Invasions of nonindigenous organisms have long been linked to trade, but the contribution of individual trade pathways remains poorly understood, because species are not observed immediately upon arrival and the number of species arriving annually is unknown. Species interception records may count both new arrivals and species long introduced. Furthermore, the stock of invasive insect species already present is unknown. In this study, a state-space model is used to infer the stock of detected as well as undetected invasive insect species established in the United States. A system of equations is estimated jointly to distinguish the patterns of introduction, identification, and eradication. Introductions of invasive species are modeled as dependent on the volume of trade and arrival of people. Identifications depend on the public efforts at invasive species research, as well as on the established stock of invasive species that remain undetected. Eradications of both detected and undetected invasive species depend on containment and quarantine efforts, as well as on the stock of all established invasive species. These patterns are estimated by fitting the predicted number of invasive species detections to the observed record in the North American Non-Indigenous Arthropod Database. The results indicate that agricultural imports are the most important pathway of introduction, followed by immigration of people. Expenditures by the U.S. Department of Agriculture and the Agricultural Research Service are found to explain the species identification record well. Between three and 38 invasive insect species are estimated to be established in the United States undetected.