A method is described for estimating the impact of a parasitoid on the abundance of a nontarget host, using the intrinsic rate of host increase, the average abundance of the host in the presence of parasitism, and the estimated mortality caused by the parasitoid. The method is applied to the braconid Microctonus aethiopoides Loan, introduced to New Zealand to control Sitona discoideus Gyllenhal in lucerne but also attacking native weevils Irenimus spp. and Nicaeana spp. The nontarget host population was modeled using discrete Ricker or continuous logistic models, tuning the models to host population data in the presence of parasitism, then removing parasitism and determining the increase in predicted equilibrium host density. In an area where up to 30% parasitism of a nontarget host population has been recorded, the model estimated an 8% reduction of the nontarget host. In another area, where the parasitoid has not established, the method was applied in reverse to predict the parasitoid’s impact if it did establish. In this case, the model predicted a 30% suppression of population density. The host’s intrinsic rate of increase, r_m, accounts for this difference in predicted impact, which was small in the low altitude area where r_m was high, and the impact was larger in the higher altitude site where r_m was smaller.