Flash Update
FLASH UPDATE
October 7, 2007
__________
Market Mania Fueled by Data Touts
M3 Annual Growth Highest Since November 1971
__________
PLEASE NOTE: The SGS-Ongoing M3 for September (based on 24 days of 30 days of reporting) has been updated on the Alternate Data Series tab at www.shadowstats.com. -- Best wishes to all, John Williams
Some years back, I would go to the horse races at Garden State Park, along with my father -- who loved the horses -- and our favorite restaurateur, a wonderful character known as Steve. Steve always bet on tips. He'd get a tip from the valet parking attendant, from the fellow who sold pencils with the racing programs, from the headwaiter servicing the parterre boxes, etc. I remember once Steve having bought $100-win tickets on five horses (they were not long shots) in a seven horse race, and he lost. He also, on occasion, had some big winners, but over time, the races were a money-losing proposition for Steve. My father's selections were based on analyzing a number of factors ranging from past performance to whether the horse was foaming at a certain part of its anatomy; he also looked at the picks of various touts he respected to see what he might be missing. Occasionally, he had some big winners, and he tended to come out a little ahead over time.
Listening to the daily Wall Street touts, who attempt to hype stock market-mania based on the latest economic data, brings Steve and his horse-race betting to mind. In the case of the monthly payroll report, the race does get fixed, on occasion, while a great deal of the touting takes place after the race already has been run. What is remarkable is that this particular number, if honestly reported, is random in nature and of no significance until it exceeds 129,000 as either a gain or a loss. Even then, it is in the wrong direction 1-in-20 times, based on the government's 95% confidence interval.
A potentially serious issue is that the Bureau of Labor Statistics can bring in the monthly payroll gain anywhere it wants to, and the Administration knows that a number in a certain range -- that can be dismissed as statistical noise or revised away the next month -- will move the markets as effectively as a Federal Reserve policy action. Accordingly, when I suggested last month that the 4,000 jobs loss reported for August was designed to help push the Fed into its easing, one indeed has to wonder what is going on in the background, when August revised to an 89,000 jobs gain in the September report. One might read the current September 110,000 jobs "gain" as a sign the Fed can hold steady at the next meeting, rather than taking action that would tumble the U.S. dollar further. Then, again, if the numbers are honest, which I do not believe, these speculations are just over random statistical noise.
With stock indices at historic highs, the equity markets look as good to me now as Steve's betting on five horses to win in a seven a horse race: not much likely upside potential and some very meaningful downside potential.
That said, the BLS reported seasonally-adjusted September payrolls up by 110,000 (228,000 net of revisions) +/- 129,000, with August gaining 89,000 in revision, after an initial report of a 4,000 jobs loss. Unadjusted year-to-year payroll growth slowed to 1.19% in September, from 1.23% in August. Consistent adjusted and unadjusted annual growth rates, in combination with the massive revision, suggest September would have shown a gain of 142,000. Again, with monthly seasonal factors being readjusted each month as needed, the BLS can generate any desired result.
These data were released along with a Bureau of Labor Statistics (BLS) announcement that the next benchmark revision will show that unadjusted March 2007 payrolls were overstated by roughly 297,000 jobs. That means the currently reported levels will be revised downward by perhaps 500,000 as the BLS models readjust history.
In the household survey, which counts the number of people with jobs, as opposed to the payroll survey that counts the number of jobs (including those of multiple job holders), seasonally adjusted employment rose by 463,000 in September, following a 316,000 employment decline in August. The seasonally-adjusted September U.3 unemployment rate was reported at 4.70% +/- 0.23%, up from 4.64% in August, while unadjusted U.3 eased to 4.5% in September from 4.6% in August. The broader U.8 rate held at an adjusted 8.4% in September and fell to 8.0% from 8.4%, unadjusted. Net of the "discouraged workers" defined out of existence during the Clinton Administration, the actual unemployment rate continues to run around 12%.
Keep in mind that the payroll data have been boosted by 1,118,000 jobs in the last 12 months by the addition of regular bias factors, a.k.a. the birth/death model. September's bias was to the upside by 17,000 jobs. This bias system never has been adjusted to handle a recession, where the biases likely would turn negative. Accordingly, in the current recessionary environment, the payroll data are being significantly overstated versus underlying reality.
As noted on Friday, these jobs data from the BLS are nearly worthless as economic indicators and remain highly suspect, given the global financial markets in ongoing crisis and given the positive ratings of U.S. President and Congress hitting historic nadirs. The reported numbers continued to run counter to better-quality employment indicators such as new claims for unemployment and the collapsing help-wanted advertising index.
M3 Growth at 36-Year High. Based on roughly 24 days of 30 days of reporting, the preliminary estimate of year-to-year growth in the SGS-Ongoing M3 measure hit 14.6%, up from 13.9% in August, and at the highest reading since November 1971. That was two months after Richard Nixon closed the gold window in August of 1971.
Annual M2 growth for September was about 6.8%, up from 6.7%. Part of the relative gain in M3 last month appears to be a movement of cash into larger time deposits and such, as market perceptions have shifted towards a further easing by the Fed. The SGS estimate will be finalized after next week's reporting.
Week ahead. Underlying reality would suggest a larger trade deficit (Thursday, October 11), a weaker retail sales change (Friday, October 12), and a stronger PPI (Friday, October 12) than consensus expectations. If the Fed's concerns in the still unfolding liquidity crisis indeed are tempered by fears of U.S. dollar dumping, both the trade and retail data will be particularly vulnerable to "positive" massaging.
Further details will follow in the October SGS newsletter.
The October SGS is targeted for posting during the week of October 15th. An e-mail advice will be made of its and any intervening Flash Update/Alert postings.