Jack Rasmus

COMMENTARY: LAST WEEK’S JANUARY JOBS REPORT FROM THE LABOR DEPARTMENT HAD SOME VERY PECULIAR AND STRANGE NUMBERS, INCLUDING A MASSIVE UPWARD REVISION OF 3 MILLION JOBS DUE TO SEASONALITY ASSUMPTIONS. WAS THERE REALLY 243,000 ACTUAL (REAL) JOBS CREATED LAST MONTH, OR WAS IT LARGELY STATISTICAL LEGERDEMAIN? HERE’S SOME QUESTIONS RAISED ABOUT THE NUMBERS. (What follows is somewhat ‘wonkish’ but for those interested, do read on).

‘Those Peculiar January Jobs Numbers; Or, When 243,000 Jobs Aren’t’ by Jack Rasmus, copyright 2012

The January 2012 jobs report released by the US Labor Department on Friday, February 3 indicated that 243,000 new jobs were created in the nonfarm sector of the US economy last month. Additionally, the U-3 unemployment rate fell from 8.5 to 8.3%. How real are those numbers? Are they actual jobs created? Whats the true unemployment rate?

First, it is important to note that the 243,000 January jobs numbers are not the actual jobs created. They represent seasonal adjustments made to the raw data for jobs, referred to as the not seasonally adjusted jobs tally for the month. The January jobs report reflects an anomalous massive upward revision of the raw jobs data, due to assumptions about seasonality and new business formations.

Lets look at trends from November 2011 through January 2012 for both the seasonally adjusted and not seasonally adjusted for purposes of comparison.

The actual (not seasonally adjusted) jobs numbers for November 2011 show there were 133.179 million nonfarm jobs in the US economy that month. The following month, December 2011, total nonfarm jobs had declined to 132.952 million, for a decline of 227,000 jobs. That makes sense, given that 203,000 jobs were lost in construction, which is typical for Decembers, while 74,000 jobs were reduced by states and another-72,000 by cities and schools that month. Offsetting the construction-public sector job losses were 142,000 jobs added in Retail, mostly department stores, which also makes sense given the holiday season. Manufacturing and other service sector jobs changed little, some up and some down slightly over the month. Again, these are the not seasonally adjusted jobs for November.

What about the seasonally adjusted jobs numbers for November? One would expect some differences in numbers here, of course. Lets look. Total nonfarm jobs increased in November, by 203,000 instead of declined (per the not adjusted numbers) by 227,000. That represents a net difference and swing of 430,000 jobs.

Now lets make a similar comparison of seasonally adjusted and not seasonally adjusted for jobs between December 2011 and January 2012 that were reported on February 3, 2012 for last month. What appears is an incredible 7 to 10-fold increase in the difference between seasonally adjusted and not seasonally adjusted.

The not seasonally adjusted, raw jobs numbers show a loss of jobs for January 2012, after the holiday season, of 2.7 million jobs. That includes about 300,000 construction jobs, which is not strange given the mid-winter slowdown typical of this sector. Plus another 600,000 jobs in retail, which makes sense after the typical holiday sales hiring surge typical in November-December. And another 400,000 in business professional services as most businesses trim their labor force at the start of the year to keep costs down and to watch when and where to add jobs back in the subsequent months. However, the adjusted, upward revised numbers for January showed a gain of 243,000 jobs instead of the unadjusted 2.7 million fewer jobs. That 243,000 gain in jobs includes adding construction jobs in mid-winter, and adding even more jobs176,000in Retail and Services after the holiday season hiring surge. This retail-services job gains for January occur, moreover, despite the dismal retail sales holiday season when, except for autos, retail sales actually declined by 0.1% compared to the previous year. Why would retail employers add jobs after that poor sales season? Why would they not reduce the huge numbers of part time and temp hires of November-December in January, as they typically do after the holidaysespecially given the poor retail sales performance? And why would the construction sector add jobs in mid-winter? And why would business professional sector companies add 1.1 million jobs, according to the seasonal adjustment assumptions, instead of trimming jobs, as reflected in the unadjusted numbers? In other words, why would professional-business services not make their typical beginning-of-the-year labor force temporary reductions?

It is interesting to note that instead of a net swing between the seasonally adjusted-not seasonally adjusted numbers of 430,000, as occurred during November-December, we get a net swing, or difference, between seasonally adjusted vs. not adjusted of nearly 3 million jobs for December-January? Does this make sense? One would expect major differences between seasonally adjusted-not seasonally adjusted numbers. But a seven-fold increase in the difference from month to month–from 430,000 to 3 million–is not credible.

Lets look at this massive difference and net swing anomaly from another perspective: the unemployment numbers. To start, forget about the U-3 unemployment number preferred by the press, with its reported reduction in unemployment rate from 8.5% in December to 8.3% in January that the administration and press have been hyping. The more accurate U-6 unemployment rate is a better indicator since it accommodates part time, discourage workers, and underemployed workers. It too underestimates true unemployment by about another 2%, per this writers calculations, but not nearly as dramatically as the U-3 number.

In December the U-6 unemployment rate for 15.2%, both for not seasonally adjusted and seasonally adjusted total employment. That means, for the unadjusted employment levels there were 20.208 million jobless in December 2011 and 20.096 million unemployment in December per the seasonally adjusted numbers. Thats a difference of only 112,000 unemployed.

But look at the December-January difference in unemployed between the two sets of numbers: The U-6 unemployment rate for seasonally adjusted fell to 15.1% in January (from 15.2%) in December, while the not seasonally adjusted number of unemployed rose from 15.2% in December to 16.2% in January. The U-6 indicates the number of unemployed rose, which makes sense for construction, retail, and other sectors per the preceding argument. But for the seasonally adjusted numbers, unemployment declined for the U-6 by 0.1% (and 0.2% for the U-3). The net swing between the two sets of data for December-January was 1.108 million, compared to the net swing for November-December of only 112,000. The difference represent a ten-fold jump for December-January.

How can the seasonally adjusted vs. not seasonally adjusted jobs numbers be so large for December-January compared to previous months? How can what appears to be a decline in jobs clearly in January end up reported, after seasonality and other statistical adjustments, as an upward revised 243,000 jobs?

Other economists have been focusing on the possible problem with the seasonality assumptions in the January jobs numbers this past week, but their commentary is not reaching the public press. The essential point is that the January jobs report is peculiar, and requires an explanation by Labor Department statisticians why there was a swing of about 3 million jobs last month between the raw jobs numbers data and the upward revisions in the seasonally adjusted numbers for the month.

Jack Rasmus

Jack is the author of An Alternative Program for Economic Recovery, available at his website, http://www.kyklosproductions.com, and the forthcoming April book, Obamas Economy: Recovery for the Few, by Pluto Press-Palgrave.