Spending on Social Welfare Programs in Rich and Poor States. Final Report.. Chapter V: Final Observations and Conclusions


Our econometric analyses and case studies of six poor states led us to several conclusions regarding state fiscal capacity and state spending on social welfare:

Finding #1: States of LESS fiscal capacity spent LESS PER CAPITA on social welfare programs than states with HIGHER per capita incomes. Federal grants did not reduce absolute spending differences between rich and poor states. Average federal grants to the wealthiest states were actually higher in dollar terms when compared to states with lower fiscal capacity. However, because state own-source spending was much lower in poor states, federal intergovernmental grants constituted a larger share of the social welfare budgets of poor states than of rich states.

Finding #2: State fiscal capacity bore a stronger relationship to spending on non-health social welfare programs than on health-related programs. Between 1977 to 2000, differences between rich and poor states were greatest for spending on cash assistance and non-health social services (such as child care, child welfare, energy assistance, transportation assistance, and programs for the homeless). Differences between rich and poor states were smaller for health-related programs, such as per capita spending on Medicaid and payments to public hospitals.

Finding #3A: Between 1977 and 2000, state spending on social welfare changed in major ways. Even after controlling for the higher levels of inflation found in health services, spending on Medicaid greatly increased throughout this period, most rapidly in the late 1980s and early 1990s. Spending on non-health social services rose gradually throughout this period. Average state spending on cash assistance rose in the late 1980s and early 1990s but fell dramatically after the mid-1990s.

Finding #3B: These trends varied greatly between rich and poor states. Medicaid grew substantially for all states, but the growth was particularly strong among those of low fiscal capacity. The correlation between state fiscal capacity and per capita spending on Medicaid declined over time, as per capita spending by poor states climbed to levels only exceeded by the wealthiest states, while Medicaid spending in wealthier states grew slowly or stalled during the late 1990s.

States of different fiscal capacities also converged in their spending on cash assistance programs. Wealthier states reduced their spending during the middle and late 1990s, while poor states on average showed little change in their per capita spending on cash assistance throughout the last two and a half decades.

By contrast, differences grew between rich and poor states in their spending on non-health social service programs, especially in the late 1990s. Growth was fairly consistent among wealthier states through the 1980s and 1990s; growth in spending on these non-health services was much weaker, however, among the poorest states.

Finding #3C: These changes produced major shifts in the composition of social welfare budgets in rich and poor states. States of all fiscal capacity have greatly increased the proportion of their spending devoted to Medicaid while reducing the proportion spent on cash assistance. Poor states, however, unlike rich states, have also reduced the proportion of their budgets spent on non-health social services. Thus, the package of benefits offered by poor states has changed markedly in recent years, toward health care and away from non-health services.

Finding #4: Econometric analyses found that different factors influenced different social programs. Spending on cash assistance was increased by federal grants, and unemployment. Spending on cash assistance was also greater in urban states, i.e., those with higher population densities. There was a negative relationship between spending on cash assistance and state fiscal capacity. This negative effect not well understood, and could be related to the positive effect of wages on personal income and the negative effect of wages on cash assistance rolls.

Medicaid spending was increased by personal income, federal grants, and unemployment. However, the effects of federal grants (measured by federal grants for social welfare programs) were particularly strong for Medicaid spending, and population density had an effect opposite to its impact on cash assistance. Medicaid spending was greater, other things equal, in comparatively rural states, i.e., those with lower population densities.

Non-health social services was most affected by overall state income. It was strongly and consistently related to state fiscal capacity. It was also significantly related to federal grants for non-social-welfare programs. Like Medicaid (and public hospital payments), it was also higher in states with lower population densities.

Thus, some variables shaped the composition and dynamics of state social welfare spending. States with higher population densities spent more on cash assistance, while rural states with low population densities spent more on Medicaid, public hospitals, and non-health social services. Unemployment, an indicator of changing social needs, increased spending on cash assistance and Medicaid but not on non-health services or public hospitals.

Finding #5A: The econometric models were most successful in explaining spending differences and changes among wealthy states; the models fared less well in accounting for spending in poor states Most of the variables — including fiscal capacity, unemployment, and federal grants — showed relatively strong effects among the wealthier states. In poorer states, fiscal capacity, unemployment, and federal grants showed little or no effects. One important exception was Medicaid. Spending on Medicaid was significantly and strongly affected by federal grant dollars in poor states.

Finding #5B: There were substantial differences among poor states in their long-run propensities to spend on programs (as captured in the “state effects of the econometric model). In particular, there were different propensities for spending on cash assistance and health-related programs (Medicaid and public hospitals). Some poor states (mostly rural southern states) spent very little on cash assistance but relatively more on health-related programs, while other poor states (mostly in the West) had larger cash assistance programs and spent less on Medicaid. This trade-off between health and cash assistance programs was not found among wealthier states. Wealthier states were, in general, less likely than poor states to reveal negative correlations between their long-run propensities to spend on different program functions. Thus, poor states showed greater specialization in their spending “packages when compared to wealthier states.

Finding #6: Case studies of six states of low fiscal capacity and high social needs indicated that the basic trends in spending found among poor states before 2000 continued after that year. Spending on Medicaid grew in most of the poor states despite fiscal downturns. Large cut-backs in Medicaid eligibility and basic services were uncommon; in fact, some major program expansions occurred. Nor did cash assistance spending decline — in fact, some increases were found in spending on TANF assistance.

Major cuts were most often imposed on non-health social services and administrative expenses, especially staffing. This may have been a departure from previous recessions, when government employment was usually not strongly affected by revenue downturns. But our case studies revealed large reductions in front-line personnel, supervisors, and managers — and sometimes the elimination of entire local offices. Cuts were also made in employment services; transportation and other work supports; fatherhood, parenting, and pregnancy prevention programs; and child welfare services. Child care subsidies were reduced substantially in the two states that had devoted considerable state revenues to such subsidies in the past.

Finding #7: The case studies visits also revealed that, at least among poor states, spending in different program areas were typically determined by different political and administrative processes. Cash assistance programs were affected by the interaction between caseload levels and the rules and benefit levels determined (and not often revised) by state legislatures. Choices affecting cash assistance spending seemed to be more influenced by racial and ideological divisions in the state.

By contrast, Medicaid policies and expenditures were, especially in the rural southern and border states, strongly affected by the active political involvement of service providers, such as nursing homes, hospitals, physicians, and many others — resulting in a relatively bipartisan politics of institutionalized interests. This public-private interdependence was especially strong in the rural southern states and may help account for their relatively higher levels of Medicaid spending.

To be sure, the generous federal matching rates for Medicaid in poor states were also critical in state budget and policy processes; and poor states were strongly affected by federal changes in mandated eligible populations or minimal services, changes which were frequent in the Medicaid program during the late 1980s and early 1990s. However, these factors alone did not seem sufficient in accounting for Medicaid’s robust support during the recent state fiscal crises. Our case studies suggested that when a program has strong and active constituencies that supported greater spending, an attractive match rate may suffice to expand spending during boom times and prevent major cutbacks during recessions. If, however, a program (such as many in the child welfare area) does not have such strong political advocates, even the same federal match rate may not prevent major services during fiscal downturns.

Finally, non-health social services were typically of low political salience and administrators were often given significant discretion over how to allocate funds across different services. State resources, program flexibility, and executive priorities seemed more important in determining how much was spent on these non-health services and which services were funded and which were not.

Recent changes in federal programs affected the ability of poor states to support and control their non-health services to low-income people. For example, TANF surpluses and the block grant’s flexibility offered administrators and other officials greater control in determining spending priorities among these services, when they needed to make cuts. This flexibility — and the observed tendency of states to fund a wider range of programs under TANF and MOE — also created a more wide-ranging competition for resources among many services that probably had not competed before.

In sum, fiscal capacity remains an important factor in understanding state spending on social services. The most remarkable findings of the study, however, concern the different influences on spending across program areas and states. These differences suggest important dynamics in state spending on social welfare. Because different states and program areas are subject to different influences, we may continue to see — as we did find in the period from 1977 through 2003 — major shifts in the levels, composition, and distribution of state spending on various social welfare functions.

View full report


"report.pdf" (pdf, 1.52Mb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®

View full report


"apa.pdf" (pdf, 2.64Mb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®