As everybody knows, unemployment rates differ widely between occupations and between regions, as well as across age, race and (sometimes) sex groups. The striking thing is how stable these differences are. In all countries unskilled people have much higher unemployment rates than skilled people. Similarly, youths have higher rates than adults. In addition in most countries (though not the United States) regional differences are highly persistent – with unemployment always above average, for example, in the North of England and the South of Italy.

The first task is to document these differences (in section 1) and then to explain them (in section 2). An obvious question is why occupational and geographical mobility does not eliminate the differences between unemployment rates in different occupations and different regions. We attempt to answer this question. Our main focus is thus on the persistent imbalance between the supply and demand for labour across skill groups, regions and age groups. But there are additional imbalances which are temporary. Suppose, for example, that there are two occupations which have the same average unemployment rate over time but in one year demand shifts from one occupation to the other; this will produce a temporary imbalance until corrected. Such ‘one-off’ structural shocks have aroused great interest in relation to the issue of real business cycles (see Lilien, 1982). They are also clearly of interest to the unemployed themselves. But they account for a fairly small fraction of the inequality among unemployment rates observed in the average year.