Casey Mulligan offers a chart of the nonseasonally adjusted teenage employment data as proof that aggregate demand effects are not the dominate factor behind low employment. It is such a silly analysis that the instinct is to immediately dismiss it without comment. I see, however, that Greg Mankiw has anointed the Mulligan analysis as an important piece of evidence:
Casey points out that there is a regular surge in teenage employment during the summer months because more teenagers are available to work (that is, the supply of their labor has increased). That is no surprise: It is normal supply and demand in action. But if aggregate demand were the main constraint on employment, this increase in supply should not translate into higher employment during deep recessions such as this one. But it does!
It really makes one understand why the public often dismisses academics as out of touch in their ivory towers. One has to imagine that neither Mulligan nor Mankiw ever held a real summer job. Nor, apparently, have they looked at any other nonseasonally adjusted data. Nor do they appear to have much understanding of the basic ebb and flow of US economic activity over the course of the year.
Mulligan and Mankiw both look at the seasonal increase in teenage employment during the summer and conclude that of course this should not happen in a period of weak aggregate demand. Therefore, the hiring must be driven by the sudden influx in teenage employment during the summer. It seems to entirely escape them that aggregate demand has a very predictable season pattern - a seasonal pattern that exists in a recession or expansion.
These seasonal patterns in demand activity are not new. Indeed, I imagine that if the data existed, we would see the pattern has remained virtually unchanged since the dawn of human existence, as least in parts of the world where seasonal weather patterns govern economic life. Indeed, it is the reason we have an influx of teenage labor in the summer. It is a throwback to the days of America's agricultural past, when the DEMAND for additional labor in the summer months necessitated closing schools for the summer.
Are these seasonal patterns still evident? Turn to the NSA retail data series:
As sure as the sun rises each day and winter turns to spring, sales spike at the end of the year as the holiday season approaches, collapse at the beginning of the year, rise in the summer, and then decline in the fall.
You can set your clock to this trend. Every retail analyst knows this trend. Every teenager who has ever held a summer job knows this trend. And a huge swath of data follows similar trends, albeit usually without the pronounced end of year impact. Building activity, waste disposal, tourism, you name it, it has a seasonal demand pattern. I only have to look out my ivory tower to see it - stuff grows faster in the summer, and the city hires crews of teenagers to cut it back. The pool down the road is open and staffed by teenage lifeguards only in the summer. Not because the lifeguards are available, but because there is no demand for an outdoor pool most of the year in Eugene.
Aggregate employment data has always followed this trend:
Recession or expansion, the demand for labor increases in the summer and winter, following the patterns of demand. In the summer, some of this demand is satisfied by a seasonal rise in teenage labor supply. This isn’t exactly brain surgery.
Mulligan and Mankiw have not found any smoking gun that refutes Keynesianism. They do unwittingly reveal that they don't really understand how to use nonseasonally adjusted data. But, more importantly, they reveals a remarkable ignorance of - and, presumably, lack of appreciation - for the ebb and flow of economic life in America. It is almost like not understanding basic elements of your own culture.