 |
|
Chapter
8Effects on Total Health Care Spending, the Scope of the Federal Budget, and the Economy
Proposals that would substantially change the health insurance market and health care systems could affect total spending on health care, the flow of payments between various sectors of the U.S. economy, and the operation of the economy. The Congressional Budget Office will consider those effects in its analysis of major health care proposals.
CBO will use data from the national health expenditures (NHE) and other sources to estimate the impact of various proposals on total health care spending. Some proposals might contain provisions that explicitly limit the rate of growth in total health care spending; such proposals might impose a global budget or budgetary cap for all or a part of that spending. In analyzing the effectiveness of such strategies, CBO will consider several factors, including the scope of the global budget, the targets selected for different categories of spending, and the mechanisms used to enforce the caps.
Changes to the health care system could also affect revenues and—more generally—the flow of funds between households, employers, and federal and state governments. Proposals that would affect the flow of payments might raise several budgetary concerns that have not been examined in earlier chapters of this report. Some proposals might assign the federal government a more active role in the health insurance market. For example, the government could be required to disburse subsidies to cover the cost of health insurance, collect health insurance premiums from policyholders, or make payments to insurers. Any of those changes might raise questions regarding who—the government, the insured, or the insurer—bears financial responsibility for any shortfalls in payments that might occur. Other proposals might require that individuals or businesses make payments directly to nongovernmental entities. Depending on the specific features of a proposal, CBO might judge that payments resulting from federal mandates should be recorded as part of the federal budget, even if the funds did not flow through a federal account.
Proposals that would make large-scale changes to the provision and financing of health insurance could also affect the operation of the broader economy in various ways. Because most health insurance is currently provided through employers, proposals could affect labor markets by changing both individuals’ decisions about whether and how much to work and employers’ decisions to hire workers. Such effects could arise in several ways:
■
Proposals that decreased the economic gains from an additional hour of work, through higher taxes or the phaseout of subsidies or credits for health insurance as income rises, could cause some people to work less or not at all.
■
Proposals that made health insurance less dependent on employment status could induce some people to retire earlier and others to change jobs more often.
■
Proposals that treated firms differently on the basis of characteristics such as size or average wages could affect the allocation of workers among firms.
■
Proposals that required employers to provide health insurance could adversely affect the hiring of employees earning at or near the minimum wage because the total compensation of those workers could exceed their value to the firm.
Proposals could also affect the size of the nation’s stock of productive capital, especially through their effects on government budgets. The net effect on the economy of a broad proposal to restructure the health care system would depend crucially on its details.
Effects on National Health Expenditures
Proposals that would significantly alter the health insurance market or the delivery of health care could also have a major impact on the amount and composition of overall spending on health care in the United States. CBO intends to include an estimate of the impact of major proposals on the total amount of national health expenditures in its analysis. For specific categories of spending, however, particularly by type of service, determining the likely impact presents greater challenges and may not be feasible.
Background on NHE Estimates and Projections
The Centers for Medicare and Medicaid Services produces estimates of total health spending in the United States—the national health expenditures—for current and future years. The largest component of the NHE is spending on health care goods and services, but the NHE also includes amounts spent on program administration, the net cost of private insurance (that is, the difference between benefits and premiums), public health activities, research, equipment, and structures. The estimates are sorted by the source of funding (private or public) and by the type of service (hospital, physician, pharmaceutical, and so forth).
CMS uses a variety of data and techniques to measure national health expenditures. For government programs like Medicare and Medicaid, CMS uses administrative records to determine current spending levels. Estimating spending on health care in the private sector presents a larger challenge because doctors, hospitals, and insurers are not required to report income to the federal government by type of service and sources of funding. To estimate that spending, therefore, CMS uses surveys from public and private sources, such as the Census Bureau, the National Center for Health Statistics, and private industry organizations (for instance, the American Hospital Association).
CMS also projects future health care spending by funding source and type of service. The projections are based on historical trends and observed relationships between health care spending and the economy. In its projections, CMS uses the same macroeconomic and demographic assumptions as those used in its annual report on the Medicare trust funds.1 CMS estimates that national health expenditures will total $2.6 trillion in 2009 and will grow to $4.3 trillion by 2017. (See Table 1-4 in Chapter 1 for CMS’s projections by funding source and type of service for 2009.)
The NHE estimates are especially useful for monitoring trends in total health care spending in the public and private sectors. They provide less insight into the impact of policy changes on particular segments of the population, however, mainly because the estimates are accounting measures of transactions and thus do not identify where the true responsibility for financing health care falls. The NHE accounts show the net cost of private insurance, for example, but they do not reveal who paid for the health insurance premiums, whether the premiums were initially paid by consumers or their employers, or whether employers’ payments for premiums were ultimately passed on to their employees through reductions in wages or in other forms of compensation. Similarly, the NHE category for medical research does not show the total amounts expended by all parties on studies and evaluations. Instead, that category includes only the amounts invested by government agencies or nonprofit organizations. Furthermore, although drug manufacturers and other commercial entities fund research efforts using at least some of the profits they earn from sales of drugs and other items, those research funds are not separately identified in the NHE data. As a result, complete information about the distribution or incidence of health care costs is missing from the current presentations of NHE data.
CBO uses CMS’s estimates and projections of national health expenditures in its analyses of the effects of proposals on health care spending. When appropriate, CBO adjusts those estimates to reflect its own baseline assumptions of spending for federal programs, differences in the technical and economic assumptions used by the two agencies, and other relevant factors. For example, CBO substituted its own projections of growth rates for drug spending for those used by CMS when producing estimates of the costs of Medicare’s drug benefit. As new information on health care spending becomes available, CBO might consider modifying the estimates from CMS in other ways.
Global Budgets for Health Care
Some proposals might try to restrain spending on health care by establishing a global budget. A global budget for health care applies the notion of a fixed payment for health care services to a health system, in whole or in part. In the same way that certain health plans cap aggregate payments to physicians, global budgets try to minimize incentives for providers to increase the volume of services in response to reduced fees. A global budget can function in a single-payer system (one in which a single entity—generally, a government agency—pays health care providers) or in a multipayer system (one in which individuals, employers, and the government jointly fund health care expenditures). A global budget could be constructed on a macro level, in which an overall spending target is set, or it could be devised on a per capita basis, with adjustments for age, sex, health status, and other determinants of health care spending.
The potential effects of such approaches, and the concerns and implementation challenges they raise, can be seen in Medicare’s current payment system for physicians as well as in examples from other countries that have adopted global budgets. In many regards, the challenges associated with global budgeting are also analogous—albeit on a larger scale—to those presented by administered pricing systems (see Chapter 5). Those experiences suggest that whether global budgets are effective depends on their scope, the targets selected for future spending, and the methods by which they are monitored and enforced. One of the challenges of global budgeting is that there can be trade-offs between some of those parameters. Government officials, for example, might find it difficult to enforce budget caps on patients’ out-of-pocket expenditures, but limiting global budgets to government programs might spur some patients and providers to move to or shift costs to the private sector, where spending was not constrained. Another challenge is selecting parameters for a global budget that would limit the growth in spending but not result in a misallocation of resources that could adversely affect patients’ health.
Examples and Potential Challenges
In the United States, budget caps have been implemented at the national level for several federal health care programs and at the local level by some state governments and communities. Those experiences may have limited application to broader efforts to implement a global budget, though; when budget caps are limited to a specific program or community, patients and their providers may turn for care to other sectors in which spending is not as constrained. The experiences of other countries—including Canada, the United Kingdom, and Germany—may provide more insight into the effects of implementing a global budget that encompasses most or all of the population.
The most comprehensive example of a global budget in the United States is found in the Medicare program. Medicare’s payment system for physicians—known as the sustainable growth rate mechanism—sets annual and cumulative spending targets for those payments.2 Doctors are paid a fixed fee for each service they provide, but if total spending exceeds the target amounts, an across-the-board reduction is supposed to be made in future fees to bring spending back into line (on both an annual and cumulative basis). The SGR targets were initially set in 1998 to reflect spending on physicians’ services at that time. For subsequent years, payments per enrollee are allowed to increase at about the same rate as growth in per capita gross domestic product and by an estimated change in fees for physicians’ services (after adjusting for certain factors, such as changes in laws and regulations, which affect health care spending but are outside the control of providers and patients). Since 2002, actual and projected payments for physicians’ services have consistently exceeded the target amounts, and as a result, substantial reductions in the nominal level of doctors’ fees have been projected in order to meet the spending targets. After the Medicare program reduced physicians’ fees by nearly 5 percent in 2002, legislation was enacted to limit or delay the impact of the scheduled reductions in each of the following years (which, in some cases, has meant that larger fee reductions will be needed in the future for spending to remain within the targets).
Another example of budget caps at the federal level is the health care system for military veterans. Because that system is funded through the annual appropriation process, the resources available in each fiscal year are capped. As a result, the Department of Veterans Affairs places veterans in different categories, weighing factors such as income and service-connected disability in determining which veterans will receive care. More than 10 million veterans who do not rank high enough on the VA’s list are not eligible for VA-funded care (see Chapter 6 for further details).
Some states and communities have also tried to establish budgets for certain subsets of health care expenditures. In the early 1990s, Oregon sought to expand its Medicaid program to a broader population—all residents below the federal poverty level—while limiting the number of services that were covered to constrain the state’s costs. Toward that end, the state developed a prioritized list of medical services and would cover only those services above a cutoff point; the list is updated every two years as part of Oregon’s biennial budget process to reflect both new information on medical care and changes in the state’s fiscal condition. Although the priority list has made decisionmaking about the allocation of public sources for health coverage more explicit, its impact on costs has been modest—partly because many of the more expensive types of treatments are included on the list. Nor has the state’s Medicaid program been able to avoid cutbacks in enrollment: In 2004, tight budget resources led the state to freeze enrollment in its Medicaid program for some low-income adults.3
The experience of Rochester, New York, during the 1980s provides some insight into the effectiveness of a global budget that was not limited to a government program. In that community, the local hospitals voluntarily agreed to an overall cap on revenues from all insurers (including Blue Cross and Blue Shield, Medicare, and Medicaid). A prospective payment program was established and monitored by a nonprofit agency, whose board contained representatives of the local hospitals. According to one analysis, hospital costs in Rochester increased more slowly than the national average, and the hospital sector’s share of health care costs fell from 55 percent to 38 percent over that decade. Rochester’s experiment ended in the late 1980s, when a federal waiver of certain Medicare regulations was terminated.4
Other countries have also established global budgets of various types. In the United Kingdom, a health care budget is set nationally and then allocated to regional organizations. As with an integrated health plan, doctors generally receive a salary or a fixed payment per patient rather than per service, and hospitals receive a budget allocation. In Germany, the payment system for physicians is similar to Medicare’s, with fee-for-service reimbursement of doctors and fee adjustments to meet spending targets. In Canada, expenditure caps have been applied in a number of provinces, both through hospital budgets and through global caps on physicians’ expenditures.
One concern about such approaches is that they may discourage the use of effective (as well as ineffective) health care. Comparisons of health indicators across countries, however, do not show any definitive differences that are attributable to the existence of global health budgets. For example, the United States compares unfavorably with some countries that use global budgets in measures such as life expectancy or infant mortality rates, but survival rates for certain types of cancer are higher in the United States than in those countries. The role that global budgets play in affecting those health outcomes is unclear, however. Given the many factors that affect health status, it would be difficult to isolate the impact of global budgeting.
The goal of a global budget is to limit total health care spending or certain types of health care spending to a predetermined amount. The experiences in the United States and other countries suggest that establishing and defining that amount would entail several steps:
■
■
■
Establishing mechanisms to monitor and enforce the spending targets.
Those steps could be specified in law; alternatively, some or all of them could be delegated to a national board or federal agency.
Scope of the Budget. A first step in establishing a global budget would be to determine its scope—that is, what services and payments it would encompass. A global budget could cover all health-related services, including those that are elective (such as cosmetic surgery) or related to health indirectly (long-term care, for example), or its scope could be limited to services typically covered by a health insurance policy. A global budget could also seek to limit only insured costs or could address out-of-pocket costs as well. Similarly, the budget could include payments for health care from all sources (including private insurers, Medicare, and Medicaid) or could cover a narrower set of payers. In general, a broader scope would make a global budget more challenging to implement and monitor, especially if it encompassed out-of-pocket spending by individuals. However, a broader scope might also reduce the "leakage" of payments from covered to exempted sectors—which could arise in an effort to circumvent the budget cap.
Targets for Spending. A second step would be to set an initial level for the global budget and a formula or process for updating it over time. One key issue is whether the initial spending target would reflect current spending levels or a lower level that was set as a policy goal. A closely related issue is how the budget target would be updated from year to year; to the extent that a growth rate was specified that was lower than projected growth in health care spending, the budgetary constraint would become more binding over time.
A similar issue is how to allocate the overall budget targets—by geographic areas (such as states), across provider types (such as hospitals and physicians), or both. In setting targets on the basis of geography, a key issue would be whether to accommodate current differences in spending among regions—which are substantial—or to seek reductions in those differentials. In setting targets by provider type, a key issue would be how to make trade-offs across those sectors or to accommodate underlying changes in medical practices over time.
In general, assigning budgets to specific hospitals or providers might increase accountability but would also raise concerns about potential misallocations of those limits. Allocating budgets on the basis of individual hospitals’ historical operating costs, for example, would in effect reward relatively inefficient hospitals and penalize relatively efficient ones. In addition, more uniform budgets could require that high-cost hospitals make difficult adjustments, particularly if those higher costs reflected having sicker patients or providing higher quality care. Proposals might also need to address the scope of services to which the budget applied, and whether and how any adjustments might be made for emergencies.
Monitoring and Enforcement Mechanisms. Choices about the scope of any global budget would not only help determine the challenges involved in monitoring and enforcing it but also shape what mechanisms might be needed to do so. In general, the less effective the monitoring and enforcement mechanisms, the more likely it is that actual spending would stay at the level projected in the absence of a global budget.
One challenge in implementing a global budget might be the lack of accurate and timely data with which to monitor spending. Although the national health expenditures include information on private health care spending, the data are estimated by CMS using surveys from public and private sources. No other agency currently monitors private spending. Even for a large public program like Medicare that receives data from actual claims, it takes several years to reconcile and settle payments. That long span highlights the difficulty of measuring health care spending in real time.
Another key question regarding implementation of a global budget is how payments would be adjusted if they exceeded the spending targets. Total payments could be adjusted prospectively, by capping the amount of premiums paid to insurers, for example, or by allocating fixed amounts to different types of providers (hospitals, for instance). Alternatively, retrospective adjustments might be needed to bring actual spending into line with the targets. Retrospective adjustment mechanisms could include penalties for insurers that have excess spending or reductions in fee-for-service payment rates for providers if total costs exceeded spending targets.
Each approach presents substantial challenges in implementation. If providers continued to receive fee-for-service reimbursement, global budgets would not create incentives for them to restrict spending; indeed, they could partly offset the effects of a reduction in fees by increasing the volume or intensity of the services they provide. Over time, the resulting increase in volume would cause further reductions in fees, which might ultimately cause some providers to cut back on services. A global budget would have more binding effects if it was combined with a system of capitation in which provider- or plan-level incentives and systemwide incentives were aligned to encourage the efficient provision of care. The challenge with a global budget based on capitation, however, is bringing the demand for services in line with the resources that are available.
Factors Affecting CBO’s Estimates
Key considerations in CBO’s analysis of any proposal for a global budget would be its scope, the severity of its limits, and the monitoring and enforcement mechanisms it includes. In particular, new monitoring systems would be needed in order to impose an effective budgetary constraint on spending that is not currently financed by the federal government. Without such a system, it is unlikely that an effective and binding budget could be devised to encompass all spending on health care—inevitably, some private spending would occur outside the scope of the cap, and it would be difficult to detect such spending. Proposals with weaker monitoring and lax enforcement mechanisms would be less likely to actually reduce spending below currently projected levels in the categories that are subject to the budget.
Impact of Proposals on the Composition of Compensation and Tax Revenues
Many proposals to modify the health insurance system could affect revenues by causing a shift between taxable and nontaxable forms of compensation. For example, capping the current exclusion for employment-based health insurance would increase revenues (including payments to the Social Security and Medicare trust funds) because premiums above the cap would be subject to individual income taxes and payroll taxes. A tax cap—that is, a limit on the amount of health insurance premiums that may be excluded from taxable income—would also raise revenues if, as a result, employers’ contributions for health insurance plans were lower than what they otherwise would be and workers received higher wages or other taxable benefits instead.5 Similarly, a tax credit for individually purchased health insurance could cause some employers to drop existing plans and instead boost other forms of taxable compensation.
Even proposals that would not amend the Internal Revenue Code would affect revenues if they caused a change in the allocation between taxable and nontaxable forms of compensation. A proposal requiring firms to provide health insurance coverage could increase employers’ contributions for health insurance, causing wages and thus revenues to fall. Conversely, a proposal to impose a global budget could constrain health care costs, lowering employers’ payments for health insurance and thus increasing wages and revenues.
Changing the tax treatment of health insurance would affect revenues not only by subjecting more wages to income and payroll taxes but also through interactions with a number of other tax provisions. For example, policies that caused changes in earned income would affect the amount that taxpayers could contribute to individual retirement accounts or claim as earned income tax credits and refundable child tax credits.
CBO will include in its analyses the effects of proposals on revenues. Estimates of proposals that change the Internal Revenue Code will be prepared by the staff of the Joint Committee on Taxation.
Flow of Payments and Budgetary Treatment
Major health insurance proposals could have a substantial influence on the flow of payments for health insurance and health care among government agencies, employers, individuals, insurers, and health care providers. If a proposal affected the flow of payments to or from the federal government, CBO would account for the timing of the outlays and receipts when estimating the net impact on the federal budget. In its analysis, CBO would also consider the effects on the budget of any federal mandates on individuals, other private entities, and state governments. If legislation imposed a federal mandate to purchase or pay for health insurance, some revenues and costs could be reflected in the federal budget—even if the money was not collected or disbursed by federal agencies.
One consideration in analyzing the effects of proposals to expand health insurance coverage would be the impact on the timing of outlays and receipts to the federal government. Another consideration would be who would bear financial responsibility for any shortfalls in funding for insurance and health care.
The various payment arrangements used in Medicare illustrate different strategies for allocating risk between health care providers and the federal government. Hospitals and doctors participating in the fee-for-service Medicare program submit claims to local intermediaries or carriers after the services have been provided. The intermediaries process the claims, verifying that they are for covered services and actual beneficiaries. Only after that adjudication has occurred does the federal government make payments. (Additional processes exist for appeals and fraud detection.) For private health plans providing Medicare’s basic and drug benefits (Medicare Advantage plans), however, the federal government makes a fixed payment per enrollee at the beginning of each month, and the plans then arrange payments to hospitals, doctors, pharmacies, and other providers. Although that approach accelerates the payments to private health plans, the plans bear the risk that the payments from the government might not be sufficient to cover their costs; conversely, the federal government bears the risk that the payments are too high. For the drug benefit, the federal government takes an additional step, reconciling federal payments with plans’ actual expenditures at the end of each year. That reconciliation could result in the plans owing the government money or vice versa, but it also causes administrative costs—such as collection costs—to increase.
Similar issues would arise if a proposal required the Internal Revenue Service to collect premiums for health insurance. One approach would build on the system used for the health coverage tax credit (described further in Chapter 2). Under that program, individuals can send the amount of the premium that they owe directly to the IRS, which bundles that payment with the appropriate federal subsidy payments and sends the total sum to insurers.6 Under an alternative approach, people would pay their portion of the premiums during the year through wage withholding and estimated quarterly payments. However, the IRS would not know until a taxpayer filed a return at the end of the year whether he or she had paid the full amount owed. In the interim, the Treasury Department would make monthly payments to insurers whether or not the payments from enrollees had been received. After the tax return was filed, the IRS would reconcile the amounts received from individuals with the payments owed and seek to collect or refund any discrepancies.
The advantage of the first approach to using the IRS is that it would limit federal exposure for unpaid premiums. The disadvantage is that it would require the IRS to actively collect and track taxpayers’ payments throughout the year. An advantage of the second approach is that the IRS would not have to create a new large-scale infrastructure to collect premiums; its disadvantages are that it would involve potentially large temporary transfers of general revenue and that the IRS would have to determine whether any collection activities related to unpaid premiums were worth the expense.
Budgetary Treatment of Federal Mandates
When proposals would affect the outlays made or receipts collected by federal agencies, CBO would account for those effects in estimating the proposals’ costs. When proposals would establish federal mandates that would not result in payments to or from the federal government, however, the issue of their budgetary impact would be less clear. In some cases, CBO would treat any resulting payments as part of the federal budget; in other cases, it would not. The extent of federal control and compulsion is a critical element in determining budgetary treatment. To assess whether transactions should be reflected in the federal budget, CBO would consider whether a proposal included these factors:
■
Payments by individuals or employers that are the result of a federal mandate;
■
Any required government payments (for example, subsidized premiums for low-income individuals);
■
■
Restrictions on the amount of discretion allowed to entities that collect or pay premiums.
In general, CBO believes that federally mandated collections—those resulting from the exercise of sovereign power—should be recorded in the budget as federal revenues, even if such amounts are not paid to a federal agency. Similarly, a mandated transfer of any such collections to others should be recorded as a budget outlay. An example of such transactions is the existing Universal Service Fund, which collects money from telecommunications carriers and spends it to subsidize telecommunication services to high-cost areas, to low-income consumers, and to schools, libraries, and rural health care providers.
In its 1994 analysis of the Clinton Administration’s health care proposal, CBO concluded that payments to and from the "health alliances" should be included in the accounts of the federal government but that they should be distinguished from other federal operations and shown separately—as is the practice for the Social Security program.7 (Although those alliances would have collected and paid health insurance premiums, they might not have been considered federal agencies because they would have been established by states and other entities.) At that time, CBO based its view primarily on the judgment that the proposal would establish a federal entitlement to health benefits and that the mandatory premiums used to finance the new entitlement would constitute an exercise of the federal government’s sovereign power.
Throughout this volume, CBO has used the term "mandate" when discussing certain types of proposals that would include new federal requirements for health insurance. The use of that term should not be interpreted to indicate a CBO conclusion that such requirements would meet the definition of "mandate" under the Unfunded Mandates Reform Act (UMRA). When analyzing specific legislative proposals, CBO assesses whether they would impose a mandate on state, local, or tribal governments or the private sector—an analysis that is required by UMRA. To the extent that a proposal would require state, local, or tribal governments or private entities to undertake some activity that they would not otherwise take on—or prohibit them from activities that they would otherwise pursue—it could constitute a mandate under UMRA. The discussions of proposals in this volume, however, do not consider the costs of potential mandates as defined in UMRA, nor do they attempt to quantify the impact of those proposals on states’ spending.8
Given that health care constitutes roughly one-sixth of the U.S. economy, any changes to the health care system could affect the operation of the broader economy. This section reviews possible effects of proposals to expand health insurance coverage on labor markets, the capital stock, and international competitiveness.
The overall economic effects of comprehensive changes to the health care system are difficult to predict. Although economic theory and experience provide some guidance about the effects of specific provisions, large-scale proposals to restructure the health insurance system may contain numerous pieces that could interact—affecting labor supply, the capital stock, and productivity in complex and possibly offsetting ways. Depending on the nature of those interactions, a comprehensive proposal might yield results that differ from those examined in this analysis.
Large-scale changes to the health insurance system could affect labor markets by changing people’s incentives to work and employers’ decisions to hire workers. The availability of health insurance options can affect people’s incentives to enter the labor force, work fewer or more hours, retire, change jobs, or even prefer certain types of firms or jobs. In addition, some proposals—such as employer mandates—could affect firms’ decisions to hire workers.
Changing Incentives to Work. Changes in health care policy that affected taxes or subsidies could have an impact on the economic gains from work. For example, proposals that would increase government spending on health care might be financed in part by additional taxes. Taxes levied as a percentage of labor income have two opposing effects on how much people choose to work. On the one hand, those taxes reduce the amount of after-tax wages earned for each additional hour worked, which tends to diminish the incentive to work. Economists refer to that as the "substitution effect" of tax rates on the number of hours worked. On the other hand, by decreasing the amount of after-tax income earned for any given amount of work, taxes tend to encourage people to work more to make up the difference. Economists refer to that as the "income effect." Most studies conclude that for simple changes in tax rates on labor income, the substitution effect typically outweighs the income effect—on average, due largely to the response of secondary earners. (Secondary earners are generally the spouse of the main earner in a household.) Therefore, increases in marginal tax rates generally reduce the number of hours worked.9
The precise impact that tax financing had on hours worked would depend on the details of the tax changes. Some tax provisions that would increase revenue would reduce after-tax income but have little or no corresponding effect on the return from an additional hour’s work. For example, a proposal to reduce the dollar amount of a flat tax credit (say, from $5,000 to $4,000) would affect after-tax income but have no impact on the after-tax hourly wage. Tax increases of that type would probably increase the number of hours worked.
Other types of proposals could have the opposite effect on people’s incentives to work. For example, some proposals would include subsidies to help low-income people pay for health insurance. A subsidy could be provided through the transfer system (possibly as a voucher) or through the tax system (as an exclusion from income, a tax deduction, or a tax credit). A subsidy represents an increase in income, which might discourage work effort, all other things being equal.
To limit costs, subsidies are often phased out as a beneficiary’s income rises. Over the phaseout range, a worker receives less compensation for each additional hour worked, because each dollar earned reduces the subsidy. That effect is sometimes referred to as an "implicit tax." That implicit tax can lead people to work fewer hours than they otherwise would, in the same way that income and payroll tax rates do.
Policymakers face a trade-off in deciding how to phase out subsidies. If subsidies are phased out quickly, the implicit tax rates, and thus the negative impact on work incentives, can be quite high. Implicit tax rates can be reduced by expanding the range over which the subsidy is phased out, but doing so increases the number of people subject to the implicit tax and, if the range is extended by raising the income level at which it is completely phased out, also boosts the total cost of the subsidy. In the extreme, a subsidy can be granted to everyone, which eliminates any effect on the economic gains from work but substantially increases costs. By contrast, a subsidy can be eliminated all at once at a certain income level (creating a "cliff" in the relationship between the subsidy and income), which eliminates the cost of phasing out the subsidy but significantly increases the disincentives to work for people whose potential income is in the neighborhood of the cliff.
Some aspects of the current health care system create work disincentives, so changes to that system could curtail or eliminate those effects. One program that creates work disincentives for its recipients is Medicaid. That program is structured so that eligibility for benefits is completely eliminated at specified income levels (a cliff).10 For individuals with income close to those thresholds, working more and earning a higher income can lead to the loss of all Medicaid benefits, creating a powerful disincentive to work. A system that made health insurance coverage independent of income would eliminate that disincentive for current Medicaid recipients. As an example of the potential effects on labor supply, one study found that a series of increases in the income limit for Medicaid eligibility in the late 1980s and 1990s increased the labor force participation of working-age single mothers by 1.4 percent.11
Some proposals would limit current tax subsidies for health insurance by reducing or eliminating the tax exclusion for employment-based health insurance. Eliminating that exclusion would make a larger share of compensation taxable. By itself, that change would reduce after-tax income, encouraging people to work more to make up for their lost earnings.12 Capping the exclusion would also affect the relative prices of goods: The effective price of health insurance would rise, making other goods appear less expensive. One such good would be "leisure"—which people "purchase" in forgone earnings by choosing to work less. Assuming there were no other changes, an increase in the price of health insurance would tend to boost the consumption of other goods—including leisure. As a result, labor supply would decline. CBO has estimated that the proposals in the President’s budgets for fiscal years 2008 and 2009 to make employment-based health insurance taxable and create capped tax deductions for individuals with health insurance would reduce the supply of labor.13
Reducing the Link Between Employment and Insurance. Proposals that would make insurance less dependent on employment status (for example, by substituting public programs or individually purchased insurance for employment-based health benefits) could induce more workers to retire earlier and could reduce the participation of younger workers in the labor force as well. (Proposals that strengthened the link between employment and health insurance—for example, by requiring coverage of employees—could have the opposite effects.)
Employment-based insurance offers a number of advantages (including lower administrative costs, favorable tax treatment, and coverage of existing conditions) that may be difficult or impossible for workers to obtain by purchasing insurance individually. For that reason, its availability can play an important role in people’s decisions to enter or remain in the workforce—especially if they are nearing retirement. People who are insured through their employer but are not offered health benefits after retirement have an additional incentive to remain employed until they qualify for Medicare at age 65. Proposals that include some kind of "bridge" coverage for early retirees would remove that incentive and increase the likelihood of retirement before age 65, thereby decreasing the supply of labor.
A review of the literature found that workers whose health insurance covers them in retirement are more likely to retire at any given age and tend to retire earlier, on average, than those without such benefits.14 Some studies found that the availability of health benefits during retirement increases the probability of retirement before age 65 by 30 percent to 80 percent.15 Studies using other estimating techniques generally found smaller effects, and a few found little or no effect. However, the weight of the evidence indicates that retirees’ health coverage probably leads to earlier retirement. Such findings suggest that proposals that would substitute other forms of coverage for employment-based insurance could cause some people to retire earlier than they would under the current system.
Increasing the availability of health insurance outside the workforce could also reduce the labor force participation of younger workers, although there is less evidence supporting that effect. The impact on participation would probably be highest among secondary earners because they tend to be more responsive to changes in compensation than are primary earners. Currently, if primary earners are not offered family coverage through their employer, other members of their household may enter the workforce in order to get the benefits of employment-based insurance. Some research indicates that spouses not covered under primary earners’ insurance are more likely to be employed than spouses who are covered through such a plan. Expanding access to health insurance could cause some of those secondary earners to stop working.
Changing the Degree of Job Lock. Some of the same advantages of employment-based health insurance that may keep more people in the labor force can also affect how often workers change jobs. People who have medical problems (or have family members with medical problems) can have an incentive to stay in a job that provides health insurance benefits in order to cover those preexisting conditions, even if more productive opportunities exist elsewhere. (Those opportunities could include working for a different employer or becoming an independent entrepreneur.) That phenomenon is sometimes referred to as "job lock."
The evidence is mixed regarding the effects of employment-based health insurance on job turnover. Although some empirical studies find that workers are less likely to change jobs when faced with the potential loss of health insurance, others find little or no effect.16 Much of that evidence is difficult to interpret, however, because many jobs that provide health insurance have other attributes that discourage turnover. Moreover, most studies to date rely on data collected before enactment of the Health Insurance Portability and Accountability Act in 1996. That law placed some constraints on the ability of employment-based plans to deny coverage for preexisting conditions, especially for workers who were previously covered under other plans, and therefore has most likely reduced the importance of job lock.
To the extent that employment-based insurance affects turnover, it can have both beneficial and adverse effects on the economy. Firms may have a greater incentive to invest in their workers (by providing training or increasing their skills or knowledge) if the probability of retaining those workers is increased. However, workers may also choose to stay in their current positions solely to retain their current health coverage rather than move to other jobs in which they could be more productive.
Mandating Insurance Coverage. Some proposals would require employers to offer health insurance to their employees. Because employees largely bear the cost of health benefits in the form of lower wages, the effects of those proposals on employment and hours worked could be relatively minor.17 However, an employer mandate could affect the amount of work available for certain categories of workers.
In particular, a plan that mandated that employers offer health insurance could reduce the hiring of low-wage workers. In a competitive market, the demand for and the supply of labor determine the total compensation—including both wages and benefits—that employees earn. If employers were required to provide health insurance benefits, their employees’ wages and other forms of compensation would be lower than what they otherwise would be by the amount of the cost of the insurance. However, wages and other forms of compensation for employees earning amounts at or near the minimum wage might not be able to fall by the full cost of the health insurance coverage required by law. That constraint could lead employers to hire fewer of those workers, increasing unemployment, although that effect is likely to be small. One study estimates that 224,000 workers (or about 0.2 percent of all private-sector workers) could become unemployed if firms were required to provide health insurance costing $2 per hour worked, on average.18 In contrast, a study of the state-level employer mandate in Hawaii found that the rate of employment grew faster in Hawaii than in the rest of the United States after the mandate was instituted (perhaps because of factors other than the mandate).19
Affecting Workers’ Choices of Firms. Some proposals to provide subsidies for health insurance or require employers to offer health insurance coverage would differentiate between firms on the basis of criteria such as the number of workers, revenues, or salary levels. That differentiation could create incentives for certain types of workers to work for certain types of firms.
The responses to those types of incentives could take several forms—some involving actions of workers, some involving actions of firms, and some involving actions of both parties. For example, new workers in the labor force could choose jobs with certain firms rather than others. Or those already in the workforce could quit their jobs and move to different firms. Firms could "outsource"—that is, lay off employees and contract with other companies for the same services. Alternatively, firms could divide themselves into subsidiaries with low and high average wages.
For example, the Clinton Administration’s health proposal included subsidies to firms with low average wages to help offset the costs of mandated health benefits. Those subsidies would have created incentives for low-wage workers to work for firms with low wages, on average, and high-wage workers to work for firms with higher average wages.20 In its analysis of the 1993 plan, CBO found that the resulting shift of workers among firms would increase the net budgetary cost of the proposal (because the shifts would reduce tax payments or increase subsidies) and also could decrease the efficiency of the allocation of workers to firms.21
Effects on the Nation’s Capital Stock
Health care proposals would affect the nation’s stock of productive capital through their impact on saving, because the amount of national saving determines the resources available for domestically owned investment. (National saving is the sum of private saving by households and businesses and government saving—federal, state, and local budget surpluses.) If a health care proposal increased government budget deficits, or reduced surpluses, it could decrease the resources available for investment and, therefore, the capital stock.
The effects of health care proposals on private saving are more ambiguous and partly depend on the specific provisions of any proposal. Proposals that expanded access to health insurance and medical care could reduce private saving, whereas those that reduced consumption of health care could increase private saving.
People who are not currently covered by insurance have an incentive to save more to guard against unforeseen medical expenses. Covering more people could reduce that motive for saving, which would lower private saving and result in a smaller capital stock. However, because many people who currently lack coverage are in low-income households, which tend to save very little regardless of whether they have insurance, the effect of expanding coverage on private saving may not be large. Furthermore, increasing the number of people with insurance would lessen their risk of incurring uninsured medical expenses, even if it reduced their saving.
Proposals that include an "asset test" for receiving subsidies or other benefits could also decrease saving. If receipt of benefits is dependent on a person’s or household’s assets being below a certain level, people may refrain from saving in order to stay below the limit. Such proposals could have an opposite effect on Medicaid recipients, however. Because the current Medicaid system includes asset tests, proposals that eliminated those tests could increase saving, particularly among households that have assets slightly below the current thresholds.22
Proposals might affect consumption, and therefore the capital stock, through their impact on income and on consumption of health care. Subsidies or other provisions that increase the after-tax income of recipients would probably increase consumption, all other things being equal, which would reduce the resources available for investment. More generally, proposals that resulted in an increase in consumption of health services would tend to reduce saving and investment unless other types of consumption fell by corresponding amounts. Conversely, proposals that reduced consumption of health services would probably boost saving and investment.
To the extent that changes in the health insurance system led to improved health status among workers, the nation’s economic productivity could be enhanced. Some research has also suggested that improved health status might increase the growth rate of productivity.
Studies have found that healthier workers work more hours and earn higher wages than those who are less healthy.23 That relationship suggests that changes to the health insurance market that lead to better health outcomes could both increase the labor supply and raise productivity (presumably, workers earn higher wages when they are healthy because they are more productive).
Other studies compare the economic output of different countries and how that output is related to various measures of health status within each country.24 Those studies generally find that countries with better health outcomes grow faster than other countries. The results must be interpreted with caution, however. Because higher income may lead to better health, the direction of causation is not clear. Greater growth may lead to healthier citizens, rather than vice versa. Moreover, those studies are based on the very wide range of health outcomes—for example, life expectancy—observed when developing countries are compared with industrialized nations. The implications for marginal advances in health outcomes in already fairly healthy countries are, therefore, unclear.
Because the impact on health outcomes from major changes to the health care system is uncertain, it is not clear whether such changes would have a substantial impact on overall economic output or productivity.
Some observers have asserted that domestic producers that provide health insurance to their workers face higher costs for compensation than competitors based in countries where insurance is not employment based and that fundamental changes to the health insurance system could reduce or eliminate that disadvantage. However, such a cost reduction is unlikely to occur, except in the short run.
The equilibrium level of overall compensation in the economy is determined by the supply of and the demand for labor. Fringe benefits (such as health insurance) are just part of that compensation. Consequently, the costs of fringe benefits are borne by workers largely in the form of lower cash wages than they would receive if no such benefits were provided by their employer.
Replacing employment-based health care with a government-run system could reduce employers’ payments for their workers’ insurance, but the amount that they would have to pay in overall compensation would remain essentially unchanged. Even though changes to the health care system could have various effects on the supply of labor, the underlying amount of labor supplied at any given level of compensation would hardly be affected by a change in the health care system. As a result, cash wages and other forms of compensation would have to rise by roughly the amount of the reduction in health benefits for firms to be able to attract the same number and types of workers.
Compensation could take some time to adjust to its market-clearing level (the point at which supply and demand are equal). During that time, firms that formerly provided health benefits—especially firms that employ workers under multiyear contracts—could experience substantial reductions in labor costs, which would boost their profits temporarily.25 But those firms would experience no permanent change in their competitive status.
Background information on the derivation of the national health expenditures can be found at www.cms.hhs.gov/NationalHealthExpendData/downloads/dsm-06.pdf and www.cms.hhs. gov/NationalHealthExpendData/downloads/projections-methodology.pdf.
The Balanced Budget Act of 1997 established the sustainable growth rate system, replacing the volume performance standard that had linked payments to the overall growth in the number and mix of services since 1992. Before 1992, Medicare paid physicians on the basis of the lowest of three charges: the physician’s actual charge, the customary charge (the amount the physician usually charged for the service), or the prevailing charge (the amount that similar doctors charged).
Oregon Health Services Commission, Office for Oregon Health Policy and Research, Prioritization of Health Services: A Report to the Governor and the 74th Oregon Legislative Assembly (Salem, Ore., 2007).
For additional description and details, see Institute of Medicine, Changing the Health Care System: Models from Here and Abroad (Washington, D.C.: National Academy Press, 1994), www.nap.edu/catalog.php?record_id=9218#toc; and William J. Hall and Paul F. Griner, "Cost-Effective Health Care: The Rochester Experience," Health Affairs, vol. 12, no. 1 (1993), pp. 58–69.
Some nontaxable benefits—such as employers’ contributions for social insurance (payroll) taxes—are linked by law to wages and thus automatically change as wages rise or fall.
Individuals also have the option of paying the full amount of the premium directly to the insurer during the year. If they choose that option, they claim the credit on their tax return, thus lowering the amount they owe or increasing their tax refund. Most HCTC claimants, however, opt to pay their share of the premiums to the IRS and thus receive the benefits of the subsidy sooner.
For a discussion of this issue, see Congressional Budget Office, An Analysis of the Administration’s Health Proposal (February 1994), pp. 41–50.
For more information on UMRA, see Congressional Budget Office, Identifying Intergovernmental Mandates, Issue Brief (January 2005).
See Congressional Budget Office, Labor Supply and Taxes (January 1996).
Although eligibility for Medicaid varies by state, all states are required to cover pregnant women and children under age 6 whose family income is at or below 133 percent of the federal poverty level, as well as children who are at least 6 and under 19 with family income of up to 100 percent of the federal poverty level.
For estimates of the size of the effect on labor supply, see Aaron S. Yelowitz, "The Medicaid Notch, Labor Supply, and Welfare Participation: Evidence from Eligibility Expansions," Quarterly Journal of Economics, vol. 110, no. 4 (November 1995), pp. 909–939.
Proposals might include provisions that offset the impact on after-tax income; for example, the measure included in the President’s budgetary proposals for fiscal year 2008 included increases in other deductions to roughly eliminate, on average, the effect on after-tax income. See Congressional Budget Office, An Analysis of the President’s Budgetary Proposals for Fiscal Year 2008 (March 2007).
See Jonathan Gruber and Brigitte C. Madrian, Health Insurance, Labor Supply, and Job Mobility: A Critical Review of the Literature, Working Paper No. 8817 (Cambridge, Mass.: National Bureau of Economic Research, March 2002).
Those studies sought to examine the correlation between health benefits and the probability of retirement, controlling for other factors.
For a discussion of those studies, see Brigitte C. Madrian, The U.S. Health Care System and Labor Markets, Working Paper No. 11980 (Cambridge, Mass.: National Bureau of Economic Research, January 2006), p. 19.
See Lawrence H. Summers, "Some Simple Economics of Mandated Benefits," American Economic Review, vol. 79, no. 2 (May 1989), pp. 177–183.
Katherine Baicker and Helen Levy, Employer Health Insurance Mandates and the Risk of Unemployment, Working Paper No. 13528 (Cambridge, Mass.: National Bureau of Economic Research, October 2007).
Norman K. Thurston, "Labor Market Effects of Hawaii’s Mandatory Employer-Provided Health Insurance," Industrial and Labor Relations Review, vol. 51, no. 1 (October 1997), pp. 117–135.
That plan included subsidies to firms with low average wages to help them pay for the mandated health benefits (because for firms paying low wages, the health benefits represent a higher fraction of the wage bill). That means low-wage workers would effectively have their wages subsidized if they worked at a low-wage firm but not if they worked at a high-wage firm. Smaller firms also would have received higher subsidies under the plan. See Congressional Budget Office, An Analysis of the Administration’s Health Proposal.
Ibid. For example, CBO estimated that the shifting of employees would have raised the cost of the Clinton Administration’s 1993 health care proposal by $12 billion per year once the shifting was complete. In addition, differential treatment based on firms’ size could also lead to an inefficient allocation of capital and could change the structure and distribution of firms. For example, subsidies to small firms could encourage large firms to break up into small ones, even if that form of organization was not the most efficient.
Jonathan Gruber and Aaron Yelowitz, "Public Health Insurance and Private Savings," Journal of Political Economy, vol. 107, no. 6 (1999), pp. 1249–1274; and Alex Maynard and Jiaping Qiu, "Public Insurance and Private Savings: Who Is Affected and by How Much," Internet draft, October 31, 2005, www.carleton.ca/economics/seminar%20papers/Alex%20Maynard-April21% 202006.pdf.
See, for example, Robert Haveman and others, "Market Work, Wages, and Men’s Health," Journal of Health Economics, vol. 13, no. 2 (1994), pp. 163–182.
For a review of such studies, see David E. Bloom, David Canning, and Jaypee Sevilla, "The Effect of Health on Economic Growth: A Production Function Approach," World Development, vol. 32, no. 1 (2004), pp. 1–13.
For purposes of estimating the impact of proposed legislation, CBO makes the simplifying assumption that total compensation is fixed and that changes in health insurance costs translate immediately into offsetting changes in wages and other forms of compensation.
|
|