This research examined the appropriateness of a 2-parameter model for crop insurance premium ratemaking. Besides conventional way of calculating crop insurance premium using normal curve theory, this study uses empirical crop yield's distribution to measure downside risk in approximating crop insurance premium. Statistical means for all selected crops by country in Ontario with respect to premium calculated under normal yield distribution assumption (NPREM) and premium calculated using empirical yield distribution (EPREM) are presented. With respect to NPREM and EPREM, a significant statistical difference between mean premiums by crops at various coverage levels are found. This study argue that this difference is mainly attributed to the differences in downside risk. This issue is particularly important, because rejecting a null hypothesis that NPREM and EPREM are equivalent, suggest that approximation of the true (i.e., empirical) distribution by a normal distribution may bias insurance premiums. However the key finding is that in determining crop insurance premium, the downside-risk measured relative to a normal yield distribution function does not necessarily violate research which determines crop insurance premium using empirical crop yield's distribution function.