Flash Update
FLASH UPDATE - Jan. 7, 2007
JOHN WILLIAMS' SHADOW GOVERNMENT STATISTICS
FLASH UPDATE
January 7, 2007
__________
December Payroll Growth Understated But Not Credible
Unusual Weather Patterns Promise Major Data Distortions
__________
Is Bureau of Labor Statistics Playing Games with the Credit
Market?
Highly unusual reporting and revision patterns for the jobs data were seen again, for December. Employment conditions are close to showing a recession, but each month the Bureau of Labor Statistics (BLS) keeps filling in prior periods with levels of upside revisions that are unprecedented outside of the annual benchmark revisions. The revisions are unusual enough for the BLS to have published a statement last month proclaiming that the changes were not unprecedented. Something very strange is going on in the reporting. Whatever it is, it more likely is tied to data manipulation than to accurate economic reporting. All questions may be answered when the annual benchmark revision is published next month. An upside revision already has been announced, and the employment universe for the year ahead will be set up by the January numbers.
While the payroll numbers were weak, but still unbelievably high, there was no obvious smoking gun in the report that showed the gimmick. The payrolls dropped by 206,000, before seasonal adjustment, so the gains were all thanks to the adjustments. That said, though, the internal data suggested that the reported adjusted growth should have been even stronger.
Even though the 167,000 gain in December's seasonally-adjusted nonfarm payrolls was tepid in nature, it was somewhat above consensus expectations, and the U.S. markets tumbled as hopes for a Fed easing were diminished, somewhat. What would the markets have done with the 226,000 gain that could have been reported?
There are ongoing serious problems with the seasonal adjustments made to the payroll data. Where year-to-year changes in both the seasonally-adjusted and unadjusted series should be virtually identical, they have not been recently and were not in December. Applying the just-reported, unadjusted annual respective November and December growth rates of 1.3487% and 1.4025%, to the seasonally adjusted reporting of 2005 indicates the adjusted monthly jobs gain should have been about 226,000. While such would have roiled the markets even further, that growth rate still would have been considered sub-par in the halcyon Clinton days, when the politically-controlled monthly jobs growth was set at 250,000.
The reported December gain of 167,000 +/- 106,000, was 196,000 net of revisions. The previously reported gain of 132,000 in November revised to 154,000, as the pattern of heavy upward revisions to prior-period data continued.
These unusual revision patterns suggest a possible effort to mask indications of otherwise non-existent economic strength from the ever-sensitive credit markets. If these numbers are being so manipulated, it likely involves either the Fed and/or the now market-savvy Treasury Department, not the traditional political hypesters in the Administration.
The historical seasonal-factor revision to the household survey took place with the December data. On the new basis, the seasonally-adjusted December U.3 unemployment rate held at 4.48% +/- 0.2% for the second month, with the unadjusted U.3 holding at 4.3%. The adjusted and unadjusted broader U.6 unemployment rates also held at 8.0% and 7.8%, respectively. As with the payroll series, the seasonal-factor changes for the household series are being used to throw current "strength" into prior periods.
Despite the twisted jobs reporting, help-wanted advertising and new claims for unemployment continued to show sharply deteriorating employment conditions. Even the gains shown in the seasonally-adjusted December purchasing managers surveys remained very soft in the employment arena.
Seasonal adjustments of monthly economic data can be terribly misleading in the best of times. Most government efforts at such adjustments are well intentioned, but do not work well when seasonal patterns are not usual. This Flash Update is being written on January 6th in central New Jersey, and I have the air conditioner running. December was snowless here for the first time in more than a century. In contrast, Denver has been hit by successive and unusually debilitating snowstorms. Both circumstances disrupt normal commerce in ways that the government cannot possibly adjust for in its economic guesstimates. Accordingly, do not be surprised with unusually screwy growth patterns being reported for series like industrial production and retail sales, etc.
Unlike the employment series, where the seasonal factors can be gimmicked monthly, most series will show periods of catch-up from seasonal distortions, since most series, in theory, total to the same level year, adjusted or unadjusted.
Further detail follows in the January SGS.
___________________________________________
January's "Shadow Government Statistics" monthly newsletter is targeted for month-end release.
FLASH UPDATE
January 7, 2007
__________
December Payroll Growth Understated But Not Credible
Unusual Weather Patterns Promise Major Data Distortions
__________
Highly unusual reporting and revision patterns for the jobs data were seen again, for December. Employment conditions are close to showing a recession, but each month the Bureau of Labor Statistics (BLS) keeps filling in prior periods with levels of upside revisions that are unprecedented outside of the annual benchmark revisions. The revisions are unusual enough for the BLS to have published a statement last month proclaiming that the changes were not unprecedented. Something very strange is going on in the reporting. Whatever it is, it more likely is tied to data manipulation than to accurate economic reporting. All questions may be answered when the annual benchmark revision is published next month. An upside revision already has been announced, and the employment universe for the year ahead will be set up by the January numbers.
While the payroll numbers were weak, but still unbelievably high, there was no obvious smoking gun in the report that showed the gimmick. The payrolls dropped by 206,000, before seasonal adjustment, so the gains were all thanks to the adjustments. That said, though, the internal data suggested that the reported adjusted growth should have been even stronger.
Even though the 167,000 gain in December's seasonally-adjusted nonfarm payrolls was tepid in nature, it was somewhat above consensus expectations, and the U.S. markets tumbled as hopes for a Fed easing were diminished, somewhat. What would the markets have done with the 226,000 gain that could have been reported?
There are ongoing serious problems with the seasonal adjustments made to the payroll data. Where year-to-year changes in both the seasonally-adjusted and unadjusted series should be virtually identical, they have not been recently and were not in December. Applying the just-reported, unadjusted annual respective November and December growth rates of 1.3487% and 1.4025%, to the seasonally adjusted reporting of 2005 indicates the adjusted monthly jobs gain should have been about 226,000. While such would have roiled the markets even further, that growth rate still would have been considered sub-par in the halcyon Clinton days, when the politically-controlled monthly jobs growth was set at 250,000.
The reported December gain of 167,000 +/- 106,000, was 196,000 net of revisions. The previously reported gain of 132,000 in November revised to 154,000, as the pattern of heavy upward revisions to prior-period data continued.
These unusual revision patterns suggest a possible effort to mask indications of otherwise non-existent economic strength from the ever-sensitive credit markets. If these numbers are being so manipulated, it likely involves either the Fed and/or the now market-savvy Treasury Department, not the traditional political hypesters in the Administration.
The historical seasonal-factor revision to the household survey took place with the December data. On the new basis, the seasonally-adjusted December U.3 unemployment rate held at 4.48% +/- 0.2% for the second month, with the unadjusted U.3 holding at 4.3%. The adjusted and unadjusted broader U.6 unemployment rates also held at 8.0% and 7.8%, respectively. As with the payroll series, the seasonal-factor changes for the household series are being used to throw current "strength" into prior periods.
Despite the twisted jobs reporting, help-wanted advertising and new claims for unemployment continued to show sharply deteriorating employment conditions. Even the gains shown in the seasonally-adjusted December purchasing managers surveys remained very soft in the employment arena.
Seasonal adjustments of monthly economic data can be terribly misleading in the best of times. Most government efforts at such adjustments are well intentioned, but do not work well when seasonal patterns are not usual. This Flash Update is being written on January 6th in central New Jersey, and I have the air conditioner running. December was snowless here for the first time in more than a century. In contrast, Denver has been hit by successive and unusually debilitating snowstorms. Both circumstances disrupt normal commerce in ways that the government cannot possibly adjust for in its economic guesstimates. Accordingly, do not be surprised with unusually screwy growth patterns being reported for series like industrial production and retail sales, etc.
Unlike the employment series, where the seasonal factors can be gimmicked monthly, most series will show periods of catch-up from seasonal distortions, since most series, in theory, total to the same level year, adjusted or unadjusted.
Further detail follows in the January SGS.
January's "Shadow Government Statistics" monthly newsletter is targeted for month-end release.