Posted 9:27 PM by Robert

Jobs? Jobs? Jobs?

If the recovery is and remains "jobless," it is possible President George W. Bush might not be re-elected, as happened to his father in 1992. But it's worth noting that after that election, the revised 1992 employment numbers were better than had been thought. Payroll employment was initially reported to have risen only 423,000 during 1992, - but that number was later revised to 1,157,000. That made for an average of 96,417 per month during 1992 - in contrast to the average 35,250 per month thought to be the case while the campaign was being waged. It would be an irony for George W. Bush to succeed in avoiding all of his father's mistakes, but nevertheless failing to be re-elected because the Department of Labor statisticians had not failed to avoid the mistakes of their predecessors from 1992. Life is not always fair - but irony in politics always comes cheap.

The Man Without Qualities believes that it is rather likely that the economy is again creating more jobs (and more payroll jobs) than shown by the government's initial reports. That phenomenon occurred every year from 1992 through 2000 - with an upward revision averaging 609,000 per year. Downward revisions of 519,000 and 579,000 occurred for 2001 and 2002, respectively, with another upward revision of 203,000 for 2003.

My suspicion that official employment numbers are now understating employment - even payroll employment - follows from the recent drop in the applications for "continuous" unemployment benefits. In the week of December 27, 2003, there were 3,278,000 "continuous" claims, but during the week of February 21, 2004 "continuous" claims were down to 3,091,000. Changes in payroll employment are generally correlated with the inverse of the number of such claims (the correlation between payroll employment and initial unemployment claims isn't as good as the correlation between payroll employment and continuous unemployment claims).

The correlation between payroll employment and "continuous" unemployment claims should lead to annualized gains exceeding 1% - or 1.3 million jobs gains (averaging 108,000 per month). In contrast, the past three months have averaged 42,000 job gains per month.

In late 1992 the number of unemployment insurance applications declined - and payroll employment growth then rose to almost 2% from about 0. Of course, there is no guaranty that the same thing is happening now - indeed, the labor market of 2004 is full of surprises. For one thing, "continuous" benefit applications may be down because they have expired for soem workers.

But, then, the whole focus on "job creation" is somewhat problematic from a political (election) perspective. The unemployment rate is now only 5.6% - and it is the unemployment rate that historically has had the strongest correlation to election results, not "job creation." It appears correct that the unemployment rate would be higher but for "discouraged workers" - but that is always the case. The number of "discouraged workers" is not particularly high. In addition, many people have observed that higher mandated worker benefits may be discouraging payroll employment. But, if that effect is significant, it should also be true that higher mandated worker benefits may be discouraging reporting of payroll employment. Indeed, some countries with very high mandated benefits (Italy, for example) have correspondingly huge "undocumented" sectors - it would make some sense that increased government-imposed labor costs would expand the "undocumented" sector of the US labor market, too. One would expect "off the books" employment especially at smaller businesses - and USA TOday, for example, says that smaller businesses don't seem to be reporting as many hires as expected. Moreover, the same article goes on to explain somethng noted here in prior posts:

[T]he Labor Department may not be capturing a key source of post-recession job gains: newly self-employed workers who aren't counted in the agency's monthly surveys. The number of these tiny start-ups is growing, says Mark Zandi, of consultant Economy.com.

The labor market of 2004 - as perused in the comfort of the future - may hold more surprises for us than we dream of today.

But all that might - or might not - help Mr. Bush.

POSTSCRIPT:

I neglected to mention that - as noted in a prior post - the employment component of the Institute for Supply Management’s manufacturing index has been very high for several months - which is also consistent with hiring possibly being more substantial than the initial government reports indicate.