Delong buys into the bogus Blanchard and Leigh research

Delong is back on the multipliers are 1.5 theme.

Growth Forecast Errors and Fiscal Multipliers 
 
This paper was passed around the blogs, Simon Lewis from England started the circulation. The paper is bogus. The authors use the time honored Romer prescription of selecting economic events that are special and separating them out as isolated instances. There is no better way to confirm one's prior beliefs. One cannot make that separation and still manage to control for exogenous events. The authors committed a basic statistical error. They have to compute their objective function over the aggregate collection of events, then separate out the classes according to the function. They separate first and compute second, a fundamental error and one that we find when an economist has priors to confirm.

 This is the old Keynes is dead debate, what are government spending multipliers really?  The problem with stimulus is that there really is a stimulus theory.  We will never arrive at the true stimulus theory because the researchers have only one goal, government multipliers must always be 1.5. The opposing economists spend so much time debunking this, and there are few economists left who can get at the facts objectively.  Where is Jim Hamilton when we need him?