Last week we were being warned that the labor market was too tight and an outbreak of inflation was imminent. Today, Edward Lazear, formerly chief economist in the Bush administration, is telling us that labor demand is plummeting as shown by a reduction in the length of the average workweek. Okay so the labor market is too tight and too weak, that looks really bad. Given that economists also tell us that robots will take all our jobs and that a growing population of retirees will leave us with a shortage of workers it is remarkable that the public doesn't insist on hanging us all.

But Lazear does raise an interesting point that deserves to be considered. The Labor Department's establishment survey shows a notable decline in the length of the average workweek from 34.5 hours in November (he picks September as his starting point) to 34.2 hours in February. In work time, this is equivalent to a loss of more than 1 million jobs. As a result of this drop in hours, the index of total hours worked actually shows a decline of 0.1 percent since September and 0.6 percent since November. What do we make of this?

There are two questions that come up. First how reliable are these data and second, are they consistent with other data?

On the first point, there are grounds for skepticism. As Lazear notes in the piece, weather could have been a factor in shortening workweeks in February. Much of the Northeast-Midwest was hit by unusually bad weather that may have led people to miss a day or two of work. If we look back over the last three decades, there are other instances of comparable declines in the length of the average workweek that don't seem to correspond to downturns in the economy.

For example, we saw the same drop from 34.5 hours in December of 1994 to 34.2 hours in May of 1995, a period when the recovery was moving along at a healthy pace. (These data refer to the hours of production and non-supervisory workers, since the all workers index does not go back before 2006.) The average workweek declined from 34.7 hours to 34.4 hours between January of 1989 and May of 1989, again a period where the economy was still growing at a healthy rate. And it fell from 34.8 hours in November of 1987 to 34.5 hours in March of 1988. There was an even sharper drop from 35.0 hours in January of 1986 to 34.6 hours in July of 1986. In short, we have seen these sorts of drops before even as the economy was growing and adding jobs at a healthy pace. 

The other question is whether the data are consistent with what we see in other series. Here the answer is clearly no. The separate survey of households shows that the number of people employed part-time (defined as less than 35 hours a week) has actually fallen by more than 0.3 percent since September. This survey is much less reliable than the establishment data. Also, it is possible that workers may working fewer hours but not crossing the 35 hour threshold to be classified as part-time. (We could also directly analyze the micro data, but that's more effort than I am prepared to put into this question just now.) Anyhow, the published data from the household survey clearly gives no support to the declining hours story.

So, what should we make of this drop in hours? For the moment, the answer should be not much. Let's see what happens in the next two months. There is a story that Obamacare will lead to a reduction in the length of the average workweek, as many people will no longer need to work a full-time job to get health care insurance. That is one of the good things about the Affordable Care Act in my book. However, a reduction in hours for this reason should be largely offset by an increase in the number of jobs. If that proves to be the case, that will be good news.


Note: Typos fixed, thanks Robert Salzberg.