Since the early 1990s, Mexican immigrants to the United States have increasingly settled in non-traditional locations outside those (such as California and Texas) with historically high Mexican density. While this diffusion of the Mexican unauthorized population is well documented, the reasons are not well understood. We assess a key hypothesis of immigrant location choice particularly relevant to the current policy debate – intensity of enforcement at the U.S.-Mexico border. This hypothesis, advanced primarily by Massey, Durand, and Malone (2002), is that increased border enforcement in traditional migrant crossing areas led immigrants to choose alternative border crossing routes, and in turn to choose non-traditional destinations. While sizeable investments are made in regulating cross-border flows, little is known about how these investments impact the destination choice of immigrants. However, where immigrants settle is a critically important “input” to a wide range of policies at the national and local level. The hypothesis is difficult to test because immigrant location decisions may respond to unobservable characteristics of destinations that are correlated with border enforcement efforts. The slowdown in Mexican immigration to the U.S. and changes in state immigration policy, in particular, also complicate analysis of border policy. Moreover, lack of data on immigrant crossing location makes it difficult to know the border enforcement intensity faced by potential immigrants to different U.S. destinations. We overcome these problems by relying on historical border crossing and settlement patterns among Mexican immigrants and by instrumenting for border enforcement with lagged growth rates of unauthorized immigrant apprehensions along the U.S.-Mexico border.