英语论文网

capture the differences in the shape of the densities of male and female wages (Breunig and Rospabe, 2007). This can be important to take into account given that female wages tend to be more clustered in the lower part of the wage distribution than those of men. Despite that it does give a good approximation for the magnitude of existing discrimination in a sample. For this reason it has been the most commonly used technique and this paper is no exception.

Pay-equity paylines

Another interesting empirical procedure is the method of obtaining regression estimates of male and female pay-lines as used in pay equity. It has not been applied in the research literature but it can provide an estimate for discrimination within organisations. The methodology is similar to that used in the traditional human capital regression decomposition. Pay-equity procedures typically involve four generic steps (Gunderson, 1995). Firstly, predominantly female job categories and predominantly male job categories in the organisation are indentified. A selected criterion such as 70% or more of either gender in the job is used. Secondly, the jobs are assessed in line with a sex-neutral job evaluation scheme where job-assessment points are attributed to the components of skill, effort, responsibility and working conditions. A total job-assessment score is then calculated by adding those points. It is then possible to weigh the importance of each component as the range of points allocated to different components reflects it. Thirdly, pay as a dependent variable is regressed on the job assessment scores as the single explanatory variable, separately for male and female dominated jobs. Finally pay adjustments, i.e. compensations, are made usually by increasing the wage in female-dominated jobs so that they meet the payline of the same job-assessment score.

Differences in the male and female paylines are taken as a measure of discrimination in the sense of unequal pay for work of equal value, where value is indicated by the job evaluation point score (Gunderson, 2006). The score could even be a more suitable measure of productivity than the traditional human capital variables, since the score is designed from elements that employers noted as being important to them in establishing the price of the job. On the downside, the pay-equity payline procedure can only be adopted on an organisation-wide basis.

Narrowly defined occupations

This empirical procedure entails the comparison of male-female pay gaps within the same narrowly defined occupation. This method is plausible in a way that productivity-related effects such as education, experience, training, and health status are controlled. It is also true for the unobserved factors such as commitment, motivation, and diligence. It becomes reasonable as well as convenient to assume that employees in such narrowly defined occupations are equally productive, because they are qualified enough for a specific job (i.e. met the requirements of that position).

A good application of this method can be found in the working paper by Bayard, Hellerstein, Neumark and Troske (1999). A large matched employer-employee data set covering essentially all industries and occupations across all regions of the U.S. was assembled. As in many other works, it was discovered that a large part of the gender pay gap can be accounted for by the concentration of f