II A – Roots of the revolution

13In the mid-1990s, the roots of this revolution were planted by two trends, one economic and one political. First, a national recession put tremendous fiscal pressure on states, especially in the area of social services. Though the recession lasted only from July 1990 through March 1991, [10] the slow recovery created major deficits in state budgets for years to follow. Twenty-nine states were forced to cut their budgets in 1991, and thirty-five states made budget cuts totaling $4.5 billion in 1992. [11] California’s budget did not pass until months past the constitutional deadline in 1993, and even as late as 1995, eight states, including New York, were still forced to make cuts. [12]

14Why did the recession hit states so hard? One reason is that nearly all American states have strict balanced budget requirements, [13] a rule that is notably lacking at the federal level. Presidents and Congress can react to deficits by increasing spending and cutting taxes, as President Obama and the Democratic Congress did with the 2009 stimulus package. States, by contrast, must balance their books through painful cuts and tax increases. A second reason is that because states administer so many of the nation’s safety net services, any financial downturn has a two-sided effect on their budgets. The supply of tax revenues goes down, but the demand for social services like Medicaid and welfare rises.

15A contemporary report by the National Association of Budget Officers at the time demonstrated how these supply and demand dynamics threw state budgets out of balance: “Both the modest budget growth and the midyear budget adjustments reflect the tepid economy as well as pressures from double-digit growth in Medicaid spending and increased welfare caseloads.” [14] The statements of state leaders testify to how they saw these trends. Medicaid was called the “Pac-man” of state budgets, a reference to the video game character who ate up everything around him, and Georgia governor Zell Miller compared it to the state’s quick-spreading kudzu plant. [15] State officials chafed at what they perceived as a lack of flexibility to make major Medicaid and welfare cuts, and called upon Washington DC to give states great power over these increasingly expensive programs.

16After the 1994 elections, Congress became much more receptive to such demands. The Newt Gingrich-led “Republican revolution”, in which the GOP took control of the House of Representatives for the first time in a generation, set the stage for the devolution revolution. This shift in party control came in the midst of the growing political polarization between the two parties in America. [16] Whether the sharp disagreement between Democrats and Republicans on issues such as the optimal size of the safety net was due more to shifts among elected leaders [17] or among voters, it was clear during the 1994 elections that the partisan divide was vast. After winning those elections, Gingrich and many other Republicans pushed for a switch from federal matching grants for social services to “block grants”, a proposal that had also been made by Ronald Reagan. This was a proposal as much about ideology as it was about federalism, since all sides expected that block grants would lead to smaller government. “Block grants aim to provide greater federal budget certainty and a stronger state incentive to contain program costs”, according to Lambrew. [18] They do so by giving states both a fixed annual sum from the federal government to run a social service program (increasing predictability) and by making states responsible for any and all funding above that sum (providing them with the incentive to cut costs). In 1995, Speaker Gingrich pushed to switch the Medicaid program into a block grant to states, but lost in a political battle that revived the electoral fortunes of President Bill Clinton.

17The Democratic president had steadfastly opposed major changes in Medicaid, but throughout his career had called for “an end to welfare as we have come to know it”. [19] He vetoed two early welfare reform bills pushed by Republicans, but as the 1996 election approached, Clinton took a centrist position by cutting a deal with Gingrich on a major welfare reform package in August of that year. The new law transformed the Aid to Families with Dependent Children program, renaming it Temporary Assistance for Needy Families, proposing a five-year lifetime limit on benefits, requiring that welfare recipients go back to work after two years on the rolls, and providing more child care for working recipients. [20] The law also changed federalism by making welfare a block grant, and giving states tremendous control over how to spend the grant. No longer did states receive matching dollars from the federal government, giving them the greater incentives to cut costs that Gingrich had favored for Medicaid. States were also given great freedom to craft their new programs; they could make their lifetime limits either shorter or longer than the federal baseline of a five-year limit, they could set the length of the workweek required of welfare recipients, and could create other limits or extensions. Whether states wished to build a program that was more conservative or more liberal than the federal guidelines, they had the flexibility to do so. [21]

18Another major deal cut between Clinton and Gingrich the next year completed the devolution revolution. The creation of the state Children’s Health Insurance Program (CHIP) as part of a budget deal in 1997 put states in charge of building a new health care program for the children of the working poor and, at state discretion, their parents. Reaching families with incomes too high to qualify for Medicaid but too low to purchase private insurance, the program could be offered through Medicaid, through the plan that covered state employees, or through a new program that met specific benchmarks. [22] State legislatures and governors were left to decide. In this way, the CHIP program gave states another great source of authority. However, in its funding structure, it resembled Medicaid, with matching rates giving the federal government the ability to influence state decisions.

19While the move in the mid-1990s to shift policy and fiscal control of social services to the states may have been an incomplete victory for the advocates of states’ rights, it still constituted a revolution. Though Medicaid remained generally unchanged, the new CHIP program gave states policy control of a program with fiscal incentives still managed by the federal government, and the TANF program transformed welfare into a system with policies and purse strings firmly under state control.

20In the battles over each of these programs, Democratic and Republican leaders in Washington DC clearly understood that the major policy dimension that polarizes American parties – the battle over whether to expand or to shrink government – was at stake in the debate over whether to devolve authority and restructure grants. Federalism, to paraphrase Robertson, [23] would determine who won and who lost the fight over health care costs and welfare generosity. Republican support for decentralization of social policy, as Béland and de Chantal have observed, [24] was driven primarily by ideological opposition to big government. To understand completely the link between devolution and the size of the safety net, it is necessary to look closely at how federal funding formulas can influence state decisions.

II B – How federal funding formulas reshape state incentives

21In the joint federal–state programs that compose America’s health care and welfare programs, the national government has used both regulations and inducements to motivate states to broaden their safety nets. The regulations have come in the requirement that states dramatically expand their welfare (beginning in the 1930s) and health care programs (starting in the 1960s) to meet new nationwide standards. The inducements have been the federal dollars sent to pay much of the cost of these new mandates, structured in a way that gave states a clear incentive to meet and often to go beyond the national baseline. Federal grants were intentionally designed to push states toward more generous services than they might otherwise provide. If Washington DC had sent states block grants of a fixed amount, most states would have spent all of the federal money while trying hard to minimize state costs paid in addition to the block grant. Instead, federal law makers designed matching rates that required states to spend their own dollars in order to bring in federal money. The more a state spent of its own, the more money it would bring in from Washington DC, a formula for growth in social service provision.

22The rationale for this system, from economic theories of fiscal federalism, is that matching grants should “be employed where the provision of local services generates benefits for residents of other jurisdictions”. [25] That is, because the advantages of providing for the welfare and health care of residents in one state spill over into other states and the nation as a whole, those benefits should be internalized into the decision calculus of law makers in each state. It did not hurt that Democrats in Congress at the time favored higher social service spending, and realized that matching grants would create a powerful incentive for states, even ones that had traditionally been frugal, to move in this direction. Importantly, the federal matching rate varied across states, and always provided at least one federal dollar for each state dollar spent. Table 1 shows the matching rates that an illustrative sample of states receives today in Medicaid and in the CHIP program. It shows that poorer states like Mississippi and New Mexico received a higher match than rich states such as California and New York, a deliberate feature of the funding formula designed to give state law makers in states with tight budgets the ability to provide generous services.

Table 1

Federal matching rates for social service programs

State Medicaid CHIP Arkansas $4.31 $4.25 California $1.60 $1.86 Kansas $2.30 $2.61 Mississippi $5.61 $4.87 Montana $3.54 $3.38 New York $1.60 $1.86 New Mexico $4.13 $3.99 Texas $2.44 $2.46

Federal matching rates for social service programs

Notes: Entries indicate how many dollars will be transferred by the federal government to match one dollar of state spending, by program and by state. Medicaid and CHIP figures are from the 2010 fiscal year, both taken from Kaiser Family Foundation (2011) and Families USA (2011). [26]

23Consider the effects of this grant structure on the decision calculus of governors and state legislators writing their budgets. For a program like the Children’s Health Insurance Program, in New York the nearly two-to-one match means that in good times, the state can increase its budget by $100 million dollars and deliver, thanks to the federal grant, an additional $286 million in health care services to poor families. When the state budget is flush, this acts like an escalator, providing a great incentive to spend a surplus on the CHIP program because law makers get so much “bang for the buck” in this program area. But just as the federal match makes it easy to increase spending in this area during good times, it also makes it hard to cut health care spending in bad times. When New York is in deficit, in order to reduce its state deficit by $100 million, law makers must eliminate $286 million worth of services. This effectively sends $186 million back to the federal government, never an attractive option to state policy makers. This makes cutting health care to working families much more difficult, in both political and policy terms, than cutting other areas of government like higher education, prisons, or primary education that are almost entirely state funded. Because matching grants push states to increase health care provision in good times and make it hard to cut in bad times, they should be thought of – whether one favors or opposes generous social services – as a lever of federal control over state decisions that reduces state autonomy.

24In a poor state such as Mississippi, the federal matching rate is even higher. This gives the state an even better fiscal deal, at the same time that it makes the federal lever of control even stronger. For $100 million in new state spending, Mississippi law makers can secure an extra $487 in federal CHIP funds, but of course they would have to sacrifice that much in order to trim their state budget by $100 million. Washington DC structured grants this way during times when Democrats controlled government, knowing that the arrangement would lead to social service expansion, while Newt Gingrich and other Republicans sought to change the formulas in order to shrink government.

25Yet sometimes lost in this debate about the overall size of government is the important redistribution of federal tax dollars across states accomplished by the varying matching rates detailed in Table 1. Pointing out the disparities in these matching grant rates paints a portrait of an “equalizing” federalism that is at odds with the conventional idea that America’s national government treats all states uniformly. “In all other federal systems”, Robertson writes, “the national government deliberately equalizes regional resources by redistributing more financial aid to the poorest regions. The United States is the only federal system that does not equalize state resources in this way.” [27] In Medicaid, a massive program in which the federal government distributes $318 billion in aid across the states annually, [28] the aid is spread out unequally, with the most generous grants going to the poorest states. At least in this realm, American federalism is not exceptional; it redistributes money from rich to poor states just as federalism does in other nations.

II C – Implementing the devolution revolution in California

26Leaders in each state capitol had little time to ponder such distributional concerns in the late 1990s, because the devolution revolution gave them important decisions to make. Every state’s autonomous control over its welfare programs dramatically increased after the federal passage of welfare reform in 1996. While the freedom granted by this devolution sent states in different directions, the experience of one state can teach general lessons about how the shift in power and the new fiscal relationship played out in state capitols around the country. The nation’s largest state, California, has often been a policy leader whose choices, especially in social services, have diffused across the rest of the nation. [29] The state’s government at the time was split between the nation’s two political parties, with a Republican governor negotiating policy with a legislature controlled by Democrats. Like the rest of America, the Golden State was just coming out of the deep recession, giving policy makers some money to spend on expanding social services but leaving them cautious about how such generosity would threaten the state budget when lean years returned. Though never a “typical state”, California’s political and fiscal conditions at this time were reflective of what was occurring in many other states.

27In California, 1997 was the year of welfare reform, with Washington DC’s actions the previous year putting this issue at the top of the agenda for state officials at the state’s capitol in Sacramento. Governor Pete Wilson, a moderate Republican who had been elected on a platform of expanding “preventive” social services like child care and pre-natal health, but who had been forced by the state’s struggling economy to push for budget cuts in these and other areas, saw this as a chance to leave a major policy legacy. The Democrats who controlled the two houses of the state’s legislature worried that any welfare reform meant welfare cuts to the constituents whom many Democrats represented. Yet while legislators have a monopoly over the power to introduce bills in California, as in all American legislatures, they felt pressured to respond to the federal mandate for reform as well as to the way that block grants changed state fiscal incentives. As the following account shows, this led to one of the paradoxical patterns of federalism: while devolution gives states freedom, it also forces decisions upon them and allows the federal government to dictate state policy agendas.

28When California’s legislature convened for the beginning of its session in January 1997, one influential newspaper columnist wrote that the capitol’s hallways were filled with talk of many issues, but that, “There was, however, one topic that overshadowed all others in terms of magnitude, uncertainty and urgency: welfare reform. Under new federal welfare-reform legislation, Republican Gov. Pete Wilson and the Democrat-controlled Legislature have six months to develop an overhaul of the system that consumes nearly a fifth of the state budget and supports nearly 5 million Californians, about 20 percent of the nation’s welfare burden.” [30] The extent to which state action was forced by federal devolution was illustrated by another columnist’s quote: “The challenge the state faces to come up with a new welfare system is enormous, dictated by the federal welfare reform bill President Clinton signed last summer.” [31]

29Does this mean that devolution pushed the state in a direction that its leaders did not want to go? In the legislature’s case, it certainly did. Prominent leaders like Senate Health and Human Services Committee Chair Diane Watson, a Democrat representing an inner city Los Angeles district with many welfare recipients, certainly would have preferred not to change the existing system. But Gov. Wilson clearly did want change, and seized the political opportunity created by federal action. In the spring of 1997, he gave speeches across the state in support of his reform plan, which went beyond the federal guidelines by putting a one-year limit on welfare enrollment, cutting the size of monthly checks to recipients, and requiring them to work 35 hours a week in job training or education programs to receive the checks. [32]

30Democratic legislative leaders, quite predictably, pushed back against the governor’s proposal. When the bills containing them, authored by Republican legislators, reached the Assembly Human Services Committee, the Democratic majority on the committee stripped out their key provisions – the one-year time limit, the grant reductions, and the 35-hour workweek – on votes that broke down along party lines. “I would suggest that’s gutting the bill to almost nothing”, said Republican Assemblyman Keith Olberg, author of one of the bills. [33] Yet legislators knew that the governor possessed the key power of the veto, giving him the power to kill any welfare deal that the Democrats themselves reached and thus assuring that the state’s reaction to federal devolution would be a compromise between what California Democrats and Republican Gov. Pete Wilson wanted. Dion Aroner, the state Assembly’s Human Services Chair, spoke about welfare reform in a tone that was at once resigned to the fact that federal action worked against her policy preferences but also optimistic that it opened up new opportunities for positive state innovations. “We will do significant damage to lots of families and seniors and kids who will be hurt by what the federal government did”, Aroner said. “We can’t fix that […] But if the system becomes more focused on helping people get off welfare and finding them jobs, […] that will be historic reform.” [34]

31In order to craft a compromise that might garner both legislative approval and the governor’s signature, top leaders created a conference committee that brought together eighteen members of the Assembly and Senate, many of whom were political moderates not closely tied to either side of the issue. After liberals on the committee pushed for a plan that “went as far as it legally could to maintain the status quo […] Speaker Cruz Bustamante intervened, and he and Senate President Pro Tem Bill Lockyer directed Democratic members of a welfare conference committee to tone down their plan and make it at least marginally acceptable to moderates”. [35]

32After Gov. Wilson vetoed the first set of welfare bills sent to him at the beginning of the summer, in August he signed a compromise package crafted in the conference committee. Passed on votes of 65-11 in the Assembly and 33-5 in the Senate, with liberal Democrats objecting, the plan included an 18-month time limit and a 20-hour work requirement, splitting the difference between the governor’s plan and the legislature’s proposals. [36] Just as importantly for the state, it slowed the growth of welfare spending, with all of the savings going to other California programs because the funding was sent through a block grant. The devolution of power from the federal government had spurred state action, but the new policy bore the imprint of both branches of state government and put all of the savings back into California’s budget.

33Welfare was not the only power passed along to the states by Clinton and the Republicans in Congress. In late 1997, after the welfare reform debate was resolved in California, state law makers were given authority over a newly created program to provide health insurance for the children of the working poor through the Children’s Health Insurance Program. Unlike welfare reform, this program was a clear expansion of the social safety net, a concession that President Clinton won from Newt Gingrich that thrilled Democrats in state capitols across the country and calmed the fears of fiscal conservatives with its generous federal matching rate. Because every dollar of California spending was matched with nearly two federal dollars, Republican Governor Wilson and legislative Democrats both agreed on expanding coverage.

34While they could have done so through Medicaid or the government employees’ health care program, California law makers took advantage of the federal option to craft a completely new program which they called “Healthy Families”. The privately administered, publicly funded insurance program would be available to 580,000 children whose parents paid a premium of about $27 per month, costing the state $478 million and the federal government $875 million. [37] The governor proposed a plan that Democratic leaders immediately embraced. While a handful of conservative law makers criticized the plan as too costly for the state and some Democrats worried that the premiums charged would be too expensive for the poor, it passed within weeks with overwhelming bipartisan support. [38]

35The lessons of this California case study are clear, and confirm the strategies that Gingrich and Clinton pursued in Washington DC. Giving states more autonomy through block grants in welfare led California to contract its welfare services and spending, even though the federal law did not strictly mandate such retrenchment. In the Children’s Health Insurance Program, by contrast, the generous matching grants combined with state autonomy was a gift that California’s Republican governor and Democratic legislators eagerly accepted. They designed a new program, putting a Californian imprint on the national idea, but happily took Washington’s money to expand their health coverage of the working poor. Devolving power through different funding structures caused the health care safety net to widen while the welfare safety net shrunk.

III A – State spending patterns since the 1990s

37Just as California’s response to powers devolved from the national government was shaped by the state’s partisan configuration, policy needs, and fiscal capacity, other states took the devolved responsibilities in different directions. Their governments crafted policies that “fit their states”, in the term often used by advocates of devolution, or at least reflected contemporary demands and the preferences of state leaders at the time. To structure their CHIP plans, 13 states chose to expand Medicaid coverage, and 2 offered a program based on the plan offered to state employees, but 35 did what California did: they created their own new systems, verified as “equivalent” to the benchmark but with state-specific details. [39] When it came to reforming their state welfare systems, 36 states adopted the federal default of a five-year lifetime limit on benefits, but 9 states adopted shorter limits and 5 states enacted longer limits. In both policy areas, states picked policies that generally fit the preferences of the parties holding their governorships and state legislatures. States with more “professional” legislatures – those paying full-time salaries, meeting year-round, and employing large staffs – were more likely to move beyond the default options and craft innovative programs. [40]

38At the same time that the devolution revolution changed the shape of state social safety nets, it also changed their scope. State spending on health care and welfare both shifted. Importantly, they went in different directions, a divergence that can be explained by the differing funding formulas chosen for each policy area. In health care, where generous federal matching grants created an incentive for states to spend a bit more of their own money to capture many federal dollars, spending rose sharply. By definition, CHIP spending rose as the new program came into being.

39But Figure 1 shows just how much the Medicaid matching rates, preserved in the 1995 budget battle that fended off the transition to a block grant, gave states both the motivation and the means to spend ever-larger sums on health care for the poor, aged, blind, and disabled.

Figure 1

Growth in federal, state, and total Medicaid spending since “devolution revolution” (in billions)

Growth in federal, state, and total Medicaid spending since “devolution revolution” (in billions)

Notes: Medicaid data taken from Table 3 of the Centers for Medicare and Medicaid Services’ National Health Expenditures Web Tables, accessed at https://www.cms.gov/NationalHealthExpendData/downloads/tables.pdf in November 2011.

40Medicaid spending was already high in the early 1990s, but after the mid-1990s it rose sharply as medical costs soared and as states frequently added recipients and services to their programs. This growth generally slowed when a stronger economy meant that fewer Americans qualified to receive health care coverage, but grew nonetheless. Total state and federal Medicaid spending more than doubled from $161 billion in 1997 to $374 billion in 2009.

41By contrast, the rise in welfare spending was sharply curtailed through the welfare reforms of 1996. As Figure 2 shows, after an initial increase in federal expenditures to provide services such as child care and job training in the immediate aftermath of welfare reform, spending on this safety net program has been flat or decreasing over the past decade.

Figure 2

Growth in federal, state, and total welfare spending since “devolution revolution” (in billions)

Growth in federal, state, and total welfare spending since “devolution revolution” (in billions)

Notes: “Welfare” spending defined as Temporary Aid to Needy Family (TANF) total federal and total state maintenance of effort spending. Welfare data taken from appropriate tables in the US Department of Health and Human Services’ TANF Financial Data Archives, accessed at http://www.acf.hhs.gov/programs/ofs/data/archives.html in November 2011.

42Since the numbers reported for both Medicaid and welfare here are not adjusted for inflation, they indicate a clear retrenchment in this portion of the American safety net. This is a function of two trends. First, the default guidelines passed down from the federal government to the states tightened up welfare eligibility through policies such as weekly work requirements and, most importantly, the lifetime limit on benefits. Although the federal welfare reform did also encourage more spending in areas such as child care, and the federal legislation did not dictate any reduction in monthly benefit checks, many states were able to contain their costs by trimming their welfare roles once recipients had reached their lifetime benefit limits. The second factor at work here was the shift to a block grant funding structure. When states cut welfare after 1997, they could (subject to “maintenance of effort” requirements [41]) keep much of the savings, and this no doubt played a role in the many benefit cuts that have come during tough fiscal times over the past decade.

43Overall, the devolution revolution has been a tale of two programs. Once Newt Gingrich’s proposal to block grant Medicaid was defeated, the program saw no major policy or funding formula changes. As a result, this program continued to grow through state and primarily federal funds, and was augmented by the Children’s Health Insurance Program with its generous matching grants. In welfare, both policy reforms and a shift to block grant funding set the program on a path to retrenchment. Though one could still debate whether the compromise reached by Speaker Gingrich and President Clinton led to “an end to welfare as we have come to know it”, both the state and national governments have been able to curtail the growth of welfare spending. States have been able to shape the program’s new rules to the policy preferences of their elected officials, and as many states meet the bare minimum spending requirements, the federal government has paid an increasingly large role in funding welfare.

III B – Changing “who decides?” changes “who pays?”

44The shares of funding for a program that is paid by the federal government and by the states is important for those concerned with redistribution. One fact of federalism that is seldom discussed but which deserves serious contemplation is that shifting the responsibility for funding programs to the states also shifts the burden of paying for them to less affluent Americans. This is a result of the way that tax structures differ across different levels of government. The federal government relies primarily on a progressive income tax and other payroll taxes to fund its programs. [42] Because money is raised this way, higher income Americans generally pay a larger share of their incomes in federal taxes than middle and low income Americans do. States vary dramatically in the sources of their funding. Legislators, governors, and direct democracy voters have set up fiscal systems that rely in different measures on income taxes, sales taxes, and property taxes. There is no single state tax structure. What is generally true, though, is that state taxes all across the country are less progressive than federal taxes.

45A “progressive” tax system is one in which higher income households pay a higher tax rate. By contrast, the less affluent must pay a larger portion of their incomes in taxes than the rich in a “regressive” system. Analyses of the tax rates paid by Americans in different income groups show that federal taxes are strongly progressive, while nearly every state has, instead, a regressive tax system. Figure 3, below, reports the tax rates paid by different income groups, beginning with the poorest 20% of Americans at the top and continuing through to the richest 1% at the bottom.

Figure 3

Fraction of income paid in federal taxes, by income group

Fraction of income paid in federal taxes, by income group

46Moving down the chart, we see that Americans who make more money pay higher tax rates. The least affluent income groups pay 3.9% and 9.1% of their incomes in federal taxes, because the tax rates on their incomes are so low and because many qualify for the earned income tax credit (EITC), a rebate given to those with low incomes to make up for the fact that Social Security taxes are withheld from their monthly paychecks. As incomes rise, eligibility for the EITC goes away and those in the third group, the middle class, pay 13.9% of their income in taxes. The upper middle class pays 17.3% and the richest Americans pay tax rates of about 20%, with the rate edging up at every income level. Overall, taxes paid to the federal government take much more from the rich than from the poor.

47The pattern in the states is radically different. As Figure 4 shows, state and local [43] tax rates are highest for the least affluent. States rely on three general types of taxes. First, many raise money from personal and corporate income taxes, but state income tax rates are generally less progressive than federal rates and some states, such as Texas and Florida, have no personal income tax at all. Second, most states [44] fill their coffers with sales taxes, which charge a flat rate on the purchase of products. Since poorer Americans spend more of their incomes on these purchases than rich Americans, who are generally able to save more money, sales taxes hit the poor especially hard. Finally, states charge property taxes, paid by homeowners and passed along, by market forces, to renters. These rates – which range from 0.40 in Hawaii to 2.57 in Texas [45] – are flat across property values, and thus generally hit lower income earners harder than the affluent. Although states vary widely in which of these three areas they rely most upon for their tax revenues, the cumulative effect is that states charge higher rates to poorer residents, acting like the folklore character Robin Hood in reverse.

48Figure 4 reports tax rates by income group for Texas (a relatively low-tax, strongly regressive state), California (a high tax, less regressive state) and for the average across all states. In each case, state residents in the lowest income group pay the highest rate: 10.2% in California, 12.2% in Texas, and 10.9% across all states. By contrast, the top 1% of earners only pay 7.4% of their income in state taxes in California, a mere 3.0% in Texas, and 5.2% nationally. Averaging across all states, tax rates go down as income levels rise, just the opposite of the pattern for federal taxes. Why is this the case? In Texas, taxes are so clearly regressive because the state imposes no personal income tax and instead taxes the sales and housing expenses that cost the poor so much of their incomes.

Figure 4

Fraction of income paid in state and local taxes, by income group

Fraction of income paid in state and local taxes, by income group

49California does impose an income tax, which explains why it is less regressive than Texas and the state average. Yet because the Golden State’s famous Proposition 13 dramatically reduced property tax receipts, it relies on a steep sales tax that taxes the poor heavily.

50Overall, federal taxes take the most from the wealthiest, while state tax policies give some of this redistribution back. The implications for federalism are clear. Any shift in program responsibility from the federal to state governments shifts the burden of paying for these programs increasingly to the less affluent. Devolution, in this respect, works against redistribution. By contrast, federal “bailouts” of the states are paid for much more by high-income earners than tax increases enacted at the state level. Recognizing this does not imply a normative judgment about such shifts – there have been strenuous arguments made in favor of both progressive and regressive tax structures. What is hard to debate, however, is that changing who decides on a policy by shifting control of it from one level of government to another also changes who pays for it.