In this paper I develop an analytically solvable and structur-ally estimable economic geography model and apply it to predict migration flows for the period following the CEE's integration with the EU. The main innovation of my ap-proach is that it endogenises both, explanatory variables and the migration rate. Model's parameters are estimated econometrically using a structural equation, which is de-rived entirely from the theoretical NEG model. My empirical findings advocate that there is enough evidence to predict a selective migration among the three Baltic States. However, labour mobility in the Baltic countries is sufficiently low to make the swift emergence of a core-periphery pattern very unlikely at this geographical level.