Philip Khor
Philip Khor

An Analysis of The Gender Wage Gap In Australia

Why does the gender gap matter?

The Australian national gender pay gap currently stands at 14.1%, and among Australian full-time employees, there is a gender pay gap of 21.3 per cent (WGEA 2019).

Paying equally is a simple matter of fairness — equal work deserves equal pay. A persistent gender wage gap results in less monetary incentive for women to invest in their human capital, and our economies will be worse off with limited access to human capital. An interesting example of how women’s limited participation in non-household work makes our societies worse off is where the presence of female legislators improved economic growth in Indian constituencies as women were less likely to be corrupt.

Reducing the wage gap also has the potential to balance bargaining power within households, since women have the financial means to live independently without their partner. There have been many explanations forwarded for the persistence of the gender wage gap:

  • The observed motherhood penalty in wages is perhaps due to discontinuities in work from childbearing and childrearing. There’s evidence even from gender-progressive countries such as Denmark [1] [2] that mothers’ earnings are adversely impacted years after childbirth.

  • Women’s uncompensated household work may not suit working hours, e.g. there’s evidence that women find it difficult to work overtime, which results in a gender earnings gap [3]. It does not help that many organizations do not offer flexible working arrangements.

  • Labour market discrimination (statistical- and taste-based). These two concepts distinguish between perception of traits of the average woman compared to an average man, and preference-based discrimination. For instance, women may be perceived as less committed workers than men.

  • Glass ceiling and limited access to mentorship. Women find it hard to access career growth opportunities in predominantly male-led organisations, where males may feel uncomfortable being mentors to women, but unknowingly make it difficult for women to progress in organisations. As a result, the men whom leaders are more familiar with are set for career growth while women are sidelined for promotion.

  • Socialisation of women in favour of lower-paying jobs. Girls are raised to believe that well-remunerated, male-dominated jobs such as engineering and computer science are unsuitable for them.

For this analysis, I’ll be looking at Taxation Statistics 2016-17, provided by the ATO and available for download [here]. Specifically, I’ll be looking at Table 14B, which records information on taxable income aggregated by ANZSCO occupational classification codes and sex for each income bracket for various taxation years. A caveat before I go any further: because I only consider totals here, the data may also reflect differences in working hours, e.g. males working more than females. As discussed, this may be a product of institutional barriers that limit women’s labour market participation, and should not be discounted as a source of gender inequity.

Distribution of salaries by gender

The following histograms show the distribution of average salaries across different occupational classes in Australia in 2013-14 and 2016-17, weighted by number of individuals. However, there may be large variation within occupations. These distributions don’t account for that variation and assume salaries are tightly distributed within each occupational category.

There seems to be a clear upward trajectory in the distribution of male salaries with the peak being around $100,000 per annum. In contrast, the distribution of female salaries seems more symmetric, and the peak for the salaries distribution for females is well under $100,000.

Overall, there’s not much insights we can get from looking at the distributions alone.

Average salaries by broad occupational category

I summed salaries by gender and took the average of salaries across individuals within ANZSCO occupational categories. On average, women earn 60-69% of males’ salaries, and this is the case across all occupational categories.

Digging deeper

For each detailed job category, I take the ratio of average salaries and wages for females and males. Across practically each broad job category, females earn less than males, sometimes as much as 50% less. In none of the professionals job categories do females earn more than males. Women earn higher wages in some of the community and personal service workers and clerical and administrative workers job categories.

Has the raw gender wage gap narrowed since 2013-14?

I calculate the gender gap as the percentage difference between male and female salaries, and then find the difference in percentages between 2013-14 and 2016-17. If it is positive, then the gender gap has widened, and vice versa.

The gender gap seems to have shrunk for the majority of jobs classified as managers, professionals, technicians and trades workers and clerical and administrative workers. However, it seems to have widened for the majority of jobs in the remaining categories.

What can we do?

  • Parental leave policies, both maternal and paternal are critical to help women resume work after childbirth.

  • Pay transparency can help reduce the gender wage gap by providing information to workers in their negotiations with employers.

  • Companies can introduce flexible working arrangements, which are paramount to facilitating women’s work. Policies such as remote working options, flexible working hours and accommodating childcare arrangements can help.

What can we do?

The Workforce Gender Equality Agency is a great resource to start with, and if you’re keen check out their Data Explorer.


This analysis is intended as an update to my earlier analysis on 2013-2014 statistics on the gender gap, available at GitHub. My prior analysis used taxable income, which I judged as too noisy. Differences between genders could be due to differences in deductibles between genders and other aspects of tax policy not related to the gender pay gap.

The code for the analysis is available a here.

Leave a Comment

Your email address will not be published. Required fields are marked *