IV. The Political Economy of Decentralization

The intellectual case for decentralization originates in the most basic arguments concerning democratic government and the effective representation of citizens' interests, to which economic arguments based on efficiency have been added more recently. Political philosophers from the 17th and 18th centuries, including Rousseau, Mill, de Tocqueville, Montesquieu, and Madison distrusted autocratic central government and held that small, democratic units could like ancient Athens preserve the liberties of free men. In several of the Federalist Papers, Madison theorized about the prevention of tyranny via a balance of powers not only among the branches of central government, but between central and regional and local governments as well.

The modern case for decentralized government is well represented by Wolman (in Bennet, 1990), to whom we now turn. Wolman groups his arguments concerning decentralization under two main headings: Efficiency Values and Governance Values. Efficiency Values comprise the public choice justification for decentralization, where efficiency is understood as the maximization of social welfare. Wolman contrasts the provision of public goods with the market for private goods. Within the private economy, individual preferences are expressed through a market mechanism which facilitates continuous signaling between supply and demand via prices. The nature of public goods, however, is such that competitive markets will not provide them.

When public goods are provided, tax and service packages should reflect as accurately as possible the aggregated preferences of community members. However, because individual preferences for public goods differ, there will be some divergence between the preferences of individual community members and the tax and service packages reflecting the aggregated community preferences. It is likely that the average divergence of individual preferences from the tax and service package adopted by the community through its government will be less in small communities of relatively like-minded individuals than it will be in larger, more heterogeneous areas. [...Allocative] efficiency and social welfare are thus likely to be maximized under highly decentralized political structures.^[4]

One counter argument to this rationale highlighted by Wolman is that decentralization will complement, or even exacerbate, disparities among communities with different economic means. In principle, however, this objection is easily addressed through grants administered centrally, designed to equalize localities' resources. Similarly, decentralization to low levels of government may reduce efficiency by inhibiting the achievement of economies of scale in the provision of some services, though this too is easily solved via different levels of government providing these services according to their technical characteristics. A more serious objection is posed by the existence of externalities^[5] in the provision of local public goods, which reduce overall efficiency for society. These can be solved by reverting to higher levels of government, which internalize the externalities in their taxing and spending decisions.

But the fundamental objection to such efficiency arguments, and one which is generalizable for this literature, is that it simply assumes that central government will produce more standardized, less-differentiated outputs less suited to local preferences than local government. Although this is intuitively appealing, the lack of an explanation for how this comes about amounts to assuming away the problem. We can easily draw up a model where central government installs agents in each community to gather information and detect local needs and preferences, which data is then relayed at low cost back to the center (see Annex 1). All outputs are produced centrally, and then distributed in the desired amounts and qualities to all localities. If we assume economies of scale in at least some outputs, and transportation costs low enough that these gains are not canceled out (neither unreasonable), then it is easy to see that such a system would dominate the decentralized solution on (productive) efficiency grounds.^[6] This could form the basis of an argument for the centralization of government which, in purely logical terms, is every bit the equal of its opposite.

Some authors have sought refuge in the position that the competing claims are a metric question, and therefore resolvable through empirical research. But this response is also highly unsatisfactory. The issue in question - whether central government is likely to produce outputs that are more or less similar to individual's preferences - is intimately tied to the structure of governmental institutions and to the incentives that these produce. As such it is tractable to theorizing about cause-effect relationships concerning structure, product and aggregate outcomes. It is, in fact, a question unsuited for empirical study, as attempts to measure the allocational efficiency of "central" versus "decentralized" government, in ignorance of plausible mechanisms, are likely to omit important explanatory variables and result in models which conflate causes and produce meaningless estimates. It is therefore important to provide a mechanism which adequately explains the central feature of the standard efficiency argument.

Governance Values comprise Wolman's second grouping, in which he includes: (i) responsiveness and accountability, (ii) diversity, and (iii) political participation. The first values, responsiveness and accountability, are easily the most important of these and comprise the political parallel to the argument about efficiency detailed above.

Decentralization, by placing government closer to the people, fosters greater responsiveness of policy-makers to the will of the citizenry and, it is argued, results in a closer congruence between public preferences and public policy. This is not only because decision-makers in decentralized units are likely to be more knowledgeable about and attuned to the needs of their area than are centralized national-government decision-makers, but also because decentralization permits these decision-makers to be held directly accountable to the local citizenry through local elections.^[7]

However this argument, as presented by the political science literature, is also a priori. Politics in decentralized governments may be more closed than national politics, and more susceptible to interest-group capture or manipulation by powerful agents. Alternately, local elections may be contested on non-local issues, such as the popularity of the national government (as occurs in the UK). Wolman asserts that the question of whether responsiveness and accountability increase or decrease with decentralization is an empirical one.

Diversity in public policy is another argument in favor of decentralization. Diversity is valued because (i) it offers citizens a greater choice in public services options when they are deciding where to reside (see the Tiebout model below); and (ii) it is thought to encourage experimentation and innovation in public policy. Decentralization will thus result in a variety of policy approaches at the local level, some of which will be more successful than others. Once given policies have been shown to work, they can be taken up by other decentralized units as well as central government. This argument obviously depends critically upon the empirical question of whether decentralized structures really do foster greater policy diversity than central government. Limited research suggests that it does, but the question cannot yet be considered closed. The question of the diffusion of successful policy experiments to other units and levels of government is also an open, empirical one. We note, however, that the organizational literature stresses that decentralized structures promote innovation, while centralized structures promote adoption.

The devolution of real power to localities is also thought to enhance political participation amongst the people, because of the increased levels of interest and involvement in local government which it brings about. This process, it is argued, enhances democratic values and performs a systems-maintenance function, thus promoting political stability. Similarly, decentralization provides the opportunity for citizens to debate and decide upon those local issues which matter most to them, thus promoting political education. Lastly, local politics provides a training-ground for local leaders, who can then progress to become national leaders. These claims are also subject to empirical investigation; initial results are inconclusive, but not favorable. Participation levels measured by turnout at elections in the United States, a country considered decentralized, are notoriously low, especially for local elections (typically around 30%). It is also low in the UK (around 40%), but much higher in countries considered more centralized, such as France (70%) and Italy (85%).^[8] Nor is the educational function of decentralization evident in the level of political discourse in the US and UK. This last point is probably related to the importance of mass media in the modern political discourse. When most political debate is mediated by a technology that displays extremely centralizing characteristics, the electorate is likely to be better informed about the intricacies of the national and international politics that fill the airwaves than they are about the state of the municipal sewerage system, or the performance of local police.

In conclusion, and in the wake of a wide variety of arguments, it is important to note that the intellectual core of the case in favor of decentralization is composed of a combination of the allocational efficiency argument with that concerning the responsiveness and accountability of local government. Other arguments regarding diversity, education and leadership development may bolster this reasoning, but are of secondary importance. The possibility that local government can be designed in such a way that it accurately perceives the needs of its electorate and faces clear incentives to attend to them, both to greater degrees than central government, constitutes a powerful hypothesis in its favor.