While much evidence supports the pacifying role of power-sharing for divided places, less is known about how it should be institutionalized. My dissertation addresses this gap. In particular, I focus on the distinction between two institutional types of power-sharing: On the one hand, corporate power-sharing, based on government quotas for specific ethnic groups, veto rights for their representatives, and ethnically-based autonomous regions. On the other hand, liberal power-sharing, based on low electoral hurdles for government inclusion, super-majority provisions, and national federal structures. I argue that the institutionalization of power-sharing poses difficult trade-offs. Specifically, corporate power-sharing should strongly reduce conflict risks for included groups in the short-term, as it provides tangible reassurances and reduces their grievances. However, it often includes some groups at the expense of others and visibly deviates from majority rule. Hence, it is prone to violent backlashes. Furthermore, it provides mobilizational resources to included groups and increasingly reinforces ethnic divisions. In this way, it engenders renewed conflict in the long-term. In contrast, liberal power-sharing avoids these side-effects. However, at the same time, it provides less tangible reassurances and hence only weakly reduces conflict risks. I test these expectations by relying on a novel, global dataset of institutionalized power-sharing. In addition to considering the link from power-sharing to conflict, I directly investigate its intermediate impact on mass grievances and ethnic salience. My findings offer partial support for my arguments. First, they indicate that corporate power-sharing indeed engenders inclusive practices, alleviates mass grievances of targeted groups, and reduces their conflict risks. Conversely, it also results in backlashes from other groups. Second, while liberal power-sharing also incentivizes inclusive practices, it does not exert strong effects on grievances or conflict risks. However, I find only limited support for the expected destabilizing effects of corporate power-sharing in the long-term.