An ex post measure of risk reduction is commonly used in the literature to predict the potential reduction of farmers’ yield risk due to weather-based index insurance and area-yield crop insurance. In this paper, we evaluate the predictive power of the ex post risk reduction for different weather-based index as well as area-yield index and farm yield insurance contracts. We compute an empirical benchmark of potential risk reduction against which the ex post estimates are evaluated by distinguishing between a training data set and a test data set. Based on data for 40 wheat producers in Kazakhstan, our empirical analysis shows that the ex post approach can overestimate farmers’ future risk reductions due to crop insurance schemes based on weather indexes or area yields. Therefore, we argue that the decision about the market launch of index-based insurance instruments should be based on more than just the common ex post approach.