Speaking differently: How dialect affects E-government adoption

Leveraging the novel and comprehensive Chinese Business Environment Survey data spanning 2019 to 2023, this paper identifies dialect difference as a significant yet unexplored determinant of electronic government (E-government) services adoption. Our theoretical model posits that dialect-induced language barriers elevate uncertainty about the time (effort) required for successful online administrative interactions, raising the likelihood of procedural failures and thus deterring enterprise engagement with E-government platforms. Exploiting the establishment of the Data Administration Bureau as an exogenous policy shock, we empirically find that, following the intervention, regions predominantly speaking local dialects were 5.1 % less likely to adopt E-government services—equivalent to approximately 5.9 million fewer enterprise users—relative to Mandarin-speaking regions. This effect is significant after controlling for conventional confounding factors such as social networks and cultural clustering, and remains robust across alternative specifications and validation checks. Our findings shed lights on frictions in government digitalization, implying welfare improvements from mitigating dialect-related access constraints.