Full bibliography

The results of research on the determinants of unemployment spell durations of individuals experiencing job separations in each year from 1978 to 1980 and from 1982 to 1985 are presented. Accelerated failure time models that incorporate explicit assumptions concerning the functional form of the baseline hazard are estimated for each year, and for a variety of functional forms. Cox's (1972) proportional hazards procedure is also employed. The results obtained are robust to the functional form assumed, but not necessarily to the year of data used. It is found that the average duration of an unemployment spell increased significantly as the economy moved into recession during the early 1980s, and that, for the most part, it decreased during the subsequent recovery. However, even though the aggregate unemployment rate fell between 1984 and 1985, the average duration of an unemployment spell increased. For some demographic groups, economic recovery does little to reduce unemployment spell durations. Most notable in this regard is the change in the relationship between age and spell duration.