Economic and Social Implications of Uninsurance Within Communities
For individuals and for families, being uninsured can be associated with a loss of income. A family’s breadwinner may lose a job that included health insurance coverage, or an individual’s inability to work because of poor health may be prolonged or made worse by a lack of coverage for treatment that would improve the individual’s health (IOM, 2002a, 2002b). For the communities in which uninsured individuals and families live, the lack or loss of health insurance coverage by some may undermine the shared economic and social foundations of the entire community. For example, uninsurance may result in higher prices charged to everyone for health services and health insurance premiums or higher state and local taxes levied in order to support local hospitals that provide substantial amounts of uncompensated care. Health insurance is not the solution to all communal ills; nevertheless, the Committee hypothesizes that its presence or absence can make a substantial difference for a community’s economic fortunes. While public health insurance programs such as Medicaid also make claims on state tax revenues, the incidence of the tax burden for federal or shared federal–state financing programs falls more broadly upon all American taxpayers and can alleviate some of the financial demands that uninsurance places on disproportionately affected communities.
The impact of uninsurance on the social and economic “glue” that binds a community together is difficult to assess because a community’s social and economic resources are as likely to affect uninsured rates as the converse. In addition, the hypothesized relationships between the local uninsured rate and community resources may feed back on themselves in a cycle that amplifies their adverse impacts. For example, declining economic stability and prosperity can foster a sense of social isolation and loss of confidence in public institutions and services,
making those who are relatively better off, namely workers with health insurance coverage, feel vulnerable and less prepared to share scarce resources with those who lack a job or health insurance by, for example, extending public benefits to them (Ehrenreich, 1989; Rubin, 1994). Communities with relatively high uninsured rates may experience economic difficulties when uninsurance places new financial demands on local health services providers who make significant contributions to the local economy. As a result, the communities most in need of taxsubsidized assistance to health services institutions and providers may be the least likely to have the tax revenues and be able to afford to finance care for uninsured and underserved persons (Lewin and Altman, 2000).
In this chapter the Committee discusses several probable economic consequences of uninsurance, first characterizing these varied potential community effects as hypotheses and then distinguishing between those for which adequate evidence exists and those requiring further documentation and research. The chapter is organized into five parts. The first three sections consider potential effects of uninsurance on the local economy, through (1) the relationship between uncompensated care locally and increases in health care costs and spending, (2) the increased pressures on states and localities to support greater public spending on health care in response to high or rising uninsured rates, and (3) the loss of fiscal stability by health care institutions, due to high uncompensated care burdens, that may lead to the loss of this major component of the local economy. The fourth section explores new ways to think about potential effects of uninsurance on a community’s social bonds and its social institutions in the context of the recent research literature on income inequality, social capital, and health. The fifth section draws together a discussion of key research questions that, once addressed, will fill in the details about economic consequences of uninsurance for communities.
THE CONTEXT: INCREASING HEALTH CARE COSTS AND TAXPAYER SUPPORT
Over the past two decades the uninsured rate nationally has grown slowly but steadily, despite an increasingly tight labor market that expanded employment-based coverage and yielded tax revenues to expand public coverage. Present economic trends have changed from boom to economic recession and, absent major policy interventions, indicate a continuing rise in the uninsured rate. Projected health care cost and health insurance premium increases are likely to add to the ranks of the uninsured (Chernew et al., 2002; Cutler, 2002). It is probable that rising numbers of the uninsured, in turn, will result in further increases in the costs of care and coverage for those with health insurance. In the face of renewed health care inflation, employers have not yet dramatically dropped coverage for their employees but have increased cost-sharing and pared benefits (Strunk et al., 2002). The Committee’s observations in Coverage Matters remain valid:
The rising uninsured rate since the late 1970s is directly related to increasingly unaffordable health insurance premiums, as health care inflation has advanced more rapidly than increases in real income or purchasing power (Cooper and Schone, 1997; Holahan and Kim, 2000). After a period of stability in the mid-to-late 1990s, health care costs are again increasing at a rapid rate (Heffler et al., 2001).
The decade of the 1990s saw a slowing of this inflationary trend, with economic prosperity and relatively low rates of increase in health care costs accompanied by increases in employment-based coverage. Declines in public coverage (reflecting declining Medicaid enrollment particularly after welfare reform), continued until public coverage increased in the late 1990s with implementation of the State Children’s Health Insurance Program (SCHIP) (Holahan and Pohl, 2002).
The first recession of the new century has already erased the gains in private coverage made in recent years. Current estimates released by the Census Bureau indicate that between 2000 and 2001, an additional 1.4 million persons became uninsured, raising the national uninsured rate for persons under age 65 from 15.8 percent to 16.5 percent, or a total of 41 million uninsured persons (Mills, 2002).
The Committee hypothesizes that those with health insurance pay more for it than they otherwise would if fewer people lacked coverage. Public support (e.g., disproportionate share payments to hospitals, subsidy of public hospitals and clinics, federal grants to health centers) and private subsidy (e.g., charity care at private nonprofit hospitals, cross-subsidy provided by employment-based coverage insurance payments) covers an estimated 85 percent of the value of the health care delivered to uninsured persons that is not paid for by the uninsured themselves (Hadley and Holahan, 2003). These costs are passed down to all taxpayers and consumers of health care in the form of higher taxes and higher prices for services and insurance. The Committee’s fifth report will take up this subject in greater depth, showing the economic effects of these costs and expenditures on society.
Locally, residents are likely to subsidize care for their uninsured neighbors (though in an arguably inefficient and invisible manner) through taxes and higher prices for health services and health insurance in their community. Because directly provided health services are largely a function of state and local governments, the tax burdens of funding that direct care for uninsured residents is more concentrated locally than is the burden of Medicaid financing or other public programs in which the federal government participates.
Current financing structures for health services are also regressive. One study of health care costs in South Carolina in 1996 finds that persons in households that earned less than 100 percent of the federal poverty level (FPL) had per capita health care costs that were, on average, 70 percent lower than those of persons in households earning more than 400 percent of FPL, yet they paid four times as high a proportion of their income for health care services in taxes and direct spending
(Conover, 1998).1In counties with hospital taxes or other subsidy-generating revenues tied to health care use, the uninsured likely bear fewer of these costs than do low-income people in general because they use fewer services.
BUDGET IMPLICATIONS FOR STATES AND LOCALITIES
Finding: Public subsidy of uncompensated care delivered to uninsured persons requires that additional public revenues be raised, resulting in a higher tax burden at the local level, the receipt of monies from federal coffers, or the diversion of resources from other public purposes. If additional revenues are not generated, budget cuts may be imposed, either for health care or across the board, that can adversely affect all members of the community and even increase the number of uninsured persons locally.
Finding: State and local governments’ capacity to finance health care for uninsured persons tends to be weakest at times when the demand for such care is likely to be highest ( i.e., during periods of economic recession).
Public support for health care services for uninsured persons places demands on state and local governmental budgets. Although the discussion in the previous section assumes that the increasing cost of health care and coverage will be met by increased spending, this is not always the case. States and localities may respond to greater health-related pressures on public budgets in a variety of ways. They may change the state’s Medicaid eligibility, benefits, or reimbursement policies; raise taxes (within the limits on their capacity to do so); or use one-time sources of revenue such as tobacco settlement dollars to meet demands for serving uninsured residents.
States’ abilities to levy taxes and their tax structures affect state responses to financing care for the uninsured. Box 4.1 provides a broad overview of the diversity among the federal and state tax systems and the proportion and nature of the expenditures by the federal and state governments, respectively. The preeminence of the federal government in raising public revenues suggests that any public finance solution to uninsurance will need to rely in large part on federal financing.
Community residents subsidize care for uninsured patients through their tax dollars. Like the private sector, public funders of health care currently have to cope with increased costs for health services in a time of economic recession. In comparisons among states, relatively higher levels of state spending on Medicaid
A brief overview of the structure of the U.S. federal system will give an appreciation of the resources available for health care at the local level. The federal government is the overwhelmingly dominant fiscal presence; it collects two-thirds of the governmental revenue in this country (Miller, 2002b). Even with recently enacted tax cuts, the federal personal income tax will yield as much revenue as all state and local taxes combined.
The tax structures of the different levels of government vary widely. Eighty-five percent of federal revenue in fiscal year 2001 came from direct taxes on individuals—the personal income tax or Social Security taxes. By comparison, state and local governments collect less than half of their revenues by directly taxing individuals. More than half of their revenues comes from indirect business taxes; some combination of sales and property taxes; and other licenses, taxes, and fees (Miller, 2002b). There are simply no generalizations to describe the different state systems: some states have no income taxes, some have no sales tax, others have neither.
The federal government does not use most of its revenue to participate in the economy as a purchaser of goods and services. Its budget is primarily a transfer mechanism—direct transfers to individuals such as Social Security and Medicare (49 percent), indirect transfers through state governments for programs such as Medicaid and child nutrition (11 percent), other grants to state and local governments (6 percent), and net interest (9 percent) (Miller, 2002b). Slightly more than one-fourth of the federal budget can be considered direct spending, divided between national defense (17 percent) and domestic expenditures (9 percent) (Miller, 2002b). By comparison, most state and local spending is either direct spending or transfers from states to local jurisdictions.
The levels of government also have very different debt constraints. State and local governments primarily borrow to finance capital construction. While there is a great deal of misunderstanding surrounding states’ balanced budget requirements, these generally can be described as forbidding borrowing for operating expenditures. The federal government, in general, does not describe the purpose of most borrowing but rather borrows to finance an overall unified federal budget deficit.
These differing governmental finance systems produce very different impacts and requirements. The federal and various state tax systems respond very differently to the changes in the business cycle. States that depend on income taxes usually tie either their definition of income or their tax liability to federal definitions and consequently are substantially exposed to changes in federal law. States with sales taxes are highly sensitive to changes in consumption patterns but also benefit when reductions in interest rates produce strong sales of durables. They have been constrained by federal prohibitions on taxing Internet and mail-order sales. Jurisdictions (mostly local) that depend on property tax revenue generally experience a lag in receipts, reflecting the delay between increases in property values and increases in assessments.
are not necessarily associated with lower uninsured rates. (This depends on the number and proportion of low-income and medically needy persons in the state, and on the breadth and generosity of the state’s Medicaid program [Holahan, 2002]). However, for a given state, lower rates of public coverage are likely to result in fewer insured persons and a higher uninsured rate. Decreasing enrollment in public insurance (due to tightened or otherwise constrained eligibility) in an environment where employment-based coverage rates are flat or declining will result in less use of health services overall (because uninsured persons use fewer services, on average). However, when uninsured persons do use services, they are likely to need more costly health services because of delays in seeking care (IOM, 2002a, 2002b). The lesser use of services by uninsured persons and their relatively high out-of-pocket spending for health care suggest that aggregate public spending would increase if they gained health insurance coverage.
Given the range of strategies employed to finance uncompensated care and safety-net arrangements from community to community there is no generalized, simple relationship between a community’s uninsured rate and its tax burden. An increase in uninsurance would be expected to be met by pressures to increase taxes and reallocate public funds (to the extent that legal structures of taxation and spending allow) devoted to activities other than the provision of personal health services. Thus, a relatively greater uninsured rate may result in higher state and local tax burdens. Because 57 percent of national Medicaid expenditures are paid for by the federal government (and 70 percent of SCHIP spending nationally has been paid for by the federal allocation),2 when states provide coverage and health care through federally matched insurance programs, the demands on local funds for uncompensated care are likely to be decreased.
Over the previous decade, the fiscal capacity and resources of the states for spending on health programs have grown, due to both the general economic boom in the mid-to-late 1990s and revenues from tobacco lawsuit settlements with the states. Most recently, however, particularly since 2001, states are increasingly experiencing hard times, with economic recession, federal cuts to Medicaid (limiting DSH payments) and Medicare, and public resistance to raising taxes (Dixon and Cox, 2002; Lutzky et al., 2002). States are in a bind because of budget shortfalls, and many plan to cut Medicaid spending in the coming years (KCMU, 2002; Smith et al., 2002). The consequences of these responses are likely to result simultaneously in lower public funding for health insurance, fewer public funds available for other purposes, and the need for higher taxes.
State and local programs and policies affecting services for the uninsured are mediated by state Medicaid program policies. Medicaid represents 20 percent, on average, of states’ budgets, and the financial incentives of the federal match as well as federal program requirements draw state funds away from more discretionary spending on the uninsured and into the Medicaid program. Total Medicaid program expenditures have increased throughout the 1990s, accounting for almost 15 percent of national health spending in 1998. At the same time, the state and local share of total health spending remained constant over the decade and was about 6 percentage points lower than the share of national health care spending that state and local governments accounted for in 1970 and 1980 (Miller, 2001). This means that there was relatively less of state and local budgets available for all non-Medicaid health spending than in earlier decades.
Changes in a state’s spending on Medicaid are likely to affect the level of community uninsurance and the demand for uncompensated care. For example, in Massachusetts, the decision by state officials to trim Medicaid costs by eliminating an expansion that had covered 50,000 adults is predicted to add to the unreimbursed expenses accumulated by hospitals (Goldstein, 2002; Powell, 2002). These hospitals will continue to provide care to at least some of the now newly uninsured state residents and will not be able to draw on the state’s uncompensated care pool, which is depleted of funds.
Medicaid program spending often takes priority over state health programs of direct services, including services for uninsured residents, in state budget allocation decisions because of its structure as an individual entitlement and the federal financial match for program expenditures by the state. Medicaid’s federal match is designed to provide relatively greater federal support to those states with lower average personal incomes. However, it only imperfectly reflects relative state needs. States still have to cope with reduced tax revenues and higher uninsured rates at the same point in the economic cycle.
Much of the federal government’s economic presence is designed to be counter-cyclical, providing automatic stabilizers both to the economy and to the institutions and individuals that participate in it. Medicaid is frequently discussed in these terms because some persons automatically become eligible when their incomes are lost or diminished during economic downturns. Thus, program enrollments and costs grow at times when state revenues may be reduced. Further, the determination of the federal match rate for the program drives more federal money to states whose incomes fall relative to the national economy.
The lags built into the federal medical assistance percentage (FMAP) calculation, however, often produce a pro-cyclical effect for many states.3 The FMAP is
based on a three-year moving average of per capita income and is published a year before it comes into effect. Thus, the FMAPs for federal fiscal year 2003 are based on per capita personal income for 1998–2000. States with longer-term secular shifts in wealth see this appropriately reflected in their FMAPs, but those with shorter cyclical shifts often see perverse results. State budget crises in recent years have led 18 states to consider cutting Medicaid eligibility, which would be likely to increase uninsurance in each state and further destabilize local health care providers and institutions (Smith et al., 2002).
Limits on State and Local Capacity to Raise Revenues and the Use of One-Time Revenues
States and localities have a number of constraints on their ability to raise new revenue sources to subsidize care for uninsured persons. Almost all states are required to balance their budgets annually, and many states have constitutional or statutory limits on tax rates or the rate of growth in tax revenues (Desonia, 2002). These governments are effectively prohibited from borrowing to finance operating expenditures and must finance these expenditures out of current tax and nontax (e.g., fees) revenue. States with low per capita income or depressed economies, characteristics that are positively related to uninsured rates, have even more constraints on financing health care for uninsured residents than do more prosperous states.
The entitlement nature of most state government support for health financing means that these programs tend to absorb whatever revenues may be available (Hovey, 1991). Once decisions for health entitlement programs have been made, substantial pressure is placed on the rest of state and local budgets. Medicaid cost pressures may or may not be met by commensurate parallel increases in spending. Governments will constrain Medicaid increases and increase revenues within their limited capacities to do so. For example, Alabama, with an uninsured rate of 13 percent, relies on a regressive tax structure (property and sales taxes) that has limited ability to raise revenues (Ormond and Wigton, 2002). Because its revenues are limited, Alabama’s Medicaid program relies on federal supplemental payments (through disproportionate share hospital [DSH] and upper payment limit [UPL] payments) to a relatively greater extent than do other states. Alabama also uses tobacco settlement monies for a Medicaid Trust Fund.
Tobacco settlement dollars have been an important new revenue stream that has cushioned the effect that increased public spending related to uninsurance would otherwise have on states. However, this means that these revenue streams
are not available for other public purposes. Since the November 1998 Master Settlement Agreement between 46 states and the tobacco industry,4 tobacco settlement payments (returning prior years’ Medicaid expenditures for smokers) have provided states with almost $30 billion (2000–2002) (Dixon and Cox, 2002). These massive paybacks have allowed states to expand public financing of and eligibility for public insurance programs or for public hospital operating subsidies, among many health-related and unrelated uses. Public health and education efforts aimed directly at reducing smoking have been crowded out by the competing demands of programs that finance personal health care services, including Medicaid, SCHIP, and state-only insurance programs for medically indigent persons and families (Dixon and Cox, 2002). Over the four state fiscal years 2000 through 2003, 29 percent of tobacco settlement revenues to the states have been allocated to health services programs (exclusive of long-term care, which accounts for another 5 percent), while tobacco use prevention programs accounted for just 5 percent of the total (Dixon and Cox, 2002). States including Arizona, New Jersey, and New York have devoted significant proportions of their tobacco settlement revenues to health care—either to their Medicaid programs, to support for safety net providers, or to expand public coverage (Bovjberg and Ullman, 2002; Dixon and Cox, 2002). For example, New York has financed expanded coverage through its broad-based Health Care Reform Act of 2000 with a combination of tobacco settlement dollars and an increase in the cigarette tax; the programs intended to provide coverage are estimated to cost about $500 million over three and a half years (Coughlin and Lutzky, 2002).
Local governments (municipalities, counties, and special-purpose health or hospital districts) bear a large share of the direct financing of public hospital and clinic services. One study of local public financing for urban safety net hospitals (in Houston, Oakland, and Miami) finds that the dependence on local revenues made for unstable funding (Meyer et al., 1999). Furthermore, local revenue sources such as sales and property taxes fluctuate with the communities’ changing economic fortunes (Clemmitt, 2000). The following vignettes exemplify both the diversity in local financing arrangements and their often provisional nature.
An innovative program serving medically indigent and uninsured persons in Hillsborough County in south Florida generated a $6.6 million deficit (out of a $90 million budget) due to increased demands for care (Nguyen, 2002). In response, the County Commission raised taxes and reallocated dollars, in addition to tightening eligibility and cutting the generosity of benefits, to try to meet the needs of the approximately 19,000 persons currently enrolled. Rejecting a proposal to levy a one-cent gasoline tax, the commissioners decided to use $3.1 million in county reserve revenues, $2.5 million in property taxes that had been
earmarked for road construction, and some funds from those used to pay for prisoners’ health care to cover the program’s deficit.
In Texas, the need for indigent health care has outstripped the resources available (Pinkerton, 2002). During 2001, two of the larger counties that border Mexico exhausted the funding for health services programs mandated by the state. After spending local tax revenues and $4.6 million in state aid, Hidalgo County shuttered its program, resulting in the increased use of local emergency rooms for care. Cameron County trimmed services and instituted cost controls to save its $2.3 million program. Federal payments for uncompensated hospital care did not cover the full tab in local hospitals; the nearly 500-bed Providence Hospital in El Paso received $3 million in compensation for the estimated $10 million to $11 million it provided in free care in 2001.
In the spring of 2001, Cincinnati’s safety net arrangements were reported to be under increasing strain due to a growing number of uninsured persons (Cincinnati Business Courier, 2001). The city of Cincinnati had budgeted $13.4 million for health clinics, but one of the safety net providers in the city, the Cincinnati Health Network, citing empty financial reserves, was going to the city for another $300,000 to keep from having to turn away new uninsured patients.
During 2001, the Kansas City Council debated whether to fund a $205.3 million public safety capital improvement plan by reducing or eliminating the general fund’s support for indigent care at the Truman Medical Center and local community health centers (CHCs) (Jaffe, 2001).
Improvised Strategies to Address a Nationwide Problem
Federal health care financing policies provide only a part of the framework that states and localities must construct in order to meet their residents’ needs for health services and insurance coverage. The programmatic structure of the federal Medicaid program places considerable fiscal demands on the states and at the same time offers states limited support to address the health care needs of uninsured residents. States and local governments, with highly disparate fiscal resources, health insurance coverage rates, and populations, have adopted a variety of revenue-raising and programmatic strategies to piece together programs of health care coverage and services. These public programs, and the community health care providers that depend on the financial support that they supply to serve low-income and uninsured residents, are far less stable and secure than such essential community services and resources should be, given our nation’s wealth.
ECONOMIC BASE AND THE POTENTIAL FOR DEVELOPMENT
Health care accounts for a significant proportion of the economy nationally. In 1999, national health expenditures accounted for 13 percent of the gross domestic product (GDP), rising from less than 9 percent of the GDP in 1980 to
slightly more than 13 percent in 1993, at which point the share stabilized (Cowan et al., 2001). Despite this prominent place in the national economy, much less is understood about how health care contributes to the economic vitality of particular communities and, more specifically, how uninsurance within a given community affects the level of health care expenditures and the local economy overall (Doeksen et al., 1997). The question of aggregate, national economic effects of uninsurance in the United States will be addressed in the Committee’s next report on the economic costs related to uninsured populations. The discussion here sets out the Committee’s hypotheses about the pathways by which uninsurance could affect local economies and communities’ economic viability, particularly in rural areas.
How Uninsurance May Affect Communities’ Economies
A community’s uninsured rate reflects the availability of employment-based insurance coverage for individuals and their families and of public insurance programs such as Medicaid and SCHIP, the viability of the market for individual nongroup insurance policies, and the costs of health care and insurance premiums in the local market. Changes in any of these factors may affect a community’s uninsured rate, and the concentration or clustering of uninsured persons in a community may in turn affect its social and economic foundations. An increasing or high uninsured rate, and the attendant high public costs, may discourage employers from locating or continuing to make their home in a community.
A high uninsured rate may both indicate and contribute to an area’s economic challenges as it reflects the relative availability of employment-based coverage. For example, health insurance is likely to be less available in areas with a lower-waged labor force (IOM, 2001a). Although there is little direct evidence of the effects that a lack of health insurance may have on workplace productivity, there is some evidence of the lower health status of uninsured workers compared to their insured counterparts (Fronstin and Holtmann, 2000; Hadley, 2002). Furthermore, there is a growing body of research that finds the uneven availability of workplace health benefits keeps workers “locked in” to their jobs if they already have coverage through their employer (Gruber and Madrian, 2001). While such job lock may actually benefit the employer who offers health insurance in terms of worker retention, it may also inhibit the worker’s propensity to move among jobs and employers, such that the local or regional economy is less dynamic and expansive than it might otherwise be.
The financial stress of relatively high local uninsured rates on health services arrangements is discussed at length in Chapter 3. While that chapter considered some of the consequences for access to and quality of care in the community, it did not address the aftermath of these effects for the economic base of the community. The health sector is a key part of most local economies, particularly in rural areas, where hospitals serve as social, historical, and civic, as well as economic, anchors. The following section focuses on the centrality of health care institutions
and services in rural communities and economies to highlight the potential impact of high uninsured rates on communities. Rural communities are featured both because of the simpler model they offer for describing interrelationships among institutions, health care coverage and financing policies, and uninsurance and because the research literature is more extensive in this area. Despite the difficulty of tracing the broader economic impacts of uninsurance in metropolitan areas, the Committee finds it plausible that somewhat similar forces could be at work in urban communities.
Health Care and Uninsurance in Rural Economies
The economic effects of uninsurance on the viability of health care providers and institutions are easier to measure in rural than in urban areas because the scale and complexity of rural economies is less than those of urban economies. Rural residents on average are older and have lower incomes than do urban residents; consequently public financing, particularly through Medicare and Medicaid payments, is particularly important for the rural health services infrastructure (Cordes, 1998; Mueller et al., 1999). Medicare funding subsidizes access to care not only for Medicare beneficiaries but also for all community members (see Chapter 2 for more discussion). Recognition of Medicare’s special role in financing rural health services is expressed with enhanced payments to designated sole community hospitals, through the DSH program and the Critical Access Hospital program (Poley and Ricketts, 2001). These dollars, spent locally, can be a source of new jobs, population growth (e.g., by making the community more attractive to retirees), tax revenues, and economic development (Cordes, 1998). For this reason, hospitals can represent a form of economic development that brings outside dollars into the community (Cordes et al, 1999). The public monies (through Medicare and Medicaid payments) spent on health services also attract additional private dollars spent by patients at local businesses (e.g., shopping, restaurants) in connection with medical or hospital visits. Local rural economies lose dollars spent on health care if residents travel out of the area to obtain services. The loss is amplified to the extent that traveling patients also spend their money on other goods and services outside their home communities, especially in more urban areas (Cordes, 1998; Doeksen et al., 1998; Scorsone et al., 2001).
One way to measure the importance of the health sector for rural economies is by means of an input–output model, with both overall employment and per capita income in a rural county dependent on the presence and size (or loss) of a hospital (Scorsone et al., 2001). Such models were developed in the 1980s to better understand the effects of changes in health care financing (specifically, the new Medicare prospective payment system for inpatient hospital services) on rural hospitals and economies during a period when many rural hospitals closed (Cordes et al., 1999; Poley and Ricketts, 2001). Researchers simulated changes in jobs, income, and tax revenues for a given county predicted by increases or decreases in payments to hospitals. Both spending by hospitals and spending by hospital em-
ployees as residents were taken into account in these simulation models. The estimated magnitude of the “multiplier effect” of the gain or loss of one hospital-based job varies by the size of the hospital.
Not surprisingly, according to these simulations, small, rural hospitals have smaller absolute effects (measured in terms of annual business sales, the number of full-time and part-time jobs, and hospital employee earnings) than do larger hospitals in urban areas. One study of rural hospitals finds a range of effects on the number of local jobs, with the smallest hospitals contributing 77 jobs and the largest more than 1,300 jobs (Cordes et al., 1999). Hospitals constitute one of the most influential economic sectors in rural areas, as trade and services in the other sectors are often influenced by the fate of a local hospital (Cordes et al., 1999). Rural hospitals are also more likely to close altogether than are larger hospitals in urban areas, so the relative impact on the local economy can be very great (Colgan, 2002). In rural counties, a hospital may be the second largest employer after the schools, with health-related jobs comprising between 10 and 20 percent of the local jobs (Doeksen, et al., 1997).
A higher-than-average uninsured rate and the corresponding burden of uncompensated care on the local hospital may threaten the survival of the institution and reduce the viability of the economic base of the community. The closure of a hospital may not be the only option for the community or hospital, however (Hartley and Lapping, 2000; Lapping and Hartley, 2000). For example, in Appalachia County, Kentucky, conversion of the local hospital into a clinic brought new funds into the community, as well as jobs and a new public insurance-based program intended to provide coverage to uninsured persons (Ziller and Leighton, 2000). The county’s high uninsured rate was one of many interrelated factors that made continued operation of the hospital impractical. These factors included the modest tax base and high poverty level of the area and the tendency of insured residents to travel to a neighboring urban hospital with more up-to-date medical technology.
Rural hospital closings are associated with the loss of community physicians (Hartley et al., 1994). Many physicians rely on the hospital to structure an economically viable practice. Because physician practices also bring new jobs into a community, if a community loses them, the local economy may suffer as well, albeit to a more limited extent than with a hospital closing (Doeksen et al., 1997). According to the director of rural health programs for the Tennessee Hospital Association, a new doctor in a community may be expected to create five new jobs and $500,000 in income, which then will produce further economic activity locally (Raiford, 2002).
SOCIAL CONSEQUENCES OF UNINSURANCE: FIRST THOUGHTS
Health care providers in a community function as social institutions as well as sources of health services. What is known from existing literature about the social
effects of uninsurance within a community? In previous sections of this chapter, the Committee has described how uninsurance can tap a community’s store of economic resources or capital, resulting in financial strain on the health care system or, to reduce this strain, the raising of new tax or other revenues. Uninsurance may also strain social relationships among community members and local institutions. In the discussion that follows, the Committee identifies social effects as a new area of inquiry that may yield important findings about community uninsurance in the future. Two promising approaches to thinking about social effects, in terms of the intersection of uninsurance with measures of social capital and of income inequality, are briefly reviewed.
Health insurance, whether offered through an employer or publicly to groups left out of the employment-based system, is part of an implicit social contract in the United States (IOM, 2001a). The lack of insurance coverage represents a breach of that contract that, when experienced by large numbers of individuals in a community, may erode the social bonds that define and nurture functioning, healthy communities, as uninsured persons are made aware of their lesser claim on services and resources that are generally valued as essential to a dignified and secure life (Walzer, 1983; Miller, 1997; Faden and Powers, 1999). At the same time, less social cohesiveness and mutual trust within a community may be associated with greater uninsurance, for example, when employers or political jurisdictions do not offer health insurance to their workers and constituents.
There has been little empirical research about how community uninsurance might influence social relationships and what the variety of potential community effects might be that are associated in this way with uninsurance. Social connectedness or cohesion, for example, may promote access to health care by strengthening local health services arrangements, similar to the positive influence that such cohesion has been demonstrated to have on local government functioning (Hendryx et al., 2002; Shi et al., 2002). A relatively strong collective ethic may spur local leaders to raise funds for uncompensated hospital care, for example, or motivate the establishment of a free clinic to serve uninsured people. A community’s values, history, and culture contribute to its success in responding to broader economic trends and market changes (Steinberg and Baxter, 1990). Steinberg and Baxter describe these responses in terms of what they call “mechanisms of accountability” that help build or foster the capability of communities to address social needs and respond to changing economic and fiscal environments. To the extent that community uninsurance diminishes these feelings of connectedness, access to health care for all community residents and the health of a community’s population in the aggregate may be affected.
One way to observe and measure these social relationships is through the construct of social capital or cohesion. Social capital refers to the stocks of resources available through social relationships, as measured by indicators such as civic engagement, norms of reciprocity, and interpersonal trust between members of a community. Definitions of social capital are usually multidimensional and
have both cognitive and social structural elements (Macinko and Starfield, 2001).5 The more social capital that a community has, the greater is its ability to undertake collective action and to “realize common values and maintain the social controls that foster health” (Sampson and Morenoff, 2000). The process of mobilizing the community resources described as social capital can be thought of as a community’s “collective efficacy,” specifically, the level of mutual trust among community residents and the degree of shared confidence that residents can and will exert local social control for the public good (Sampson and Morenoff, 2000).
There are several plausible connections between uninsurance and social capital within a community:
Low social capital may lead to greater uninsurance, for example, if a lack of social cohesion leads to decisions not to offer public coverage. Researchers have demonstrated an association between indicators of social capital at the state level and mortality rates (Kawachi et al., 1997) and poor self-rated health (Kawachi et al., 1998).
Uninsurance may deplete a community’s stock of social capital. This may be manifested in several ways. Lochner and colleagues (1999) propose four overlapping aspects of social capital: collective efficacy, a psychological sense of community, neighborhood cohesion, and community competence. For example, in a community with many uninsured persons or with a high uninsured rate, residents may experience a weakening of their beliefs or confidence in their ability to effectively take care of themselves or their community (collective efficacy), because of perceived or genuine difficulties in obtaining coverage or in gaining access to health services that health insurance usually facilitates.
Social capital and uninsurance may function as markers or indicators of other social processes within the community, for example, social stratification and economic or income inequality. Some researchers have found an association between high degrees of income inequality and low social capital, measured by indicators of civic engagement and social trust. For example, areas with a high level of income inequality have a difficult time maintaining community bonds (Kawachi et al., 1997). Conversely, the erosion of social cohesiveness tends to lead to policies that widen the economic gap within communities (Kaplan et al., 1996).
Another way to think about how uninsurance affects social relationships is to consider the phenomenon of income inequality within a community. For individuals and families, health insurance not only facilitates access to health services but also constitutes a component of socioeconomic status.6 The lack of financial access to health care and relatively high out-of-pocket expenses (among the uninsured) may exacerbate existing inequalities in income and wealth among a community’s residents. Kawachi and colleagues (1999) have proposed that greater income inequality in a community or society, through the mechanism of diminished social cohesion, leads to higher mortality (see also Kennedy et al., 1996; Kawachi and Kennedy, 1997a, 1999; Kawachi, 1999). Others have pointed out that income inequality also correlates with other community characteristics such as racial composition or individual educational attainment that may equally explain area variations in mortality rates (Mellor and Millyo, 2001, 2002; Deaton and Lubotsky, 2002).
The relative income inequality hypothesis posits that the range and steepness of the socioeconomic gradient among residents in a community can affect the health of all members of the community, not only those at risk for poor health due to poverty or other barriers to receipt of appropriate health care (Wilkinson, 1997). Three pathways have been proposed as ways that income inequality may adversely affect the health status of all community members, through
the disparity or relative deprivation created between the interests of lower-income and higher-income persons (expressed in lower tax rates and levels of public spending on education and health services) (Kaplan et al., 1996);
the erosion or diminution of social cohesion (e.g., trust in other community members and in the health care system); and
stress invoked by social comparisons among community members (Kawachi and Kennedy, 1999).
Limited and mixed evidence points to a small adverse effect of community-level inequality on the health of all community members, independent of each individuals’ social characteristics, across a variety of measures of income distribution (Kawachi and Kennedy, 1997b; Lynch et al., 1999; Kahn et al., 2000; Sampson and Morenoff, 2000). The size of relative inequalities in average household incomes at the state and county levels contributes to differences in age-adjusted mortality rates and chronic disease morbidity for a number of conditions. The size
of the effect seen at the state and county levels tends to be small (Sampson and Morenoff, 2000).
The cross-sectional, observational nature of much of the evidence means that little can be said with certainty about the direction of causality between income inequality and health-related outcomes and the size of any effect (Macinko and Starfield, 2001). Social inequalities, health disparities, and uninsurance tend to cluster geographically, potentially amplifying the effects on each. The observed relationships between income inequality and health may be related to unmeasured influences on the patterns of residential location (Sampson and Morenoff, 2000; Shi, 2000; IOM, 2001a; Kawachi and Blakely, 2001). Measures of income inequality may also function as a proxy for another, unmeasured aspect of societal or economic inequality that adversely affects health in the aggregate, for example, the effect of economic growth or decline (Mellor and Milyo, 2001; Deaton, 2002). The Committee draws no conclusions about the role of disparities in health insurance status at the community level on population health but, rather, points to emerging developments in the social science literature that may someday allow for a more robust exploration of potential relationships between community uninsurance and the social dimensions of community life.
4.1 Effects of Local Uninsured Rates on Health Costs
Does the local uninsured rate, independent of other factors, affect the cost of health services and insurance premiums within the local market area?
Cross-sectional studies are the basis for our limited knowledge about how local uninsured rates contribute to the increasing cost of health services in local health services markets. Longitudinal small-area studies are needed to look at the hypothesized chains of events at the state and local levels to establish or disprove a causal relationship between uninsured rates and health insurance costs locally and to develop estimates of the size and strength of any such relationships.
4.2 Uninsurance and Tax Burdens
What is the burden on local taxpayers of uninsurance within states and localities? How does that burden compare to the tax burden that would be imposed if public insurance programs or subsidized coverage were expanded to reduce uninsured rates? How does the impact of public financing of health care vary across economic cycles?
As discussed in Chapter 2, public support for care for uninsured persons is substantial, yet documentation is inconsistent and exact estimates are difficult, if not impossible, to derive. More consistent reporting is needed across the states and localities of revenue and expenditure streams for financing the array of insurance-based and direct health services programs. A better understanding of state-level
budget allocation decisions and the fiscal incentives operating in Medicaid and other programs that support care for uninsured persons could inform proposals to improve the target efficiency and equity of federal and state health financing programs. Specifically, programs of institutional support for uncompensated care, such as Medicare and Medicaid DSH payments, should be evaluated in terms of efficiency and equity.
A systematic analysis of the incidence of the federal, state, and local tax burdens for financing health care generally and for low-income and uninsured Americans in particular should be undertaken as part of any impact analysis of health care financing reform initiatives.
4.3 Uninsurance Effects on the Local Economy
Does the adverse financial impact of uninsurance on local health services and institutions extend to the local economy overall?
Finally, if little is known about the role that the health sector plays in rural economies, even less is known about the economic relationships that bind local health services to urban neighborhoods and metropolitan areas. Research is needed to identify the ways in which a changing uninsured population affects the financial viability of health care providers and institutions and to explicate the complicated series of financing relationships among public and private payers. Are there instances in which a high uninsured rate triggered a funding or political crisis that led to the transition of a local health care system toward a more efficient and effective use of limited resources? To what extent are an uninsured population’s uncompensated health care expenses matched by public funds with the objective of keeping a local economy afloat?
Although potential economic and social spillover effects of community uninsurance are not as well documented or described as the effects on access to health care, the Committee finds the preliminary evidence on economic consequences, and the promise of new research on social consequences, a compelling start. In Chapter 5, the Committee considers community effects related to population health.