Data Manipulation through Patronage Networks: Evidence from Environmental Emissions in China

Abstract

Abstract Data manipulation is considered one of the most pernicious forms of information distortion that challenges the effectiveness of the principals in monitoring agents. This study investigates factors influencing the behaviors of data manipulation from the perspective of patronage networks. Patronage networks feature a reciprocal exchange between patrons who control appealing resources and clients who ask for protection and benefits from the patrons. We hypothesize that in a three-tier principal–supervisor–agent model, agents favored in the patronage network would receive tacit permission and protection from the supervisor, and hence be encouraged to falsify performance data and game the principal. The empirical results, using panel regression models based on officially reported and satellite-observed environmental emissions data in China, suggest that cities with leaders favored in the patronage network present a stronger tendency for data manipulation. This study has theoretical implications for understanding the impact of informal institutions on performance management.