Friday, September 4, 2009

Job Loss Revisionism

by the Sandwichman

The biggest problems with the "better than expected" or "slowing pace" interpretations of the monthly BLS employment situation is that no one really knows what the final numbers will be. An underlying factor in this month's not-as-bad-is-good news is that last month's not-as-bad-is-good news turns out to be worse than earlier reported. (See here for a handy table of BLS monthly revisions).

Warning, you can get whiplash from all the stops and starts in the data. With a total -49,000 revision to the June and July numbers, the total *additional* job losses announced today for the past three months was -265,000, which is higher than the originally announced figure for July. In July, the revisions of May and June employment were upward by a total of +43,000 jobs, so the aggregate job losses announced last month came to -204,000. So, all told the news was worse for August than for July by -61,000 jobs.

The preliminary report for June was -467,000 jobs. Last month that was revised to -443,000 jobs. Today it was revised again to -463,000. There's nothing wrong with refining the numbers as more complete data come in. What is objectionable is the trumpeting of supposed "trends" that are entirely within the bounds of the survey's margin of error.

1 comment:

Martin Langeland said...

Walt Kelly might describe this as "From here on down its uphill all the way."*
However some might take mordant solace in Tom Lehrer's observation that we are "sliding down the razor blade of life."**
* Incompleat Pogo
** Bright College Days