And now for the major economic news of the week, the U.S. employment report for February. It was another solid affair—the third month in a row of over 200,000 job gains, with plenty of supporting details. I do have some worries about the quality of these new jobs, not to mention their durability, but for now things are looking better than they did even a few months ago.
The headline gain of 227,000 came with upward revisions of 61,000 to the back months (41,000 to January and 20,000 to February). Gains were widespread through the economic sectors. Manufacturing extended its impressive recovery, adding a very respectable 31,000 jobs, with what the Brits call the metal-bashing industries in the lead. Mining and logging were also strong, reflecting a little-noticed (and environmentally worrisome) increase in domestic U.S. energy production. (Factoid: North Dakota has lately been producing more oil than Ecuador, a member of OPEC. Transportation and finance were modestly in the plus column.) Temp firms, bars and restaurants, and health care posted strong gains—the first two tenuous and not-well-paying sectors; health care has some good jobs and some bad jobs mixed in. Retail and construction were lost workers, among the few sectors that did. Government job losses slowed considerably—local gov was even modestly in the black.
Despite those decent job gains, it’s important to point out that they’re slightly below the long-term historical average, including recessions. It’s about a third below the rate typically seen in expansions (leaving out recessions, that is). But considering how bad things have been, modestly subpar seems exuberant.
Average hourly earnings were up a very weak 0.1% for the fourth month in a row. For the year, hourly earnings were up 1.9%. They’ve been running between 1.8% and 2.0% for more than two years now—below the rate seen during the Great Recession, and barely keeping up with inflation. This weak growth rate suggests that low-wage jobs figure disproportionately figure in the recovery.
The figures I’ve been quoting come from the monthly survey of several hundred thousand employers. The BLS also does a simultaneous survey of about 60,000 households. It’s rich with demographic detail, but the smaller sample size makes it more error-prone, especially in the short term. But the establishment survey can miss new startups, which can be especially important in times of economic recovery. So disparities between the two surveys can be analytically interesting. This month, the household survey was considerably stronger than its establishment counterpart. Employment rose 428,000—or 879,000 when adjusted to match the payroll concept. The unemployment rate was unchanged, but only because of a significant influx of new and returning workers into the labor market. The number of job losers was down for the month. And the number of people quitting jobs voluntarilty rose. Clearly the public is feeling increasing confidence in the job market. “Hidden” unemployment measures were either flat or down. Those classed as not in the labor force but wanting a job now rose trivially, and those working part time for economic reasons fell strongly. The broadest measure of unemployment, U-6, which includes unwilling part-timers as well as those who’ve given up the job search as hopeless, fell 0.2 point to 14.9%. Though that’s still tragically high, it’s still this measure’s lowest level since February 2009.
Only the extreme duration of unemployment categories saw an increase – those jobless for less than 5 weeks, and those without work for 99 or more weeks. Though down from its highs, the number of long-term unemployed remains very sticky, which is both a serious social and economic problem, as millions hang on the margins with deteriorating skills and attachment. But the rise in the number of short-term unemployed is what you’d expect as frictional unemployment—people briefly unemployed between jobs—begins to take precedence over the recessionary kind.
More on the household survey’s outperformance. On one hand, it’s encouraging, and suggests that the payroll survey may be missing some business startups, as it often does early in recovery/expansion periods. But the adjusted household survey has been outperforming since 2006. It showed fewer job losses in the recession and has staged a stronger recovery over the last two years. While the missing new establishment story no doubt figures here, it’s also likely that some serious structural issues are at play.
For example, a 2009 NBER working paper (pdf) by Katharine Abraham, John Haltiwanger, Kristin Sandusky, and James Spletzer finds that workers at the demographic extremes are most likely to be categorized as employed in the household survey but not by the establishment survey. That is, poorly educated, immigrant, and minority workers are more likely than average to be employed off the books, and highly educated, skilled workers are more likely to be employed as independent contractors. Given employers’ reluctance to make lasting commitments these days, it’s quite likely that growth in these sorts of less-than-formal employment arrangements is an important factor in the labor market recovery (and also cushioned the recession’s blows). These are jobs, but not stable ones with benefits.
So, all in all, a decent employment report. But in 2010 and 2011 we saw a run of strength early in the year, followed by a deterioration as summer arrived. There was a similar pattern in 2004, but it did not repeat in 2005, as strength was sustained beyond spring. Perhaps this pattern of growth followed by a fade is a seasonal adjustment quirk, or a more structural feature of early recoveries. But maybe 2012 will be the year when recovery finally takes hold, and we can start putting all the post-crisis wobbliness behind us. Of course, we’ve still got over 5 million jobs to recover that we lost in the Great Recession, not even allowing for population increase. We also still have all kinds of long-term structural problems, like a lack of economic dynamism, polarization, a debt overhang. But things are getting better, slowly and tenuously better.