Friday, November 6, 2009

October Employment Data: Disappointing But Not Uniformly Bad

Although the spike in the unemployment rate to 10.2% is attracting most of the attention because of its presumed psychological implication, the rest of the employment data for October paints a somewhat more complex picture of the labor market.

As we have emphasized repeatedly in this space, despite its high visibility as a barometer of labor market conditions, the unemployment rate is a lagging indicator and typically tends to drift higher in the early stage of an economic recovery (and, as such, it is not a useful gauge for detecting relatively changes in labor market conditions around turning points in the cycle). The reason for that seeming paradox is that the rate tends to move higher due to an expansion of the civilian labor force as a result of previously discouraged workers that return to it when the economic environment starts improving. Still, in October, the spike in the unemployment rate was not caused by an increase in the size of the labor force but by a huge decline (-589,000) in household employment. (This is a survey completely separate from the establishment survey that produces the nonfarm payroll data).

What is becoming an increasingly curious pattern here is that there is a very big gap that has opened up between the household employment data and the payroll data, as the former have declined by a total of 1.77 million in the last three months alone versus a total decline of "only" 563,000 in payroll employment during the same period. Such gaps between the two surveys are not totally unheard of over a number of a few months but, ultimately, that gap will need to narrow moving forward. Given that the improving trend in payrolls has been well established by now and is unlikely to be reversed suddenly (despite the fact that they are still turning out monthly declines), the most likely outcome is that the gap will diminish with household employment improving somewhat and the unemployment rate dipping in the next couple of months.

In fact, the payroll trend continues to improve, as the 190,000 decline in October was accompanied by a 91,000 net upward revision to the previous two months. This brings the average decline in payrolls in the most recent 3-month period to 188,000 versus 357,000 in the prior three months and 690,000 in the first quarter of the year. In other words, things are getting steadily, and appreciably, better on the labor market front, but not fast enough to project a confident turnaround of the employment picture.

One element that was relatively disappointing in today's data was that the average workweek remained at its cycle-low of 33.0 hours. Normally, an upward drift in the workweek (despite its heavily choppy pattern on a month-to-month basis) is supposed to precede a pick up in hiring patterns, as companies tend to utilize their existing labor force more intensely first (when the economic environment starts improving) before they hire more workers.

All in all, despite the eye-catching pop in the unemployment rate, the employment report did little to challenge the premise that an economic recovery is slowly getting underway but it did reinforce the view that this recovery will be of the moderate variety for some time.

Anthony Karydakis