After all the optimism embodied in the Fed’s recent minutes, at least as I read them, the employment report this morning was a shocker. Instead of the expected 200,000 or so gain in jobs, the figure came in at 74,000, well below the lowest estimate. What the heck happened?
Making matters more confusing, along with the very low job growth came a decrease in the unemployment rate, from 7 percent to 6.7 percent. At the same time, the underemployment rate remained the same, at 13.1 percent as revised. So, we have a much lower growth rate than expected, a substantial improvement, and status quo—all at the same time, about what is apparently the same thing. What?
Part of the explanation is that the job growth and unemployment rate numbers aren’t actually the same thing; they’re based on different surveys, with different criteria for inclusion. The job growth number is based on what is called the establishment survey, which covers about 145,000 businesses, with more than half a million workers. To be counted, a job has to be paid for at least one day in the survey period—no pay, no job—which affects how hourly workers, for example, are counted.
The household survey, on the other hand, is the basis of the unemployment rate. This number comes from a survey of about 50,000 households, and counts as employed those who worked 15 hours per week or more during the survey period. This is a smaller survey, and the results are therefore more variable.
The difference in the methodologies suggests that the drop in the job growth number may well be weather-related. The household survey includes a data point for workers who could not work due to weather. At 273,000, this was well above the normal average of 166,000, suggesting that the gap, of 110,000 or so, could be due to weather. Declines in hours worked and construction employment also support the weather thesis. Adding that 110,000 to the actual number brings us back close to expectations. The weak figure, then, might very well be a result of the terrible December weather, and not an indication of real economic weakness.
If the job growth number isn’t as bad as it looks, though, the unemployment number is not as good as it looks. Despite the decline in the headline unemployment rate, the job growth of 143,000 was much less than the decline in the participation rate—that is, people reporting themselves to be part of the labor force—of 347,000. The decline in unemployment was therefore not due to job growth and can’t be interpreted as a sign of strength. But, that said, the decline in the participation rate isn’t necessarily a sign of weakness either. The stability of the underemployment rate, at 13.1 percent, suggests that those least attached to the labor force are still hanging in there.
Given the strength of the employment reports leading up to this one, which included multiple positive revisions (including revising November up to +241,000), it’s hard to see what could have caused such a significant decline—other than the weather.
These data series are noisy, and in this case, there are good reasons to believe this weak number is an outlier. It does, however, suggest that the foundations of the recovery may not be quite as solid as the market, for example, may have thought.
One closing thought about the Fed minutes. After wrapping up yesterday’s post, it occurred to me that the minutes actually suggested both of the things that the market should want to see: stronger economic growth and continued Fed stimulus. The stock market’s tepid reaction to what in many respects was a very favorable report was interesting, and consistent with the weaker performance we’ve seen since the start of the year.
All in all, we’re in a somewhat more uncertain place today than we were yesterday. Right now, this report looks like an anomaly, but I will be watching closely as the data comes in to see if that continues to be the case.