To know only one country is to know none.
S.M. Lipset
Introduction
Federalism, as compared to what? * Most students of federalism focus on only one country. Occasionally they compare differences between or among federal systems. Rarely do they compare federal systems with other types of governmental structures. Despite this, advocates of federalism make all sorts of claims about federalism. But rarely are their claims tested in any systematic manner. In this article we draw on a small but important body of work that has begun to contrast federal with unitary and other systems in order to determine how well federal systems fare by comparison. We review these findings, contribute some additional analysis of our own, and then use the discussion to clarify distinctions between federalism and decentralization which are often confused in some of this literature. Our findings are far from conclusive, but they are instructive, in large part because they all point in the same direction and to the same conclusion: unitary systems out-perform federal systems on almost all measures of government effectiveness and efficiency, and citizen well-being. Part I compares federal systems with unitary systems on a number of dimensions. Part II examines a variation on the federal/unitary distinction by exploring the claimed benefits of consociational government in contrast to either federal or unitary structures – though it does show that consociational structures are only likely to flourish in unitary systems. Part III examines the managerial concepts of centralization and decentralization and tries to sort out the confusion often associated with decentralization and federalism. Part IV extends the issues set out above by exploring the misunderstandings and misapplication of theories of fiscal federalism to political federalism. We show that fiscal federalism and various allied theories are properly understood as theories of decentralization, not federalism. The Conclusion underscores what the paper has demonstrated: the value of empirical studies that move beyond comparing federal systems with each other, and compare federal systems with other types of systems.
I. Federal systems compared to unitary systems
Two political scientists whose work stands out in this area are John Gerring and Erik Wibbels. Their work is concerned with a number of structural features of governments and is not designed to focus on differences between federal and unitary regimes. Nevertheless, they both compare these two structures in a great many ways, and their findings are of immense value to students of comparative federalism. Individually and with various colleagues, both Gerring and Wibbels have produced an impressive body of work that reveals comparisons between unitary and federal structures[1]. This comparison is a central theme both in their large N study based upon data provided by the International Monetary Fund (IMF) and the World Bank, and their more intensive follow-up studies of individual or small groups of countries.
Still other comparative politics scholars whose work bears directly on our concern of « federalism as compared to what? » are Arend Lijphart[2], John Loughlin[3], and their various colleagues. They too have compared federal and unitary structures. Lijphart focuses on thirty-six democratic countries, and he is preoccupied with identifying the requisites for stable democratic governments. But his data set allows us to compare federal and unitary systems. John Loughlin, along with various colleagues, has undertaken a vast number of detailed case studies on individual states, some with unitary and some with federal systems, which ask the same sorts of questions. Thus his case studies, while not directly comparing federal and unitary systems, nevertheless facilitate close analysis of paired comparisons of federal and unitary systems.
This paper draws on the work of Gerring, Wibbels, Lijphart, and Loughlin. We report on their findings with respect to comparisons of federal and unitary systems, and we also re-analyze some of their data in order to address our concerns more directly. Two strong findings emerge from this work. First, the comparisons consistently show that unitary systems out-perform federal systems on just about every dimension of government performance, public participation, and citizen well-being that is considered. Second, it reveals the limits of such research. We deal with this second concern immediately below.
Structural models of government are not and well specified, and are not likely ever to be so. There are only a small number of countries, and they come in complex bundles and mixtures, so that sorting out various possible factors or combinations of factors, and examining the importance of each separately is a challenge. The effort to isolate and identify the effects of federalism (and unitarism) is challenging because of possible endogenous (and exogenous) factors that cannot easily (if at all) be carefully identified and isolated from the object of concern the political structures. Indeed, there are often more theoretically important variables in the models than there are Ns in the data sets. So conclusions must be understood as modest, and findings must be interpreted with caution. Still, this work, even with its limitations, is a marked improvement from generalizations based on the examinations of single setting, as is the case with virtually all claims made about the consequences of federalism.
With these caveats, the research to date has produced several findings that should inform the federal-unitary debate. First, the comparative studies have generated virtually no findings suggesting that showing that federal systems perform more effectively than unitary systems. The studies examined any number of different indicators of governmental effectiveness, political stability, political participation, and social welfare and found that virtually none of them supported the claim that federal systems are more effective or efficient. Indeed, the evidence runs in the opposite direction: governments with unitary structures consistently out-perform federal systems on almost all dimensions almost all the time. Second, centralization and decentralization vary widely in both federal and unitary states, so that what might appear to be due to a distinctive feature of federalism is better understood as consequence of decentralization. Third and related to the issue above, authority of subnational governments to tax and spend is more closely associated with issues of decentralization and centralization than federal and unitarianism. Fourth, contrasting the literature about the benefits of federalism with these systematic cross-national studies, it appears that the federalism literature tends to cherry-pick decentralized federal states to compare with centralized unitary states, despite the fact that both federal and unitary systems vary widely in terms of centralization and decentralization. That is, scholars conflate distinctive traits of federalism with effects that may actually be more closely associated with decentralization. These issues are explored in the pages below.
This section asks the question, which type of governmental system – unitary systems or federal systems – are most closely associated with effective and efficient government processes, effective public policies, and social welfare indicators among the general population? The inquiry draws together the findings of seven sets of comparative studies that contrast federal and unitary systems. The studies are: 1) a large N study involving over one hundred countries by Gerring e.a.[4]; 2) a similar large N study by Wibbels[5], Both of these studies examine a number of oft-made claims about the advantages of federal over unitary structures; 3) several auxiliary studies by these two authors and various colleagues; 4) relevant portions of the sustained work of Arend Lijphart[6] that contrasts consociational systems with both federal and unitary systems; 5) a more recent and related study by Taylor, Shugart, Lijphart, and Grofman[7] that builds on Lijphart’s early work; 6) several carefully constructed comparative case studies that examine decentralization in unitary systems by John Loughlin and colleagues[8] and 7) our own re-analysis of data provided in the work of Gerring e.a.[9], Wibbels[10], and Taylor e.a.[11].
As we shall see, while each of the studies focuses on somewhat different issues, nevertheless each points in the same general direction, and to the same conclusion: all things equal, unitary systems fare better than federal systems on just about every dimension on which they have been compared. Furthermore, consociational unitary systems and unitary systems with proportional representation (PR) fare much better than any type of federal arrangement. Both large N studies, as well as closely textured case studies of paired comparison states point to this. Of course, for reasons discussed earlier, each of these studies has its limits and constraints, and cannot be taken as « the » definitive assessment. Still, despite limits, the findings in the several studies are consistent and robust.
1. Comparing federal systems with unitary systems: consequences for diverse populations
In « Centripetal Democratic Government: A Theory and Global Inquiry »[12], Gerring and colleagues ask, how can a diverse constituency best be governed – through a unitary system or through a federal system? Pursuing this question in a large N cross national study of both unitary and federal systems, they report that unitary systems (with proportional voting) are far more effective than federal systems at achieving inclusion, and the establishment of a responsive, and accountable bureaucracy. We note that instead of using unitary and federal structures, Gerring substitutes centripetalism and centrifugalism respectively. Centripetal system a unitary system that is almost always accompanied by proportional representation, few veto points, strong political parties, and responsive bureaucracy. In contrast, a centrifugal system in his terminology is a federal system that is characterized by single-member-district plurality (SMD), multiple veto points, diffuse consensus based policy making, and sluggish bureaucracies.
Gerring and his colleagues compared centripetal (unitary) and centrifugal (federal) governments on a variety of policy-relevant factors, including: (1) bureaucratic quality, (2) tax revenue, (3) investment rating, (4) trade openness, (5) gross domestic product (GDP) per capita, (6) infant mortality, (7) life expectancy, and (8) illiteracy[13]. They selected bureaucratic quality and tax revenue because these factors represent variations in the degree of centralized power in a state; and investment rating, trade openness, and GDP per capita because they typically are used as indicators of economic performance. And they explored infant mortality, life expectancy, and illiteracy because they are standard indicators for social well-being.
Drawing on data from over seventy countries, they ran two regressions of these and a variety of other factors, controlling for still other factors such as geography, resource endowment, colonial heritage, « years a democracy ». For each indicator, they ran a regression for a full model, i.e. with all the variables; they then ran regressions for a reduced model, which dropped out variables that did not meet the minimum threshold for statistical significance (p <. 10). Their most salient finding was that centripetal (unitary) governments consistently out-perform centrifugal governments on almost all governmental effectiveness and quality of life indicators[14].
Gerring and colleagues acknowledge the limits of their analysis. They did not draw a random sample of states, and thus the data are subject to endogenous factors that the model may not have accounted for. Still, given these limits, they are confident in findings, since parts reinforced each other. Overall states with centripetal (unitary) structures were consistently found to perform better than centrifugal (federal) states on almost all standard government-effectiveness and social well-being indicators they examined. Had their findings been weak or mixed, they might have hedged their bets. But almost all of the many indicators pointed in the same direction. Of course, they do not offer a deterministic model and indeed emphasize that endogenous factors can and are important in ways not captured in their analysis. And so there may be exceptions that confound the pattern. Still, what they found is a strong general tendency, strong enough to lead the modest conclusion that it is not obvious that federalism is more effective in promoting many of the effectiveness claims that are so often made for it. Their findings are reinforced by any number of small N and case studies which compare welfare benefits in various types of countries[15].
2. Comparing federal systems with unitary systems: the promotion of peace, political stability, liberal policies, and protecting minorities among a heterogeneous population
Others have expanded the comparison of the relative benefits of federal and unitary systems. Employing a multi-method, multi-technique research strategy, Wibbels[16] and colleagues[17] employ both large N, cross-national data analysis and detailed comparative case studies, address claims that federal systems are better able than unitary systems to maintain peace, foster political stability, guarantee liberal rights, and enhance protections of minorities in polities with heterogeneous populations. In this series of papers from 2005-2010, Wibbels and colleagues examined a large N-study assessing the importance of various structural factors of governments (most notably unitary and federal structures), and then explored the federal/unitary distinction as it affected a number of more focused issues, such as the proposed federalization of Iraq; regional inequalities, conditions on ethnic groups, the resource « curse » in U.S. states, and business activity in several federal states. Their main metric for identifying outcomes was the distribution of tangible resources at the disposal of the government. Their main findings are that ethnic compartmentalization, regional economic differences, and urbanization as powerful predictors of state instability, and that federal systems are less able than unitary systems to mitigate the problems associated with these conditions. In sum, their conclusion is that federal systems are less able than unitary systems to distribute resources at their disposal proportionately across the national population, and that this in turn undermines their ability to foster stability and protect basic rights.
The central weakness of federal states, this body of work suggests, is that federal states over-represent voting minority interests by using single-member district electoral rules, and in doing so over-represent rural regions in the national legislature. This in turn contributes to civil unrest and political instability. Federalism also creates incentives for groups to advocate for redistribution from the central government to the regions. This further skews resource distributions, and as well fosters conflict and instability that undermines federalism’s central premise of inclusiveness.
Wibbels follows this up this large N cross national data analysis with detailed case studies of three federal systems (the United States, Argentina, and India), contrasting his findings with the large N cross-national study with more fine grained analysis of these three sites. Here he finds that the nature and structure of constitutions are influenced primarily by factor endowments, and regional wealth distribution. « Factor endowments » refer to the built-in economic resources such as population size, natural resources, climate, and the like. Regional wealth distribution refers to the general level of wealth inequality between regional populations. He finds that federal systems are associated with settings in which there is wide variation (wider on balance than unitary systems) in regional factor endowments, such as wealth distribution across regions prior to federalization. Adopting a federal structure, he continues, leads to a national government that represents regions rather than people, and as a consequence establishes a built-in and permanent arrangement that over-represents interests of rural and less well-endowed units. Furthermore, constitutional constraints on the national government inhibit national democratic majorities from enacting their agenda, and enabling rural elites disproportionate influence in policymaking. This in turn can lead to political crisis and instability.
This poses the question: why are federal systems established in the first place? He suggests that the inequality is built in at the outset; if there are inequalities among regions, the elites will favor federalism because it weakens majority rule in the central government. His analysis in both his large N studies and his detailed cases studies reinforce each other. Both point to the same conclusion: central governments with federal systems tend to have poorer outcomes (on a host of social policy scales) than do unitary states[18]. Put another way, weak central governments in federal systems are less effective at implementing policies, and more likely to generate opportunities for waste, corruption, and inequitable redistribution of resources than central governments in unitary regimes. These features, he emphasizes, undermine claims about federalism is a means of empowering the citizens and promoting accountability.
In sum, both Gerring and Wibbels emphasize that central governments in unitary systems provide more effective governance and administration, better social welfare benefits, and support a better quality of life as measured by various indicators than do a mixture of national and state governments in federal systems. These findings hold when the data are examined in a variety of different ways and with various controls. They are also reinforced by his case studies that look more intensely at selected issues. While robust enough, these findings nevertheless are subject to caveats standard in this type of work. There are a host of endogenous factors that complicate analysis and cannot adequately be controlled for. For instance, does a distinct political culture lead to federalism, or does federalism lead to a distinct political culture?
Of course, for reasons set out earlier, neither of them offers a definitive theory of the strengths and weaknesses of unitary and federal structures. At best they offer statements about patterns and tendencies. Still, their findings suggest that the pattern is strong, general, and enduring. In sum, their findings strongly suggest that unitary systems are more effective than federal systems, at least as measured by a variety of standard quality of government and quality of life indicators.
II. Contrasting unitary and federal systems with consociational arrangements and with each other
In considering the relative advantages of various structural arrangements, not all comparativists have focused exclusively on the federal/unitary distinction. Arend Lihphart has spent his career challenging the assertion put forward by political scientist Robert Dahl, that democracy is possible only in small homogeneous societies. This led Lijphart to examine the possibility of stable effective government in societies with heterogeneous populations. More particularly, he has sought to understand why his own country, the Netherlands, is democratic, stable, and prosperous despite having a heterogeneous population. His work has consistently shown that neither unitary nor federal systems by themselves are especially effective in representing the interests of diverse citizens in a heterogeneous society. However, his analysis of the Netherlands and a few other countries had led him to conclude that consociational arrangements are highly effective in terms of promoting stability, public services, citizen well-being, and the like[19]. Distinctive features of consociational arrangements include a parliamentary system (which almost invariably means a unitary rather than federal system); proportional representation; organized, disciplined, and strongly ideological political parties; and government by grand coalition rather than minimum winning coalitions. Although the dominant party or coalition has disproportionate power, Lijphart argues that grand coalitions, even in the face of outright majority victory, foster stability, and assure that winners do not neglect losers. Lijphart maintains that this arrangement contributes to political stability, to the development of effective and efficient public services, and to economic prosperity, in at least the Netherlands and Belgium and elsewhere where it has been used during the Twentieth Century. But it too, he warns, has its limits. Consociation is a social arrangement, and not a formal constitutional structure. It works only if citizens are organized into political parties and a social consensus encourages them to participate in grand coalitions, and fails if people are not well organized and represented.
A number of commentators have linked Lijphart’s consociationalism with the classical Medieval pillar system in which the different center of social power – church, military, great land owners, merchants – sat at the king’s table and had a role in constructing policy. Historically this arrangement is said to have worked well, and if one discounts unorganized interests (e.g. peasants, workers), perhaps it did. But during periods of change, the arrangement is problematic, for instance when peasants suddenly demand land, labor organizes and presses for collective bargaining, or an influx of immigrants reshapes the political landscape.
Others have offered the idea of « functional federalism » as a variation on Lijphart’s consociationalism. In « functional » federalism, « interests » rather than territories are represented. Thus it may be that members of Congress from Kansas represent not only Kansas but agricultural interests more generally, or that members of Congress from New Jersey’s not only represent that state and its pharmaceutical industry but pharmaceutical interests wherever they might be. But Lijphart’s claim is more restricted; he maintains that a grand coalition in well-organized and highly segmented societies such as the Netherlands once was (there is dispute about the applicability of his theory to post-immigration Netherlands) can provide a great many advantages. Thus he uses his work to refute the claim by Robert Dahl that democracy requires a fairly high degree of social homogeneity. His findings provide the exception (one of several, it turns out) that challenge the rule laid down by Dahl and a host of democratic theorists following in his wake. As important as the idea of consociationalism is, it should be noted that Lijphart’s theory rests upon the foundation of a unitary system with proportional representation in its parliament.
As Lijphart’s project expanded, he and colleagues began to explore the politics of accommodation in democratic countries more generally, and as well to explore other correlates of effective democratic governance[20]. In a series of studies examining the structural feature of thirty-six democratic countries[21], he sought to determine what structural factors and combinations of factors are most likely to foster inclusiveness, political participation, political stability and social and economic well-being. In one extended study, Lijphart and colleagues[22] analyzed data from the thirty-six liberal democratic countries in order to locate the United States with respect to a number of indicators of public policy and social well-being. They found that United States does not stand out among the other democracies on any of a several measures of political effectiveness and efficiency, and economic and social well-being of its citizens. Indeed, on almost all of their measures the United States ranked in the middle or the lower middle on the continua.
To illustrate, Taylor et. al compare the U.S. to other states in terms of several aspects of constitutional design like: division of powers, electoral systems, political parties, legislative design, and judicial design. They note that many features of U.S. constitutional design create bad incentives for good public policy formation. For instance, they note that the U.S. division of powers allow U.S. states an uncommon level of discretion in implementing federal policy. Furthermore, the two-party system diminishes voters’ choice on particular issues, and a policy-oriented judiciary encourages legislatures to avoid settling difficult political questions. In sum, they point to these host of features that make enacting effective social policy considerably more difficult than is the case in other advanced democracies.
A natural extension of this work was not only to compare the United States with the other thirty-five democracies, but to divide these states into unitary and federal systems and to compare the two types with each other. Surprisingly, Lijphart had never done this. So, drawing on the cumulative concerns about government effectiveness and citizen well-being that Gerring and Wibbels examined, we divided Lijphart’s thirty-six countries into unitary and federal structures to more fully explore the hypothesis that federal systems are more effective on both dimensions. The data set provides an especially good source with which to address these questions. All thirty-six of the countries are liberal, stable and well-functioning democracies. None of them is an oligarchic, kleptocracy, dysfunctional, or authoritarian state where formal structure might be made irrelevant by one or more of these factors. That is the inquiry is restricted to well-functioning, prosperous democracies, not failed, failing, or even weak states. We asked: which system – federal or unitary – is more effective in providing responsive and effective government, and in fostering policies designed to enhance social well-being? Using Lijphart’s data, and Gerring and Wibbel’s criteria for federal and unitary states, we contrasted the two types of systems. We then correlated various measures of government effectiveness and social well-being with the type of structure, unitary or federal. Figure 1, below, presents the results of the investigation.
Figure 1. Linear Model Predictor for Federal/Unitary States[23]
In terms of variables, we incorporated the most important factors Gerring and Wibbels included in their models. Specifically, we selected variables that captured both quality of governance and quality of social policy outcomes. We select on these variables because they are central to most arguments about the relative merits of federalism, and especially American federalism. For instance, with regards to quality of governance, we include « democratic quality», « proportion of public debt » and « tax percentage » in our model. These indicators give metrics on how efficient a state’s bureaucracy is, and how responsive the government is to its constituents. These are important because federalism scholars argue that federalism fosters good governance and the benefits this provides.
Social policy indicators, we included are « health care expenses », « military expenses », « life expectancy », and « Gini coefficient ». These are standard variables in studies of governmental effectiveness and social welfare. Per capita expenditure a state spends on healthcare is a good proxy for social welfare expenditures more generally. Conversely, military expenses can be seen as prioritizing guns over butter. Life expectancy is another useful omnibus measure of « quality of life ». And the Gini coefficient is now the standard measure for macro-level income inequality. Overall, these several variables provide a set of rough but useful measures of governmental social policy effectiveness.
Unfortunately, we cannot not specify the model in more detail owing to the sparseness of data. Furthermore, the model might reasonably be discounted because it uses a relatively small number of states to look for general relationships. Still, if we included other variables in the model, we would be trading off breadth for depth, and lose the ability to discover potentially interesting relationships. At any rate, our concern here is to see if federal systems enjoy the advantages so many claim for them, rather than provide a theory for governmental effectiveness. Whatever the case, there is no evidence federal systems are effective. Our work here, although fraught with the same statistical challenges that Gerring and Wibbel’s faced, shows no evidence to support the claim.
III. Centralization and decentralization in strong unitary systems
A confounding factor in a lot all discussions of federalism is decentralization. Many scholars (see the discussion of fiscal federalism below) assert that federalism and decentralization are synonymous, and many constitutional scholars see no need to distinguish the two concepts[24], though traditional federalism scholars following Riker[25] and Elazar[26] insist upon the distinction. But even federalism scholars have been slow to appreciate vast variations in decentralization within unitary systems. One of the few scholars who has explored the variety of decentralization in unitary systems both cross-nationally and over time is John Loughlin[27] and with colleagues[28]. He has been examining national, regional, and local taxing authority and expenditures within European Union (EU) countries since the late 1980s, and in so doing has traced dramatic shifts in both taxing and spending powers of local and regional governments vis à vis national governments in both federal and unitary states. They have found dramatic shifts—from centralization to decentralization, and vice versa, in both taxing and spending authority in both federal and unitary systems, and have examined the impact the European Union (EU) has had on national, regional, and local governmental authority. Although Loughlin (and colleagues) did not set out specifically to address federal/unitary differences that are the focus of this article, their many careful articles are helpful in revealing just how independent the concept « decentralization » is from the concept « federalism ». Their work reveals that unitary systems can vary dramatically in the degree to which they are centralized or decentralized. Their work shows that local and regional governments in both federal systems and unitary systems vary widely with regard to both taxing and spending powers. And it shows that these powers change dramatically over time – in both federal and unitary systems.
These conclusions are illustrated in two of their detailed cases studies on Sweden and France. In a 2005 report on Swedish tax policy, Loughlin e.a. traced changes in local income tax policy in Sweden from the early 1980s and unto the late 1990s. Local authorities in Sweden had long had considerable discretion to tax and spend, but in 1991, Parliament expanded the policy still further when it adopted the « Local Income Tax Policy Act ». The new law significantly expanded local authority both to tax and spend, and was justified, the authors report, as an « effort to strike a balance between the central government’s concerns with high standards, and the efficiency of devolved administrative responsibilities to local subdivisions[29] ». The Act reinforced this policy by vesting local governments with still more primary responsibilities for operating a number of public programs, including primary, secondary, vocational education, and health care[30]. However, almost all local locally raised taxes went to programs which were nationally mandated and had to meet strict and uniform national standards. The result is in effect centralized policy making, and decentralized administration.
The 1991 Act represented the outer limits of Swedish decentralization. A few years later in the mid-1990s, fiscal fortunes let to a new alignment. As the economy declined, it became more painful for Parliament to raise enough funds to effect income transfers to and enhanced oversight pm the poorer regions of the country, communities where local authorities could not meet national standards, and areas per capita service delivery is more expensive (e.g. rural areas). The result was a shift away from local authority and more national authority. Loughlin notes that this a « pendulum effect » operates in Sweden[31]. It shifts with the robustness of the national economy. The mid 1990s changes represented a « swing » of the pendulum back towards greater centralization, though Loughlin and his colleagues note that economic recovery will probably mean be a return to greater decentralization. The lesson: in good times the central government shrinks, and in bad times it expands[32]. But through all these changes, the central government continues to set the ceiling for tax rates, to identify the programs to be supported, and to establish standards for these programs. Although local authorities have considerable discretion to manage and innovate within these constraints, their authority is constrained. This pattern is found in a number of other unitary EU countries. So, while Loughlin and colleagues identified a distinct trend towards greater decentralization of taxing and spending authority, they find that arises from national parliamentary authority not constitutionally entrenched local or regional autonomy[33].
Loughlin finds a similar pattern in France. Indeed, France he maintains, has long been even more decentralized than Sweden. Decentralization is more deeply entrenched, both by law and by tradition. According to some[34], despite its reputation for being militantly hierarchical, France may be the West’s most decentralized country. France has 37,000 communes (municipalities), another 100 « departments », twenty-six « regions » and 2700 (and rapidly) growing special districts (the numbers keep changing). All have both taxing and spending authority[35]. In contrast, according to a 2015 Census Bureau report, the United States has 30,000 incorporated cities and towns, plus another 3000 plus counties, fifty states, and thousands of special districts. One commentator has even used the term « fragmentation » to describe France’s governing structure[36].
Almost all communes (municipalities) in France predated the formation of the modern state and continue to hold pride of place in French political tradition. While some are tiny and perform little more than symbolic functions, others are sizeable cities or conglomerates of urban areas, exercise considerable power, and possess substantial political clout on the national scene. The other important subnational units in France are « departments », geographical units that do not conform to any « natural » boundaries, and were created under Napoleonic rule, both to weaken traditional forms of local authority, and to rationalize national governance. France’s twenty-six regions are an even newer creation. They have been established in part as a consequence of the EU’s policy of a « Europe of Regions » which some see as an attempt by EU officials to expand transnational power at the expense of national authority. Whatever the reasons, it has this effect. Increasingly, regions frustrated by policies in their nation’s capitals turn to Brussels and the European courts for support. This is most visible with Scotland, Catalonia, and a few other places, but rumblings are heard throughout Europe. Still, it is too early to detect any long-term international consequences from this effort (Jolly).
But clearly regionalization has created a new layer of decentralization in several countries, including France. The French state now transfers an ever increasing portion of nationally generated taxes to the new regional councils, and the rapidly growing number of special districts[37], which have quickly become competitors to both long-established communes and departments. In turn, each of these units has taken steps since 2003) to expand authority to levy taxes. Still, as in Sweden, policy continues to be set by the national government[38].
In looking for a general trend over the past several decades, Loughlin and colleagues[39] detect what might be appear to be federal-like developments in countries with unitary structures. Some national governments (the principals), like France, operate on an agency model; they transfer funds to local and regional governments (their agents), who then spend funds for programs established by national directives. Other national governments with choice models take in less revenue and create fewer programs, allowing sub-national units to raise a greater share of revenues, create their own programs, and then implement them locally. Despite these differences, Loughlin and colleagues report that most of the new decentralized programs rest on agency and not principal models. Furthermore, in many European countries where local authorities are granted discretion, they are nevertheless held accountable to a web of national standards.
Some might call these developments soft-federalism. Certainly, local and regional units with taxing and spending authority are likely to develop powerful constituencies and quickly develop an aura of permanence. But, they are not constitutionally entrenched, and as Loughlin’s research on Sweden demonstrates, what central governments bestow they can also take away. Perhaps the greater lesson is that some unitary governments have discovered the value of decentralization « all the way down », at least when it is convenient, and all the way up when it is not.
Intrigued by the variation in taxing and spending powers Loughlin and colleagues found in their case studies, we drew on data from the International Monetary Fund’s (IMF) decentralization indicators (the same dataset used by other political scientists working on federalism) to systematically probe for an association between federalism and decentralization, and unitarianism and centralization with respect to taxing and spending powers. Focusing exclusively on data on taxing and spending that is collected by the IMF, we break down government activities into three types, « General Government Revenues/Expenditures », « Central Government Revenues/Expenditures » and « Budgetary Government Revenues/Expenditures ». The relevant measures for this study are General Government (GG) and Central Government (CG). GG refers to the overall level of revenue or spending undertaken by a government, and CG refers to the portion of spending accrued by the central government. Our hypothesis is that federal states are operationally different from unitary states with respect to the degree of decentralization – unitary states on balance are more centralized that federal states.
We can calculate decentralization; by subtracting CG from GG for each state, then dividing that value over the GG, we get the percentage of revenue or spending undertaken by the central government and sub-national units respectively. This is obviously an imperfect measure because it probably overestimates the relative size of the central governments. Again, because the IMF does not distinguish central government spending that is truly national from central government spending in the form of grants, it becomes difficult to do rigorous cross-country analysis without overlooking important nuances. That being said, this measure is at least as good as the accepted ones in the current literature, so we can proceed with analysis. The results of our analysis are presented in Table 1.
Table 1. Ratios of Sub-Unit Spending/Revenue to Central Government Spending/Revenue in Selected Federal and Unitary States[40]
The chart above represents a small N analysis of a few select states that provide a useful baseline from which to analyze the question of « are federal and unitary states substantially different from each other, with regards to economic decentralization ». These states were chosen primarily because they provide the best examples of careful data collection in economies that are also large enough to make meaningful comparisons with one another. Other important states such as Brazil, India, and China were excluded because of lack of IMF data on either the revenue or the spending.
The results, or rather the non-results, send mixed signals – and therefore undermine any grand theory that links federalism to decentralization. In a nutshell the figures reveal that there is as much variation with respect to centralization/decentralization (as measured by taxing and spending) within both federal and unitary systems as there is between the two systems. For instance, both France (unitary) and Germany (federal) have sub-unit spending that far outsizes sub-unit revenue – central governments raise much of the revenue and then reallocate it to sub-national units. There is also no strong relationship between the ratio of sub-unit revenue generation and spending and the country’s classification as either federal or unitary. To illustrate, Italy’s (federal) sub-unit/central government ratios are about .22, and South Korea’s (unitary) ratios are about .24. While obviously two handpicked observations cannot be used to extrapolate to the entire dataset, they do illustrate the pattern.
The discussion about centralization and decentralization here does not provide any concrete answers about the conditions under which countries centralize or decentralize. But the discussion does underscore that both types of systems can vary widely in the degree to which they are decentralized or centralized in any of a number of areas, and that this variation is independent of the unitary/federal difference. The discussion of both Loughlin’s case studies and the systematic comparison of several federal and unitary systems should be considered in light of an important difference between the two systems. « Federalism grants subunits of government a final say in certain areas of governance[.] », while « decentralization […] is a managerial strategy by which a centralized regime can achieve the results it desires in a more effective manner[41] ». Although « federalism generally results in a fairly high level of decentralization, decentralization does not necessarily lead to federalism[42] ».
IV. Economic Theory, Credible Commitments, and the Confusion between Decentralization, and Federalism
Since World War II public finance and political economy scholarship has transformed the study of public institutions. The theory of public goods not only revolutionized public finance, it has transformed many political science and public law. This development has not only demonstrated that institutions matter, it has generated powerful theories to show how they act. This work as had enormous influence on students of federalism and constitutional law. As powerful and as helpful economic theory has been in understanding governmental institutions, it has also led to considerable misunderstanding, and perhaps nowhere as great as in the analysis of federalism. In developing their theories of public finance, economists have used the term federalism interchangeably with decentralization. In terms of the power and reach of their « federal » economic theory, the distinction is unimportant. But for political scientists and constitutional scholars, the distinction is crucial. For economists, it may make little difference is a configuration of authority flows from federalism or decentralization. But for political and constitutional analysis, it can make all the difference. We explore our concerns below.
Economic theory models institutional designs that aim at specific or well-defined purposes: If you want to achieve certain objectives, here’s how you should pursue them. If you want to optimize the production of a public good, create a special distinct for taxing and spending that eliminates free riders. If you want to foster the development of markets, restrain government taxing and spending so that investments can flourish. If you want to foster innovation in public policy, create incentives for governmental experimentation. Economists have developed powerful models, with both normative and predictive powers.
To take one example, theories of « fiscal federalism » have been enormously successful in specifying institutional design and predicting the optimal production of public. The theory has revolutionized public finance and helped establish the field of public choice. We do not challenge the power of this theory, but we do want to emphasize that it is not really a theory of federalism. Rather, it is a theory of decentralization. Consider. We want to reduce pollution and smog, and improve air quality. Fiscal federalism informs how we should go about doing so: we should establish a pollution abatement district whose boundaries are defined by the catchment area as by hills and valleys, prevailing winds, and population, industry, and the like. As population increases and technologies improve and problems take on new shapes, the boundaries can be adjusted accordingly. The organizing principle is to align costs with benefits. Indeed, to the extent that fiscal federalism scholarship is concerned with traditional federal units – states and provinces, as opposed to functionally established special districts – it is likely to emphasize the suboptimal production of public goods, the failure to align costs and benefits, and the never-ending struggle to shift costs to other units[43]. It is true that some political economists try to square federalism with the production of public goods[44], but even they tend to dwell on the inefficiencies of federal systems, and favor the establishment of special districts expressly created to pursue specific objectives – that is, decentralization.
Of course, it may be perfectly fine for economists to call such an arrangement federalism – or that matter anything else. The theoretical model is equally powerful or weak whatever it is called. But for constitutional lawyers and political scientist’s truth in packaging is important since in these fields, the distinction between federalism and decentralization is crucial. After suffering much confusion and criticism from political science colleagues, Wallace Oates[45], the dean of the School of Fiscal Federalism, has acknowledged that the term « fiscal federalism » is confusing, that the theory is in fact about specialization and decentralization, and that if he had it to do over again, the economists would have selected another term.
But the problem is not primarily for the economists. Political and constitutional theorists have adopted the language of economic theory wholesale, and economists’ claims about « fiscal federalism » are among the most oft-used arguments for constitutional federalism as well. The reader is invited to pick up standard defenses of federalism in the United States, and peruse their notes and references. He will find copious references to economists writing about fiscal federalism and other allied areas of public finance[46]. And a perusal of the text will find transplantation with few if any adjustments. Economic theory exerts a siren call that has led to careless embraces. Some have associated the special purpose districts fostered by fiscal federalism the principle of subsidiarity, which they see as a basic rationale for federalism. But this too is wrong. As David Golembowski has shown[47], subsidiarity is and has always been a central, and a decidedly un-federalist principle. It privileges central authority, which grants or withdraws authority to subunits as circumstances and conditions permit. It is a classic decentralization.
What is obvious for fiscal federalism holds for the use of the term, federalism, in other, related areas of political economy. For instance, Douglass North and Barry Weingast have developed a powerful account of « market-enhancing federalism » which is both elegant and compelling. It asks, what can governments do to foster competitive markets and economic development? Weingast has provided a convincing account of « market enhancing » federalism in England in two different eras, first following the Glorious Revolution in the late Seventeenth Century, and again in the early Nineteenth Century[48]. Market enhancing events included the victory of parliament over the king, who was wont to extract money from the noblemen to pay for wars they did not support, and a life-style of which they did not approve; the creation of a robust separation of power; expansion of the franchise; institutionalizing the principle of parliamentary sovereignty and judicial independence: and the embrace of the rule of law. The consequences were enormous: a more fiscally disciplined central government, and greater stability and security for investors and credit markets. Weingast’s theory is convincing, but one is hard-pressed to see much that is federal in it. He describes a situation in which members of parliament wrested power from the king – this is separation of powers. He describes process which limited the powers of the central government – this is limited government and judicial independence. While he describes a complicated process of creating credible commitments to limit the crown’s authority, to institutionalize law making, and the like, there is little if anything that is federal in nature, unless all aspects of « limited government » are understood as federalism.
Similarly, Weingast describes a similar dynamic in China following the Cultural Revolution, which he also labels « market enhancing federalism ». The key institutional actor here was the Central Committee of the Communist Party. It developed polices that provided for a stable banking system, recognized the rights of private property, and guaranteed the enforceability of contracts and the independence of courts. Hesitant about its abilities and its plans, the Central Committee pursued these arrangements in selected regions of the country, one step at a time. If they worked, they were continued. If not, they were disbanded and other methods tried.
Weingast describes this policy as « decentralizing one step ahead ». Indeed, this is just what it is. It is classic decentralization. The Central Committee had an objective, had some as to how to achieve it, and rather than establishing one nationwide policy, generated an experiment. It expanded economic autonomy incrementally, one region at a time in order to see how things worked. It an assessment was positive, they expanded the operation[49]. While it is true that the experiment involved local officials and established special economic regions, the local autonomy that was granted what not « federal » in any meaningful sense of the term, it was delegated discretionary authority within the parameters of the experiment. Even Weingast’s own term – « engineering decentralized reform » – envisions a central planner, the engineer, who designs and then implements the plan through trial and error, one step at a time[50]. This is a classic example of experimental decentralization. This description is at odds with the federal political principle – the existence of autonomic territories which within their spheres of autonomy can do or not do as they wish. In contrast the Chinese experience is a classic example of top-down central planning that leads to managerial decentralization.
What holds for fiscal federalism and market-enhancing federalism, holds for a host of other forms of federalism inspired by economic theory: competitive federalism, experimental federalism, liberty-enhancing federalism, all-the-way-down federalism, and still other forms of adjectival federalism which specify specific policy objectives and are concerned with institutional designs that foster them. Institutional designs constructed by economists may or may not foster such objectives, but it is careless in the extreme to attribute such objectives to federal systems generally since they were often established in the crucible of war or conquest or as the by-products of colonial policies whose sub-units are charged with or assume a variety of general functions of government. Rarely – never – are they tailor-made to focus on the objectives that appropriators of economic theories attribute to them.
If the idiosyncratic use of the term federalism were restricted to economists, there would be no issue. But the problem arises when these theories are casually transplanted into political and constitutional theories of federalism. Perhaps political and constitutional scholar are seduced by the siren call of the rigors of economic theory, or perhaps they find that the economists; theories fit their ideological temperament. Whatever the case legal theorists of governmental federalism now cite economists the way court cases are often cited – as forms of authority that substitute a symbol for substantive argument.
A number of federalism scholars have reacted to this misappropriation of economic theory. Feeley and Rubin have catalogued the various ways economic theory has been misapplied and interpreted by federalism scholars[51]. Jenna Bednar (2007)[52], a thoughtful scholar of federalism who finds public choice theories congenial, nevertheless points out that for the most part economic theories of federalism are really theories of decentralization, and better suited for examination of institutional design of industrial organizations than governments. Richard Briffault (1994)[53], a scholar of urban law, has challenged the casual embrace by legal scholars of Tiebot’s[54] classic economic theory of local government. Susan Rose-Ackerman (1980)[55], an economist and legal scholar, has showed convincingly that claims based on economic theories that federalism promotes innovation (the « states as laboratories » metaphor), and that federalism preserves markets[56]. Still others make similar arguments. In most of these cases, economic theories appear to have been employed as ways to try to legitimate (by association with the rigor of « economic theory » and « rigor ») an underlying yet unarticulated normative argument they advance.
We do not mean to play a game of semantics. Although federalism is both a protean and multi-dimensional concept, and like obscenity, may be more difficult to define than to recognize, it is a long-standing term of art that distinguishes one important governing structure from others. The comparative analysis undertaken in this article has been an effort to distinguish important structural differences between unitary and federal political systems, and distinguish them from the ideas of centralization and decentralization. The core of the article was to assess evidence about the differential consequences of unitary and federal structures. The limited evidence suggests significant differences. Unitary systems fare better on almost all the measures. Our suggestion here is a for a more careful application of economic theories to the analysis of federalism, for more research on differences between federal and unitary states, and for greater care in distinguishing between the terms, limited government, centralization and decentralization.
Summary and Conclusion
This paper has tried to identify the distinctive nature of federalism and claims about it by exploring federalism in comparative perspective and in theoretical context. Despite recognition of the importance of comparative legal analysis, the vast majority of research on federalism consists of single-state studies. When comparative studies of federalism are undertaken, they are almost always comparisons among federal systems. Almost no comparative research on federalism has focused on differences between federal and unitary systems. Our objective was to rectify this imbalance by focusing on the comparison of federal with other types of systems, particularly unitary systems. Our review turned up only a few political scientists who have conducted a handful of studies that systematically explore structural differences among nation states, including differences between federal and unitary systems. They included two sets of large N studies assessing a variety of governmental structures in a large number of counties in light of various measures of government effectiveness and efficiency and citizen well-being. We also found another group of smaller N studies that contrasted different governmental structures. We drew on the data in some of these studies to explore further issues central to our concerns.
What we found in all these studies, including our own, is that their conclusions almost all point in the same direction and to the same conclusion: unitary systems out-perform federal systems on almost all indicators of governmental effectiveness and efficiency, and quality of life. Of course we noted a host of caveats. But we also noted that studies using different indicators, different data sources, and different research designs nevertheless came up with the same findings, that unitary systems out-perform federal systems.
We then examined variations on this theme. We found that there may be as much variation in terms of centralization/decentralization in federal systems as there is in federal systems. We found that there may be as much if not more variation in centralization and decentralization within federal and unitary systems as between them. Furthermore, we found that unitary polities are more effective at tailoring decentralization to deal with particular problems.
Finally, we suggested that in their haste to associate with rigor and theory, some legal scholars have been too quick to appropriate the findings of economic theories of federalism. Our point is that almost all these theories, whatever their intrinsic value, are actually theories of decentralization, and thus have little if anything to tell us about federalism.
What lessons are to be learned from all this? It was suggested by the quote from Seymour Martin Lipset we posted at the beginning of this article, « to know one country is to know none ». Applying his admonition to the study of federalism, he invites us to ask, « Federalism as compared to what? ».
Malcolm M. Feeley
Claire Sanders Clements Professor of Law at the University of California at Berkeley. He is the author of numerous books and articles, including (both with Edward Rubin) Federalism: Some Notes on a National Neurosis (1994), and Federalism: Political Identity and Tragic Compromise (2009).
Aniket Kesari
PhD Student in the Jurisprudence and Social Policy program at the University of California, Berkeley and a rising JD student at Yale Law School.
Bibliography
Althusius Johannes, Politica Methodice Digesta, C.J. Friedrich ed., Cambridge, Harvard University Press, 1932.
Anderson Benedict, Imagined Communities: Reflections on the Origins and Spread of Nationalism, London, Verso, 1991.
Bakke Kristin M. & Wibbels Erik, « Diversity, Disparity, and Civil Conflict in Federal States », World Politics, 59, 2006.
Bednar Jenna, « The Political Science of Federalism », Annual Review of Law and Social Science, 7, 2011.
Bentley Arthur F., The Process of Government: A Study of Social Processes, Chicago, University of Chicago Press, 1908.
Beramendi Pablo (2007), « Federalism », in C. Boix & R.E. Stoker (dir.), The Oxford Handbook of Comparative Politics, London/New York, Oxford University Press, 2007.
Bonoli Giuliano & Trien Philipp, « Cost-Shifting in Multitiered Welfare States: Responding to Rising Welfare Caseloads in Germany and Switzerland », Publius, 46, 2016.
Briffault Richard, « What about the ‘Ism? Normative and Formal Concerns in Contemporary Federalism », Vanderbilt Law Review, 47, 1994.
Cole Alistair, « France between Centralization and Fragmentation », in J. Loughlin, F. Hendriks & A. Lidstrom (dir.), The Oxford Handbook of Local and Regional Democracy in Europe, Oxford, Oxford University Press, 2011.
Cooter Robert D., The Strategic Constitution, Princeton, Princeton University Press, 2002.
Elazar Daniel, American Federalism: A View from the States, 3rd ed., New York, Harper and Row, 1984.
— Federal Systems of the World, 3rd ed., London, Longman Current Affairs, 1994.
Erk Jan, « Does Federalism Really Matter? », Comparative Politics, 39, 2006.
— « Comparative Federalism as a Growth Industry », Publius, 37, p. 262-278, 2007.
Feeley Malcolm & Rubin Edward, Federalism: Political Identity and Tragic Compromise, Ann Arbor, University of Michigan Press, 2008.
Filippov Mikhail, Ordeshook Peter, Shvetsova Olga, Designing Federalism: A Theory of Self-Sustainable Federal Institutions, New York, Cambridge University Press, 2004.
Franck Thomas (dir.), Why Federalism Fail, New York, Oxford University Press, 1968.
Friedrich Carl, Trends in Federalism in Theory and Practice, New York, Praeger, 1968.
Gerken Heather, « Federalism as the New Nationalism », Yale Law Journal, 123, 2014.
Gerring John, Thacker Strom C., Moreno Carola, « Centripetal Democratic Government: A Theory and Global Inquiry », American Political Science Review, 99, 2005.
Gerring John & Thacker Strom C., « Political Institutions and Corruption: The Rise of Unitarism and Parliamentarianism », British Journal of Political Science, 34, 2004.
— A Centripetal Theory of Democratic Governance, New York, Cambridge University Press, 2008.
Golembowski David, « Federalism and the Catholic Principle of Subsidiarity », Publius, 45, 2015.
Greve Michael S., The Upside-Down Constitution, Cambridge MA, Harvard University Press, 2012.
Hills Roderick M., « Federalism and Public Choice », in A. O’Connell & D. Farber (dir.), Research Handbook on Law and Public Choice, Cheltenham, Edward Elgar, 2010.
Holt Robert & Turner John, The Political Basis of Economic Development: A Comparative Analysis, New York, Van Norstrand, 1968.
Inman Robert, « Federalism’s Values and the Value of Federalism », CESifo Economic Studies, 53-4, 2008.
Inman Robert & Rubinfeld Daniel, « The Political Economy of Federalism », in D. Mueller (dir.), Perspectives on Public choice: A Handbook, Cambridge, Cambridge University Press, 1997.
Jolly Seth K., The European Union and the Rise of Regionalist Political Parties, Ann Arbor, University of Michigan Press, 2015.
Kaufman Herbert, Are Governmental Organizations Immortal?, Washington D.C., The Brookings Institution, 1976.
Kincaid John, Federalism, 4 vol., London, SAGE, 2011.
Lidstrom Anders, « Sweden: Party-Dominated Subnational Democracy under Challenge? », in J. Loughlin, F. Hendriks & A. Lidstrom (dir.), The Oxford Handbook of Local and Regional Democracy in Europe, New York, Oxford University Press, 2011.
Lijphart Arend, The Politics of Accommodation: Pluralism and Democracy, Berkeley, University of California Press, 1968.
— Democracy in Plural Societies: A Comparative Exploration, New Haven, Yale University Press, 1977.
— Patterns of Democracy: Government Forms in Thirty-Six Countries, New Haven, Yale University Press, 1999.
— « Constitutional Design for Divided Societies », Journal of Democracy, 15, 2004.
Loughlin John, Regional and Local Democracy in the European Union, Brussels, Committee on the Regions, 1999.
— Subnational Government: The French Experience, London, Palgrave Macmillan, 2007.
— « Reconfiguring the State: Trends in Territorial Governance in European States », Regional and Federal Studies, 17, 2007.
Loughlin John, Lidstrom Anders, Hudson Chris, « The Politics of Local Income Tax in Sweden: Reform and Continuity », Local Government Studies, 31, 2005.
Loughlin John, Hendriks Frank, Lidstrom Anders (dir.), The Oxford Handbook of Local and Regional Democracy in Europe, New York, Oxford University Press, 2011.
North Douglass & Weingast Barry, « Constitutions and Commitment: The Evolution of Institutions of Governing: Public Choice in Seventeenth-Century England », Journal of Economic History, 49, 1989.
Oates Wallace E., « An Essay on Fiscal Federalism », Journal of Economic Literature, 37-3, 1999.
Olson Mancur, The Logic of Collective Action, Cambridge, Harvard University Press, 1965.
Pierson Paul, « Fragmented Welfare States: Federal Institutions and the Development of Social Policy », Governance, 8, 1995.
Riker William, Federalism: Origin, Operation, Significance, Boston, Little, Brown, 1964.
Rodden Jonathan & Rose-Ackerman Susan, « Does Federalism Preserve Markets? », Virginia Law Review, 83, 1997.
Rodden Jonathan & Wibbels Erik, « Beyond the Fiction of Federalism: Macroeconomic Management and Multi-tiered Systems », World Politics, 54, 2002.
— « Fiscal Decentralization and the Business Cycle: An Empirical Study of Seven Federations », Economics & Politics, 22-1, 2010, p. 37-67.
Rose-Ackerman Susan, « Risk Taking and Reelection: Does Federalism Promote Innovation? », Journal of Legal Studies, 9, 1980.
Feeley Malcolm M. & Rubin Edward L., Federalism: Political Identity and Tragic Compromise, Ann Arbor, University of Michigan Press, 2008.
Rubinfeld Daniel, « The Economics of the Local Public Sector », in A. Auerbach & M. Feldstein, Handbook of Public Economics, vol. 2, Amsterdam, North-Holland, 1987.
Skocpol Theda, Evans Peter, Reuchmeyer Dederich (dir.), Bringing the State Back In, New York, Cambridge University Press, 1985.
Smith Anthony, National Identity, Reno, University of Nevada Press, 1991.
Taylor Stephen e.a., A Different Democracy: American Government in a 31-Country Perspective, New Haven, Yale University Press, 2014.
Tiebot Charles, « A Pure Theory of Local Expenditure », Journal of Political Economy, 64, 1956.
Triesman Daniel, « Decentralization and the Quality of Government », 20000 [http://www.imf.org/external/ubs/ft/seminar/2000/fiscal/triesman.pdf]
Weber Eugen, Peasants into Frenchman, Stanford, Stanford University Press, 1976.
Weingast Barry, « The Economic Role of Political institutions: Market-Preserving Federalism and Economic Development », Journal of Law, Economics, and Organization, 111, 1995.
— « The Political Foundations of Democracy and the Rule of Law », American Poetical Science Review, 91, 1997.
— « Second Generation Fiscal Federalism: Implications for Decentralized Democratic Governance and Economic Development » (Sept. 7, 2007).
Weare Kenneth C., Federal Government, New York, Oxford University Press, 1953 [3rd ed.].
Wibbels Eric, Federalism and the Market: Intergovernmental Conflict and Economic Reform in the Developing World, New York, Cambridge University Press, 2005.
— « Decentralized Governance, Constitution Formation, and Restribution », Constitutional Political Economy, 16, 2005.
— « Madison in Baghdad? Decentralization and Federalism in Comparative Politics », Annual Review of Political Science, 9, 2006.