Origin community networks and diffusion of remittance behavior among Mexico-U.S. Migrants

Filiz Garip, Harvard University

Remittances from migrants constitute an increasingly important component of livelihoods in developing countries, and of scholarly debate in the literature. Prior research suggests altruism, insurance or investment as potential motives for migrant remittances. This paper suggests an alternative view: Remittances may result from social interactions in origin community networks, and migrants may be influenced by other migrants' behavior. Using longitudinal survey data from the Mexican Migration Project, this paper analyzes the remittance flows between Mexico and the United States from 1970 to 2000, and argues that the increase in the amount of remittances over time can be partially attributed to peer effects. To separate peer effects from other confounding influences, the paper utilizes an identification strategy outlined by Manski (1993), and introduces nonlinear and lagged measures of aggregate group behavior. Preliminary findings suggest that, even controlling for the most important determinants of remittances suggested by prior work, migrants' behaviors are strongly correlated with aggregate remittance patterns in the origin community. These results seem robust to the addition of community fixed-effects. These encouraging findings indicate the potential cumulative nature of remittance behavior, and invite the application of additional analysis to determine future predictions and policy directions.