Two Essays on Government Behavior

The main theme of this dissertation is government’s strategic behaviors. We show that different budget structures give governments incentives to behave differently, and that the Leviathan model and the Bureaucratic model are better in modeling government behaviors than the median voter model. We first discuss theoretically the design of an optimal tax system, promoting the Leviathan government to maximize social welfare in order to maximize its own revenue. Then we examine empirically how government behaviors vary with different budget structures. In essay I, we apply the Buchanan-Brennan (B-B) rule to examine the effects of a tax system on the efficiency of agricultural production in the context of Chinese local governments, which receive insufficient control from the central and are free from the pressure from local residents due to asymmetric information and lack of horizontal accountability. We extend the B-B rule to include the incentive issues and the risk sharing, and also their trade-off. Farmers and the agricultural sector assume significant roles in the national economy of China, while the under-provision of public infrastructures and the risks involved negatively affect agricultural production and therefore impede economic growth. Within the principal-agent framework, we illustrate how the problem is inherent in the agriculture tax system in China and propose our solution of special earmarking. In essay II, we test empirically for the government’s strategic behaviors. We argue that the environmental performance is affected by government policy. Therefore it relates inherently to the budget structure and government incentives. With an illustrating model between structure of revenue and expenditure and pollution level, we propose three hypotheses, which state that the lower the ratio of business related tax in total revenue, the higher the ratio of property tax in total revenue, the higher the share of health expenditure in total expenditure, the government will have higher incentive to control pollution and thus the pollution level is lower. Our empirical evidence provides support to our hypotheses, which show that structures of revenue and expenditure do affect the government’s incentives to control pollution. Therefore, changes in the budget structure might be helpful to achieve better environmental performance