Fiscal Decentralization And Economic Growth: Theory And Evidence

Focusing on relations of fiscal decentralization and economic growth, this thesis includes theoretical and empirical studies on fiscal decentralization, which has been a cutting-edge issue on public economics area. Different with existing literatures, the theoretical framework of this thesis combined endogenous growth models with principal-agent theory. First, in an endogenous growth model we introduced in chapter 3, government expenditures were classified according to different contributions to output and different applications, this required optimal ratios among them on the balanced growth path. Solution of the model showed an inverted U shape between fiscal decentralization and economic growth. An extension of this model discussed about fiscal autonomy of local-level governments, also showed the inverted U shape between fiscal decentralization and economic growth. However, fiscal decentralitioan is an exogeneous variable in the model, this garuanteed the optimal economic growth on the balanced growth path. From the perspective of principal-agent theory, fiscal decentralization was choosed by the parliament as an instrument to incentive central politicians and local politicians simultaneously. The principal-agent model introduced in chapter 4 provided an endogenous treatment on fiscal decentralization as a fuction of economic growth. Transitional dynamic analysis gave different multiple equilibria under two different conditions, where contribution of central government expenditures on output was different from that of local-level government expenditures on output.As well as the theoretical work on fiscal decentralization, this thesis reported our empirical work using Chinese data from country-level governments. Using fix effect estimation with LSDV weighted, fiscal decentralization of country-level governments has significant positive relations with economic growth. Adding squared fiscal decentralization into estimation model reported a U shape between fiscal decentralization and economic growth. Compared between "poor countries"("fu pin gong zuo zhong dian xian") and others, estimation showed a smaller coefficient of fiscal decentralization of "poor countries". District Distictions still exist among country-levels. Comparing with the situation on the west part of China, fiscal decentralization showed greater positive relations with economic growth on the central part and the east part, caused a bigger gap among distinctions.Further research of this thesis can go along three branches. First, in the principal-agent model of chapter 4, agents are risk neutral. Other types of agents behave different in the incentive systems, thus caused a diffent solution of the model. While considering the incentive systems among governments, multiple equilibria would still exist. Secondly, our model assumed that there were no externalities among government expenditures. This may be impractical, while it is useful to focuse on relations between fiscal decentralization and economic growth in a simple model. Actualy, there are several ways to internalize the externalities of government expenditures. If the number of local-level governments is limited, or there are some leaders with large population or large economy, local-level governments react with each other. Loss of efficiency from externalites will be partly offseted by scale effects from simulations among local-level government. Another way is to reallocate the expenditures among governments. Some public good with strong externalites shoule be provided by central government will minimize the cost of efficiency loss. The third branch is along the empirical work of this thesis. Lack of data of transfer payments giving to country-level governments may cause some estimation errors. Specificaly, expenditure decentralization on the west part may be lowers when cutting a big sum of correspondent transfer payment, and income decentralization on the east part may be higher when adding a big sum of non-conditional transfer payments. How to estimate fiscal decentralization is still a hard work for further research.