Abstract
This paper adapts the canonical New Economic Geography model for experimental testing of the model's behavioral assumptions by developing a finite-player, finite-horizon dynamic game of migration. Our analysis gives distinctive predictions when migration is consistent with myopic behavior (MB) and when it is consistent with sequentially rational or perfect forward-looking behavior (FB). These alternatives are tested in an economic laboratory experiment with increasing number of agents in different treatments. Results show that perfect FB loses ground against MB as the number of agents and periods increases, and this number may be surprisingly small.
Highlights
The core-periphery (CP) model, which launched New Economic Geography (NEG) as a separate field, assumed that migrants made myopic adjustments and based their migration choices on the current real wage differences between locations (Krugman, 1991b)
Østbye form solution for the inter-regional real wage differential. To adapt this model featuring an infinite number of agents to experimental testing with a finite number of subjects, we develop a finite-player, finitehorizon dynamic game of migration, in which one subject effectively represents a positive mass of population in the CP model
We study migration dynamics in the CP model of New Economic Geography
Summary
The core-periphery (CP) model, which launched New Economic Geography (NEG) as a separate field, assumed that migrants made myopic adjustments and based their migration choices on the current real wage differences between locations (Krugman, 1991b). Since the long-run equilibrium outcome under MB always coincides with the outcome of one of the expectations-driven equilibria under FB, we study a related question of how good an approximation the assumption of MB is for the equilibrium selection problem We address these questions by developing a migration framework that can be tested in an experimental setting. It would not be feasible to obtain the real wages expected by potential migrants under different agglomeration conditions in the future periods with field data, and this common knowledge of information is fundamental for our test of the behavioral assumption of FB This fact has tilted the balance towards an experimental approach. Proofs and instructions for the experiment are given in Appendix A and Appendix B, respectively
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
