Unemployed Americans attend a National Career Fair. (Mark Ralston-AFP/Getty Images)

Why is unemployment still so high? Are people still just not spending enough? Or is there a structural problem at work?

Most economists think it's the former. They will argue that the unemployment crisis is all about "aggregate demand" — that is, people just aren't buying enough stuff. Consumers aren't spending and businesses aren't buying from each other, so businesses don't have money to hire people and mass unemployment persists. This explanation was first put forth by John Maynard Keynes in "The General Theory of Employment, Interest and Money" and continues to be accepted by current Keynesians like Paul Krugman and Larry Summers. Yet even monetarists, the Keynesians' main critics, agree that most recessions are due to a lack of demand.

Other economists, however, have challenged this picture, arguing that even if the underlying recession is due to insufficient demand, structural factors are also keeping unemployment high. Nobel laureate Michael Spence is the most prominent advocate of this view. The real problem, Spence argues, is that too many of America's jobs are in  "non-tradable" sectors like nursing or construction. That leaves our "tradable" sector, which produces goods and services for export, at a disadvantage at a time when other countries' tradable sectors are growing. We need to invest heavily in education to make our tradable sector more productive and increase employment there, Spence argues. Sachs agrees that increasing the workforce's education and skill level is more important than spurring greater demand.

If this theory is true, then current unemployment is due to a skills mismatch — people don't just have the skills to do the jobs we need them to do. The solution to this is to increase education so that people can handle those jobs, rather than to increase demand. There's one problem: This isn't the main reason unemployment is high now. A new paper (pdf) from New York Fed economists Aysegul Sahin, Joseph Song, Giorgio Topa and Giovanni L. Violante finds that a skills mismatch is responsible for at most a 1.5 percentage point increase in the unemployment rate during the recession. Given that unemployment shot up by about 5 points, that leaves the other 3.5 percentage points explained by a shortfall in demand. What's more, the effects of skill mismatch only really appear during the recession:

Source: Federal Reserve Bank of New York

Both lines show the increase in unemployment for each period caused by a skills mismatch. The red dotted line adjusts for differences between economic sectors, while the black line doesn't. In either case, both measures find that skills mismatches have a very large effect only in the aftermath of the 2001 and 2008 recessions, whereas in 2007 mismatches only increased unemployment by about 0.2 percentage points. This suggests that addressing the underlying lack of demand would prevent mismatches from hurting unemployment. This makes intuitive sense. When there are a lot of jobs due to healthy demand, the likelihood that a given person can find a job that matches their skill set is much higher. When there's low demand and jobs are scarce anyway, chances of finding a job matching one's skill set are low.

All of which suggests that the best response to the jobs crisis is one that spurs aggregate demand, through looser monetary policy, fiscal stimulus or another means, rather than one that tries to build up skills for workers. The latter policy may be desirable for other reasons, but it's really neither here nor there on the jobs crisis.