Posted - Filed under Household Survey, Release Observations.

Whereas the Payroll Survey counted all furloughed, Federal Government workers as employed, the Household Survey, which generates the unemployment rate aimed to count such workers as unemployed.

Hence the uptick in the unemployment rate reported today, from 7.2% in September to 7.3% in October.

However, those numbers are rounded (by the BLS) to one decimal place.  Calculating the September and October rates to two decimal places yields,  7.24% and 7.28% respectively.   Less than half of the headline rise.


Click to enlarge

As to what the rate would have been, absent the partial government shutdown, the hidden complexities of the seasonal adjustment process do no allow us to make a good estimate.  The BLS themselves have not chosen to try.

Similarly, any analyses of labor force participation and employment ratios would not prove very rewarding.

The November numbers will snap back into line, and in addition will reveal the true, revised seasonally-adjusted historical series for the prior months of this year.


An Interesting Note on why the October unemployment rate should have been higher.

The BLS points out today that many furloughed government workers responded incorrectly to the survey questionnaire and classified themselves as employed but absent from work.  The BLS in today’s release points this error out, but in keeping with the statistical sampling methodology, will not correct it.

In an additional document, they reveal that roughly equal numbers of furloughed workers (200,000) reported correctly and incorrectly  The effect on the unemployment rate – whatever it was – would have been roughly doubled, but for this survey response error.