Growth

Site Index
Main Page
Issues in Program Design


Contributor: World Bank
Author: Decentralization Thematic Team
Contact: Jennie Litvack


Decentralization and Economic Growth

We know very little about the relationship between decentralization and growth. Empirical evidence for the way in which decentralization affects growth has been contradictory and is plagued by measurement, specification, and analytical problems. There is stronger evidence for a relationship in the other direction -- from growth to decentralization-- but the interpretations of this correlation between high income and decentralization have varied. In the absence of strong, unambiguous, empirical evidence, researchers have put forward the following three hypotheses concerning the relationship between decentralization and growth: In each of the hypotheses, growth has only a secondary relationship to decentralization and the nature of this connection -- growth enhancing, growth-impeding, or growth-requiring- depends on what one sees as the primary effects of decentralization. These primary effects, in turn, have much to do with the specific design of decentralization policy. While the rest of the KMS node discusses the methods for reaching certain primary effects -- ideally more efficient, responsive delivery of services -- this section provides an outline of the ways in which the commonly discussed primary effects of decentralization can shape decentralization's effect on growth and desirability in less developed countries.

Three Alternatives: Hyotheses for How Decentralization Can Indirectly Impact Growth Hypothesis 1:Decentralization increases economic efficiency in public spending, therefore its dynamic effects should be growth-enhancing.
Local governments' information advantage allows them to respond better than central governments to constituents' preferences. Central governments would have to invest far more resources to obtain the same information. Although constituents' preferences do not always serve as a guide for productive investment, they can be useful as reviews of program design and guidelines for useful expenditure allocation. Citizens in decentralized systems are able to and, to varying extents, do move across jurisdictions to settle where the package of public goods provided matches their preferences. High mobility may force subnational governments to be more competitive (thus more efficient and possibly innovative) in providing services. It is an intuitive leap from these stylized static effects to the dynamic effect of increasing growth. Several studies of public expenditure and growth show that certain kinds of public expenditure have more significant growth-enhancing effects than others. Decentralization might improve the quality of proven growth-enhancing inputs such as education.

Hypothesis 2: Decentralization can lead to macroeconomic instability which can, in turn, inhibit growth.

Decentralization reduces the tax and expenditure base that the central government can use to carry out the stabilization function. Subnational governments are incapable of taking over this national macroecomic stabilization because their individual jurisdictions are so open. The practical importance of this effect, however, depends on the extent to which the central government actually uses these bases for stabilization. Poorly designed decentralization policies can create incentives for subnational governments to spend irresponsibly and unsustainably. Loosely defined, unpredictable, transfer arrangements, particularly those which allow liberal intergovernment borrowing, can lead to fluctuating, unsustainable subnational government spending. In a worst-case scenario, the central government runs the risk of having to increase its own deficit or bail out the lower levels. Decentralization of revenues and expenditures to subnational governments can take scarce resources away from national projects, such as infrastructure, that could make greater contributions to growth. The evidence concerning decentralization's destabilizing effect is, however, mixed. Research in Russia and China has clearly linked subnational government autonomy to growing central government deficits, but these cases may not be generalizable because of their unique system of using local governments to collect central government revenue. Other, more "typical" subnational governments in Argentina and Brazil have expanded expenditures beyond revenues and then turned to the central government for bailouts or debt relief. Switzerland, on the other hand, has been both highly decentralized and highly stable in the post-war period. Research on Canada has shown that increasing subnational governments' budgets may have actually had a stabilizing effect. Subnational governments have coordinated spending and may be more able to offset region-specific business cycles.

Hypothesis 3: Developing countries have significantly different institutional and economic environments than developed countries and will not reap the benefits or suffer the consequences of decentralization in the same ways.

Institutional arrangements in developing countries do not necessarily give subnational governments the incentive to use their information advantage in responding to their constituents. Leaders may be appointed or elected on the basis of their social standing. Subnational governments in developing countries may not have the economic resources attract sufficiently trained people or the human resources to manage larger budgets. The decentralization of Peru's water supply and sanitation sector, for example, reduced the quality of service because local governments did not have the technical ability to maintain or plan infrastructure. Nevertheless, Colombia's success in decentralizing some services suggests that these ability constraints can be overcome by private-public partnerships. While hypothesis one might be true for developed countries, hypothesis two might be a more accurate description of developing countries.

Conclusion: Design Matters

There is no clear, automatic, relationship between decentralization and growth. Direct empirical studies have not been satisfying and a return to the literature on decentralization's primary effects may be a more useful way to think about the relationship between growth and decentralization. From this point of view, the design of decentralization becomes the key factor in determining whether policies will lead to the efficiency linked to higher growth, exacerbate the deficits and instability connected to lower growth, or simply become mired in institutional constraints.