Recent changes enacted by the Immigration Reform and Control Act of 1986 (IRCA) create the potential for two large temporary worker programs, one constructed from the existing H-2 program and the other an outgrowth of IRCA's amnesty provisions. Prior experience with guestworker programs in Europe and the United States suggests, however, that temporary labor migration ultimately will engender a flow of immigration substantially in excess of the number of temporary visas originally allocated. In this paper, we outline a theoretical rationale to explain this observation and test it using microdata gathered from former participants in the Bracero Program, a US-sponsored temporary worker program that ran from 1942 to 1964. Our results indicate that bracero migrants were very likely to make repeated trips, both with and without legal documents; that they were quite likely to introduce their sons and daughters into migratory careers; and that they were eventually likely to settle in the United States in substantial numbers, We argue that, in the long run, there is no such thing as a temporary worker program.