China has been severely criticized on its poor performance in public sectors although it has achieved great economic performance since 1980s. To interpret this paradox, economists tend to emphasize the key role of local government competition by relying on the Tiebout Model. On the other hand, political scientists and sociologists prefer to address the role of sectors other than the government and the power of voice. In this paper, we construct a synthesized game theory model based on the “exit-voice” theory proposed by Hirschman (1970), combining both the “vote with your feet” mechanism and the “vote with your hands” mechanism to identify the critical conditions that decide the equilibrium. According to the model, the possible outcomes depend on the relative cost/benefit of exit and voice. Specifically, the model suggests that exit would become the equilibrium strategy when its cost is lower than that of voice. Yet, exit would turn to be the last resort after voice if voice cost is lower than exit cost. In other words, exit and voice are substitutes for each other when voice cost outrides exit cost, otherwise they are complementary. However, if the cost of exit is so high that it discredits the threat of exit, then keeping silence would become the equilibrium outcome. Based on the model, we further explore the mechanisms of voice in promoting local public governance when voice becomes the equilibrium strategy. We employ the concept of co-production (Ostrom, 1996) to the whole process of local public governance, including public policy making, public goods provision, as well as performance evaluation and monitoring public administration, and propose an integrated framework to capture the indispensability of the residents’ contributions in updating the performance of public sectors. We argue that the contributions of residents are indispensable to the following aspects: first, in public policy making because they possess important local knowledge which is necessary for high quality policy making; second, in public goods provision because the inputs of residents are essential for the successful delivery of most of the public goods; and third, in the performance evaluation and monitoring public administration because residents possess the most valuable information that is necessary for performance evaluation and they can substantially improve the detection of local government’s misbehavior by involving countless potential monitors. Using internet penetration rate as a proxy variable, we test the effect of voice on anti-corruption by using China’s provincial panel data from 1998 to 2010 with a GMM model. We identify a significantly negative relationship between internet penetration rate and corruption level after controlling for important variables indicated by previous literature. We thus argue that there is both theoretical and practical significance in combining “vote with your feet” with “vote with your hands”. In theory, how to define the role of the citizen in a society is one of the central questions of the modern theory of democracy and the public administration. In this paper, by proposing the citizen as co-producer, we show how the role of the citizen could be expanded in public governance. From the perspective of co-production, citizens are not only entrustors, voters, clients, consumers or even volunteers, but also active and indispensable subjects of rights who could make constructive contributions to the welfare of the society. Nevertheless, it is worthy of noting that the power of voice relies heavily on the threat of exit. On the other hand, moving toward theoretical synthesis indicates that we should balance exit and voice when considering relevant institutional designs in practice. Typically, in order to strengthen the power of voice, we should place voice before exit, support voice with the threat of exit, and commit ourselves to designing institutions that can improve the public’s willingness to voice and reduce the cost of voice.