Undocumented immigration from Mexico into the United States : modeling spatial structure of push factors at the municipio level

Includes bibliographical references (p. 23-24)In order to identify the underlying factors associated with undocumented Mexican migration into the United States traditional and spatial statistical applications were applied to explain and model the phenomena. In this study data from the U.S. Border Patrol was used to create a migration propensity index (MPI) which is a ratio of detainees from each municipio in Mexico. MPI was then related to data from Mexican Census and sociological studies to identify the variables associated with undocumented migration. By applying statistical methods such as an; principle component analysis, ordinary least squares regression, and geographically weighted regression the underlying factors and spatial structure of the variables associated with undocumented immigration were able to be identified. The final model incorporated a geographically weighted regression that was able to identify seven unique regions in Mexico, each with unique variable coefficients associated with MPI. This provides evidence that the factors associated with undocumented migration into the U.S. are not homogeneous and that different regions throughout Mexico send migrants to the U.S. for different reasons