تمرکز زدایی مالی جزئی

The fiscal decentralization impulse now sweeping the world often leads to partial decentralization, where subnational governments are funded by central transfers, rather than leading to full local autonomy. Despite the practical important of this arrangement, the literature contains no economic analysis of a partial decentralization regime in a Tiebout-style model. This paper provides such an analysis, relying on the key assumption that public-good provision requires effort on the part of government officials. By choosing different degrees of effort, localities can then provide different public-good levels even when a fixed, common transfer constrains them to spend the same amount. A number of useful results are derived.

Under fiscal decentralization, subnational governments are granted autonomy in the provision and financing of public goods. While subnational autonomy has long been a feature of the fiscal environment in the United States and several other countries, a greater degree of central control over the public sector is common elsewhere, especially in the developing world. However, partly in response to advice from the World Bank and other international agencies, many developing countries are seeking greater fiscal decentralization by attempting to reallocate spending and taxing authority to subnational governments. This movement has been motivated in part by the ideas of Tiebout (1956), who argued that local control of spending allows the public sector to respond more effectively to varied consumer preferences for public goods. Despite this impulse toward decentralization, an obstacle to the achievement of true local autonomy in developing countries lies on the tax side. Even when spending authority is passed downward to subnational governments, a lack of adequate tax capacity (especially at the local level) often prevents these governments from funding expenditures out of their own revenues. Instead, spending relies heavily on transfers from the central government. As explained by Shah (2004),

The fiscal decentralization impulse now sweeping the world often leads to partial decentralization, where subnational governments are funded by central transfers, rather than leading to full local autonomy. When it occurs, the choice of partial fiscal decentralization is often dictated by a lack of subnational tax capacity. Despite the practical importance of this arrangement, the existing literature contains no economic analysis of a partial decentralization regime in a Tiebout-style model. This paper has provided such an analysis, relying on the key assumption that public-good provision requires effort on the part of government officials. By choosing different degrees of effort, localities can provide different public-good levels even when a fixed, common transfer constrains them to spend the same amount. The paper shows that, as long as the transfer under partial decentralization can be chosen in a close-to-optimal fashion, that regime is superior to full central control, where the center dictates common public-good levels along with a fixed amount of spending. This conclusion provides a key practical lesson: when localities are able to adjust public-good levels despite a fixed spending requirement, the center should relinquish control of this decision, allowing localities to make their own z choices. The analysis also shows that public-good variety is lower under partial decentralization than under the preferred but infeasible full decentralization regime, muting the consequences of Tiebout sorting. However, if localities can complement the fixed central transfer with limited local tax revenue, constrained by a cap, then public-good variety is constrained only for high and low demanders, with the center of the distribution unaffected relative to full decentralization. Finally, Leviathan behavior on the part of local officials leads to overspending under full decentralization, so that partial decentralization may become the preferred arrangement even when full decentralization is institutionally feasible.