Digital Surveillance and Self-Censorship in Autocracies: Evidence from a Survey Experiment in Kazakhstan

Abstract Digital surveillance technologies are increasingly employed, especially in authoritarian regimes seeking to monitor and shape online communication. Yet we have little empirical evidence about how such surveillance affects citizens’ expression of political views. Theory suggests that awareness of being surveilled induces self-censorship, discouraging individuals from voicing opinions on sensitive topics. This paper tests this proposition using a survey experiment conducted in Kazakhstan in November 2023 (N = 5,025). Participants were randomly exposed to a reminder of government surveillance, an assurance of privacy, or a control condition before answering sensitive and non-sensitive questions. Exposure to the surveillance reminder reduced responses to sensitive items by about three percentage points, while the privacy assurance had no effect. The effect is driven by respondents who consume foreign media, suggesting that politically informed individuals are more responsive to surveillance cues. These findings provide experimental evidence that perceived surveillance discourages political expression and reinforces authoritarian stability.