Central-Local Relations in the PRC under the Tax Assignment System: An Empirical Evaluation, 1994-97

Abstract

In 1993, Chinese scholars Wang Shaoguang and Hu Angang registered a warning that the state capacity of the PRC was declining. According to Wang and Hu, this declining capacity is exemplified by two phenomena in the mid-1980s: the declining share of the national fiscal revenue in the gross domestic product (GDP) and the decline of the central government`s share of the national fiscal revenue. Wang and Hu warned that this declining capacity of the state will also lead to a widening disparity of wealth distribution and economic development, which eventually will produce social chaos. Many believed that the PRC government was affected by the above analysis and thus adopted a tax reform measure in 1994. The new tax system between the central and local governments, the tax assignment system (fen shui zhi), was designed to allow the central government to regain its lost share in national fiscal revenue and to alleviate local disparities. By using official statistical data, this study intends to evaluate the effect of the practice of the tax assignment system in the years between 1994 and 1997. Different from what is claimed by the Chinese officials, this study finds that the system has not been able to successfully achieve its intended policy goals. This paper is able to reach such a different conclusion as it takes into account the ”tax return”(shuishou fanhuan) from the central government to local governments. As an integral part of the tax assignment system, the ”tax return” mechanism carries a strong compromising institutional character. The system, designed as an incentive for the local governments to cooperate, also hindered the central government in raising its extractive capacity. The result of reviewing the practice of this system in the years from 1994 to 1997 shows that the central government seems unlikely to be able to phase out the ”tax return” mechanism and put in place a full-fledged tax assignment system.