Cartel operations still exist worldwide despite the recent development and enforcement of antitrust laws in scores of countries that previously had no such legal framework. Since 1990, 283 international cartels were discovered by antitrust authorities around the world (Connor 2003). According to a 1990s' sample of U.S. Department of Justice and European Commission prosecuted cases, the discovered cartels operated in a wide variety of industries such as chemicals, metals, paper products, transportation, communication, food, textiles and services (OECD 2002). Were it not for the efforts to deter cartels by the antitrust agencies, there would likely have been many more such arrangements (Connor 2005). The presence and societal impacts of domestic and international private cartels is an important policy issue that has drawn the attention of various key public stakeholders including justice, trade and commerce departments, policymakers, and industry and consumer advocacy groups. Because only the detected cartels are observed, it is difficult to measure the impact of antitrust policy on collusion (Harrington 2006). One central issue is then concerned with whether antitrust agency’s spending on cartel detection is optimal to dissolve cartels under market uncertainty. Therefore, the purpose of this paper is to examine theoretically the optimal spending by the antitrust agency when cartel’s profit is uncertain, link the theoretical findings with the observed real world cases, and to provide useful antitrust policy implications. Cartel firms operate in an uncertain environment, facing market demand uncertainty and the possibility of being detected by the antitrust agency. We develop a dynamic model in which an incumbent cartel decides whether or not to dissolve the cartel based on the observed profit and the antitrust agency’s budget level used for cartel detection. At the beginning of each period, the cartel observes its potential short run profit and the budget level that the antitrust agency allocates on cartel detection, and decides whether to keep the cartel or dissolve it. By keeping the cartel in operation, there is a probability that the cartel is detected and hence penalized. By dissolving the cartel, the firms receive competitive profits in the current and future periods. The Bellman equation is solved using the Newton’s method and based on the given parameters, we found the optimal values and the optimal actions when different levels of cartel detection budget allocation are observed. The post-optimality analyses show that the value of the cartel is an upward-sloping function of the profit for any antitrust budget. The critical cartel profit below which the cartel dissolves is calculated when the antitrust agency’s budget allocation increases each year. The higher the observed antitrust budget, the lower the value for the cartel and the higher the profit required to sustain the collusion. Obviously, governments can set budgets sufficient to deter cartels. Our paper sheds light on the effectiveness of current spending levels in assessing whether antitrust enforcement is effective enough to limit cartel formation to an acceptable social level.