With Unemployment, Don't Forget Holiday Hiring

There are times during the year when more jobs are created in the U.S. than we are told about. The holiday season is one of those times.

When the media reports monthly changes in seasonally adjusted job numbers, they often don't say that the numbers are seasonally adjusted. Even the government agency that releases the monthly employment data, the Bureau of Labor Statistics (BLS), no longer says "seasonally adjusted" in its leadoff on job change.

But there's a lot of seasonal job growth that takes place in late fall and early winter - even during recessions - but this growth gets wiped out by the application of seasonal adjustment factors to the raw data.

That's not to say that seasonally adjusting the jobs data, or for that matter any time series affected by seasonal fluctuations, is wrong or undesirable. Seasonal adjustment is important and enables data users to more clearly observe the cyclical and secular behavior in time series; behavior that's critical to understanding the economy and guiding economic policy.

Nevertheless, significant changes in month-to-month changes in employment go unobserved. Even if seasonal jobs are temporary, that does not make them unimportant.

In its latest release on the employment situation, the BLS reported that nonfarm payroll jobs were "little changed" in November (+ 39,000). The average monthly increase for the past three months, September to November, was a modest 62,000.

Prior to seasonal adjustment, however, nonfarm jobs rose by 217,000 in November, and were up by an average 559,000 a month in the September to November build-up for the Thanksgiving and Christmas holidays. Historically, job changes in these months have been consistently more positive for the unadjusted data than for the seasonally adjusted data, not unexpectedly.

Looking at the October to November change in jobs going back to the year 2000, the average monthly difference (net increase) was 144,000 greater for non-seasonally adjusted jobs than for seasonally adjusted jobs.

Last month private sector jobs, seasonally adjusted, were up a meager 50,000. Before seasonal adjustment, they were up by 101,000, twice as much. Health care jobs, after seasonal adjustment, increased by 19,000, but were up 41,000 prior to adjustment. But the big difference was in retail trade. After adjustment, retail jobs dropped by 28,000 compared to a 300,000 jump before seasonal adjustment, the difference mainly reflecting Christmas hiring.

True, temporary workers hired for the holidays will be let go by year-end, and over the course of a year seasonal hirings and firings are offsetting. But we lose information when real-world monthly changes in jobs are given a back seat and go unnoticed.
Consider a jobseeker unemployed for six months. If that person finds a temporary job for the holidays in a department store, even knowing that he or she will soon lose that job, that worker is nevertheless better off earning temporary wages than no wages at all. More income will be available for spending and unemployment compensation is suspended, even if only for a while.

The 300,000 jobseekers who found work in retail trade last month amounted to fewer than the usual number who became employed for that month in past years, so the gain got seasonally adjusted into a job decline. Still, that's a lot of new jobs.

We don't always see the reality underlying the data that officials and the media choose to emphasize. But there are a lot of holiday workers a lot less unhappy today than when they were unemployed, and we are all the better off for it.