Models for short-range forecasts differ from those for intermediate and long-range forecasts because of the possibility of introducing lagged exogenous factors as explanatory variables. It is widely believed that certain exogenous variables, in particular estimates of state income, are useful leading indicators of migration rates. In this paper, panel-data (or longitudinal-data) models are used to represent the relationship between destination-specific out-migration and several explanatory variables. The introduction of this methodology into the migration literature is possible because of some new and improved databases developed by the US Bureau of the Census. In this paper, data from the Bureau of Economic Analysis are used to investigate the incorporation of exogenous factors as variables in the model. One aim in the paper is to use graphical techniques in the understanding of the relationships between migration and these exogenous factors. These techniques provide insight into the strong relationships suggested by the many gravity models that have appeared in the literature. However, when one also includes additional parameters that are estimable in longitudinal-data models, it turns out that there is little additional information in the exogenous factors that is useful for forecasting.