January Employment: A Tale of Two Surveys

by: The Fundamental Analyst

[click to enlarge charts]

How do you describe the January employment report? Messy, inconsistent, contradictory, all of the above? Non-farm payrolls rose a lackluster 36,000 in January according to the the BLS, also the benchmark annual revision wiped 483,000 jobs from 2010 payrolls. After the benchmark revision, it turns out 6 months of 2010 had negative payroll numbers and the average monthly gain was just 76,000. Of course there is the possibility that the BLS establishment survey is under-estimating job growth in this recovery just as they underestimated the job losses during the recession. But since we’d have to wait until next year’s benchmark revision to find out, that kind of speculation doesn’t help much.

The unemployment rate plunged for the second straight month and now sits at 9.0%. However, that plunge is not due to robust employment growth. The unemployment rate is calculated from the household survey which is different from the survey that calculates non-farm payrolls. Even so, the household survey also did not show huge growth in employment, adding a modest 117k jobs. The big kicker was another big drop in the civilian labor force. Apparently 504k people dropped out of the labor force in January. Thus you add the two numbers together and the number of unemployed dropped by more than 600k. As noted last month, this trend cannot continue, sooner or later the civilian labor force will start to grow which is a drag on the unemployment rate. Thus you cannot extrapolate the current fall in the unemployment rate and you shouldn’t be surprised to see it tick up again in the next few months.

37 months since the peak in employment and we are nowhere back to that level. We often hear estimates of how long it will take to get back to the previous peak or back to a ‘normal’ unemployment rate…whatever that means. Such estimates range anywhere from a few years to a decade. What this kind of analysis overlooks is that it is unlikely that we recapture previous peaks in employment before we have another cyclical downturn in employment, or in other words a recession. As has been pointed out by Lakshman Achuthan of the ECRI, we are likely headed into a period of shorter economic cycles.

Like last month the unemployment rate overstates the strength in employment but we should also acknowledge that the establishment survey could be underestimating job growth as they did in 2003-4. Further to this point, the nfp survey does not count self-employed persons whereas the household survey does. Also to add to the bullish case was the 0.4% rise in average hourly earnings, almost double what was expected.