[This paper proposes an econometric methodology for estimating the conditional probability of an individual leaving unemployment in any particular week of his spell out of work. We use this method with cross-section data on uncompleted unemployment spells to investigate a number of questions concerning unemploymentduration. In particular we look at the impact of unemployment benefits and how this changes over the course of an unemployment spell. We also consider how our results are affected by unobserved sample heterogeneity and present estimates which, to some extent, take account of this problem.]

The relationship between education and unemployment incidence, although theoretically clear-cut, has rarely been quantified. Some estimates of this relationship are obtained for Britain, and these are split into the impact of schooling and qualifications on the probability of entering unemployment and on the expected duration of spells within that state. Finally, estimates of the rate of return to schooling are adjusted to take account of unemployment incidence.

The relationship between education and unemployment incidence, although theoretically clear-cut, has rarely been quantified. Some estimates of this relationship are obtained for Britain, and these are split into the impact of schooling and qualifications on the probability of entering unemployment and on the expected duration of spells within that state. Finally, estimates of the rate of return to schooling are adjusted to take account of unemployment incidence.