The Economic Rationales For Narrowing The Gender Pay Gap - Appendices
Appendix One: Terms of Reference
The paper was tasked with addressing the following questions:
- What does evidence point to as the key factors influencing the pay gap and pay and employment equity and the key ways to improve the position for women?
- What are the economic rationales for pay and employment equity?
- What are the economic benefits of pay and employment equity at both the enterprise and economy-wide level?
- What other theoretical or emerging issues are likely to impact on the future need for, and sustainability of, interventions and strategies to narrow the pay gap and improve pay and employment equity, particularly at the lower paying segment of the labour market?
Appendix Two: A note on measurement and decomposition issues in analysing the gender pay gap
The gender pay gap varies depending on whether the comparison is made between median hourly earnings, the arithmetic mean of hourly earnings, or the geometric mean hourly earnings. In 2003, these measures produced gender pay ratios of 88.1 percent, 86.4 percent, and 88.3 percent respectively (Dixon, 2004: 4). Dixon justifies using the geometric mean as it is less sensitive to changes in the upper tail of the earnings distribution than are arithmetic means. Hyman, on the other hand, questions the discounting of very high incomes, given the over-representation of males amongst the group (Hyman, 2001).
The focus of decomposition techniques is to separate out differences in the level of human capital and differences in returns to human capital. Differences in human capital are regarded as a justified reason for pay differences whereas differences in return to human capital are seen as potentially discriminatory unless there are other explanations.
Most statistical analyses of the gender pay gap have used the Oaxaca-Blinder decomposition technique which emphasises individual characteristics, such as women's lower levels of "human capital" (education and experience) and the individual choices they make about the jobs they do and the industries in which they work.
The Oaxaca-Blinder approach has been criticised for assuming a clear dichotomy between human capital and discrimination, as the two important contributors to the gap, and has been described as "focusing primarily on ... women's deficiencies relative to the attributes of men" (Rubery et al 2002: 132). This focus on human capital theory is seen as not adequately accommodating important institutional, sociological and organisational factors that impact on the gender pay gap (Todd and Eveline, 2004: 24).
Walby and Olsen identify four main problems with the Oaxaca-Blinder method:
- The decomposition uses the individual as the unit of analysis and therefore does not adapt to factors that do not take a human capital form, such as segregation and union membership
- The model assumes the factors of analysis can be added, however, some overlap
- The model does not deal well with situations where a variable has a significant presence for one sex, but not for the other, for example breaks from employment and part-time work
- By separating out the women's and men's equations, women are compared with women, and men with men, which makes it difficult to explore discrimination. For example, separate male and female analyses of the association between education and wages suggest that women receive a higher return from education. This is misleading as, at equal levels of education, women get paid less than men on average (Walby and Olsen, 2002: 108-9).