Third-party security guarantees are one of the most prominent techniques mediators and other third parties can utilize if they seek to enhance the prospect for peace (Walter, 2002). Yet, one of the main theoretical puzzles of the civil war termination process is why belligerents believe in thirdparty promises to guarantee peace. Third parties do not necessarily live up to their commitments. For instance, a peace agreement was reached in the civil war in Afghanistan in 1996, called the Islamabad Accord. The agreement was brokered by Pakistan, and also sponsored by Saudi Arabia and Iran. The agreement stipulated that a joint commission comprising representatives of the Organization of the Islamic Conference should be formed in order to monitor the ceasefire and the cessation of hostilities. However, the ceasefire commission never materialized, the ceasefire broke down and the conflict continued unabated.1 Yet, in other cases, third parties have followed through on their promises despite the fact that acting as guarantor is a costly endeavor. For example, in the final stages of the devastating civil war, the United Nations (UN) decided to mitigate the fears of the belligerents, Frelimo and Renamo, by promising to verify and monitor the ceasefire and the demobilization process stipulated in the peace agreement. The international engagement was one of the driving forces behind the combatants’ decision to end their conflicts through a settlement in 1992. After the deal was reached, the UN also implemented its promise to guarantee the agreement, which has been identified as one important reason why the peace in Mozambique became durable, and the case turned into a success-story of peacemaking in civil wars (Schneidman, 1993; Synge, 1997; Walter, 1997). This variation in the guarantee-giving process leads to the following research question: under what conditions are third parties credible as guarantors of peace in civil wars?