The result of this experiment would be mostly useless unless the results were stark and causation can be traced back to this policy via other methods. It doesn't control for the millions of other factors that could affect the outcome.