We test different classic migration theories by using incentivized laboratory experiments to investigate how potential migrants decide between working in different destinations. We test theories of income maximization, skill-selection, and multi-destination choice as we vary migration costs, liquidity constraints, risk, social benefits, and incomplete information. The standard income maximization model leads to a much higher migration rate and more negative skill-selection than occurs when migration decisions take place under more realistic assumptions. The independence of irrelevant alternatives assumption mostly holds when decisions just involve wages, costs, and liquidity constraints, but breaks down once we add risk and incomplete information.