Friday, February 19, 2010

Estimating the stimulus, the accuracy of macro

How do macro econos measure the effect of the stimulus? Macro Advisers compare the predicted path of GDP from weighted history with the measured GDP path after the stimulus.

Going along with them, then, I must assume that GDP (money) measurements are valid right after the crash. These numbers come from the BEA, how have they been doing on the revisions of late? Late last year a revision from 3.5 to 2.2, over quarterly periods. GDP is not very self similar over this time.

I have a better instrument, miles driven, as discussed in this Calculated Risk post. From Calculated Risk, lets look at his miles driven graph, it seems very self similar over the period. If we think the stimulus nudged miles driven, then, OK.

I am looking for a match, some inventory measure mostly self correlated, and overly correlated with the stimulus.

No comments: