Abstract:
Recent empirical studies of gender discrimination point to the importance of accurately controlling for accumulated labor market experience. Unfortunately in Australia, most data sets do not include information on actual experience. The current paper, using data from the National Social Science Survey 1984, examines the efficacy of imputing female labor market experience via the Zabalza and Arrufat (1985) method. The results suggest that the method provides a more accurate measure of experience than that provided by the traditional Mincer proxy. However, the imputation method is sensitive to the choice of identification restrictions. The authors suggest a novel alternative to a choice between arbitrary restrictions. Copyright 1997 by The Economic Society of Australia.