A Ballot for a Bullet

• In Event: Opposition and Protest: Participation and Consequences

Abstract

Repression and pseudo-democratic elections are two pillars of most modern autocracies. We argue that the interactions between these two pillars tend to create a slippery path for electoral authoritarian regimes—repression may stimulate strong transient electoral opposition, inducing dictators to commit to less democratic institutions.

We empirically examine this argument in the context of Hong Kong, where the police actively used tear gas to suppress the protests and riots in the Anti-Extradition Law Amendment Bill Movement in 2019, especially in the months before the district council election. We explore the unintended effects of tear gas on the electoral outcomes and the ensuing government response by comparing upwind and downwind election districts. Matching real-time wind data to approximately 10,000 tear-gas locations during the massive street protests, we discover that tear gas drifts contributed to the historical defeat of the pro-regime camp in most of the districts. The tear gas effects on electoral outcomes mainly concentrate in the downwind districts close to the tear gas shooting sites and remain robust to the exclusion of the districts containing the shooting sites, suggesting that the electoral opposition arose from genuine exposure to tear gas drifts. The landslide win of the pro-democracy camp prompted the Chinese government to revamp the electoral system, toppling almost all elected pro-democracy councillors, and to effectively ban future legislative oppositions. We also find that the elected councillors in downwind districts are more likely to be disqualified or forced to resign earlier than their counterparts in other districts. Our finding demonstrates that state repression in electoral authoritarian regimes causes unexpected electoral backlash, thereby establishing a novel linkage between repression and democratic backsliding. Our evidence casts doubt on the long-term sustainability of multiparty pseudo-democratic elections that currently exist in almost all modern autocracies.

Similar to Acemoglu and Robinson (2000), we also emphasize transient mass movement opportunities as the origins of more permanent institutional changes. Extending their theory to a world of electoral authoritarian regimes, we reach a more nuanced conclusion. When social movement is strong enough to overthrow regime-backed candidates in pseudo-democratic elections, but not strong enough to overthrow the regime itself, the rational choice for the dictator is democratic backsliding. Only when the social movement is strong enough to threaten the survival of the regime itself will the original thesis of Acemoglu and Robinson (2000) – the elites embrace democratization to avoid social unrest – kick in. For electoral authoritarian regimes, democratic backsliding is more likely than democratization, simply because the latter requires a higher bar of collective action.

Authors

• TING CHEN, Hong Kong Baptist University

• Li Han, Hong Kong University of Science and Technology

• Tao Li, Hong Kong Metropolitan University