Jobs

September job numbers come out on Friday. The payroll number will get a lot of attention; anything under 200,000 will mean uncomfortable questions for Bush on Friday night. Anything over 200,000 will cause Bush to point to the future with optimism while Kerry reminds everyone of the dismal numbers in previous months. (Really, this sort of predicting is too easy!) Another important number coming soon is the annual revision. Greg Mankiw thinks that payroll employment will be revised up by as much as 288,000 for the year (with a margin of error of +/- 140,000 -- that sort of predicting isn't much harder!) Objectively speaking though, Mankiw is probably right. Economists who follow this sort of thing realize that there is usually a positive revision in the first couple years of a recovery. 2003 was the year the jobs recovery started to get off the ground, so I would expect about 200,000 based on past experience. Mankiw might be a little optimistic, but my estimate is in the same range as he gives.

The reason for this is that while the payroll survey is a much better measure of jobs than the household survey, the payroll survey tends to miss newly created companies and self-employed workers for a while. Eventually, the new firms work their way in the survey, but in periods when more new firms are being created than usual (such as the early part of an expansion), it underestimates the actual number of jobs.

Whatever the revision ends up being, it will be exaggerated by the Republicans (assuming it is positive) and dismissed by the Democrats. If it ends up being close to my estimate, I won't give it much thought either way since I've already figured it in to my own assessment of the employment situation.