The objective of this research is to use publicly available data to develop predictive models for transport mode and Mexican destination decisions for shipments from various U.S. regions. Aggregate logit models have been calibrated for three commodities: machinery, electronics, and automobiles. A profile of Mexico and its industries is presented along with a review of past efforts in freight demand forecasting. The data set of aggregate shipments used in the model estimation is comprised of origin, destination, commodity type, mode of transport across the border, and value. Destination attributes, such as population, employment, number of firms in the industry, and number of shippers and warehouses, are also included.