Suitably Flip

Mixed Job Data and the Recession That Still Ain't

Following the unexpected increase in jobs revealed earlier this week in the ADP report, a lot of folks were optimistic that today's employment situation report from the Labor Department would similarly outshine expectations.

Well, it did and it didn't.

The report showed a loss of 51,000 jobs in July, but that was better than economists' expectations of a loss of 75,000.  On the other hand, the unemployment rate ticked up from 5.5% to 5.7%, beyond expectations of 5.6%.  (Of course, it's worth noting that President Clinton - who, we've been assured, was some kind of prosperity genie - saw monthly unemployment at this level or higher on average more than three times per year during his administration.)

So how can the economy add more jobs (lose fewer jobs) than expected while the unemployment rate rises more than expected?  Since unemployment only looks at active job seekers, it suggests more people entered the labor force in July than expected (and that those new workers found it unusually difficult to find jobs).  This is a trend that's been suggested by the last couple months' worth of data and is not likely wholly unrelated to the fact that a 12% increase in the minimum wage kicked in during the month in question.

The timing of the release of the July employment report (just one day after the latest report on GDP growth, which showed a slightly negative reading in the revision to fourth quarter 2007) has a lot of people puzzling over whether we are indeed in a recession (and if so, if it's merely as bad or infinitely worse than the Great Depression).

Of course it's neither.  If you like the "two quarters of negative GDP growth" metric, it's plain to see we're not in a recession, unless you've summarily decided the growth estimates for the first half of 2008 will be revised downward hugely in the future (significantly more than the 4Q revision).  The official recession arbiter, the NBER, defines a recession thusly:

A recession is a significant decline in economic activity spread across the economy, lasting more than a few months, normally visible in real GDP, real income, employment, industrial production, and wholesale-retail sales.

"More than a few months" still requires this blended contractionary condition to infect more than an isolated quarter, and the decline-qualifier "significant" suggests that a particularly small decline (for instance, the smallest decline that would even register, such as the one now observed in last year's fourth quarter) wouldn't even trip the NBER trigger if it persisted for "more than a few months" (which appears not to have).

A post up yesterday at Econbrowser (via Odd Numbers) makes use of a recession indicator index of the blogger James Hamilton's own formulation.  The numbers crunched in his algorithm (which uses only public data and the methodology for which is explained using enough Greek characters that you know it's legit) yield an index that's been remarkably accurate in mirroring the NBER's retroactive designation of recessions over the last 40 years.  The chief complaint about Hamilton's index seems to be that it's a significantly lagging indicator (i.e. by the time it recognizes a recession, the implications on policymaking, investing, etc., may be less useful since the event it's identifying will be months-old).

But in this case, where we've got a potentially troublesome revision to a quarter that's more than half a year old (and some folks are using that as the basis of an argument that a recession did indeed take hold months ago), Hamilton's indicator takes on new relevance.

If you were to scream "Recession!" only when Hamilton's index exceeds 80%, you wouldn't miss any since 1968, nor would you register any false positives.  For the most recent quarter (in this case, 1Q08), it reads a measly 38.4%.

That's the highest reading since the mild recession that began at the end of the Clinton administration, but it's nowhere near typical recession territory.

One has to notice that the indicator's peaks have been trending lower over the last few recessions, but so has the actual severity of American recessions.  The astoundingly good correlation between the index and NBER-designated recessions strongly suggests that (while it's still possible for the U.S. to enter one during this business cycle) we weren't in one by the end of the first quarter.  And with our brand new advance second quarter growth reading of 1.9% from the Commerce Department, it seems particularly unlikely that we entered a recession during the three month period to which Hamilton's index is blind.

Comments

Mixed Job Data and the Recession That Still Ain't

Following the unexpected increase in jobs revealed earlier this week in the ADP report, a lot of folks were optimistic that today's employment situation report from the Labor Department would similarly outshine expectations.

Well, it did and it didn't.

The report showed a loss of 51,000 jobs in July, but that was better than economists' expectations of a loss of 75,000.  On the other hand, the unemployment rate ticked up from 5.5% to 5.7%, beyond expectations of 5.6%.  (Of course, it's worth noting that President Clinton - who, we've been assured, was some kind of prosperity genie - saw monthly unemployment at this level or higher on average more than three times per year during his administration.)

So how can the economy add more jobs (lose fewer jobs) than expected while the unemployment rate rises more than expected?  Since unemployment only looks at active job seekers, it suggests more people entered the labor force in July than expected (and that those new workers found it unusually difficult to find jobs).  This is a trend that's been suggested by the last couple months' worth of data and is not likely wholly unrelated to the fact that a 12% increase in the minimum wage kicked in during the month in question.

The timing of the release of the July employment report (just one day after the latest report on GDP growth, which showed a slightly negative reading in the revision to fourth quarter 2007) has a lot of people puzzling over whether we are indeed in a recession (and if so, if it's merely as bad or infinitely worse than the Great Depression).

Of course it's neither.  If you like the "two quarters of negative GDP growth" metric, it's plain to see we're not in a recession, unless you've summarily decided the growth estimates for the first half of 2008 will be revised downward hugely in the future (significantly more than the 4Q revision).  The official recession arbiter, the NBER, defines a recession thusly:

A recession is a significant decline in economic activity spread across the economy, lasting more than a few months, normally visible in real GDP, real income, employment, industrial production, and wholesale-retail sales.

"More than a few months" still requires this blended contractionary condition to infect more than an isolated quarter, and the decline-qualifier "significant" suggests that a particularly small decline (for instance, the smallest decline that would even register, such as the one now observed in last year's fourth quarter) wouldn't even trip the NBER trigger if it persisted for "more than a few months" (which appears not to have).

A post up yesterday at Econbrowser (via Odd Numbers) makes use of a recession indicator index of the blogger James Hamilton's own formulation.  The numbers crunched in his algorithm (which uses only public data and the methodology for which is explained using enough Greek characters that you know it's legit) yield an index that's been remarkably accurate in mirroring the NBER's retroactive designation of recessions over the last 40 years.  The chief complaint about Hamilton's index seems to be that it's a significantly lagging indicator (i.e. by the time it recognizes a recession, the implications on policymaking, investing, etc., may be less useful since the event it's identifying will be months-old).

But in this case, where we've got a potentially troublesome revision to a quarter that's more than half a year old (and some folks are using that as the basis of an argument that a recession did indeed take hold months ago), Hamilton's indicator takes on new relevance.

If you were to scream "Recession!" only when Hamilton's index exceeds 80%, you wouldn't miss any since 1968, nor would you register any false positives.  For the most recent quarter (in this case, 1Q08), it reads a measly 38.4%.

That's the highest reading since the mild recession that began at the end of the Clinton administration, but it's nowhere near typical recession territory.

One has to notice that the indicator's peaks have been trending lower over the last few recessions, but so has the actual severity of American recessions.  The astoundingly good correlation between the index and NBER-designated recessions strongly suggests that (while it's still possible for the U.S. to enter one during this business cycle) we weren't in one by the end of the first quarter.  And with our brand new advance second quarter growth reading of 1.9% from the Commerce Department, it seems particularly unlikely that we entered a recession during the three month period to which Hamilton's index is blind.