This estimate is based on a rather aggregative view of the labor market. It is important to dig deeper to get a better understanding of the problem, and there is a considerable amount of research under way exploring the quantitative importance of the various forms of mismatch. For example, the International Monetary Fund has recently released a special study based on a new state-by-state measure of the gap between demand and supply for workers with different levels of educational attainment. The study examines the impact of this variable and the foreclosure rate on state-level unemployment. It estimates that 1.5 percentage points of the national unemployment rate is due to these two sources and their interaction. Thus, according to this study, these two types of mismatch alone can account for a significant fraction of my estimate of 2.5 percentage points.
Good economic policy is about using the right tool for the problem at hand. The mismatch problems in the labor market do not strike me as readily amenable to the kinds of monetary policy tools currently available to the Fed.
Oh well.