How does Shielded (vs. Open) Voting Shape Collective Actions? Evidence from Decentralized Autonomous Organizations

Abstract

Voting serves as a fundamental pillar of governance, supporting collective decision-making. Different from traditional one-person-one-vote systems, blockchain technology-enabled decentralized autonomous organizations (DAOs) allow users to participate in governance via token-weighted voting. Some DAOs employ open voting where all votes cast are visible in real-time. However, given the unique characteristics of DAOs and the automated enforcement mechanisms of smart contracts, open voting in DAOs can unintentionally invite strategic behaviors, such as misaligned voting and cascade initiation. Some other DAOs implement shielded voting, concealing real-time results until the voting period ends. Shielded voting theoretically reduces tacit collusion by preventing voters from reacting to ongoing voting outcomes. This study empirically investigates the efficacy of shielded voting (vs. open voting) on key outcomes in collective actions: free-riding (measured by voter turnout and large token holder dominance) and strategic behaviors in DAOs. We leverage a natural experiment on a voting platform for DAOs. Our results show that shielded voting decreases overall voter turnout. Notably, shielded voting disproportionately affects smaller token holders, whereas larger ones with more voting power (the top 5%) are relatively unaffected, thereby exacerbating the dominance of large token holders. Further analyses reveal that shielded voting increased average voting response time per voter and per token, suggesting that shielded voting raised information search costs, discouraging voter turnout. Meanwhile, the negative relationship between shielded voting and voter turnout is attenuated by the proceeding voting distribution. In others words, a clear majority in the earlier votes suppresses voter turnout more in open voting than in shielded voting. This indicates that shielded voting increased perceived outcome uncertainty, motivating voter turnout. Moreover, we show that shielded voting influences voters' strategic behaviors by reducing misaligned voting, resulting in a more balanced distribution of votes across options. Also, shielded voting prevents cascade initiation, leading to less competition in initial voting. Our results reveal a central trade-off with shielded voting in DAOs: Shielded voting curtails strategic behaviors but significantly lowers voter turnout by demotivating smaller token holders' participation.