Friday, December 04, 2009

Relatively Strong Employment Report

Today's employment report could be characterized as relatively strong-without sarcasm. Indeed, by the indicators that I look at, it was even stronger than the headline number of a mere 11,000 lost jobs.

That is first of all because for the first time since July, the household survey showed a stronger employment number than the payroll number. The payroll number is more reliable when it comes to short term fluctuations, but if the household survey number deviates consistently then this suggests that the payroll survey number could under- or overestimate the strength of the job market. Still, this was just one month, and seen over several months the household survey number still indicates greater weakness.

And secondly, the average work week rose as much as 0.2 hours, or 0.6%, increasing hours worked by the same amount. Average hourly earnings were on the other hand week, rising just 1 cent, or 0.05%. That was still enough to increase average weekly earnings by 0.65%, something which means that real weekly earnings likely rose after several months of declines.

Are there any negative aspects in the report? No, not really, except for maybe the aforementioned weak average hourly earnings number. And given the increase in hours, not even that was really significant.

However, part of the strength could be the result of seasonal adjustment distortions. November last year saw a big drop, something which could be reinterpreted by the seasonal adjustment formulas to mean that employment should be weak in November, as opposed to reflecting cyclical weakness and a big real wage schock. Given how consistently weak other economic numbers, including other labor market related numbers, have been recently it seems reasonable to assume that this

This conclusion is confirmed if we compare November seasonal adjustment this year compared to previous years. If seasonal adjustment adds more or subtracts fewer than in the past than it is likely last year's weakness have been interpreted as seasonal rather than cyclical. And if we look at the employment report for November 2007, seasonal adjustment reduced job growth by 156,000, whereas now it reduced it by only 91,000. If seasonal adjustment had been the same as in 2007, job losses would have numbered 76,000 instead of 11,000.

Even so, this report strengthened the bullish case as job losses were still smaller than earlier in the year.