Friday, February 15, 2013

Compared to What?

There’s an older (slightly offensive) acronym SSDD. It’s used for situations in which the same mistake is made over and over again.

When a macroeconomist digests what the media or politicians say, this acronym comes to mind a lot.

As in the article entitled “In Shovels, A Remedy for Jobs and Growth” by Eduardo Porter in the February 13 issue of The New York Times. Porter is a reporter, so he can be forgiven for repeating an official viewpoint:

At the end of last year, according to the nonpartisan Congressional Budget Office, the economy was still about 5.5 percent smaller than it would have been had it avoided the recession and kept growing along its long-term potential path,

Ah … see … actual real GDP compared to potential real GDP. The issue is: how do we get potential real GDP?

There are a number of ways, but the important points are that there is more than one way, and all claims about the size of that gap are based on choosing one method and rejecting the others. But, what if you’re wrong?

Porter provides a graphic.

13-02-13 New York Times Graphic of Real and Potential GDP

Graphics look authoritative, but they don’t typically show or explain the underlying assumptions.

Here’s the problem: this shows real GDP returning to potential real GDP, and there are assumptions that produce that result. In macro, we need to question the assumptions.

What this graphic shows is called trend reversion. And trend reversion is a result that follows only from the assumptions of certain models of trend. Porter’s 5.5% claim also results from that assumption.

It doesn’t take much to note that this graphic is showing something that isn’t cool at all. Look at the 2000-1 recession. It’s not there. Look at the 2001-2007 expansion: it has a recession in the middle of it in 2003.

If your model misses stuff we know in the past that badly, why trust it to make forecasts about the future?

No comments:

Post a Comment