The Economic Rationales For Narrowing The Gender Pay Gap - Section One: The Gender Pay Gap
Trends in the Gender Pay Gap
The gender pay gap is one important indicator of gender differences in employment outcomes.
A commonly used measure of the gender pay gap in New Zealand is the gap between male and female average hourly earnings from wages and salaries. In the June quarter 2005, female average hourly earnings from wages and salaries were $16.80, 85.8 percent of male average hourly earnings of $19.58. The average hourly wages and salaries of Maori and Pacific women were substantially lower than those for women overall. Maori women's average hourly earnings were $15.22 or 77.7 percent of average male hourly earnings and the average earnings for Pacific women were $13.34 or 68.1 percent of average male hourly earnings. Pakeha women's average hourly earnings were $17.34 or 88.6 percent of average hourly earnings for men. An ethnic pay gap also exists for men. Maori men earned an average of $15.80 an hour, or 80.7 percent of the average male hourly earnings. The average hourly earnings of Pacific men were $15.13, or 77.3 percent of average male hourly earnings (Statistics New Zealand, 2006: Table 10).
Both male and female part-time workers earn less than their full-time counterparts. In the June quarter 2005, women part-time workers earned, on average, 82.1 percent of the earnings of part-time men, and 76.7 percent of the earnings of full-time men (Statistics New Zealand, 2006: Table 6.07).
The gender pay gap narrowed substantially, from 72 percent to 78 percent, during the five year implementation of Equal Pay Act 1972 (Hyman, 1994:84). Since that time, the pay gap has narrowed unevenly and slowly (Taskforce on Pay and Employment Equity, 2004:23). The size of the gender pay gap varies for different age groups, and between men and women with qualifications, compared with those who do not.
Why is there a gender pay gap?
Discrimination
Much analysis of the gender pay gap has focused on explaining the gap in terms of gender differences in human capital, i.e. the education, skills and experience men and women bring to their employment. Differences in human capital are regarded as a justified reason for pay differences, whereas differences in return to human capital are sometimes a consequence of discrimination.
The findings of pay gap studies on the extent of gender discrimination are debated, as are the "reasonable" explanations. For example, many studies find being a married woman and/or working part-time are correlated with lower pay, but there remains the question of why these factors should reduce pay.
There is agreement that direct wage discrimination occurs when workers of equal productivity to an employer are paid unequally. In addition, the existence of non-wage labour market discrimination through unequal access to jobs, training, promotion and other human-capital enhancing factors, is generally accepted.
Education and work experience
Researchers have typically found that around one-third of the pay gap can be attributed to productivity-related factors such as education and length of past work experience (Dixon, 2000: 16). Relative increases in women's human capital compared with men's, through higher education levels and more time spent in the workforce, has contributed to the narrowing of the pay gap.
Occupation and industry
Occupation and industry, i.e. job characteristics, account for up to a third of the pay gap (Dixon, 2000: 16). Walby and Olsen's review of studies found that, in addition to occupational segregation and industry, other characteristics of the firm, region, attitudinal differences and differentials in commuting times also contribute to the gender pay gap (Walby and Olsen, 2002: 49). The assumption that the occupations and industries where men and women work are freely chosen is controversial, and the elements of indirect discrimination that contribute to gender difference in occupations will be discussed later in this paper. Also contributing to this part of the pay gap is lower average pay rates in female dominated occupations:
From the point of view of equal pay for work of equal value, proportions of the gender pay gap attributable to occupation/industry breakdowns are simply a starting point to be problematised (Hyman, 2001).
Marriage, motherhood and part-time work
Recent studies show that, in addition to the amount of time individuals have been in the workforce, interruptions to labour force participation, and working part-time, independently exacerbate the gender pay gap.
Pay gap studies have also found evidence of a motherhood and marriage wage penalty that has been variously attributed to a mix of human capital (work experience), job tenure, and lower pay rates in part-time jobs (Dixon, 2000: 17).
Several studies have found a bifurcation in the characteristics of mothers that work full-time and part-time. In one longitudinal study in the United States (US), low income mothers were more likely to work part-time, less likely to have qualifications, and worked in a narrower range of occupations. In contrast, the higher earning mothers were more likely to be qualified, work full-time and work for government. Job turnover was twice as high for the low income group (Lee, 2004). Another US study found evidence of a widening pay gap between married women and other women, the "family gap", that could only be partly explained by differences between the groups (Waldfogel, 1998).
Impact of wage dispersion
Studies have found considerable variation in the pay gap at different points on the earnings distribution. Some studies have drawn attention to the fact that the greater the wage dispersion is, the higher the pay gap is likely to be, due to the concentration of women amongst low earners. This point is captured by the "swimming upstream" metaphor used to explain how a widening dispersion of wages reduced the expected contraction of the gender pay gap in the US at a time when women's earnings rose rapidly (Blau and Kahn, 1997). A corollary is that mechanisms that reduce wage dispersion, such as increases in the minimum wage, tend to narrow the pay gap as disproportionately more women than men are paid at or near the minimum wage.
Institutional factors make a difference to wage dispersion. The presence, and legitimacy, of trade unions and laws governing human rights, equal employment opportunities (EEO), industrial relations and minimum wages, all impact on wage dispersion. A narrower dispersion of wages may also be more likely to occur in countries where benefit provisions or restrictive labour market regulations exclude less productive workers from the labour market (Glaeser, 2005).
Employment rates
Olivetti and Petrongolo concluded that if Ireland, France and Southern Europe had similar female employment rates[1] to the US and the United Kingdom (UK), the gender wage gaps would widen and tend to match the higher US and UK levels. This is because, in the former group of countries, lower female employment rates are also associated with employed women having higher levels of education and experience on average, than the women not in employment (Olivetti and Petrongolo, 2005).
The Gender Pay Gap in New Zealand
Dixon (2000) undertook the most detailed analysis of the gender pay gap in New Zealand, using data from the Household Economic Survey and Household Labour Force Survey Income Supplement. Her analysis of the gender pay gap changes between 1984-1988, and the further analysis of the narrowing of the gap between 1997-2003 (Dixon, 2004), took a human capital inspired approach and use Blinder-Oaxaca decomposition techniques. This approach is focused on individual attributes, and Dixon notes data limitations do not enable robust analysis that takes into account firm-level, societal and institutional factors (Dixon, 2000: 23). Appendix Two discusses some of the more technical issues and debates around the wage statistic chosen to measure the gap, and assumptions underpinning the decomposition of the pay gap.
Dixon's analysis concluded, from a number of different approaches, that four variables; education, past work experience, occupation and industry, accounted for between 40-80 percent of the gender wage differential, leaving 20-60 percent not accounted for. Proxies for skill (educational attainment and experience) accounted for 30-60 percent, and occupation and industrial distribution 20-40 percent. Dixon commented:
Although a large proportion of the contemporary make-female gap in average hourly earnings can be attributed to these differences in measured skills and job characteristics, that should not be taken to imply that the 'explained' portion of the gender pay gap is fair, efficient or justified (Dixon, 2000: 9).
Dixon found that mothers in the 25-29 and 30-34 age groups earn substantially less per hour than women without dependent children, with sole mothers have the lowest average hourly pay (from 1997-98 IS). For these women, wage regression (holding constant difference in age, educational qualification and marital status) showed the hourly earnings penalty for one child is around - 7 percent and for two or more children - 10 percent. Sole mothers in waged employment were estimated to experience wage penalties that were 1.5 times larger than those experienced by partnered mothers. When experience levels were imputed into the equations to account for the time mothers take out of the workforce, however, hourly earnings penalties reduced to insignificant levels in some cases. Dixon cautioned against drawing firm conclusions given the limitations of the data and methods used.
Dixon did not, however, find evidence in New Zealand of mothers in part-time work being paid less than similarly skilled and experienced mothers in full-time work, although this has been reported in a number of qualitative studies. Twenty percent of women in an Equal Employment Opportunities Trust online survey reported they returned from parental leave to a lower level job (EEO Trust, 2005). A study of 42 mothers with one employer found that of the 43 percent who had resigned, 33 percent returned to lower level positions, 40 percent to positions at the same level and 27 percent had a change of career (Mason, 2000). Bryne (2002) who studied women combining paid work and parenting, similarly found examples of occupational downgrading, even for some women returning from parental leave, as a consequence of restructuring and technological change.
The judgments involved in separating out discrimination from the other factors that have led to women and men largely working in different occupations, and having different patterns of paid and unpaid work, are highlighted by two New Zealand pay gap studies that came to quite different conclusions about discrimination.
Alexander, Genc and Jaforullah (2004) concluded from four years of Income Survey data (1997-2000), and controlling for productivity characteristics including age, household type, qualifications, occupational class, marital status and location, that there was evidence of significant ethnic and gender wage differentials. They acknowledged that part of these differentials were related to disadvantage in accessing employment, i.e. less time in work and more breaks from work, but regarded this as being part of the discrimination picture.
In contrast, a study of the public sector found that productivity related factor accounted for almost all the gender pay gap within that sector (Gosse and Ganesh, 2004). This study accepted occupational segregation as an explanatory variable. It found that there was no gender bias in how job size points were translated into salaries. It did not, however, address the issue of gender bias in job sizing which is a key contention of pay equity advocates.
The Gender Pay Gap in the United Kingdom
Walby and Olsen's analysis of the UK gender pay gap drew on British Household Panel Survey longitudinal data from 1990-1999. The authors recognised discrimination as one element in the "unexplained" portion of the gender pay gap, but identified indirect discrimination, such as constraints facing part-time workers, as key to the differences in women's and men's skill acquisition and progress once in the workforce.
The table below summarises Walby and Olsen's findings.
| Component | Women's levels compared with men's | Percent of gap |
|---|---|---|
| Full-time employment experience | -7.7 years | 26 |
| Interruptions due to family care | +3.2 years | 15 |
| Part-time employment experience | +4.1 years | 12 |
| Education | -0.3 years | 6 |
| Segregation | .34/.70* | 13 |
| Discrimination and other factors associated with being female | 29 |
* This is an index of the extent to which women and men work in male-dominated occupations and shows that women work in occupations that are 34 percent male whilst men work in occupations that are 70 percent male.
What is striking here is that differences in employment experience account for nearly 40 percent of the gap, with interruptions to labour force participation accounting for a further 15 percent, compared with only 6 percent of the gap being attributed to differences in women's and men's education and skills. This suggests wastage or obsolescence of skills, that is, what happens after women join the workforce, is key to both using skills fully, and ensuring continued access to skills over working life.
Previous page | Contents | Next page
[1] Employment rates refer to the proportion of the working age population in employment, that is, they are neither unemployed nor out of the workforce.
