Unemployment spells and unemployment experience

This paper finds that the average length of single spells of unemployment is longer than the average length of multiple spells. The conclusion appears robust, being due neither to the underreporting of unemployment experience by teenagers and women, nor to the interruption of uncompleted spells by the beginning or end of the calendar year, nor to the finite period of observation.
Consequently, this paper strongly suggests that statistics on average unemployment durations of completed spells seriously underestimate the unemployment experience
of all groups of persons considered in this paper: persons with single spells, because on
average single spells are longer than multiple spells; and persons with multiple spells,
because the unemployment experience of these persons includes the multiplicity of their spells.