An Stochastic Method to Forecast the Mexican Migration

Victor M. Garcia-Guerrero, El Colegio de México

In this work the net migration by age and sex is adjusted and forecasted with the method proposed. The data source is the most recent estimation of the Mexican population according to the Census 2010, the American Community Survey, the National Survey of Occupation and Employment among others (SOMEDE, 2011). The model proposed is inspired by the previous works of Lee (1993) and Bijak (2011), and it is based in a logistic function which allows to forecast migration consistently with the population dynamic. That means the model proposed allows to get sustainable migration levels for the entire demographic dynamic. One of the most interesting results is that the best time series model for migration is a differentiated moving average random process -i.e. an ARIMA(0,d,q). It means that, with one difference-process, after q-years in the future, the trend remains constant along the forecasting period.