Tuesday, July 6, 2010

Linear Projections

FLG is going to leave aside everything in this post by Matt Yglesias besides one line in this graph, which it appears Matt took from Econbrowser:
FLG knows that the CBO is filled with bright, competent, and generally disinterested people, but it always worries him when a projection looks linear, as the potential GDP appears here.  Econbrowser indicates the source as "potential GDP as estimated by CBO in January 2010," but FLG hasn't been able to find the actual source or a description of how the estimation was made.   He'll keep looking.

Maybe this is all standard stuff that is taught in Grad 101 macroecon, but FLG doesn't like its linear nature.  Linear always make FLG think a few things.  First, very short time frames, like when estimating a complicated function using a tangent line close to the point of tangency, but that doesn't seem to apply here.  Second, the person just assumed linearity because they're dumb and only know basic algebra.  That doesn't seem to apply.  Third, people assume linearity in a kind of run-home-to-momma because they don't have anything better.  FLG worries that third one is the case and we're having a public policy discussion based upon it.

Update:  The blue line is log GDP, which makes sense.  GDP grows exponentially, which makes comparison over time difficult.  So, the log is then used to linearize it.    Perhaps the same thing is happening with the potential GDP, but it looks too much like a line to make FLG want to trust it.  If that makes sense.

1 comment:

Thomas said...

GDP is historically roughy linear over a handful of years, why wouldn't a projection show that?

Creative Commons License
This work is licensed under a Creative Commons Attribution-No Derivative Works 3.0 United States License.