This study of migration in Mexico is based upon a modified Todaro approach and utilizes census data in a simultaneous equations model of 13 variables. It is unique in several ways: (a) it introduces proxy variables for employment probability and cost of migration that have not heretofore been found in the literature, but for which data are often available; (b) it illustrates that in some important cases census data have advantages not available through surveys; and (c) it finds that in some sense land reform in Mexico may have operated perversely. Finally, the study opens a new avenue for the study of informal sector growth.