The Definition of Part-Time Employment: A Switching Regression Model with Unknown Sample Selection

Rejecting the standard less-than-thirty-five-hours-per-week categorization of part-time workers, this paper offers an empirically-justified definition of the part-time employed. The definition is based on the distinction through wage payments that employers make between a high-hours (full-time) and a low-hours (part-time) group of workers. A switching regression model of wage equations with deterministic, but unknown, sample selection is estimated resulting in a significant split between high-hours and low-hours workers that is higher than the standard definition. Wage differentials are calculated and decomposed, and estimates from the structural choice model are presented.