FLASH UPDATE - December 5, 2008

JOHN WILLIAMS’ SHADOW GOVERNMENT STATISTICS

FLASH UPDATE

December 5, 2008

__________

November Jobs Plummet 732,000 Net of Revisions,
Down 873,000 Net of Concurrent Seasonal Factor Bias

Official Recession Start Is Late, As Usual

Required Reserves Surge Anew

__________

PLEASE NOTE: Any major systemic, market or economic surprises will be addressed in the days ahead with a Flash Update or Alert. Otherwise, the next planned update is for Friday, December 12th, following the release of the November retail sales report. The U.S. Treasury plans to release it 2008 GAAP-Based financial statements on the U.S. government on Monday, December 15th. Quick detail will be advised by Flash Update, and the full newsletter will follow shortly thereafter.  
– Best wishes to all, John Williams

NBER Proclaims Economy in Recession Since December 2007. The Business Cycle Dating Committee of the National Bureau of Economic Research (NBER) — official arbiter of whether or not the U.S. economy is in recession — has determined that the U.S. economy peaked in December 2007 and has been in recession since. The December 2007 peak in payroll employment appears to have been the key factor in setting the official timing.  As typically has been the case, however, the NBER is late in its recession start date, and it likely will call the end of the recession too early, as it has in the past. Consider, for example, a year before the now-formal recession onset, real (inflation-adjusted) growth in first-quarter 2007 GDP was flat, but it was in contraction for both Gross Domestic Income and Gross National Product.  The recession preceded, and indeed helped to trigger, the systemic solvency crisis that formally surfaced in August 2007. This topic will be looked at more closely in the upcoming newsletter.

Employment Picture Deteriorated Faster than Indicated by Headline Numbers. As bad as the headline November jobs numbers appeared to have been, the data indicate that consistent reporting would have shown an even bleaker picture. Perhaps someone in the incoming Congress or Administration might raise a question with the Bureau of Labor Statistics (BLS) as to the nature of the increasingly obvious positive biases built into the monthly headline payroll numbers. The problem is evidenced by ongoing massive revisions to prior history (exceeding the 95% statistical confidence interval for monthly change), after the markets have absorbed much happier headline numbers published the month before,  as well as by the consistent, overly positive nature of the headline jobs numbers indicated by the Concurrent Seasonal Factor Bias (CSFB) discussed below.

For example, the pre-election headline estimate of September payrolls was a decline of 159,000 +/- 129,000, but the CSFB suggested the reporting should have been a decline of about 219,000. In the first post-election revision, September’s decline deepened to 284,000; in today’s revision, it dropped to a decline of 403,000. October’s headline jobs loss was 240,000, while the CSFB suggested it should have been about 308,000. October’s jobs loss revised to 320,000 today. The headline November jobs loss was 533,000, but the CSFB suggested it should have been a loss of about 873,000.

Payroll Survey. The BLS reported a statistically-significant, seasonally-adjusted jobs loss of 533,000 (down 732,000 net of revisions) +/- 129,000 for November, following a revised 320,000 (previously 240,000) jobs loss in October. Annual contraction (unadjusted) in total nonfarm payrolls continued to deepen, down 1.47% in November, versus a revised 0.79% (previously 0.85%) decline in October. The seasonally-adjusted series also contracted year-to-year, down by 1.35% in November, versus a decline of 0.93% (previously) 0.78% in October.

Concurrent Seasonal Factor Bias. The pattern of impossible biases (see the Reporting/Market Focus in SGS Newsletter No. 43 of June 10, 2008) being built into the headline monthly payroll employment changes intensified sharply with November reporting. Instead of the headline jobs loss of 533,000, consistent application of seasonal-adjustment factors — net of what I call the concurrent seasonal adjustment bias — would have shown a more-severe monthly jobs loss of about 873,000. This upside reporting bias has been seen in 11 of the last 12 months, with a rolling 12-month total upside headline-number bias of 955,000.

Birth-Death/Bias Factor Adjustment. A minor element that added upside pressure to the latest payroll number was the monthly bias factor (birth-death model). Never designed to handle the downside pressures from a recession, the model added a 30,000 upside jobs bias to November 2008 (versus an upside bias of 17,000 in November 2007), and followed a net upside bias of 71,000 jobs in October 2008. The process boosted financial-activities by 5,000, but subtracted 7,000 from construction. The largest benefactor was the trade, transportation and utilities sector, which picked up an extra 17,000 jobs. Although the adjustments are made to the unadjusted series, they flow through at roughly the same magnitude in the seasonally-adjusted series. 

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), showed household employment fell by 673,000 in November, following a 297,000 decline in October.

The November 2008 seasonally-adjusted U.3 unemployment rate showed a statistically-insignificant increase to 6.68% +/- 0.23% from 6.50% in October.  Unadjusted U.3 rose to 6.5% in November, versus 6.1% in October.  The broader November U.6 unemployment rate jumped to an adjusted 12.5% (12.2% unadjusted) from 11.8% (11.1% unadjusted) in October.  Refigured for the bulk of the "discouraged workers" defined away during the Clinton Administration, actual unemployment, as estimated by the SGS-Alternate Unemployment Measure, rose to 16.5% from 15.8% in October.

Note of Caution. Keep in mind that any comments in the popular media as to historical comparisons of current unemployment data to 1994, are going against the first month published in the current series.   Accordingly, any reference to the "worst level since 1994," could well be against a much earlier period, if only the data were comparable.

In 1994, the BLS completely redesigned and redefined the unemployment series and all its measures, broad and narrow, so that the new series going forward could not be compared with the old series. I still am struggling to take my alternate measure back before 1994, where finding consistent and good data is a major problem. That said, the U.6 broad measure of 12.8% unemployment is the highest since before January 1994. 

Monetary Base Annual Growth Holds at 75%. The reporting for the two weeks ended December 3rd showed that the seasonally-adjusted St. Louis Fed’s adjusted monetary base "stabilized" in the most recent period, holding at 74.9% year-to-year growth, versus 75.4% annual growth in the prior two-week period. The sharp growth in the monetary base has been due to the recent sharp expansion in total reserves of depository institutions. That growth also stalled in the most recent period, showing a minor contraction.

Key here, however, is required reserves. Most growth in reserves has been in excess reserves, where banks have deposited funds with the Fed — earning interest — instead of lending the funds into the normal stream of commerce. In the latest two-week period, however, average daily required reserves increased sharply on a seasonally-adjusted basis, where annual growth in the last four periods has varied from 18.4% to 28.9% to 17.9% to 29.8%. Such growth, though erratic, still is at the highest level of the post-World War II era and suggests some minor movement toward a more normally functioning system.

As to reported seasonally-adjusted M3 components, the Federal Reserve’s latest H.6 weekly data (week ended November 24th) showed M2 rose by $16.3 billion to $7,945.7 billion (up by $22.3 billion the prior week), and institutional money funds rose by $13.9 billion to $2,251.2 billion (up by $14.3 billion the prior week). Tonight’s H.8 data (week-ended November 26th) will indicate, however, if declining large time deposits have continued to offset other gains.

If the weekly data related to M3 continue their recent patterns, the SGS-Ongoing M3 estimate for the November monthly average would show continued slowing year-to-year growth, down to roughly 9% in November, versus 10.8% in October, with the month-to-month change flat-to minus. Annual M2 growth likely will increase from 7.4% in October to something just shy of 8% in November, while M1 growth likely will jump from 7.6% to something over 11%. The relative gains in M1 and M2 are due primarily to the shifting of funds out of M3, not to growth in the monetary base. 

Assuming there are no shocks in tonight’s H.8 reporting. These volatile weekly data will be updated next weekend, along with the monthly average for the SGS-Ongoing M3 estimate of annual growth.    

Other Reporting. Both the manufacturing and nonmanufacturing purchasing managers surveys took further hits in November levels, sinking deeper into recession territory (36.2 manufacturing, 37.3 nonmanufacturing), where the diffusion-index measures suggest contraction when readings are below 50.0. Key components fell to new cycle lows: employment (34.2 manufacturing, 31.3 nonmanufacturing), new orders (27.9 manufacturing); including extremely sharp declines in the levels of prices paid components (25.5 manufacturing, 36.6 nonmanufacturing). The data confirm both a contracting economy and a decline in monthly inflation pressures. These and other economic series will be explored in more depth in the upcoming full newsletter.

Week Ahead. Trade Deficit: Due December 11th (Thursday), the October trade number could show some narrowing in nominal terms, as it increasingly picks up declining oil prices. PPI: Due December 12th (Friday), the November PPI will take a further hit due to collapsing oil prices, but annual inflation should remain strong. Retail Sales: Due December 12th (Friday), November retail sales likely will sink by more than the already-negative consensus expectations. Net of inflation, it also should remain negative on both a monthly and annual basis, despite what will be another sharp monthly decline in the upcoming CPI report.

__________