Why Are block grants better for citizens

In 1996, Congress and President Bill Clinton agreed on fundamental changes to Aid to Families with Dependent Children (AFDC), the safety-net program of cash assistance to poor families, transforming what had been a federal-state funding partnership into the Temporary Assistance for Needy Families (TANF) block grant to states. Turning this entitlement safety-net program into a block grant was not a new idea in 1996. Beginning in the 1940s, numerous administrations, members of Congress, commissions, and scholars recommended block grants of federal funds to lower levels of government. In 1980, Robert B. Carleson, President Ronald Reagan’s senior welfare advisor in the White House, stated his intention to achieve Reagan’s goal of returning responsibility for financing welfare to states by eliminating the guarantee of federal matching funds to assist poor parents and replacing it with a system of state block grants.

Block Grant History

In the late 1970s, the Advisory Commission on Intergovernmental Relations (ACIR) developed a list of characteristics of block grants, saying, “a block grant may be defined as a program by which funds are provided chiefly to general purpose governmental units in accordance with a statutory formula for use in a broad functional area, largely at the recipient’s discretion.” In contrast, categorical funding from the federal government to states and localities is generally limited to narrowly defined purposes and targeted populations, and typically comes with reporting obligations designed to ensure accountability to the federal agency charged with oversight of the program. Proposals to block grant programs that previously guaranteed benefits to individuals who met a set of defined qualifications (called entitlements) present special issues because shifting from the guarantee to fixed funding ends the individual entitlement, as occurred with the welfare legislation in 1996.

Federal categorical funding for a variety of targeted needy groups and communities grew in the 1960s. By the late 1960s, many scholars called for consolidation and devolution of the funding and program oversight to lower levels of government as a means to reduce the complexity and size of the federal government. Some state and local officials began to support a block grant mechanism because they believed it would reduce administrative and reporting burdens.

Related Books

In 2003, President Bush proposed converting Medicaid from an entitlement to a block grant program. Similar ideas from President Reagan in 1981 and Congress in 1995 were introduced but not enacted. Block grants aim to provide greater federal budget certainty and a stronger state incentive to contain program costs. This paper compares the preestablished funding levels proposed in 1981 and 1995 with what actually happened to federal Medicaid spending. Its results show that previous block grant proposals’ funding levels at the national and state levels were quite different from what was anticipated and what occurred. As a result, Medicaid probably could not—and cannot—maintain existing health coverage under a block grant financing structure.

Keywords: Medicaid, health care financing, entitlements, health reform

In 2003, in the shadow of the intense focus on creating a Medicare prescription drug benefit, President George W. Bush proposed an arguably more profound change in health policy: the conversion of Medicaid from an entitlement to a block grant program. Block grants are programs for which the federal government gives state or local governments a fixed amount of funds for administering and providing certain services. Compared with “categorical” grants, block grants are generally larger and less restrictive regarding how the grantees use them. Both block and categorical grants are much different from entitlement (or mandatory) programs. These programs create a government obligation to finance a benefit or service for a prescribed set of people, with no aggregate limit on funding. Examples of such programs are Social Security, Medicare, food stamps, unemployment insurance, and, by extension, tax benefits like the deduction for mortgage interest.

As one of these entitlement programs, Medicaid guarantees that certain low-income and disabled persons receive a set of comprehensive health benefits defined by federal and state law. The funding for such services rises and falls according to need. Under a block grant plan, however, the situation would be reversed. Federal (and typically state) spending would be set in advance and capped, so that eligibility and benefits would rise and fall based on the predetermined amount (Finegold, Wherry, and Schardin 2004; Kaiser Family Foundation 2004a). This idea of a Medicaid block grant is hardly, having first been proposed by President Ronald Reagan in 1981. It accommodates a conservative approach to social programs by limiting federal cost liability and encouraging state autonomy and innovation in program design. But a Medicaid block grant has yet to be enacted, in part because of an equally strong political commitment to the guarantee of Medicaid coverage of vulnerable populations. The Medicaid block grant also foundered over concerns about its impact. Simulations of how beneficiaries, health care providers, and some states would fare under the proposals proved to be effective ammunition in past battles (Families USA 2003; Holahan and Liska 1995).

Beyond simulations, the idea of converting Medicaid to a block grant has not been well studied. Indeed, it is difficult to assess the implications of a policy that does not now exist and for which there are few analogues. However, precisely because they predetermine federal funding, we can retrospectively examine the proposed amounts for Medicaid block grants. Accordingly, this article compares what actually happened in Medicaid to what would have happened had the funding limits of past proposals been enacted. The results aim to inform future inevitable debates over the block grant financing of Medicaid.

Advocates of block grants make two claims. The first is the idea of federalism, or giving control of the grants to the states. Since its beginning, the United States has struggled to balance national and state responsibilities. From President Franklin D. Roosevelt's New Deal in the 1930s through President Lyndon B. Johnson's Great Society in the 1960s, the proponents of a common, national interest prevailed. Having made a shared sacrifice, the World War II generation supported programs like Medicare that provide uniform eligibility and benefits throughout the United States. Then Ronald Reagan changed this trend, bringing back the idea of the states’ taking responsibility for welfare programs (Brown 1982). His rationale was that the states are more likely to be efficient and innovative because they are both “closer to the people” and held more accountable by them, according to a leading proponent, former U.S. Secretary of Health and Human Services Tommy Thompson (Thompson 1996, 2003). Therefore, if they were relieved of the federal program requirements associated with federal funding, the states might be able to produce better outcomes at a lower cost.

A second argument for block grants is that they would eliminate the “uncontrollable” aspects of entitlement programs (Etheredge 1983). Congress sets in advance the maximum amount of federal block grant spending, which offers both predictability and a relatively easy way to adjust the program's spending to meet broader budget goals. Whereas federal policymakers could reduce the costs of entitlement programs through specific eligibility or benefits changes, it is probably politically easier for them to cut spending on block grants, thereby delegating these difficult decisions to states. Predetermined funding also limits the extent to which the states’ actions influence federal outlays.

In terms of spending and the number of people served, Medicaid is the nation's largest means-tested program and its largest single health coverage program. Today, Medicaid serves around 50 million people, including low-income children, parents, people with disabilities, and seniors. It pays for nearly one-fifth of the nation's spending on drugs, in part because it fills in Medicare's benefit gaps for low-income seniors and people with disabilities, and it helps with the costs of about 60 percent of all nursing home residents (Kaiser Family Foundation 2004b). Certain populations are eligible for coverage because they have a great need for health care. For example, Medicaid insures more than half of all persons living with AIDS, including up to 90 percent of children with the disease (CMS 2004).

Medicaid is structured as a partnership between the federal and the state governments. The federal government requires those states participating in the program to cover certain categories of people and benefits. The states largely control the decisions about how Medicaid is administered and how to cover additional populations and services. About 70 percent of all enrollees are in mandatory groups, such as poor children and a subset of their parents, persons with disability, and seniors. Even so, 65 percent of total Medicaid benefits spending is associated with the states’ optional coverage, for two reasons (Kaiser Family Foundation 2001). First, more than half of all Medicaid enrollees in mandatory groups are children, for whom the average cost is low. Second, two of the most expensive services and populations—prescription drugs and nursing home residents—are optional. Regardless of whether they are mandatory or optional, Medicaid costs are split between the federal and state governments based on a formula known as the Federal Medical Assistance Percentage (FMAP). The federal government pays no less than half and no more than 83 percent of each dollar of Medicaid costs. States with lower per capita incomes receive higher FMAPs (Schneider et al. 2002).

Medicaid has been the target of block grant proponents in part because of the inherent tension in the federal-state partnership. Although the states already have great control over how their Medicaid programs are designed, some state officials still claim that this control is insufficient (Bush 2003). For example, the states are prohibited by federal law from charging beneficiaries more than nominal copayments for services, capping enrollment, or providing different benefit packages to different subsets of enrollees. Conversely, some argue that states have too much control, since their nearly unilateral actions can significantly increase federal spending, sometimes with the sole purpose of securing federal funding in order to reduce their own spending on health and other services (Frogue 2004; Greve 2003). These disparate views find common ground in the idea of a Medicaid block grant that would trade greater state flexibility for federal budget certainty.

Perhaps the most important factor that has rekindled the Medicaid block grant debate is cost. Not surprisingly, in the years that Medicaid block grants were proposed (1981, 1995, and 2002), the annual growth of Medicaid costs was 20 percent, 9 percent, and 12 percent, respectively. In fiscal year (FY) 2003, total Medicaid costs were $289 billion, higher even than the $246 billion for Medicare (U.S. OMB 2004), even though in most states, Medicaid's payment rates for health care providers are lower than those for Medicare. Moreover, this gap is projected to widen, and the states’ share of this total, about $124 billion, has recently grown faster than the states’ revenue (Kaiser Family Foundation 2004b).

If costs are the reason for the proposed Medicaid block grant, why isn't a Medicare block grant being debated, since it has similarly large and rapidly growing costs? The technical answer is that there is no state partner to which program responsibility can be delegated, which is a key component of block grants. But the more probable answer is politics. Capping federal spending on Medicare would affect a stronger and broader constituency. In explaining why Medicare was not included in Reagan's attempts to limit the federal government, an official described Medicare as one of the “sacred” programs (Etheredge 1983).

President Reagan proposed changing Medicaid into a block grant program in 1981, Speaker of the House of Representatives Newt Gingrich proposed it in 1995, and President Bush proposed it in 2003. President George H.W. Bush also considered the idea but withdrew it before it was drawn up as a specific proposal (Ifill 1991).

President Reagan's proposal (S. 1377, 97th Congress) would have placed an upper limit on the federal contribution to Medicaid expenditures beginning in 1982. This limit would have been set on a state-by-state basis, with the federal cap being the sum of all the states’ caps. For 1982, the state-specific federal spending cap would have been set at the federal share of a state's FY 1981 spending plus 9 percent. In subsequent years, the cap would have been the previous year's cap increased by the Gross National Product implicit price deflator, which is a measure of the economy's growth. The U.S. Senate passed this proposal but dropped it in the final budget conference, which instead adopted a policy for growth rate reduction that maintained Medicaid's entitlement. In 1982, Reagan replaced the block grant idea with a “swap” in which the federal government would assume all Medicaid costs in return for the states’ assuming all welfare costs. Congress did not take up this idea.

In 1995, the 104th Congress, under the leadership of Representative Newt Gingrich, proposed a different type of block grant, called Medigrant (H.R. 2425, 104th Congress). The proposal was introduced in the House Commerce Committee and underwent subsequent changes when considered by the full House and Senate. It would have based the federal funding caps on a complicated formula intended to measure each state's need. Specifically, each state's annual cap would have begun with its historical spending and would have, through growth rate adjustments, determined the amount based on the national spending for poor residents, adjusted for the state's input costs (e.g., local wages), case mix (e.g., the level of sickness of enrollees), and number of poor people. Despite efforts to link the caps to the concept of need, the actual amount of each state's allotment would have been reduced on a prorated basis to ensure that the total of all states’ allotments would meet an annual aggregate cap on federal Medicaid spending, which was specified in the law. The policy would also have changed the states’ contribution to Medicaid, thereby lowering it for many. This policy, with modifications, was passed in both the Republican-dominated House and the Senate as part of a larger effort to balance the budget. But it was vetoed by President Bill Clinton in 1995 and put to rest during the 1996 presidential election.

Unveiled in January 2003, President Bush's plan proposed giving the states the option of accepting federal block grant funding in return for higher federal funding in the near term, greater program flexibility, and reduced state spending. Described vaguely in the president's FY 2004 budget, the plan would have placed an aggregate cap on the federal Medicaid and State Children's Health Insurance Program (SCHIP) spending linked to budget targets over the next ten years (U.S. OMB 2003). Relative to the administration's January 2003 projections of Medicaid spending growth in the absence of a change, the proposed federal “allotments” or block grants would have been higher in FY 2004 through 2010 but lower in FY 2011 through 2013, making the proposal “budget neutral” to the federal government over the ten-year period.

To receive federal funding under this option, the states would have had to spend a predetermined amount on health services each year, called a “maintenance of effort.” This maintenance of effort would have been equal to the states’ spending in FY 2002 increased annually by the inflation in medical costs. Estimates suggest that over ten years, the spending from the states using this option would be hundreds of billions of dollars less than it would be under current law (Kaiser Family Foundation 2003a). The states taking this option would also have been subject to fewer federal guidelines on the use of federal funds for optional populations and services.

Rather than working with Congress, the Bush administration turned to the National Governors Association (NGA) to develop the details of the proposal. The NGA created a bipartisan task force with the specific charge of producing a plan that it could support. But in June 2003, the task force disbanded, having failed to achieve a bipartisan agreement on the principles and parameters of such a plan (Connolly 2003). Congress, preoccupied with trying to pass a Medicare drug benefit, did not take up the idea. The president's budget, released in 2004, stated that “the Administration remains committed to enacting legislation, which will reform Medicaid and SCHIP” (U.S. OMB 2004). However, neither the 2004 budget tables nor the Bush campaign documents refer to the idea.

The reasons for the failure in the past to enact block grant proposals may go beyond Medicaid. For example, Reagan's plan was introduced in the same year in which there was a Social Security financing crisis and an assassination attempt on the president. Newt Gingrich's plan was part of a larger balanced budget plan that also deeply cut Medicare, rolled back environmental policies, and led to a government shutdown and showdown with Clinton. In each case, there was also strong ideological opposition. As President Clinton said when vetoing the Gingrich proposal, “I will not permit the repeal of guaranteed medical coverage for senior citizens, for disabled people, for poor children and pregnant women. That would violate our values, it is not necessary, and therefore, if it continues to be a part of the budget, if necessary I would veto it again” (Clinton 1995). Proponents of the current funding structure argue that it is needed to fulfill the government's commitment to protect vulnerable populations.

The estimated impact of the Medicaid block grant plans played a major role in defeating such initiatives in the past. In 1995, a heated debate centered on three major issues: the size of the federal spending reductions (e.g., was it a “cut” or just a “slowdown in the rate of growth”?); the state-specific implications of the funding formula (e.g., were efficient states penalized; did rapidly growing states get less?); and the impact of those reductions on coverage (e.g., would poor children lose coverage; would nursing home residents lose Medicaid assistance?). Opponents cited analyses based on simulations and projections to make their claims, among which was that under the proposal, 8.8 million beneficiaries, including 4 million children, would lose Medicaid coverage in 2002 (Holahan and Liska 1995). Similar analyses and concerns were raised in 2003 (Families USA 2003; Holahan and Weil 2003; Mann, Nathanson, and Park 2003). These analyses, however, were criticized because of the inherent difficulty in projecting Medicaid spending, let alone the changes that the proposal would make (Cantwell 1995).

Given the time that has passed since the Reagan and Gingrich plans were introduced, it now is possible to conduct similar analyses using actual, past data. We now know both the maximum federal funding that would have occurred under the 1981 and 1995 proposals and the actual Medicaid spending for the years in which the block grants would have been in effect, had they passed. This article examines (1) how federal funding under the plans compares with the actual amounts; (2) how state-specific funding caps in one year compare with the actual amounts; and (3) the potential impact on coverage and benefits. It also explores past Medicaid projections and spending growth, to determine whether it is possible to design a federal spending cap that avoids excessive funding cuts. While this approach still has the weaknesses inherent in this type of analysis, its use of past rather than projected spending as the point of comparison eliminates a major source of uncertainty.

This article used data from several sources. Past Medicaid and Medicare spending data came from the U.S. Office of Management and Budget (OMB) and the U.S. Congressional Budget Office (CBO). The breakouts for the disproportionate share hospital (DSH) payments came from the Centers for Medicare & Medicaid Services (CMS) (HCFA 2000). Provider payments made through the upper payment limit (UPL) came from the CBO (U.S. CBO 2002, 2003). The DSH and UPL payments are, simply stated, extra payments made through Medicaid, primarily to public providers that are not always used for Medicaid beneficiaries and services (Schneider and Rousseau 2002; U.S. GAO 1994). Medicaid baseline spending is shown with and without the DSH and UPL payments, since these funds probably would have not been spent had Medicaid been funded through block grants.

Details of the proposals come from the legislation: S. 1377 in the 97th Congress and H.R. 2425 in the 104th Congress. For the 1981 proposal, the cost growth index comes from the U.S. Department of Commerce, as specified in the proposal (BEA 2004). For the 1995 plan, the aggregate levels of annual federal Medicaid spending were specified in the legislative language (which covered 1997 through 2002). The U.S. General Accounting Office estimated the state-specific caps, which this article uses (U.S. GAO 1995). Note that the article focuses on the effect of changing federal, rather than state, spending. Since most block grant proposals would allow the states to reduce their contribution, the total reduction in spending could be larger than described here.

In addition to comparing actual spending with what the legislation proposed, this article addresses the argument that a “more perfect” block grant could be designed by examining the predictability of Medicaid spending and growth. It does so first by examining experts’ ability to predict Medicaid costs by comparing actual spending with the projections made five and three years earlier (e.g., actual spending in 2000 compared with the 2000 projection from both 1995 and 1997). It uses the projections from the CBO's Economic and Budget Outlooks from 1985 to 1999 (U.S. CBO). The analysis uses the CBO's rather than the administration's projections mainly because the CBO is the final arbiter of the cost implications of legislative proposals, although supporting analysis used the administration's data, producing similar results (available upon request). In addition, the analysis used data from the CMS's national health accounts and the Bureau of Labor Statistics for a multiple linear regression analysis of how much of the past growth of Medicaid costs can be explained by the medical component of the consumer price index, actual enrollment growth, and a proxy for utilization changes. This analysis used the national health accounts data from 1968 to 2001 and focused on the R-squared statistic, a measure of the goodness of fit.

Tables 1 and ​2 show the difference between actual federal Medicaid spending and what it would have been had the 1981 and 1995 caps been implemented. The analysis of the 1981 block grant shows that the proposed Medicaid spending levels would have reduced federal Medicaid expenditures by 6 percent in the first five years and 26 percent over the ten-year budget window (Table 1). For several reasons, the rate of Medicaid growth during this period was considerably higher than that of the gross national product (GNP), which was used to adjust the federal funding cap. First, because of federal and state policy changes (Schneider et al. 2002), the number of Medicaid beneficiaries increased by an average annual rate of 1.6 percent from 1980 to 1990 (CMS 2002), more than 75 percent higher than the growth of the U.S. population (0.9 percent) (U.S. Census Bureau 2003). The growth in enrollment alone would have consumed more than 40 percent of the average annual increase in federal funding under the 1981 caps. Second, health care prices generally grew faster than the GNP from 1982 to 1991; the medical component of the consumer price index rose by an average of 4.8 percent compared with the GNP average increase of 3.7 percent. Third, Medicaid spending began to accelerate in 1989, in part because of a financing scheme involving DSH payments (U.S. GAO 1994). But even subtracting the DSH payments from actual spending (assuming that block grants would have lowered the incentive to use this option), the caps would have reduced federal Medicaid funding by 25 percent during the ten-year budget window (and 49 percent in 1991). Thus, had the 1981 block grant been enacted, federal Medicaid spending would have been much more constrained than anticipated at the time.

Actual versus Proposed Medicaid Spending under the 1995 Block Grant Plan, 1996–2002

Proposed Cap on Federal Spending ($ billions)Actual Federal Spending ($ billions)Difference (Proposed Cap Minus Actual Spending)aFiscal Year($ billions)(%)

1996	95.7	92.0	3.7	4
1997	102.1	95.6	6.5	7
1998	106.2	101.2	5.0	5
1999	110.5	108.0	2.5	2
2000	114.9	117.9	−3.0	−3
2001	119.5	129.4 (122.8)b	−9.9 (−3.3)b	−8 (−3)b
2002	124.2	147.5 (140.0)b	−23.3 (−15.8)b	−16 (−11)b
5-year total	529.4	514.7	14.7	3
7-year total	773.1	791.6 (805.7)	−18.5 (−4.4)	−2 (−1)

Actual versus Proposed Medicaid Spending under the 1981 Block Grant Plan, 1982–1991

Fiscal YearProposed Cap on Federal Spending ($ billions)Actual Federal Spending ($ billions)Difference (Proposed Cap Minus Actual Spending)($ billions)(%)

1982	18.3	17.4	0.9	5
1983	19.0	19.0	0.0	0
1984	19.7	20.1	−0.4	−2
1985	20.4	22.7	−2.3	−10
1986	20.8	25.0	−4.2	−17
1987	21.5	27.4	−5.9	−22
1988	22.2	30.5	−8.3	−27
1989	23.0	34.6 (34.2)a	−11.6 (−11.2)a	−33 (−32)a
1990	23.9	41.1 (40.5)a	−17.2 (−16.6)a	−42 (−40)a
1991	24.8	52.5 (49.5)a	−27.7 (−24.6)a	−53 (−49)a
5-year total	98.3	104.2	−5.9	−6
10-year total	213.7	290.3 (286.2)a	−76.6 (−72.5)a	−26 (−25)a

Table 2 shows the impact on the federal budget of the 1995 Medicaid block grant proposal. Federal Medicaid savings would have been $18.5 billion from 1996 to 2002 (the time period for the plan) and $4.4 billion if the UPL payments were subtracted from the actual spending (assuming that the states would not have created UPL programs under the block grants). Unlike the 1981 proposal, which underestimated the projected federal savings, the savings from the 1995 proposal were overestimated. In 1995, the CBO projected its federal savings to equal $182 billion over the seven-year period (O’Neill 1995). In fact, block grant funding would have exceeded the current law's funding from 1996 to 1999, a trend not anticipated by the CBO. During this period, the enrollment in Medicaid unexpectedly declined, and the inflation in health costs slowed, resulting in record-low federal Medicaid cost growth in 1996 (Holahan, Bruen, and Liska 1998). But by 2000, these trends had been reversed, driven by higher Medicaid enrollment owing to the slowing economy, rapidly rising drug and hospital costs, and the growth of long-term care costs (Bruen and Holahan 2002). In FY 2002 alone, federal payments under the proposal could have been 16 percent, or $23 billion, less than they actually were ($15.8 billion lower when subtracting UPL payments).

Because of the numerous state options in Medicaid and the geographic differences in health care costs across the United States, no two state Medicaid programs or cost structures are alike. Thus, the effects of the block grant proposals would have varied considerably across the states. For example, while the 1995 block grant would have reduced the national federal Medicaid funding for FY 2002 by 15 percent, Table 3 shows that the percentage differences in the various states ranged from a drop of 45 percent to caps that exceeded actual federal spending in two states (Texas and West Virginia). Generally, the states that would have had the greatest reductions because of the caps expanded their coverage during this period (e.g., Vermont) or had high proportions of elderly beneficiaries for whom the costs grew rapidly (e.g., Iowa). The federal funds for those states that expanded their coverage (e.g., Arizona, Tennessee, and Washington) would have fallen more than 20 percent. Similarly, the block grant would have disproportionately reduced funding in predominantly rural states like New Mexico and Idaho, which is consistent with the fact that a greater proportion of rural than urban people rely on Medicaid (Kaiser Family Foundation 2003b). Federal spending in about 30 states would have dropped more than the national average, suggesting that even if the proposed federal spending had been set equal to what it actually was in 2002, most states would have lost some federal funds.

Actual versus Proposed Medicaid Spending by State under the 1995 Block Grant Plan, 2002

StateProposed Cap on Federal Spending ($ millions)Actual Federal Spending ($ millions)Difference (Proposed Cap Minus Actual Spending)a($ millions)(%)

AK	275	500	−225	−45
AL	2,145	2,285	−140	−6
AR	1,430	1,779	−349	−20
AZ	1,913	2,516	−603	−24
CA	13,678	15,418	−1,740	−11
CO	1,030	1,237	−207	−17
CT	1,688	1,837	−149	−8
DC	579	811	−232	−29
DE	285	377	−92	−24
FL	5,680	5,866	−186	−3
GA	3,301	4,224	−923	−22
HI	373	465	−92	−20
IA	1,108	1,687	−579	−34
ID	426	608	−182	−30
IL	4,684	4,854	−170	−4
IN	2,589	2,774	−184	−7
KS	946	1,203	−256	−21
KY	2,230	2,812	−582	−21
LA	2,909	3,563	−654	−18
MA	3,312	4,156	−844	−20
MD	1,810	1,937	−127	−7
ME	801	1,020	−219	−21
MI	4,595	5,073	−477	−9
MN	2,070	2,397	−327	−14
MO	2,484	3,553	−1,069	−30
MS	1,929	2,255	−326	−14
MT	422	471	−49	−10
NC	3,432	4,449	−1,017	−23
ND	320	341	−21	−6
NE	542	859	−317	−37
NH	375	520	−145	−28
NJ	3,294	4,181	−887	−21
NM	970	1,365	−395	−29
NV	394	462	−68	−15
NY	14,888	18,340	−3,452	−19
OH	5,350	5,919	−569	−10
OK	1,393	1,801	−408	−23
OR	1,444	1,680	−236	−14
PA	5,818	7,204	−1,386	−19
RI	630	812	−182	−22
SC	2,291	2,465	−174	−7
SD	349	386	−38	−10
TN	3,342	4,283	−941	−22
TX	9,102	8,705	397	5
UT	683	729	−46	−6
VA	1,607	2,205	−598	−27
VT	305	461	−155	−34
WA	2,035	2,833	−798	−28
WI	2,267	2,782	−515	−19
WV	1,534	1,329	205	15
WY	178	191	−13	−7
USa	127,233	149,975	−22,741	−15

Even though it would have lowered federal spending less than expected, the 1995 block grant would have had a significant programmatic impact on Medicaid. After subtracting the estimated $7.5 billion in UPL spending (conservatively assuming that these payments did not fund services in 2002), federal Medicaid funding in 2002 would have been $15.8 billion less than it was under current law. To put this into perspective, the one-year reduction in federal funding under the block grant would have exceeded federal Medicaid spending on prescription drugs or home- and community-based services and would have equaled half of all Medicaid spending on nursing homes (U.S. OMB 2002). If this loss of federal funding had been absorbed through eligibility restrictions, then more than 6 million people would have lost Medicaid coverage in 2002, using actual spending and enrollment data and assuming proportionate cuts across beneficiary groups (data from U.S. CBO 2003). The states could have responded to these reductions in federal funding by increasing their own spending in order to maintain services for their Medicaid enrollees. However, the amount required to do so would have been equivalent to a 40 percent increase in the $40 billion shortfall that was estimated for 2002, prior to the states’ actions to close this gap (NGA and NASBO 2002). It is important to note that this analysis does not take into account the amount of the states’ share of spending that would have been reduced under this proposal. Had this been included, the reductions just described would have been larger in some states.

The potential negative effects of the block grant proposals might have resulted from the proposals’ details, not from the concept of block grants. If only the architects of the Medicaid block grant plans had better captured the factors driving Medicaid cost growth and variation, one might argue, they could have achieved the goals of greater federalism and federal funding predictability without the unintended consequences. Although it is impossible to resolve this argument, two analyses can address how well the past projections predicted the future. First, we examined for their accuracy the Congressional Budget Office's annual federal Medicaid spending projections, made as part of the budget process. Figure 1 compares the actual federal Medicaid spending with what the CBO projected it would be three years and five years before each given year (not taking into account the policy changes in between). The projections of what Medicaid spending would be three years into the future ranged from 28 percent higher than the actual spending for the year 1996 to 31 percent lower for 1992. The five-year projections differed even more from the actual spending (44% too high for 1998 and 37% too low for 1992). Some of this difference resulted in changes in federal and state law between the time of the projection and the actual year. Nevertheless, the results suggest not only large differences between Medicaid projections and actual spending but also inconsistency in the difference: Sometimes they were positive, and sometimes negative.

A second analysis examines how well Medicaid spending growth was predicted by individual factors measuring changes in price, enrollment, and utilization. I conducted simple regression analyses (using an R-squared test) on total (federal and state) Medicaid spending growth to determine the extent to which major indicators of price, utilization, and enrollment changes explained the growth of costs. The results suggest that if medical inflation in the past were used as the sole predictor of growth, it would have accounted for only about 11 percent of the actual growth in Medicaid costs over a 30-year period (from 1968 to 2001, data not shown). Under President Bush's optional block grant program, medical inflation was to be the only growth factor used to set the annual state payments. Adding enrollment growth should, and does, improve the explanation of the growth of Medicaid costs, raising it to about 30 percent of total Medicaid cost growth. The last factor added to the cost equation is utilization changes. Using private health spending growth per capita as a proxy for both price and utilization changes, it appears that this factor plus enrollment changes still explain less than one-third of the growth in overall Medicaid costs. This analysis demonstrates that changes in Medicaid program costs are difficult to predict and thus that a “more perfect” block grant proposal would be hard to design.

My analysis of the implications of Medicaid block grants is limited in a number of ways. First, past Medicaid spending is not necessarily “ideal” Medicaid spending or an accurate measure of need. As described earlier, some of the federal spending may have been related to the states’ Medicaid “maximization” efforts, and thus it inflated the baseline. While my analyses attempted to subtract this type of spending, no one really knows how much of federal Medicaid funding is spent on such activities. Second, my analysis does not take into account potential efficiencies that might have been prompted by the proposals. Those states facing constraints on federal funding could find new ways to contain costs in order to maintain current levels of service. Third, the CBO and the CMS use complex models to project spending, whereas I considered only top-line projections, in part because policy proposals typically cap spending in the aggregate.

Notwithstanding these limitations, the analyses described here provide several new insights into the potential implications of Medicaid block grants. The first is that neither the 1981 nor the 1995 proposal came close to mirroring what actually happened in Medicaid spending. The federal funding caps in the 1995 proposal were higher than the actual spending in the first four years and then, by the proposal's final year, considerably below it. The 1981 proposal would have limited federal Medicaid spending to well below the actual amount—so much lower that Medicaid as it is currently structured could not have been sustained had it been enacted. This suggests that although a Medicaid block grant's federal spending may be predictable, its impact relative to current law is not. Fiscal conservatives aiming to reduce federal outlays may end up spending more than they anticipated, and moderates intending to maintain current coverage may find that the funding limits are too low to do so.

A second finding, consistent with previous estimates based on simulation models, is that the effects would have been dramatically different in each state. Medicaid spending varies across states not only because of their different local costs, covered services, and mix of beneficiaries but also because of their different growth rates. Although not analyzed here, the 1981 proposal would have had especially negative effects on states in which costs were growing rapidly (e.g., those with expanding elderly populations or high health cost inflation), since its annual caps were linked to the general growth rate of the U.S. economy. The 1995 proposal would have reduced federal spending differentially across the states, ranging from zero to 45 percent. This would have resulted in political as well as policy concerns. At the time, the fairness of funding, based on estimates, was questioned by the states and their members of Congress, creating a “formula fight” that ultimately contributed to the proposal's demise (Medipork 1995; Pear 1995). The danger of this type of fight may have been why the Bush administration delegated to the governors themselves the task of coming up with the details of his block grant proposal, although this may have added to their inability to do so.

An equally contentious point in the 1995 debate was whether the Medicaid block grant would cause millions of beneficiaries to lose their coverage. Looking at the actual 2002 spending compared with what it would have been under the block grant, it appears that some people would have lost coverage. The reduction of nearly $16 billion in that year is equivalent to the cost of covering more than 6 million people and half the cost for all federal Medicaid nursing homes. Even if the overall federal cap had been considerably higher than the actual spending, the experience in SCHIP shows that a $12 billion surplus nationwide still leaves some states with insufficient funding, thereby forcing the enrollment of fewer children (Park, Ku, and Broaddus 2002). Recent state budget problems indicate that even though the states are trying to cut costs by paying providers less and controlling drug costs, 18 states reported considering cutting eligibility, and 17 states reported considering benefit cuts for 2004 (Smith et al. 2004).

Could perhaps a more accurate formula for a Medicaid block grant ensure that the federal limit would be both sufficient to maintain services and constrained enough to meet federal budget targets? My examination of the accuracy of past governmental projections of actual Medicaid spending showed that predicting Medicaid costs just three years into the future was difficult, with estimates off by 30 percent in both directions. I also reported that the actual growth of enrollment and medical inflation explained less than one-third of the growth in Medicaid costs over the past 30 years. Medicaid's multiple roles—as an insurer of low-income people, payer of long-term care, safety net for the uninsurable, and supplemental insurer to Medicare—likely contribute to the variability in its cost growth. It also may be that health programs in general have unpredictable costs. A similar analysis that I conducted of projections of both Medicare and private health spending produced similar results (available on request).

These results corroborate previous but contested assessments of the potential problems with a block grant approach to Medicaid. As such, they suggest a reexamination of the motivation for such a policy. Some of the goals of a Medicaid block grant could be achieved without changing the program's entitlement. Greater program flexibility and stronger Medicaid oversight to limit states’ schemes to maximize federal payments can be—and have been—accomplished under the current financing structure. Laws enacted in 1991, 1993, 1997, and 2000 addressed these priorities (Schneider et al. 2002). The goal that cannot be achieved without a block grant is making federal Medicaid spending predictable.

As this analysis suggests, achieving predictability would come at a cost. It would limit the extent to which state health programs could respond to epidemics, cover new but costly medical breakthroughs, and accommodate demographic shifts. It would end the federal contribution to state-initiated coverage expansions for low-income working families. And it would prevent Medicaid from compensating for the loss of private coverage during recessions. Indeed, the increase in the rate of uninsured Americans that occurred between 2000 and 2003 (from 14.2% to 15.6%) would have been even greater without a larger increase in the proportion of the population covered by Medicaid (from 10.6% to 12.4%) (U.S. Census Bureau 2004). Moreover, as this article shows, it is unlikely that the current level of Medicaid enrollment could be sustained under most block grant policies. Since individuals losing Medicaid probably have few affordable alternatives, block grants could result in raising the number of the uninsured, making Medicaid part of the problem rather than part of the solution to the coverage gaps in the United States. As such, the future debates over Medicaid block grants should focus on whether federal budget certainty is more important than covering low-income and vulnerable populations, since this is ultimately what is at stake.

The research for this article was supported by the Kaiser Commission on Medicaid and the Uninsured. The author thanks the commission's staff for their support and input. Funding for this study was provided by the Henry J. Kaiser Family Foundation. The views expressed are solely those of the author and do not necessarily reflect those of the Kaiser Family Foundation or George Washington University.

• Brown GD. New Federalism, Old Federalism, and the Fiscal Constitution. ABA Journal. 1982;68(August):946. [Google Scholar]
• Bruen B, Holahan J. Acceleration of Medicaid Spending Reflects Mounting Pressures. Washington, D.C.: Kaiser Commission on Medicaid and the Uninsured; 2002. [Google Scholar]
• Bureau of Economic Analysis (BEA), U.S. Department of Commerce. Table 1.1.9: Implicit Price Deflators for Gross Domestic Product. 2004. [accessed August 28, 2004]. Available at //www.bea.gov/bea/dn/nipaweb/TableView.asp?SelectedTable=13&FirstYear=1999&LastYear=2004&Freq=Qtr.
• Bush J. Testimony before the U.S. House Energy and Commerce Committee, Health Subcommittee; 2003. Medicaid Today: The States’ Perspective. March 12. [Google Scholar]
• Cantwell JR. Reforming Medicaid. Dallas: National Center for Policy Analysis; 1995. [Google Scholar]
• Centers for Medicare & Medicaid Services (CMS) 2002 Data Compendium. Baltimore: 2002. [accessed April 4, 2004]. Available at //www.cms.hhs.gov/researchers/pubs/datacompendium/2002/02pg34.pdf. [Google Scholar]
• Centers for Medicare & Medicaid Services (CMS) FACT SHEET: Medicaid and Acquired Immunodeficiency Syndrome (AIDS) and Human Immunodeficiency Virus (HIV) Infection. Baltimore: 2004. [accessed April 4, 2004]. Available at //www.cms.hhs.gov/hiv/hivfs.asp. [Google Scholar]
• Clinton WJ. Washington, D.C.: U.S. National Archives and Research Administration; 1995. [accessed August 28, 2004]. Remarks by the president in meeting with Democratic governors. December 8. Available at //clinton6.nara.gov/1995/12/1995-12-08-pres-remarks-in-meeting-with-democratic-governors.html. [Google Scholar]
• Connolly C. Governors’ Efforts to Revise Medicaid Stalls; GOP Group Looks to the White House. Washington Post. 2003:A4. June 13. [Google Scholar]
• Etheredge L. Reagan, Congress, and Health Spending. Health Affairs. 1983;2(1):14–24. 10.1377/hlthaff.2.1.14. [Google Scholar]
• Families USA. Slashing Medicaid: Hidden Effects of the President's Block Grant Proposal. Washington, D.C.: 2003. [Google Scholar]
• Finegold K, Wherry L, Schardin S. Block Grants: Details of Bush Proposals. Washington, D.C.: Urban Institute; 2004. New Federalism Series A, no. A-64. [Google Scholar]
• Frogue J. Medicaid's Perverse Incentives, The State Factor. Washington, D.C.: American Legislative Exchange Council; 2004. [Google Scholar]
• Greve MS. Washington and the States: Segregation Now, Federalist Outlook. Washington, D.C.: American Enterprise Institute for Public Policy Research; 2003. [Google Scholar]
• Health Care Financing Administration (HCFA) A Profile of Medicaid, 2000 Chart Book. Baltimore: Centers for Medicare & Medicaid Services; 2000. [accessed April 4, 2004]. Available at //www.cms.hhs.gov/charts/medicaid/2Tchartbk.pdf. [Google Scholar]
• Holahan J, Bruen B, Liska D. The Decline in Medicaid Spending Growth in 1996: Why Did It Happen? Washington, D.C.: Kaiser Commission on Medicaid and the Uninsured; 1998. [Google Scholar]
• Holahan J, Liska D. The Impact of the Budget Resolution Conference Agreement on Medicaid Expenditures. Washington, D.C.: Urban Institute; 1995. [Google Scholar]
• Holahan J, Weil A. Block Grants Are the Wrong Prescription for Medicaid. Washington, D.C.: Urban Institute Health Policy Online; 2003. [Google Scholar]
• Ifill G. States, Following Bush Plan, Ask Congress to Consolidate U.S. Aid. New York Times. 1991:A21. April 9. [Google Scholar]
• Kaiser Family Foundation. Medicaid “Optional” and “Mandatory” Eligibility and Benefits. Washington, D.C.: Kaiser Commission on Medicaid and the Uninsured; 2001. [Google Scholar]
• Kaiser Family Foundation. Bush Administration Medicaid/SCHIP Proposal. Washington, D.C.: Kaiser Commission on Medicaid and the Uninsured; 2003a. [Google Scholar]
• Kaiser Family Foundation. The Uninsured in Rural America. Washington, D.C.: Kaiser Commission on Medicaid and the Uninsured; 2003b. [Google Scholar]
• Kaiser Family Foundation. Medicaid and Block Grant Financing Compared. Washington, D.C.: Kaiser Commission on Medicaid and the Uninsured; 2004a. [Google Scholar]
• Kaiser Family Foundation. The Medicaid Program at a Glance. Washington, D.C.: Kaiser Commission on Medicaid and the Uninsured; 2004b. [Google Scholar]
• Mann C, Nathanson M, Park E. Administration's Medicaid Proposal Would Shift Fiscal Risk to States. Washington, D.C.: Georgetown University Institute for Health Policy Research and Center on Budget and Policy Priorities; 2003. [Google Scholar]
• Medipork. Editorial. Washington Post. 1995:A24. November 6. [Google Scholar]
• National Governors Association (NGA) and National Association of State Budget Officers (NASBO) The Fiscal Survey of States. Washington, D.C.: 2002. [Google Scholar]
• O’Neill J. Washington, D.C.: U.S. Congressional Budget Office; 1995. Letter to Congressman Bliley Regarding CBO Cost Estimate for the Medicaid Reconciliation Recommendations of the House Committee on Commerce, as Approved on September 22, 1995. [Google Scholar]
• Park E, Ku L, Broaddus M. Washington, D.C.: Center on Budget and Policy Priorities; 2002. OMB Estimates that 900,000 Children Will Lose Health Insurance Due to Reductions in Federal SCHIP Funding. [PubMed] [Google Scholar]
• Pear R. With Millions and Billions, G.O.P. Budget Takes Shape. New York Times. 1995:A1. October 28. [PubMed] [Google Scholar]
• Schneider A, Elias R, Garfield R, Rousseau D, Wachino V. The Medicaid Resource Book. Washington, D.C.: Kaiser Commission on Medicaid and the Uninsured; 2002. [Google Scholar]
• Schneider A, Rousseau D. Upper Payment Limits: Reality and Illusion in Medicaid Financing. Washington, D.C.: Kaiser Commission on Medicaid and the Uninsured; 2002. [Google Scholar]
• Smith V, Ramesh R, Gifford K, Ellis E, Wachino V, O’Malley M. States Respond to Fiscal Pressure: A 50-State Update of State Medicaid Spending Growth and Cost Containment Actions. Washington, D.C.: Kaiser Commission on Medicaid and the Uninsured; 2004. [Google Scholar]
• Thompson T. Block (Grant) Party. Heritage Foundation's Policy Review no. 75. 1996 [Google Scholar]
• Thompson T. The Future of Medicaid. Spectrum: The Journal of State Government. 2003:5–6. [Google Scholar]
• U.S. Census Bureau, U.S. Department of Commerce. 2002 Statistical Abstract of the United States. Washington, D.C.: U.S. Government Printing Office; 2003. [Google Scholar]
• U.S. Census Bureau, U.S. Department of Commerce. Income, Poverty, and Health Insurance Coverage in the United States: 2003. Washington, D.C.: U.S. Government Printing Office; 2004. [Google Scholar]
• U.S. Congressional Budget Office (CBO) The Economic and Budget Outlook. Washington, D.C.: U.S. Congressional Budget Office; Annual editions are published in January or February of each year. Note that beginning in 2000 the name was changed to The Budget and Economic Outlook. This article used the annual reports beginning in 1985, and it used the December 1995 edition (there was no report in January due to the unusual budget circumstances) [Google Scholar]
• U.S. Congressional Budget Office (CBO) Fact Sheet for CBO's March 2002 Baseline: Medicaid and the State Children's Health Insurance Program. Washington, D.C.: U.S. Congressional Budget Office; 2002. [Google Scholar]
• U.S. Congressional Budget Office (CBO) Fact Sheet for CBO's March 2003 Baseline: Medicaid and the State Children's Health Insurance Program. Washington, D.C.: U.S. Congressional Budget Office; 2003. [Google Scholar]
• U.S. General Accounting Office (GAO) Medicaid: States Use Illusory Approaches to Shift Program Costs to the Federal Government. Washington, D.C.: U.S. General Accounting Office; 1994. [Google Scholar]
• U.S. General Accounting Office (GAO) Medicaid Formula Proposal Summary of State Changes from 1996 to 2002. Washington, D.C.: U.S. General Accounting Office; 1995. (c:\medicaidexpndh2.wb1) [Google Scholar]
• U.S. Office of Management and Budget (OMB) FY 2003 Budget of the United States Government, Analytic Perspectives. Washington, D.C.: U.S. Government Printing Office; 2002. [Google Scholar]
• U.S. Office of Management and Budget (OMB) FY 2004 Budget of the United States Government. Washington, D.C.: U.S. Government Printing Office; 2003. [Google Scholar]
• U.S. Office of Management and Budget (OMB) FY 2005 Budget of the United States Government. Washington, D.C.: U.S. Government Printing Office; 2004. [Google Scholar]