The Campaign Spot

Did GDP and the Unemployment Rate Become De-linked?

The Tuesday edition of the Morning Jolt features signs you may be a Rom-neck, the worst month for the national deficit ever, and then this curious development in the economic numbers:

GDP, Unemployment Rate… You Two Used to Be So Close. How Did You Grow Apart?

In the Wall Street Journal, Jon Hilsenrath asks:

Many economists in the past few weeks have again reduced their estimates of growth. The economy by many estimates is on track to grow at an annual rate of less than 2% in the first three months of 2012. The economy expanded just 1.7% last year. And since the final months of 2009, when unemployment peaked, the economy has expanded at a pretty paltry 2.5% annual rate.

How can an economy that is growing so slowly produce such big declines in unemployment?

And now, Morning Jolt readers, I will surprise you by quoting some liberal bloggers. No, really, stay with me, because I think they articulate what you’re going to hear a lot of in 2012: a sense that the economy as we experience it does not match up to what the official numbers say.

Ezra Klein:

Something odd is happening in the economy. Jobs are coming back, and relatively quickly. But growth is lagging. Or, at the least, we think it is. Virtually every estimate of GDP growth for the first quarter of 2012 is below two percent — that’s a third lower than it was in the fourth quarter of 2011, when payroll growth was lower — and many of those estimates are being revised downward as new data streams in.

A couple of things could be going on here. One possibility is that the preliminary GDP data is wrong. That happens. In the fourth quarter of 2008, the early GDP data said the economy shrunk by 3.8 percent. Later on, we learned the real number was closer to nine percent. A smaller, more positive discrepancy might explain this riddle, too… Then there’s the possibility that the previous three months of job growth turn out to be a tease, and the recovery will falter in the middle of the year. Call that the “2011 scenario”: Back in February, March and April of 2011, payrolls rose by an average of 239,000 jobs a month. In May, June and July, that fell to an average of 78,000 a month. So far, this economy has not been kind to those who try and extrapolate a self-sustaining recovery from a few months of strong job growth.

David Dayen, writing at FireDogLake:

It could be that GDP data is just wrong, and will get reassessed upwards. That’s the optimistic scenario. Or it could be due to slower productivity growth, meaning that businesses must hire more workers to perform the same amount of tasks. Christina Romer believes that businesses “overfired” during the low points of the economy, and are now compensating by hiring to satisfy current demand (this may partly explain her prediction back in 2009 that unemployment would top out around 8%, also based on Okun’s Law and GDP projections at the time). Jared Bernstein says that trend growth is actually around 1-2% now, not 2.5%, and so Okun’s Law is actually working.

These opinions may explain the discrepancy, but they also portend bad signs for the economy going forward. Because with either explanation, it would mean that job growth will soon stall out, either as businesses finish rebalancing to meet demand, or as the labor force returns to normal growth. 

A lot of people will read this and go, ‘ah-ha, this means the Obama administration is cooking the books on the unemployment figures!’ I’m not sure it’s quite so simple; I think they were always an imprecise measurement – they’re essentially monthly little mini-census surveys of employers and random workers – that have grown less precise as the economy changed. How many Americans are getting paid cash under the table? How do you count freelancers in a slow period?

David Stockman wrote recently, “I don’t particularly believe in tin foil hats, but all of these mainstream economists treat the BLS and BEA data like it’s holy writ—when it’s evident that the reports are so massaged, estimated, deemed, revised, re-bench marked and seasonally adjusted that any month-to-month change has a decent chance of being noise. What deep secret might they be hiding? … the mainstream narrative never gets to the trend. In this case,  the plain fact is that we are warehousing a larger and larger population of adults who are one way or another living off transfer payments, relatives, sub-prime credit, and the black market. 

For amusement, Stockman found a fascinating example of the numbers seeming… a little too stable, year by year:

Since 2000, the January job loss against a December payroll of between 130 and 135 million has varied within a tiny range of about 150,000. Other than January 2009 when the economy was being smacked by the post-Lehman melt-down in the financial markets, this means that the unadjusted January payroll count declined within a super-tight range of 2.00% to 2.20% of the December payroll.

Really? Granted the U.S. economy is a regular fellow, but how could there be such astounding uniformity every January, year after year in the raw numbers, as in the following sequence for January 2001 thru January 2012, respectively: 2.16%, 2.19%, 2.05%, 2.03%, 2,03%, 1.96%, 2.03% 2.19%, 2.73% (2009 outlier), 2.20%, 2.18% and 2.02%.

After all, you have weather aberrations, huge fluctuations in year to year economic conditions, the weak, random nature of the establishment survey, the constant fiddling with the birth-death adjustment which is carried in the raw numbers, the Christmas shopping season variation from red hot-to-punk across the years, the timing of the survey week and much more. And the dice always lands on almost exactly a 2.03% change from December. Right.

This is meant to be a long-winded encouragement to you to apply your patented numbers forensic skills to the monthly BLS reports or any of the other market movers. In the last 7 years, for example, the Christmas shopping season has been all over the lot and presumably, retail hiring, too. But the unadjusted retail jobs reduction in January vs. December has not varied by much more than 150,000 from a base count of 15 million.  That’s a 1% variation, notwithstanding the huge shopping season differences they report on bubble vision.

Pretty weird, huh?


The Latest

Rat Patrol

Rat Patrol

Illegal leaks of classified information should be treated as a serious offense. But they would be easier to prevent if less information were classified.