s for paid vacation leave during the summer CANADIAN PUBLIC POLICY ANALYSE DE POLITIQUES, VOL. XXVIII, SUPPLEMENT/NUMERO SPECIAL 1 2002 This content downloaded from 207.46.13.111 on Tue, 13 Dec 2016 04:36:47 UTC All use subject to http://about.jstor.org/terms Can the Workplace Explain Canadian Gender Pay Differentials? S57 months. For example, it is unclear if total employment captures the number of persons at work during the summer months OR if total employment is the number of workers employed by the firm whether these workers are on paid vacation or not. '0Related to the coincidence of needs hypothesis, Reilly and Wirjanto (1999b) suggest that establishments with stable product/service may be better able to plan for the of labour and to accommodate women's employment patterns. If this is the case, then women workers may crowd into these workplaces with attractive (stable employment) but costly (lower wages) job characteristics. Reilly and Wirjanto create a variable for stable product based on the following question: On a scale of 1 to 9, would you say that this establishment faces a that is 1 = highly stable to 9 highly unstable? The WES does not contain information directly on the of product but does include performance measures (i.e., increased/decreased/same) for productivity, sales growth and profitability over the previous 12 months. These measures are based on a short period of time (12 months) and a trend of the stability in product demand is impossible to determine with the given data. For this reason, in product is not included in this analysis. The measure used by Reilly and Wirjanto (1999b) can be criticized since it is unclear whether all establishments used the same reference period and measures to characterize stability. is considerable debate over the inclusion of industry and occupation as explanatory variables since occupational segregation is a mechanism whereby wage discrimination may occur and their inclusion may undervalue the importance of labour market constraints on wages or alternatively, may over-justify gender-pay differentials. However, excluding occupation and industry may neglect the importance of background or individual decisions with respect to wage outcomes. 12The regression results on worker characteristics are included in Appendix Table Al. '3The F test (adjusted Wald test) suggests that the workplace characteristics are jointly and significantly different from zero. When industry is included among the workplace characteristics hypothesized to be zero, the F statistics are F(28,4508) = 24.98 for men and F(28,4062) = 19.46 for women. When industry is excluded from the set of workplace characteristics hypothesized to be zero, the F statistics are F(15,4521) = 25.39 for men and F(15,4075) = 11.92 for women. 14These elasticities are the product of the proportion of part-time workers in the workplace and the coefficient on the percent working part-time from Specification 4. For example, the elasticity for women is calculated as -0.055 = -0.163*0.3371. 15The exception is for women. 16The F test (adjusted Wald test) reveals that the variables on the degree of participation in self-directed workgroups are jointly and significantly different from zero with the exception of Specification 4 for women. '7The gender difference in the returns to the foreignheld workplaces is statistically significant at the 5 percent level. 18Table A2 in the Appendix documents the differences in the characteristics that men and women bring to the labour market. Roughly 19.5 percent (13.7 percent) of men and 11.7 percent (18.6 percent) of women are employed in managerial (professional) occupations. The receipt rates of additional compensation for men in women are 41.7 percent and 38.5 percent in managerial occupations and 29 percent and 23.2 percent for professional occupations. There is no significant gender difference in the receipt of additional compensation for managers or professionals. 20Roughly 25.4 percent of married women, 29.2 percent of single, never married women, 25.3 percent of women with dependent children and 28.9 percent of women with no dependent children earn any form of additional compensation. 21The WES asks respondents if they received any classroom or on-the-job the training in the previous 12 months. About 28.8 percent of men and 31.4 percent of women participated in on-the-job training while 36.9 percent of men and 37.4 percent of women participated in classroom training. The incidence of training is not used in the regression analysis since the impact of training on wages may not be realized immediately (i.e., within the previous 12 months). The impact of training on wages would be more accurately reflected in a measure of training that incorporates all training taken throughout the worker's career. CANADIAN PUBLIC POLICY ANALYSE DE POLITIQUES, VOL. XXVIII, SUPPLEMENT/NUMERO SPECIAL 1 2002 This content downloaded from 207.46.13.111 on Tue, 13 Dec 2016 04:36:47 UTC All use subject to http://about.jstor.org/terms
Read full abstract