Theory and practice

[unixronin]

According to Donald Luskin of Trend Macrolytics, writing in SmartMoney, Robert J. Gordon, “an acclaimed macroeconomist and professor at Northwestern University”, says the recession is over.  He cites as evidence for this that there appears to have been a peak on jobless claims, and asserts that “in every recession since 1974, the peak in jobless claims came within weeks of the bottom of the recession.”

One notices, however, that this excludes the 1970-71 recession, which had a double peak.  And in the 1980 recession, the peak of jobless claims came slightly before halfway through the recession.

But there’s another problem with this theory.  In this graph, the best example of it comes in the 1982-83 recession, where the jobless rate peaked, then dipped ... then rose again to higher levels, and peaked, and dipped ... then rose again to still higher levels.  At neither of those two interim dips — or even after the third peak — would it have been possible to say with certainty that the jobless rate had hit its highest point and was now headed monotonically back down.

So why does Professor Gordon think he can predict that now?  Yes, the jobless rate has dipped.  Unless he has a crystal ball or a time machine, he can have no way of telling whether it will rise again.  He could be right, and it would be nice if he’s right.  But at the moment, it’s a premature judgement.  Time will tell whether he’s right or wrong.

Comments

Speaking from the Heartland perspective

[luxobscura.livejournal.com]

I don't think most people could care less at what exact moment in time the recession technically recedes. They want jobs. That's really the only thing that matters.

But I guess if you have "PhD" after your name, you have to while away the time somehow, usually in ways that have absolutely no tangible value.

[rbos.livejournal.com]

Those numbers look a little suspect, right off. They're not in "per 100,000 population", they're a numerical total that doesn't correct for population growth.

[ithildae.livejournal.com]

Economists have successfully predicted nine of the past five recessions. It gives them something to do with their otherwise productive hours.

[vetpsychwars.livejournal.com]

Has he compensated for changes in how unemployment is measured? I seem to recall such a change in the Reagan era.

[jmkelly.livejournal.com]

So glad to get this news. Please have Professor Gordon call my employer and order up a few months' worth of work for us.

[unixronin.livejournal.com]

Seriously....

[johnkzin.livejournal.com]

So, we have a correlation between the fall in initial jobless claim rates and the end of recessions... wow, this guy's a genius.

But, does it make more sense that the fall in initial jobless claim rates predict the end of the recession ... or that the end of the recession predicts the fall in initial jobless claim rates? Just because one predicts the other, doesn't mean the converse is true. It makes much more sense that the end of a recession predicts a fall in initial jobless claim filing rates (as the improving economy ends a need for layoffs, and may even allow companies to start hiring again) ... than it makes for a fall in initial jobless claim rates to suddenly cause the economy to bounce back.

In fact, I might go so far as to say that no only is the end of the recession a predictor of the fall in initial jobless claim rates, but it is probably also the CAUSE of the fall in initial jobless claim rates in those events that correlate with the ends of recessions. But also note that the graph plainly shows plenty of downturns in initial jobless claim rates ... that are unrelated to recessions at all, and I can see at least 5 places on the graph where it looks like they happened in the MIDDLE of recessions. Therefore, a brief peak in initial jobless claim rates is not a predictor of the end of a recession, it's just noise (with the respect to recessions).

I think this economist needs to learn a few things about cause and effect, correlations, and how to analyze what events predict other events.

[unixronin.livejournal.com]

I can see you're about as impressed with his reasoning and rigor as I was. ;)

In addition to all the above, not until this is over will we know for sure when the recession actually ends and where the highest peak of new unemployment starts is. You can't correlate events meaningfully when you don't know whether they've occurred yet. As ithildae pointed out, economists have correctly predicted nine of the last five recessions. Predictive economics is not a mature science, and if there's one thing US economic history has shown us, it's that economists and financial analysts don't fully understand the problem yet.