The topline of this morning’s jobs report shows nonfarm payrolls expanding by 214,000 jobs in October. It’s the sort of number that suggests a narrative of a continuing, albeit disappointing, economic recovery.
But dig a little deeper and you find a much sunnier story. When read carefully, today’s report provides more evidence that the labor market is improving quite rapidly, and that the recovery has gained momentum over the past year.
To start with, realize that there are two separate surveys of employment. The survey that garners the most attention is based on asking (nonfarm) firms how many people are on their payroll, and this is the count that suggests employment grew by 214,000.
The Labor Department statisticians separately survey households, asking people if they have a job. This alternative measure of employment receives less emphasis, because it tends to be extremely volatile. Bearing this disclaimer in mind, it is notable that the household survey suggests that employment grew by an impressive 683,000 in October. While this number should be treated with substantial caution — it surely overstates the actual pace of employment growth — it should not be discounted completely.
My preferred rule of thumb is to put 80 percent weight on the nonfarm payrolls survey, and 20 percent weight on the household survey. This weighted average, which probably better describes underlying labor market conditions, suggests that employment grew by a very healthy 308,000 jobs in October.
There are also good reasons to think that the nonfarm payrolls estimate understates the underlying strength of the labor market. The data that the government statisticians report is seasonally adjusted, which means that it accounts for the fact that hiring is usually stronger in the run-up to Christmas. But research by Jonathan Wright, an economics professor at Johns Hopkins University, suggests that the algorithm that the government statisticians use is not properly tuned. (Full disclosure: He published this research in the Brookings Papers on Economic Activity, which is a journal I coedit with David Romer.) Each month, Mr. Wright calculates an improved estimate based on a better-tuned algorithm, and publishes it on the Brookings website. His improved estimate suggests that nonfarm payrolls actually grew by 233,000 jobs, rather than the 214,000 estimated by the government statisticians.
Finally, today’s numbers are simply the first estimate made by the Bureau of Labor Statistics, and they will be revised as more information comes to light. For reasons that remain largely mysterious, these initial estimates have tended to be highly cyclical, underestimating job losses during the recession, and then systematically underestimating job gains during the recovery. During the current economic recovery, the initial estimate of nonfarm payrolls has understated monthly employment growth by an average of 28,500 jobs.
If this pattern continues, it is highly likely that the latest estimates will be revised upward. Indeed, the initial estimate for job growth in August has been revised from a disappointing 142,000 to 203,000, and the estimate for September has already been revised from 248,000 to 256,000.
The old saying is that the devil is in the details. Today’s report suggests just the opposite — dig into the details, and you’ll find a much sunnier story about the labor market. All told, I read today’s report as providing more evidence that the labor market recovery is gaining steam.