The U.S. crop insurance has two distinct features that set itself apart from insurance in other areas: (i) it is explicitly subsidized with an average premium subsidy rate of about 60 percent in recent years; and (ii) the law requires the premium rate be set at actuarially fair level with the federal government paying the administrative and operational costs related to the sale and service of insurance policies. Bearing in mind these features, we examine to what extent farmers’ crop insurance choices conform to economic theory and estimate the implications of changes in premium subsidy structure. A standard expected utility maximization framework is set up to analyze the trade-offs between higher risk protection and larger subsidy payment. We show that, given actuarially fair premium, a rational farmer will choose the coverage level with the highest premium subsidy or a higher coverage level. With a large insurance unit level data, we fail to find empirical support for this theoretical results, which suggest a possible “anomaly” in insurance decisions. Estimation through mixed logit models reveals that out-of-pocket premium has a negative impact on the probability of an insurance product being chosen.