The Economic Impact of China's Anti-Corruption Campaign

The Economic Impact of China's Anti-Corruption Campaign. Nan Chen and Zemin (Zachary) Zhong. University of California Working Paper, June 2017, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2996009

Abstract: Political corruption is a major problem in governance and is pervasive especially in developing countries. Depending on pre-existing distortions, corruption may benefit economic growth by "greasing the wheel", or it may distort supply of public goods and create inefficiency. Empirically testing the effect of corruption on efficiency and distribution is difficult due to the evasive nature of corruption. We take an alternative approach by investigating the economic impacts of governments' anti-corruption efforts. Our analysis is performed in the context of China's recent anti-corruption campaign under the new president Xi Jinping, the largest of its kind in recent history. As an important initiative of this campaign, the CPC's Provincial Committee of Discipline Inspection (PCDI) send inspection teams to investigate county-level government for potential corruption. The variation in the timing of PCDI visit allows us to use a difference-in-difference design to identify the impact of anti-corruption on local economy. Using a unique administrative dataset of vehicle registration, we find that PCDI visits cause car sales to drop by 3.4% at county level. The effect is surprisingly uniformly distributed across different price tiers. Luxury brands exhibit a similar drop as domestic brands, suggesting corruption's impact permeates a wide income spectrum. Over time, the effect is strengthening: We observe a 2% drop in the first three months of PCDI visit, and a 10% drop one year afterwards. The especially large impact cannot be explained by decline in government officials' consumption behavior, suggesting anti-corruption efforts also affect the private sector.

We validate our empirical strategy by showing that:

(1) the timing of PCDI visits cannot be predicted by observable county characteristics and,
(2) car registrations exhibit parallel pre-treatment trends.

The results are robust to placebo tests and alternative specifications. We find that the effect of PCDI visit on car sales cannot be explained by local economic indicators or monitoring cost as measured by distances to provincial/prefectural governments.

Keywords: Corruption, Political Economy, China
JEL Classification: D73, P16, H70, L62