Mixed January Employment and Unemployment Signals. The Bureau of Labor Statistics’ (BLS) payroll benchmark revision showed a 19% greater jobs loss for this downturn than previously reported, along with an unexpected, albeit small, decline in January 2009 payrolls. At the same time, the BLS reported that the headline U.3 unemployment rate declined by 0.3% to 9.7% in January, an unexpectedly strong positive result.

The "improved" unemployment rate likely was helped by seasonal-adjustment distortions, which can be meaningful at this time of year. Holiday Season layoffs can be difficult to assess, even in normal economic times, and the severity of the current business downturn has skewed seasonals in other data. The unadjusted U.3 rose from 9.7% in December to 10.6% in January, but the seasonally-adjusted rate fell from 10.0% to 9.7%, a positive swing in change of 1.2 percentage points due to seasonals. In a more normal period, such as December 2005/January 2006, unadjusted U.3 rose from 4.6% to 5.1%, with the adjusted rate falling from 4.9% to 4.7%, a positive swing of 0.7 percentage points. How much the difference in results reflects distortions from the higher unemployment rate and severe contraction that mark the current period is hard to tell. To the extent there are seasonal distortions at work, such will become obvious in the months ahead, as mis-adjustments to the unemployment rate balance out over time.

In any event, the signal from the annual contraction in inflation-adjusted M3 for renewed downturn is intensifying, as will be discussed in the next Commentary. Accordingly employment and unemployment are highly likely to show significant deterioration in the months ahead, irrespective of any relative strength or minimal weakness seen in current reporting.

Payroll Employment Benchmark Revision. The benchmark revision lowered the previously-reported seasonally-adjusted payroll employment level in December 2009 by 1.363 million (1.390 million unadjusted), roughly 1.1%, and the lowered unadjusted payrolls in the benchmark month of March 2009 by 902,000, versus an initial BLS downside estimate of 824,000. The heaviest downside revision was concentrated in the August to November 2008 months, with an average monthly downward revision to the month-to-month change of about 150,000 jobs. Even so, downside revisions to the month-to-month changes in 2009 (all but November were in negative territory) averaged 50,000.

Related downside revisions to historical GDP will not be published until July 2010.

The accompanying graphs reflect the revised data (the first two show changes from the current as well as the prior benchmark revisions), both in terms of level and year-to-year percent change, as well as the revised month-to-month and net changes in the month-to-month reporting over the last two years.

Nonetheless, the BLS continues to overstate the payroll employment level and to understate the month-to-month declines in the seasonally-adjusted payroll levels, much as it did following the prior benchmark revision (see the first graph). As discussed in the Birth-Death Model section below, although the upside monthly bias has been reduced, it should be negative. I estimate that after these revisions the BLS still is underestimating monthly jobs loss by about 150,000. More will follow.

Payroll Survey. The BLS reported a statistically-insignificant, seasonally-adjusted jobs loss of 20,000 (down by 1,383,000 net of benchmark revision) +/- 129,000 (95% confidence interval) for January 2010, following a revised 150,000 (previously an 85,000) jobs loss in December. Details of the benchmark revision are discussed above.

From peak-to-trough (the peak month was December 2007; the current month of January 2010 is the short-lived trough of the current cycle), payroll employment has declined by a seasonally-adjusted 8,424,000 jobs, or 6.1%. That had been a decline of 7,242,000 jobs, or by 5.2%, in pre-benchmark December reporting. While the benchmark did not go as far as it likely should have, what is in place now will not go through major revision again until this time next year.

As the pace of reported monthly decline continued to slow against year-ago comparisons, year-to-year contraction (unadjusted) in total nonfarm payrolls narrowed to 3.0% in January from a revised 3.6% (previously 3.0%) decline in December, and from a post-World War II record 5.0% decline in July. The July 2009 decline was the most severe annual contraction seen since the production shutdown at the end of World War II, which reflected a trough of a 7.6% contraction in September 1945. Disallowing the post-war shutdown as a normal business cycle, the current annual decline would be the worst since the Great Depression.    

Concurrent Seasonal Factor Bias (CSFB).  The pattern of impossible biases built into the headline monthly payroll employment of recent years may have ended with the benchmark revision, after mixed results in the last six to seven months. The January 2010 distortion was minimal. Instead of the headline jobs loss of 20,000, consistent application of seasonal-adjustment factors — net of what I call the concurrent seasonal factor bias (CSFB) — would have shown a monthly jobs gain of 8,000, which is nothing but minimal statistical noise.

The biggest distortions shown by the CSFB were in the November 2008 to February 2009 timeframe, the period in which the BLS initially indicated its biggest problem for the now just-published benchmark revision. The BLS, however, shifted the most hefty downside revisions to the August to November 2008 period. A worksheet on this is available upon request. (See SGS Newsletter No. 50, for further background.)

Birth-Death/Bias Factor Adjustment.  As discussed in previous writings, the Birth-Death Model biases tend to overstate payroll employment during recessions. The flaws here initially were confirmed when the BLS first announced the massive 2009 benchmark revision, with the BLS indicating that the underlying assumptions to the Birth-Death Model were missing certain jobs losses.

Nonetheless, the Birth-Death Model survives, albeit scaled down minimally. The unsupportable premise that jobs created by start-up companies in this downturn have more than offset jobs lost by companies going out of business, continues. So, if a company fails to report its payrolls because it has gone out of business, the BLS assumes it still has its previously-reported employees and adjusts those numbers for the trend in the company’s industry.

The "surplus" jobs created by start-up firms, which get added on to the payroll estimates each month as a special add-factor, have been revised lower. Prior to the benchmark revision, the Birth-Death Model appears to have been adding an average of about 72,000 extra jobs per month (roughly 861,000 per year), but that appears to have been revised now to an average of about 42,000 per month (roughly 509,000 per year). This monthly bias should be negative, on average. Since it is not, the BLS continues to overestimate monthly growth in payroll employment.

That said, the unadjusted January 2010 bias was a monthly subtraction 427,000 (January is the one month of big subtraction) versus a pre-benchmark 356,000 subtraction the year before, and a revised 25,000 (pre-benchmark 59,000) addition for December 2009.

Household Survey.  The usually statistically-sounder household survey, which counts the number of people with jobs, as opposed to the payroll survey that counts the number of jobs (including multiple job holders) was restated in January for new population estimates. Given the series change, the BLS estimates a consistent January-over-December Change in employment of a 784,000 gain, versus the 589,000 monthly employment decline reported in December before the annual population adjustments.

Also due to the new population controls, the January 2010 seasonally-adjusted U.3 unemployment of 9.69% +/- 0.23% (95% confidence interval) was not strictly comparable to the 9.97% unemployment rate reported in December. Yet, given the BLS estimates of impact from the controls, a reported decline of roughly 0.3% would be statistically meaningful. Unadjusted U.3 was reported at 10.6% in January, up from 9.7% in December.

The broader January U.6 unemployment rate fell to an adjusted 16.5% (rose to 18.0% unadjusted), from 17.3% (17.1% unadjusted) in December.

During the Clinton Administration, "discouraged workers" — those who had given up looking for a job because there were no jobs to be had — were redefined so as to be counted only if they had been "discouraged" for less than a year. This time qualification defined away the long-term discouraged workers. The remaining short-term discouraged workers (less than one year) are included in U.6. 

Adding the excluded long-term discouraged workers back into the total unemployed, unemployment — more in line with common experience as estimated by the SGS-Alternate Unemployment Measure — dropped to about 21.2% in January from 21.9% in December. While there likely were some seasonal aberrations in the January reporting, the SGS measure is based on the reported U.6 measure and usually varies with it. See the Alternate Data tab at www.shadowstats.com for a graph and more detail.

While 21.2% unemployment might raise questions in terms of a comparison with the purported peak unemployment in the Great Depression (1933) of 25%, the SGS level likely is about as bad as the peak unemployment seen in the 1973 to 1975 recession. The Great Depression unemployment rate was estimated well after the fact, with 27% of those employed working on farms. Today, less that 2% work on farms. Accordingly, for purposes of a Great Depression comparison, I would look at the estimated peak nonfarm unemployment rate in 1933 of 34% to 35%.

Week Ahead. Given the underlying reality of a weaker economy (and likely re-intensifying downturn in the coming months) and a more serious inflation problems than generally are expected by the financial markets, risks to reporting will tend towards higher-than-expected inflation and weaker-than-expected economic reporting in the months ahead. Such is true especially for economic reporting net of prior-period revisions. 

Trade Balance, Goods and Services (December 2009). The December trade deficit is due for release on Wednesday, February 10th. Briefing.com shows a consensus estimate for some narrowing of the monthly deficit, and such would be necessary to prevent a downside revision to the initial estimate of 5.7% annualized real (inflation-adjusted) growth in fourth-quarter GDP. My betting is for some further deterioration in the deficit, which would knock some growth off the estimated GDP, due for revision at the end of this month. The reported fourth-quarter "improvement" in trade guesstimated by the Bureau of Economic Analysis added 0.5 percentage points to the quarterly GDP growth rate.

Retail Sales (January 2010). Due for release on Thursday, February 11th, January’s monthly retail sales are expected to show a 0.4% monthly gain, following a 0.3% contraction in December (per Briefing.com). Reporting risk is to the downside, and inflation-adjusted growth likely will be flat-to-minus on both a month-to-month and year-to-year basis.

Heads Up: 2009 GAAP-Based Financial Statements of the U.S. Government. The Treasury’s Financial Management Service advises that the GAAP statements are scheduled for release on Tuesday, February 16th. The statements were delayed, without explanation, from the regularly scheduled release date of December 15, 2009. A Reporting Focus Commentary will follow after I have had a chance to assess the report.