Overcoming authoritarian information constraints : disincentivizing falsification of local statistics in China

Abstract

Governance under authoritarianism relies on the regime’s ability to gather accurate information about the actions of government officials. In recent years, while a growing body of literature demonstrates how authoritarian regimes such as China are increasingly capable of collecting public opinion or supervising societal perceptions of the regime, monitoring the information distortion of regime agents remains a challenge. The latter seems to be contingent on administrative structure and is thus difficult to tackle. One of the most prevalent expressions of local information distortion in authoritarian regimes is the collection of economic-related data. Conventional work questioning the credibility of China’s sub-national GDP statistics presumes a reporting system that allows local governments to collect and report their own GDP one-level up. In this context, local authorities are often motivated by better career prospects to exaggerate their GDP figures. The rationale is deeply rooted in Max Weber’s theory on bureaucracy, which highlights expertise and knowledge as the sources of bureaucratic power vis-à-vis their “political masters.” A competing reality is that starting from 2018, various provinces and cities admitted their failure to reach growth targets. How can this be explained? Based on fieldwork in Zhejiang and Shanghai from 2018 to 2020, an original database and secondary materials, this thesis explores the mechanisms through which China manages to curb local statistical falsification. Specifically, as a pioneer of China’s statistical reform, Zhejiang implemented a “one-level-down” statistical calculation system to replace a traditional “bottom-up” reporting one, such that prefectural GDP figures will be announced directly by provincial superiors. In addition, the establishment of a national-level platform for firms to directly report data diminishes the ability of local statistical bureau to conceal falsification behavior. Meanwhile, the recently established possibility that cadres in local statistical bureaus could be disciplined directly by upper level statistical agencies, have de-linked the formerly shared incentives between local statistical agencies and their corresponding local governments. Hence, even if local governments are still inclined to pressure local statistical bureau officials, cadres who possess an informational advantage over governmental authorities are incentivized to avoid being punished by their statistical bureau superiors. This thesis sheds lights on the potential of curbing intra-regime information manipulation introduced by a centralized performance-setting system.

Date of Award	2020
Original language	English
Awarding Institution	The Hong Kong University of Science and Technology