As regulators increasingly embrace cooperative approaches to governance, voluntary public-private partnerships and self-regulation programs have proliferated. However, because few have been subjected to robust evaluation, little is known about whether these innovative approaches are achieving their objectives and enhancing regulatory effectiveness. In the context of a federal government program that encourages companies to voluntarily self-police and self-disclose regulatory violations, we examine how participation affects the behaviors of regulators and regulated facilities. We find that on average, facilities that committed to self-police experienced a decline in abnormal events resulting in toxic pollution, and that regulators reduced their scrutiny over self-policing facilities. Upon closer examination, we find strong evidence of these effects among facilities with clean past compliance records, but find no such evidence of among facilities with more problematic compliance histories. These findings support the theoretical promise of meaningful self-policing practices and suggest that voluntary disclosure can serve as a reliable signal of future compliance-but only among a subset of facilities.