Over the past few decades, collaboration has flourished in the public administration and policy fields as a rational means to solve complex issues and improve public service performance. Through a meta-analysis of 26 studies with 251 effect sizes, this investigation provides novel perspectives for understanding the effects of different collaborative partnerships on performance. To test these mechanisms, we applied various social science theories, such as institutional theory, resource dependence theory, a resource-based view, and transaction cost theories. Our findings indicate that the overall effect of collaborative performance is positive and significant. Moreover, meta-regression results show that public–public collaboration results in better performance than public–nonprofit or public–business collaboration, while involving all three entity types in collaborative efforts yields similar outcomes to public–public collaboration. Several implications of these findings are outlined for researchers and practitioners.