Friday isn't just the release date for the September jobs numbers. It's also the day the Feds announce their estimated yearly revision to 2004's jobs. In one day's report, Bush's fabled "jobs deficit" could be halved, cut in three, or even entirely eliminated.

My source Deep Stoat tells me that Friday will be even more interesting -- and possibly important -- than I'd thought. On that day, the Fed will announce by how many jobs it thinks it's either overstated or understated job growth throughout fiscal year 2004.

Could be a good day, could be a bad day. Deep Stoat won't say. (Actually, Deep Stoat probably doesn't know, but it makes me feel more important to hint that he knows but he's holding back.)

He sends along this National Journal piece (can't find the link, alas) and suggests that, if the revision is positive, the Democrats will scream that Bush is "misleading us" by "manipulating the economic intelligence."

Administration - On Jobs, an October Surprise?

By John Maggs
National Journal Magazine

9/25/04

Mark October 8 on your calendar. On the same day that John Kerry and George W. Bush are scheduled to clash in a debate, one of the most potent criticisms of Bush's economic record may lose most of its bite.

On that day, the Labor Department will announce September's unemployment figures, which either will show the job-creation slump of the past several months continuing, or will signal that employment growth is back on track.

More important, October 8 is also the date that the Labor Department squares the previous year's employment estimates with actual experience. And based on hints from other data on pay for workers, some Wall Street economists are expecting a hefty increase in the number of jobs created over the past year under Bush. If so, the final month of the presidential race may play out in a very different economic environment.

...

As a result, that 2.6 million job deficit since Bush's inauguration has fallen to 900,000 through August and sets the stage for a possible reversal next month. The first step toward having that happen will be September's job market. If Alan Greenspan is right and the economy has "regained some traction" in recent weeks after what he called a "soft patch" earlier in the year, then September may end up as more of a typical month for job creation during a recovery. That could mean from 250,000 to 350,000 more jobs for the month -- March had 353,000 and April 325,000.

...

A look at past revisions reveals a pattern -- the survey of businesses tends to overcount jobs created during a recession and undercount them during a recovery. (See chart, p. 2905.) ...For the two years that covered the 1990-91 recession, the overcount was nearly 900,000 jobs. But the math goes the other way during expansions -- look at the 1990s boom and you see that Labor undercounted the jobs being created, as revealed by how many people applied for unemployment benefits. From 1993 through 1995, Labor undercounted the number of jobs created by more than 500,000 a year, or about 15 percent of the total.

If the real recovery in employment began in the summer of 2003, as it now seems, then history would suggest that a revision upward is likely on October 8. And there is another reason that some economists who look closely at these numbers are expecting a big increase. So far, compensation for workers has grown healthily, while workers' average wages aren't rising as much. That would suggest that there are more workers out there than have been counted so far. ....

Ethan Harris, chief U.S. economist for Lehman Brothers, is among those who closely watch the job figures, and he said experience shows that the adjustments can be large. David Wyss, chief economist with financial publisher Standard & Poor's, said it would be expected that the benchmarking by Labor would raise the estimate of job creation in a recovery.

There is yet another reason to expect a revision upward in the job numbers.

The most widely cited number for job creation is based on what's called a "payroll" survey of businesses. But another survey by the Labor Department estimates job creation by polling 60,000 households. Most economists consider the household survey less accurate, and it isn't cited nearly as often, but it has gained some adherents in the Bush administration because it suggests that the number of people working has increased under Bush. Economists have been debating for decades about the advantages and shortcomings of the two surveys, but the gap has grown recently between the two estimates. In fact, it has never been wider.

So, how big could the upward revision be on October 8? The average yearly revision since 1979 has been 257,000, and the average of all upward revisions is 308,000. In 2000, there was a positive revision of 468,000.

What this means is that a decent September for job creation, added to an average-sized revision for a recovery, could wipe out half or two-thirds of the remaining 900,000 deficit in jobs under Bush. It is conceivable that almost all of the deficit could disappear next month.

Absolutely no cowbell for speculation, so don't even ask. Don't you guys believe in jinxes?

Update: Link here.