Spells without health insurance: the distribution of durations when left-censored spells are included.

In a previous paper (1990), we estimated the distribution of durations of uninsured spells for which the beginning could be observed. We raised the concern that uninsured spells that were in progress when the Survey of Income and Program Participation began might not be proportionally represented in a sample of uninsured spells restricted to spells with observed beginnings. This paper includes spells in progress when SIPP began. Including left-censored spells, and correcting for length bias inherent in such spells, does not fundamentally change our previous estimate of the distribution.