Veto players

One approach, most common in cross-national quantitative work, focuses on impediments to state-level legislation. Whether they use the term or not, the concept explaining the relationship between decentralisation and the welfare state is usually the same: veto players (Tsebelis, Reference Tsebelis2002), although some call it ‘institutions’ or ‘institutional fragmentation’ and roll it into a composite variable with variables such as the presence of bicameral legislatures (Brooks and Manza, Reference Brooks and Manza2007: 43; Huber et al., Reference Huber, Ragin and Stephens1993: 735–7; Huber and Stephens, Reference Huber and Stephens2001: 2; Swank, Reference Swank2002: 35). The argument is simple: federal states have more institutional veto players (groups that are capable of blocking legislation). This empowers opponents of welfare state growth and perhaps all welfare state change.

Studies of major welfare state legislation tend to back up this argument: Canada, the United States and Switzerland all show clear evidence that large numbers of veto players were major obstacles to both small and large expansions of the welfare states (Immergut, Reference Immergut, Steinmo, Thelen and Longstreth1992; Maioni, Reference Maioni1998; Quadagno, Reference Quadagno2005; Tuohy, Reference Tuohy1999). Indirectly, federal states can shape both electoral systems and party politics in ways that fragment politics and slow legislation (for example, Gibson, Reference Gibson2004; Skocpol, Reference Skocpol1995: 23–6).

The problem with this argument is that the number of veto players in a state can obviously vary independently of its territorial organisation. Switzerland's referenda, the major obstacle to welfare state expansion that Immergut identifies, are not a function of its federalism. The states as corporate actors (as against electoral circumscriptions) have not had a formal role in United States federal legislation since the Seventeenth Amendment to the Constitution made the Senate a directly elected body in 1911. The veto points in the US federal government are elsewhere – in the separation of powers within the federal government. Nothing in the vast literature explaining why the US lacks universal health care makes the role of the states a major explanation. Instead, scholars focus on the multiplicity of veto players within the federal government (Hoffman, Reference Hoffman2006).

So rather than simply count federalism and decentralisation as expansions of the number of veto players, it is better to bypass them and ask directly: how many veto players are there for the relevant level of policy-making, and where are they? They might be regions or federal states, but they might be courts or legislatures as well. Regions will only be veto players if they are ‘built into’ the central state: that is, given their own formal representation and veto capacity (Watts, Reference Watts1999). Of course, in many policy areas they control necessary resources or are governed by parties with statewide power, giving them informal veto capacity. If they are not veto players, then they might make government more complex but should not explain why a given central state adopts a particular level of welfare provision. Coding federalism as a simple veto player obscures more than it reveals.

Autonomy

The second approach focuses on the uses that decentralised jurisdictions can and do make of their autonomy. Decentralisation permits greater divergence in standards, with regions able to go above and below the level of provision that a unitary government might provide (Béland and Lecours, Reference Béland and Lecours2008; McEwen, Reference McEwen, Banting and Kymlicka2006; McEwen and Moreno, Reference McEwen and Moreno2005). In other words, decentralisation produces divergence. This argument can be used in both directions: either arguing that parts of a state can have higher standards (which would raise standards for at least those parts, and might start races upward) (Cameron, Reference Cameron1978: 1253; McEwen, Reference McEwen, Banting and Kymlicka2006; Moreno, Reference Moreno and Taylor-Gooby2001), or arguing that the state is the relevant community, and that community is divided by different standards and is less able to redistribute (Alesina and Glaeser, Reference Alesina and Glaeser2004: 78). It is the divergence between Bogdanor and Keating quoted above, and is easy to see in the many arguments from the 1990s that devolution to Scotland and Wales would permit them to have a superior or at least stable welfare state in the face of attacks by Conservative governments with only English majorities (Béland and Lecours, Reference Béland and Lecours2008).

The problem is that the actual ability of regional governments to vary in their levels of provision is itself highly variable. Federalism scholars have a number of terms for the degree to which states and regions constrain and support each other, differentiating between ‘competitive’ and ‘cooperative’ federalism (Watts, Reference Watts and Greer2006), or ‘centralised’ and ‘decentralised’ (Lijphart, Reference Lijphart1984: 176–7) or ‘interstate’ and ‘intrastate’ federalism (Obinger et al., Reference Obinger, Leibfried and Castles2005a). But it can be useful to go a step further down for some more specific variables: laws and money (Banting and Corbett, Reference Banting, Corbett, Banting and Corbett2002; Greer, Reference Greer and Greer2006). First, regions can be legally constrained (Poirier, Reference Poirier2001). Basic laws such as those found in Germany and Spain, for example, require certain standards and even procedures of the regional governments.

Second, they can be financially constrained (Watts, Reference Watts2000). They might not have the money to spend. The most common presentation of this argument comes from the economics of fiscal federalism and uses a line of reasoning that originated with Charles Tiebout (Oates, Reference Oates1999; Tiebout, Reference Tiebout1956). It basically argues that governments compete for business by offering a product (their services) at a price (their taxation). Its practical consequence is that decentralised fiscal arrangements can impede egalitarian welfare policies. That is because governments will be forced to compete for the more footloose taxpayers, and governments with large burdens can find themselves in a vicious circle of service cuts, tax increases and capital flight. Miserable cities in the United States such as Detroit or Saint Louis, ringed as they are by attractive and low-tax suburbs, are monuments to this ‘free enterprise in government’ (Wilensky, Reference Wilensky2002: 93).

But ‘Tiebout traps’ are by no means universal, and literature that implies that they are is literature that is too focused on the United States. That is not just because other factors affect the economics of a locale (neither a natural resource nor a city centre is portable). It is fundamentally because the nature and extent of intergovernmental competition and the connection of governments with any particular tax base are contingent on intergovernmental finance systems. The financial system determines the dependence of any one government on any tax base. The most common mechanism, present in every rich decentralised country but the United States, is territorial equalisation, with a formula allocating funds to governments in order to enable them to provide more equal services (the United States operates some forms of grants that have a similar effect at times). German inter-regional equalisation, underpinned by a constitutional provision, substantially reduces the consequences of fiscal migration by taxpayers from Hamburg and Bremen to Lower Saxony, avoiding zero-sum competition between the city-states and the surrounding federal state.

Of course, those intergovernmental financial systems can also impose new obligations. So financial limitations, or legal obligations born of financial limitations, can exist because of ‘earmarked’ or ‘ring-fenced’ funding that central states require regions spend on certain tasks, or they can exist because regions lack tax-raising capacity and cannot set part or all of their budgets. The sources of money and the legal constraints that come with them constrain and shape welfare policies (Fargion, Reference Fargion2000). For example, Canadian provinces that violate the Canada Health Act lose federal monetary support for their health systems (Maioni, Reference Maioni, Bakvis and Skogstad2002) and US states that do not fulfil co-financing or policy obligations can lose federal funds. In short, legal and financial constraints shape how far regions can diverge, including the extent to which they are exposed to competition.

Summary

Decentralisation is too often used as if it were a single variable. It is not. Instead, it is at best a basket of variables, and at worst a red herring. But if we read closely the arguments explaining its correlation with weaker welfare states, we find that they specify very clear mechanisms that would explain the relationship between decentralisation and welfare spending. So the real lesson of these literatures should be to go beyond simple ‘decentralisation’ or ‘federalism’ and instead look to the mechanisms that they identify:

• • the allocation and number of veto players in the central state, which is relevant to territorial politics principally if regional leaders are veto players in central state politics;

• • the intergovernmental relations that govern the legal autonomy of the regions;

• • and the financial system that constrains regional spending.

The following sections provide two examples of how these three variables illuminate the relationship between territorial politics and the welfare state.

The United Kingdom

The United Kingdom is a Westminster state and EU member with a highly asymmetric system. Northern Ireland, Scotland and Wales, which have a total of about 15 per cent of the population, each have their own elected governments with responsibility for a substantial range of social policy issues, including health, education, local government and social care. The elected governments came into being with 1998 legislation, but the three systems had been administratively and politically distinct long before then (Mitchell, Reference Mitchell2006).

Institutional arrangements

The central government, the model Westminster system, has very few veto players. A Westminster government with a majority is very powerful as governments go, with only a generally deferential judiciary and the prospect of a challenge under EU or European human rights law brooking its power. This Westminster government makes social policy for England. The devolved governments have more veto players, which is partly because of strong committee systems, but is principally due to electoral systems that virtually guarantee coalitions or minority government in Scotland and Wales. Explicit rules about formation of the Northern Ireland Executive demand proportional distribution of government seats. Nevertheless, Westminster, the Scottish Parliament, and the National Assembly for Wales (now) all have clear governing and opposition parties and no different branches to check the governing parties. The three devolved administrations are in no sense ‘built into’ Westminster; they have no formal representation there and no veto power over it.

The legal regime of devolution is, depending on the angle of vision, either very permissive or very constraining. The Northern Ireland, Scottish and Welsh arrangements are not the same; each has its own complex legal framework. The gradations in their powers reflect the differing strength of pro-devolutionist political mobilisation in Scotland and Wales, and the complexities of the peace process in Northern Ireland (this discussion draws on the work of the Constitution Unit at University College London and especially: Hazell and Rawlings, Reference Hazell and Rawlings2005; Trench, Reference Trench2007). The Scottish Parliament's powers are set out in a ‘negative list’: the Scotland Act 1998 lists the powers held by Westminster and leaves the rest to Scotland. The Northern Ireland Assembly has a similar arrangement, but is bound into complex relationships with the Republic of Ireland and UK that constrain its politicians’ freedom relative to Scotland (Carmichael et al., Reference Carmichael, Knox and Osborne2007). The most dramatic constraint on Northern Ireland, of course, is that devolution was often suspended after 1998. As a part of the peace process, devolution was vulnerable to its problems. During times of suspension, the UK government governed Northern Ireland. Wales has a less stable, more cumbersome, and more constraining, regime in which the National Assembly can make secondary legislation and can receive additional powers subject to Westminster's agreement (Trench, Reference Trench2006). In most areas of social policy, the lack of primary legislative powers for Wales has not had a major impact; much UK health, education and social policy works through secondary legislation anyway, and to date the (Labour) UK government has agreed to requests for primary legislation from the (Labour or Labour-led) Welsh governments.

The financial system is based on a formula-allocated block grant, set by a ‘Barnett formula’ that allocates new money to the devolved administrations on an unadjusted per-capita basis, so if another pound is spent in England, the Scots get about 10p added to their block grant. The formula should thereby (very) slowly erode the historically higher per-capita spending in Northern Ireland, Scotland and Wales (Aldridge, Reference Aldridge and Trench2008; Heald and McLeod, Reference Heald, McLeod, Adams and Robinson2002; McLean et al., Reference McLean, Lodge, Schmuecker and Greer2009). They are then free to spend it as they like; Scotland could, within UK law, abolish its NHS and spend the money on roads or culture. The Barnett formula has no legal status and is largely supported by the political difficulty of changing it (although it was changed at the margins in order to keep devolved administrations’ spending decisions from violating the UK government's rules about permissible levels of borrowing for running costs). Northern Ireland also has its own social security system, but its benefits are always set at UK levels, and the UK covers any resulting fiscal gap.

This combination of block finance and extensive scope for policy change produced a ‘fragile divergence machine’ (Greer, Reference Greer and Trench2007): a combination of party systems that spur policy divergence, with institutions that accept divergence but that are vulnerable to breakdown (especially in the case of the very vulnerable financial system). The UK has three elected governments with few serious institutional veto players within their territorial patches and competencies, and one with more internal veto players (Northern Ireland). The financial regime constrains overall spending but not specific budgets, and the legal regime permits considerable variation, especially for Scotland. What is unclear is how stable that system truly is.

Welfare state

The system permits some changes to welfare benefits in the devolved administrations. Probably the largest differences are in higher education (Trench, Reference Trench2008). No devolved administration followed England in introducing variable tuition fees for university students; higher education remains cheaper at devolved institutions. In health, the ‘basket’ of services of any NHS system is both expansive and generally ill-defined. The differences have consequently been fleeting – as with some medicines – or confined to two areas: long-term personal care for the elderly, which is universal in Scotland and means-tested elsewhere (Simeon, Reference Simeon and Hazell2003), and prescription charges.

These are not very dramatic changes, and the reason is clear: as the UK is presently constructed, the devolved administrations do not have the authority to vary the level of welfare provision significantly. Welfare (especially health) spending has increased across the UK, but that is because of taxation and budgeting decisions made by the UK government for England with ‘Barnett consequentials’ that increased devolved budgets. Blair government budgets created the fiscal space for Scotland and Wales to create their new entitlements.

This limited divergence in overall levels of provision comes with major changes in the organisation of social services, which is what we should expect of governments that do not control taxation, overall spending or benefits, but do have the power to organise and reorganise social services of many kinds. The UK government, which makes policy for England, is obviously less institutionally or financially constrained. We see that in the vast expansion of the English NHS budget under the Blair governments. The UK government has not, however, opted to create major new entitlements; its emphasis in social policy was on reorganisation (as with its cross-policy interest in ‘choice’) and, in some areas, tightening eligibility with welfare-to-work initiatives (see Powell, Reference Powell2008). The UK government had more options than the devolved administrations, but chose to use them to reorganise eligibility and provision. Its activity in doing so means that in many policy areas the divergence has been England from the rest of the UK.

These reorganisations plausibly have effects on the level of provision. There certainly have been major and documented differences in the rhetoric and organisation of public services between the four jurisdictions (Adams, Reference Adams2005; Adams and Robinson, Reference Adams and Robinson2002; Greer, Reference Greer2004, Reference Greer2009, forthcoming; McGarvey and Cairney, Reference McGarvey and Cairney2008; Stewart, Reference Stewart2004; Trench and Jarman, Reference Trench, Jarman and Trench2007; Williams and Mooney, Reference Williams and Mooney2008; and 2005 special issues of Regional and Federal Studies, issue 15(4) and Critical Social Policy, issue 25(4)), but there has been less research into the effects on services. This is often due to data that are not comparable, and so comparisons are high quality but on narrow topics (Bevan and Hamblin, Reference Bevan and Hamblin2009), very cautious (Alvarez-Rosete et al., Reference Alvarez-Rosete, Bevan, Mays and Dixon2005), or dependent on ignoring major differences in data collection, population characteristics or both (as with comparisons of school tests or waiting for medical treatment).

Summary

In other words, the devolved administrations create very slightly better entitlements to services. There is considerable debate and very few data about which systems have better outcomes, be they in health, education or social policy. It is easy to see how this flows from the UK's specific institutional arrangements. The devolved governments are free to develop new forms of social policy organisation, and are in large part social services providers, but are less able to develop new social citizenship rights because of their constrained fiscal positions. The result is that they altered the organisation of social services substantially, and their entitlements at the margins. If they had to sustain themselves on taxes, or negotiate their subsidies from the UK state on a programme-by-programme basis, the effects might be very different – with their organisation more to Westminster's taste, possibly, and their levels of services tailored to their different taxing capacities (reliance on own revenue without subsidies would be a disaster for Northern Ireland and Wales; Scotland's future fiscal stability is more contestable because it depends on assumptions about oil reserves, offshore drilling rights and oil prices).

The United States

US politics is a Russian doll of veto players: each one, upon examination, turns out to contain another one.Footnote ³ And the United States is known for a relatively fragmented and inegalitarian welfare state. But federalism in the United States does not, directly, explain its distinctive welfare state because federalism does not directly constitute a veto player. The United States has ample veto players, but the states are not always among them. Indeed, the states might welcome initiatives that the federal government cannot take because of its internal veto players.

Institutional arrangements

The United States has three branches of government, each with a range of powers over the others: the powerful federal courts, the Congress (divided into two co-equal houses) and the Presidency. None completely controls the federal government. Bureaucrats are divided into many agencies and constrained by extensive rules of administrative procedure. An agency will often manoeuvre strategically between its executive branch ‘boss’, its Congressional committee and potentially litigious interest groups. Each branch then contains further checks and balances. Politicians, each with their own territorial and personal bases, have a variable level of party unity. It is worthy of notice when a party achieves such a high level of party unity as to be able to legislate with anything like the decisiveness of an ordinary UK government (Hacker and Pierson, Reference Hacker and Pierson2005). It is practically the opposite of Westminster.

This multiplicity of veto players in Washington is at most indirectly due to federalism. The states are no more than legislative constituencies for the Senate. Some states have Senators of opposite parties who can generally cancel each other out. Whether either Senator chooses to represent the interests of the State government is up to that person, and the interests of the State can be hard to identify if the Governor and two (save in unicameral Nebraska) houses of the state legislature are not under the control of a unified state party.

Federal-level veto players are not due to federalism, so we must look to the effects of state autonomy to see how they affect US aggregate outcomes. How autonomous, then, are the states? US intergovernmental relations, built up over more than two centuries, lack the cleanness of the recent UK legislation. While the Tenth Amendment to the Constitution grants residual powers to the states (in the manner of Scotland's ‘negative list’), state power has been broken down and interpenetrated with federal power at least since the New Deal. This has been through roughly three methods: ‘pre-emption’ by Congress, such as expansive interpretation of the Commerce Clause (which permits the federal government to regulate interstate commerce, and which has been expanded to include issues that touch on goods and services that cross state lines) (Zimmerman, Reference Zimmerman2005); expanded standing for citizens to challenge state actions in federal courts; and use of the federal budget to bring states into specific programmes.

This use of the spending power reflects the states’ distinctive, messy, combination of dependence and independence. The United States, uniquely among the rich countries, does not have a system of formal redistribution between governments. States with weak tax capacities or lower taxes simply have less money than states with strong taxing capacities or higher taxes. Individual federal programmes (grants-in-aid) have adjustment formulas built in that give more money to poorer or more troubled states, but that is specific to each programme. There is no overall effort to equalise the fiscal resources of states in the name of social citizenship or equal opportunities. Forty-nine of the states, moreover, have balanced budget requirements that make them procyclical and prone to cut social services during downturns, and many states restrict tax bases – with no income tax in some states and no sales taxes in others. This means that the United States creates a nearly ideal environment for races to the bottom, and states do suffer from migration of both taxpayers and service receivers. Of course, it is still striking how small the effects are – a salutary reminder, given the ease with which we can imagine races to the bottom ending far below even Mississippi (Hansen, Reference Hansen2006; Robertson, Reference Robertson and Parsons2003; Rom, Reference Rom and Harrison2006).

The states’ permanent fiscal constraint, and the obvious attractiveness of outside money, means that the federal government has an enormous ability to use its spending power to shape state policy. Perhaps the exemplary case was its decision in the Federal National Minimum Drinking Age Act 1984 to deny highway matching funds to states that did not raise their drinking age to 21. The states complied. The state university systems were likewise given an enormous boost by the Civil War-era federal land grants, and federal cofinancing explains much of the US transportation infrastructure now. There are also purely federal social programmes without much state involvement, such as Social Security, the Medicare programme of health care for the elderly and the Veterans Administration, which provides social and health services to veterans. They demonstrate that the federal government need not work with states, although most federal programmes that deliver services work with or through states and local governments.

Welfare state effects

The United States has an enormous welfare state (Page and Simmons, Reference Page and Simmons2000). The difference vis-à-vis Europe is not so much in the sheer scale of spending as in its distribution: per-capita public spending on health care, for example, is higher than most of the OECD and is higher than the UK's total per-capita health care spending (OCED, 2007: 5.2). That the United States does not achieve universal health care coverage with that amount of money has been due to the overall structure of a system in which public spending often props up a very profitable private and non-profit sector and, by covering many groups, takes off enough political pressure to allow the system to muddle onwards (Brown, Reference Brown, Morone and Brown2005). What does federalism contribute to this situation, and similarly remarkable American welfare outcomes?

First, there is enormous territorial variation. The striking fact about American educational performance is less its averages than its staggering variance, with Iowa and North Dakota tied with South Korea and Taiwan as the most numerate jurisdictions in the world, Missouri tied with Canada and Scotland, and Mississippi students scoring below those of Jordan (NCES, 1996, indicator 25; the study has not since been repeated). The percentage of the under-65 population without health insurance in the states ranges from 11 per cent in Minnesota to 30.4 per cent in Texas (Commission on a High Performance Health System, 2007: 22). Massachusetts’ 2006 health reform, if fiscally sustainable, might bring its uninsurance rate down to nearly zero. It is reasonable to suggest that much of this territorial variation in outcomes is explained by the autonomy of states to make policy. Their financial constraint in a system that leaves them to self-finance much of their systems is a fact of American federal institutions. But the federal government can influence the extent to which that variation matters. US health insurance statistics only count the population under 65, because over that age Americans are covered by the federal Medicare programme. The scope for state variation is a function of federal social policy decisions.

So does federalism explain the limits of federal social policy? Probably not. The states are not federal veto players. The experience of recent US social policy suggests that they rarely if ever veto social policy (and that is without even considering their lobbies for intergovernmental transfers). To take one Clinton-era policy, the State Childrens’ Health Insurance Program (SCHIP), the federal government agreed to co-finance state programmes to provide health insurance for children of the working poor. The states signed up despite some misgivings about federal commitment (which proved justified when George W. Bush vetoed a bill renewing an expanded SCHIP in 2007). Some states did the minimum required to comply with the federal requirements, but many experimented with ways to include the children's parents or even unrelated uninsured adults, and raised or lobbied for the funds to do it. SCHIP does not just explain why most under-18s in the US have health insurance; it is also a centrepiece of state and federal coverage expansion proposals.

The tendency to use federal resources to change state policy and overall welfare outcomes was not unique to Clinton. George W. Bush came to office with an agenda of educational reform that led to the No Child Left Behind (NCLB) law of 2001. This law reauthorised a number of federal programmes supporting state and local educational policy, but made them contingent on states testing students and bringing schools up to certain standards or closing them. Such was the existing dependence on federal resources that compliance was financially necessary for the states and school systems. The result was to harness a variety of federal educational programmes to standards, albeit to standards set by states. It was a clear use of the federal spending power to make policy as well as an instance of bipartisan policy-making that overcame Washington's many veto players (a style for which Bush would not subsequently be famous) (Peterson, Reference Peterson and Greer2006). The barriers that NCLB had to overcome were not the states; they were in Congress (Rudalevidge, Reference Rudalevidge, Peterson and West2003). In fact, behind NCLB was not so much a federal imposition on states as an alliance of state and federal education bureaucrats who shared an interest in imposing standards on local school districts (Nitta, Reference Nitta2008: 186; see also Manna, Reference Manna2006).

The initial burst of activity following the inauguration of Barack Obama further underlined the point, with hundreds of millions of dollars being routed to states for their own projects (such as infrastructure) or for joint state-federal projects such as the children's health programme SCHIP. Once Washington overcame its internal vetoes with a new administration, unified party control, and a crisis to help it along, the states accepted the help, with only a few governors (in Alaska, Louisiana and South Carolina) even trying to reject any part of the aid.

Summary

Overall, the United States has a very fragmented political system and a much less egalitarian, comprehensive and coherent welfare state than the UK. But this analysis suggests that the reason is nothing as simple as ‘federalism’. The history of major welfare reforms has often been one of problems in Washington, not intergovernmental relations. States historically sign up for, and lobby for, federal welfare programmes (Smith and Evans, Reference Smith, Evans, Deakin, Jones-Finer and Matthews2004 [2002]). Medicare, Social Security and the Veterans’ Administration are all examples of federal-only programmes whose fate was always determined in Washington. Of course, Southern representatives historically opposed many welfare state programmes, and federalism helped preserve Southern distinctiveness (most notably when it allowed states to enslave and later disenfranchise blacks), but it is difficult to argue that the South would not have been a distinctive area in a more unified United States (for a sensitive treatment of the issue, see Bensel, Reference Bensel1990).

SCHIP and NCLB both show that if the federal government surmounts its internal divisions, it can create quite substantial and (at least theoretically) egalitarian programmes. The ordinary run of state politics, without such federal intervention, produces extremely variable welfare regimes that grant different levels of support and carry out policies with different levels of effectiveness (Greer and Jacobson, forthcoming). In other words, the number of veto players in Washington and interstate divergences in policy contribute to the overall shape of the US welfare state, but not in a simple way that allows us to blame ‘federalism’ or draw conclusions about the desirability of decentralisation. Instead, the conclusions that we can draw are primarily about the effects of veto players in the central government and of interregional transfers that would reduce state dependence on their competitive and very different economies.