Monday, January 28, 2008

Battle of the Recession Graphs!

According to my former student, friend and co-author Mark Perry (and endorsed by N. Gregory Mankiw), The job market is strongly signaling that we are not yet close to a recession.

Here's Mark's graph:




On the side of gloom and doom comes Floyd Norris of the NY Times arguing that the job market shows we are likely already in a recession.

Here's his graph:




Looking at the NY Times graph, it's pretty clear the prediction rests on joblessness continuing to rise strongly. The implication seems to be that hitting 13% means the rise to 30% and above is pretty much inevitable. Looking at Mark's, it's pretty clear that the prediction rests on January's new jobless claims being a good predictor of Feburary's March's and April's. I'd hate to think we were one bad month away from a recession!

Which graph do you like best? and why?

7 comments:

Shawn said...

well, the one on the top has colors, so it's clearly more betterer.

Anonymous said...

It would be interesting to see the jobless claims in per capita terms

John Thacker said...

I dislike the second graph because it leaves out the non-recession years, although since it claims that "such increases have always been associated with recessions," that would perhaps reinforce its point. OTOH, showing that there were quite a few incidents where 10-12% increases didn't lead to recessions would undermine the point, were that the case. I dislike the omitted data.

The key disagreement is between raw numbers and the year over year percentage. What's more important, a 13% jump from 4.5% to 5%, or the fact that 5% is still low by historical standards? Since recession is a number about growth, the percentage change may seem more important than the raw numbers. It doesn't matter if right now is still historically good if you're asking whether things are getting worse. Yet on the other hand, a high employment rate means that much less productivity growth is needed to still see net GDP growth.

Angus said...

Anonymous: if you follow the link to mark perry's post and read the comments, one of the commenters does exactly that (adjusts for the size of the labor market).

John Thacker said...

Oh, I should say that I also strongly dislike the first graph for using the nasty "cut off the bottom of the graph to make the differences between the bars look larger" technique. Way to make January 2008 look less than half the size of the others, despite being more like seventy-five percent.

Just a Thought said...

It takes 6 months before you can say any period of time is truely a recession (by definition) so both of these are premature.

Tom said...

Mark's graph has a weird selection of dates, showing gaps of 18 months, 9 years, 11 years, then 7 years. Perhaps there's an explanation, but it begs a question about cherry picking. I agree with John Thacker on the cut off bottoms issue.

This doesn't mean I like the NY Times graph, though. A 13% increase from a rate of 4% gives us 5.2% -- things just getting back to "normal." But a 13% increase from 13% (paging Jimmy Carter!) would put us near a disastrous 17%. Focusing on this figure is viewing the economy through a keyhole.

So... what's the latest on Britney?