, in some cases, from working at all This disincentive for work arises from two features of the proposal--the treatment of nonworkers and the phaseout of family subsidies as family income increases. It is important to note, however, that work disincentives are an inherent feature of all health proposals that target subsidies toward the poor and near poor, and those subsidies could improve the well-being of many low-income people by assisting their purchase of health insurance.

Treatment of Nonworkers. The proposal would create an implicit tax on work because it would make health coverage universal without charging many nonworkers for the full cost of their insurance. Specifically, nonworkers in low-income families would receive sizable subsidies for the purchase of insurance; their coverage would not depend on whether they worked and paid the premium or stayed at home and paid much less. The premium would simply reduce take-home pay without, from the point of view of the individual worker, buying anything. The current system also discourages some of these people from working at firms that pay for insurance, but by requiring more firms to provide insurance coverage and granting full coverage to nonworkers, the proposal would increase the number of people who were affected.

Of course, the vast majority of workers would nevertheless remain in the labor market because they need wage and salary income to support themselves or their families. But people whose spouse is employed are more responsive to changes in work incentives because they can rely on their spouse's income. This proposal would reduce the participation of these workers in the labor force.

Phaseout of Subsidies. The proposal would reduce subsidies to low-income families as their income increased, creating an implicit tax on their economic advancement. With some exceptions, families with income below a threshold amount, which would be set roughly at the poverty level, could receive a full subsidy for any portion of their premium not paid for by an employer. In 1999 through 2001, the subsidy would be phased out as family income rose from the threshold amount to twice that amount. The upper end of this range would increase somewhat in later years, reaching roughly 240 percent of the poverty level in 2004.

Workers who earned more money within the phaseout range would have to pay more for health insurance, which would cut into the increase in their take-home wage. Rough calculations suggest that in 2000, the effective marginal tax on labor compensation would increase by 7 to 9 percentage points for workers at firms that paid 80 percent of the Medicare Part C premiums. Workers at firms that paid a larger share of the premiums would face a lower tax rate, and workers whose insurance premiums were lower than the Medicare Part C premiums would face the same rate over a smaller income range.

In 2004, the increase in marginal tax rates would be slightly lower than in 2000 because the subsidy would be phased out over a wider range of incomes. At the same time, the expansion of the phaseout range means that more workers would be affected by the increase in marginal rates.

Reallocation of Workers Among Firms

Like several other health care proposals, this one would encourage some reallocation of workers among firms in ways that would increase its budgetary cost. This sorting would occur because small firms with low average wages would receive a credit for some of their required payments; therefore, workers employed by such firms could receive larger take-home salaries than if they were employed by a firm that did not receive a credit.

Nevertheless, two features of the credit make the incentive for worker reallocation in the Gephardt proposal much smaller than the corresponding incentive in most other reform proposals. First, the amount of the credit is not very large. The maximum credit is 50 percent of health insurance costs and applies to the smallest firms with the lowest average wages; the credit is phased out for firms that are larger or have higher average wages. Second, the credit is temporary, so the benefit of worker reallocation would not persist for very long.

This document is available in its entirety in PDF.

1. This percentage is lower than the percentage in the bill and reflects a revised CBO estimate of the long-run rate of enrollment in Medicare Part C.

2. See Congressional Budget Office, Estimates of Health Care Proposals from the 102nd Congress,CBO Paper (July 1993).

3. See Congressional Budget Office, The Budgetary Treatment of an Individual Mandate to Buy Health Insurance, CBO Memorandum (August 1994).

4. For each class of service, a joint per capita target would be set for Medicare Parts A and B combined. Throughout this discussion, the joint target is referred to as the Medicare A/B target.