This paper discusses first various ways of measuring unemployment and, borrowing ideas from the poverty measurement literature, proposes four more general unemployment indices which are parallel to Sen poverty index, to its generalization by Shorrocks, to the FGT, and to the Watts poverty indices. It then presents an empirical illustration based on Swiss data at the level of the “canton.” More precisely, using the so-called Shapley decomposition, it computes the contribution to the difference between the value of each of these four unemployment indices in a given “canton” and in Switzerland as a whole, of three components measuring, respectively, the impact of differences in the traditional unemployment rate, in the average unemployment duration, and in the inequality in the unemployment durations. The paper ends by discussing the impact on the results obtained of assumptions made concerning the maximum unemployment duration.