[pzone]

Backtesting is a poor substitute for experimentation. We can't run back history and check what happens if the Bush tax cuts never happened. We certainly can't create a thousand parallel economies, implement the tax cuts in half the sample and not in the other.

Economists use historical data to calibrate models. Of course data from the Reagan years was used to are for the Bush tax cuts. But simple backtesting can never tell you with certainty wherever your theories worked out or not.