A key prediction of a large class of theoretical models is that the location of economic activity is not necessarily determined by fundamentals. To test the empirical relevance of these ideas requires a natural experiment in which a large but ultimately temporary shock dislocates economic...

The present paper reports a repeated experiment on decision making under risk where subjects have to tackle the same choice problems in several rounds. We fit a simple error model and investigate how behavior changes in the course of the experiment. Our analysis complements and extends Hey...