Does a leader's self-assessed integrity matter?

We investigate the impact of leaders’ integrity on group behavior in a single-shot collective action game. Leaders have exclusive knowledge of their group projects and lead via cheap-talk investment suggestions. Followers observe the suggestion but not their leader's actual decision. Leaders can sometimes benefit from deceiving their followers (suggesting invest without actually investing themselves). In our Baseline condition, followers only observe their leaders’ suggestions. We have three treatments: Integrity, followers additionally observe leaders’ self-assessed integrity category; Random, followers additionally observe a randomly chosen group member's integrity category; Driving, followers additionally observe leaders’ self-assessed driving skill category. On average, the likelihood that leaders follow their own investment suggestions in the treatments does not differ significantly from the Baseline. Likewise, on average, the frequency at which followers invest, having observed an invest suggestion, does not differ significantly in the treatments from the Baseline. When an integrity categorization is revealed, the category revelation influences followers’ investment decisions. Followers receiving invest suggestions from leaders categorized as average/low integrity invest at approximately half the rate as followers receiving invest suggestions from leaders categorized as high integrity. Revealing that a random member is categorized as average/low integrity, however, does not negatively affect followers’ investment rate.