Factors That Determine the Wage Rate of Women in the Australian Labour Market

Introduction

Research Problem

In Australia, the gender pay gap hovered between 14 per cent and 19 per cent (Workplace Gender Equality Agency (WGEA), 2019). Its  influenced by factors such as 1) discrimination in hiring and pay decisions; 2) women and men working in different occupations and female-dominated occupations paying lower wages; 3) women’s double workload between paid work and unpaid caring and domestic work; 4) lack of workplace flexibility to accommodate caring and other responsibilities and 5) women’s career break due to caring purposes. Long term trend between 2001 and 2017 shows average weekly wage rate for full-time female and male employees increase by 24.0 per cent and 20.9 per cent respectively (Geoff, 2019).

The overarching research problem is to analyse the factors that determine the wage rate of women in the Australian labour market. To address these factors, three dimensions of the labour market were measured: education attainment, occupation category and total hours of unpaid care work. Despite the importance of issues relating to women, there is no up to date data evidence supporting factors determine the wage rate of women in the Australian labour market. This research aims to fill this gap using first seventeen waves (2001 – 2017) of Household, Income and Labour Dynamics in Australia (HILDA) survey.

Preceding literature analysed some of these labour market dimension, but this is the first analysis that incorporates all these three dimensions using first seventeen waves in my knowledge. To accomplish this, I will conduct a quantitative study based on a random effect model[1] and estimate the result using STATA software. This model will estimate the hourly wage rate[2] of women in the Australian labour market.

Literature review on the factors that determine the wage rate of women in the labour market: Hypotheses Statement

Education attainment and wage rate of women in the labour market

The relationship between the education level and wage rate of women in the labour market has been studied previously with contrasting results since evidence exists for both positive and negative.  An increase in women education is a favourable determinant of women in the labour market (Hajnalka, 2012) and level of education directly affects the relative price of working at home relative to the labour market (Raquel, 2007). However, increase in the women education lower the female fertility resulted in a negative impact on the formation of human capital in the next generation (Cabeza-Garcia, Del Brio, and Oscanoa-Victorio, 2018) and therefore lead to lower participation of women in the labour market.

In view of these arguments, I present the following hypothesis:

Increases in the educational level of women increases the wage rate of women in the labour market.

Occupation category and wage rate of women in the labour market

Women have a stronger preference for work environments that provide social services (Kossek et al., 2017 as cited in Diekman et al., 2010; McCarty, Monteith & Kaiser, 2014) such as education and healthcare. Performance of ‘female-typical’ work tasks in caring and nurturing is the main reason why ‘female-typical’ occupations are valued less and therefore paid less (Leuze, K., & Straub, S. 2016 as cited in Busch, 2013; Liebeskind, 2004). Women with the same level of performance as men are less likely to receive similar pay and promotions (Kossek, Su, and Wu, 2017 as cited in Joshi et al.,2015). In an Australia, women’s average weekly ordinary full-time earnings were $1,484.80 compared to men’s earnings of $1,726.30 across all industries and occupations in May 2019. Across the industry, the gender pay gap is highest in financial and insurance services with 24.4%, followed by professional, scientific and technical services with 24.3%, health care and social assistance with 23.9%, lowest in retail trade with 3.9% and public administration and safety with 4.9%  (WGEA, 2019).

Based on these arguments, the following hypothesis is posed:

Higher occupation category increases the wage rate of women in the labour market.

Total hours of unpaid care work and wage rate of women in the labour market

From a gender standpoint, women spend more time on unpaid care work[3] and men spend more time on paid work. The Australian Human Rights Commission (AHRC) in 2013 estimates 72.5 per cent of women from 5.5 million Australians between 15 and 64 years had unpaid caring responsibilities for the young, the sick and the elderly in their families, friendship groups and communities (McAllister, 2017). Previous evidence revealed the existence of a “gender stereotypes” and women less committed to their paid jobs due to their family engagements (Hajnalka, 2012 as cited in Nagy, B.2001, 36). There is a negative relationship between first childbirth and women paid work (Argyrous, Craig and Rahman, 2017). Furthermore, women involvement in employment reliant on young children and 92 per cent of those re-entering the workforce a year after giving birth work in a part-time capacity (Fleming and Kler, 2014). This relevant studies based on the Australian context. The conclusions were clear cut, “it is evident that the women spent more hours of unpaid care work and lower participation in the paid labour market”, so this is reflected in the hypothesis.

Increases in the total hours of unpaid care work decreases the wage rate of women in the labour market.

Research Design – Methods and Procedures

This research using a deductive process to test the hypotheses. A large number of observations will make the result is more trustworthiness. Therefore, in this research, I choose a quantitative method using secondary data extracted from the HILDA survey. The HILDA survey is a household-based longitudinal survey covering a broad range of social, demographic and economic questions.

Data set and sample collection

Panel data will be constructed over seventeen years from the period of 2001 to 2017 in which the behaviour of an individual is observed over time. This enables the repeated observations of the dynamics education of individuals, their labour market experiences and income they receive over the same individual.

The survey sample has been supplemented at various times due to attrition of individuals. In wave 17, interviews obtained from with a total of 17,571 of which 13,972 were from the original sample and 3,779 were from the top-up sample.  The survey is topped up to make up for those who no longer wish to be interviewed and include more migrants. Hence,  the sample and survey result is more representative of the changing nature of the Australian population. However, Indigenous and Non-Indigenous Australians living in remote communities are excluded from the survey sample due to a logical issue.

I limit the sample in this research to casual, part-time and full-time employment of women. Women sub-group divided into four groups; total aggregated group, young women[4] with no children, young women with children and educated[5] group.

Variables and methodology

Table 1 shows the variables that will be used in this analysis. This research will estimate the hourly wage rate of women in four different sub-groups using a REM. Log(wage) is the dependent variable and education attainment, occupational category and its square, total hours of unpaid care work and its square are explanatory variables.

All three hypotheses require the use of control variables. The HILDA dataset contains control variables that would support the testing of each hypothesis. Demographic and socio-economic characteristics of individuals such as age, gender, place of birth, English proficiency, marital status and number of dependency are controlled in this model. Place of birth and English proficiency is measured for persons who are born outside Australia.

Table 1: Definition of variables

Note: Wage and occupation are evaluated only for employed persons.       (Adapted from HILDA, 2019)

Random effect model panel data

A panel data allows changes over time to be analysed at the unit rather than the aggregate level. This greatly enhances the data available for individual-level modelling; allows phenomena to be analysed such as the duration and enables researchers to make inferences about causality based on the ordering of events in time (Zlatko et al., 2010).

Panel data provide information on individual behaviour, both across individuals and over time. It contains N individuals observed at T regular time periods. Panel data can be balanced when all individuals are observed in all time periods (T_i = T for all _i). The wages for the same individual is correlated over time but it is independent across individuals.

Women are randomly selected from a larger population and individual differences are treated as random effects. Presence of random effect will be tested using the Lagrange Multiplier (LM) test. If the null hypothesis is true, then u_i= 0. If the null hypothesis is not rejected, then no evidence to conclude that random effect is present. If the value of the test statistic exceeds a  critical value, it will conclude there is strong evidence on individual heterogeneity.

The REM has the following specification:

y_it=

βi̅

+ β₂x_2it + β₃x_3it + e_it

y_it =

βi̅

+ β₂x_2it + β₃x_3it + v_it

where:

βi̅

is a fixed population parameter and the error term (v_it) is composed of a component u_i that represents a random individual effect and the component e_it which is the usual regression random error. The combined error is v_it = u_i + e_it.

The transformed  REM is:

y_it^* =

βi̅

x₁^*_it+ β₂x₂^*_it + β₃x₃^*_it + v^*_it

where: y_it^* represents the endogenous log(wage) of women.

Dummy variables are included in the wave analysis. The wave analysis will establish whether the previous wave performs differently compare to the recent wave and the extent to which the wage will change over time.

Panel data models are more complex and sophisticated (Lester and Fitzpatrick, 2008). Panel data allow researchers to exploit both inter-individual variation and intra-individual dynamics to more fully understand complex behaviour. Panel data has several statistical and econometric advantages, such as controlling for omitted variables, allowing for the construction and testing of more complicated behavioural hypotheses and more precise estimation and consequently more accurate inference (Deborah, 2010 as cited in Hsiao, 2005).

A major advantage of REM in the panel data is that they provide a way to control for all time-invariant unmeasured variables that influence the dependent variable whether these variables are known or unknown to the researcher. The REM assumes that the omitted time-invariant variables are uncorrelated with the included time-varying covariates and has the advantage of greater efficiency relative to the fixed effect model which leading to smaller standard errors of coefficients and higher statistical power to detect effects (Kenneth and Jennie, 2010 as cited in Hsiao, 2003).

Wage is one of the most frequently cited examples of the importance of panel data for labour market analysis (Deborah, 2010). It is important to take into account unobserved individual characteristics (unobserved heterogeneity) in any labour market analysis. Unobserved heterogeneity is a form of omitted variable bias which leaves researchers unable to correctly interpret labour market phenomenon.

However, there is a potential problem when using RE estimation. If the random error v_it = u_i + e_it is correlated with any of the right-hand side explanatory variables in a RE model, then the least squares and generalised least squares estimators of the parameters are biased and inconsistent (Hill, Griffiths and Lim, 2011). Individual specific error, u_i may correlate with some of the explanatory variables. An individual’s ability and spouse’s income are variables not explicitly included in the wage equation. Therefore, these factors included in u_i. These characteristics may be correlated with women’s years of educational attainment and total hours of unpaid care work. In this case, RE estimators are inconsistent and will attribute the effects of the error component to the included explanatory factors.

Significance of the study

The rationale for this study is that earning from work for women is very important for the comfort of family and wellbeing of an economy.

References

• Argyrous, G., Craig, L., & Rahman, S. (2017). The Effect of a First-Born Child on Work and Childcare Time Allocation: Pre-post Analysis of Australian Couples, Social Indicators Research, 131(2), 831-851.
• Australia’s Gender Pay Gap Statistics (2019). Retrieved from https://www.wgea.gov.au/data/fact-sheets/australias-gender-pay-gap-statistics
• Barriers and Incentives to Labour Force Participation, Australia, July 2016 to June 2017. Retrieved from https://www.abs.gov.au/ausstats/abs@.nsf/Latestproducts/6239.0Media%20Release1July%202016%20to%20June%202017?opendocument&tabname=Summary&prodno=6239.0&issue=July%202016%20to%20June%202017&num=&view=
• Bubonya, M., Cobb-Clark, D., & Ribar David, C. (2019). The reciprocal relationship between depressive symptoms and employment status, Economics and Human Biology, 35, 96-106.
• Brady, M., & Perales, F. (2016). Hours of Paid Work Among Single and Partnered Mothers in Australia: The Role of Child Care Packages. Journal of Family Issues, 37(3), 321-343. DOI: 10.1177/0192513X14531416.
• Cabeza-García, Laura., Del Brio, Esther B., & Oscanoa-Victorio, Mery Luz. (2018). Gender factors and inclusive economic growth. Sustainability (Switzerland), 10(1), 121. DOI:10.3390/su10010121.
• Clark, A.E. (1997). Job satisfaction and gender: Why are women so happy at work?, Labour Economics, 4(4), 341-372.
• Damaske, S., & Frech, A. (2016). Women’s work pathways across the life course. Demography, 53(2), 365-391.      doi:http://dx.doi.org.libraryproxy.griffith.edu.au/10.1007/s13524-016-0464-z.
• Deborah Cobb-Clark. (2010). Using longitudinal analysis to better understand labour market transitions and evaluate policy settings. Retrieved from:  https://www.dss.gov.au/sites/default/files/files/settle/taking/following-migrants-forward_access.pdf.
• Ekkehart, S. (2010). Selection Wages and Discrimination, Economics, 4(6), 1-30A.
• Fleming, C., Kifle, T. & Kler, Parvinder. (2016). Immigrant occupational mobility in Australia, Work, Employment & Society, 30(5), 876–889.
• Fleming, C. & Kler, P. (2014). Female overeducation, job satisfaction and the impact of children at home in Australia, Economic Analysis and Policy, 44(2), 143-155.
• Geoff, Gilfillan. (2019). Findings of the 14th Annual Statistical Report of the HILDA Survey. Research Paper Series, 2019–20. Retrieved from: https://parlinfo.aph.gov.au/parlInfo/download/library/prspub/6824567/upload_binary/6824567.pdf.
• Following Migrants Forward: Exploring the benefits and challenges of a new longitudinal survey of immigrants. (2010). Retrieved from https://www.dss.gov.au/sites/default/files/files/settle/taking/following-migrants-forward_access.pdf.
• Francisco, P., Jose, P., Jesus., P., & Borja, P. (2014). On the associations between physical activity and quality of life: findings from an Australian nationally representative panel survey. Quality of Life Research, 23(7), 1921-1933. DOI 10.1007/sl 1136-014-0645-4.
• Green, C., Kler, P. & Leeves, G. (2007). Immigrant overeducation: Evidence from recent arrivals to Australia, Economics of education review, 26 (4), 420–432.
• Gorard, Stephen. (2012). Research design: creating robust approaches for the social sciences. SAGE Publications.
• Hajnalka, Fenyes. (2012). Social Mobility of Graduate Men and Women. Women’s Advantages in Higher Education and Disadvantages on the Labour Market, Social Analysis, 2(1), 31-53.
• HILDA (Living in Australia): New Person Questionnaire (NPQ) W17M. Retrieved August 28, 2019, from https://melbourneinstitute.unimelb.edu.au/__data/assets/pdf_file/0010/2409679/NewPersonQuestionnaireW17M.pdf.
• Hill, C., Griffiths, B., & Lim, G. (2011). Principle of Econometrics, 4^th edition., Pearson Australian Group Pty Ltd, NSW.
• Jennifer, Baxter (2013). Australian mothers’ participation in employment: Analyses of social, demographic and family characteristics using the Household, Income and Labour Dynamics in Australia (HILDA) survey, Australian Institute of Family Studies,   Research Paper, 52. Retrieved from https://aifs.gov.au/publications/australian-mothers-participation-employment/executive-summary
• Jennifer, Baxter (2013). Employment characteristics and transitions of mothers in the Longitudinal Study of Australian Children, Australian Institute of Family Studies, Research Paper, 50. Retrieved from https://www.dss.gov.au/sites/default/files/documents/12_2013/op50_final_tagged_pdf.pdf
• Johnston, C., McDonald, J., & Quist, K. (2019). A generalized ordered Probit model.  Communications in Statistics: Theory and Methods. 1-18. DOI: 10.1080/03610926.2019.1565780
• Johnstone, M., Lucke, J. & Lee, C. (2011). Influences of marriage, motherhood, and other life events on Australian women’s employment aspirations, Psychology of Women Quarterly, 35(2), 267-281.
• Kenneth, A., & Jennie, E. (2010). A general panel model with random and fixed effects: A structural equations approach. Social forces, 89(1), 1-34. DOI: https://doi.org/10.1353/sof.2010.0072
• Kifle, T., Kler, P. & Shankar, S. (2014). Are women really that happy at work? Australian evidence on the ‘contented female’, Applied Economics, 46(7), 686-697.
• Kingsley, M. (2018). The influence of income and work hours on first birth for Australian women. Journal of Population Research, 35(2), 107-129. doi:http://dx.doi.org.libraryproxy.griffith.edu.au/10.1007/s12546-018-9200-4.
• Kler, P., Potia, A. & Shankar, S. (2018). Underemployment in Australia: a panel investigation, Applied Economics Letters, 25(1), 24-28.
• Kossek, E., Su, R. & Wu, L. (2017). “Opting Out” or “Pushed Out”? Integrating Perspectives on Women’s Career Equality for Gender Inclusion and Interventions, Journal of Management, 43(1), 228–254.
• Lester, L., & Fitzpatrick, D. (2008). Econometric evaluation of women’s labour supply in WA: A report prepared for the Western Australia Department of Consumer and Employment Protection. Retrieved from https://www.voced.edu.au/content/ngv%3A21803.
• Leuze, K., & Straub, S. (2016). Why do occupations by women pay less?: How ‘female-typical’ work tasks and working time arrangements affect the gender wage gap among higher education graduates., Work, employment and society, 30(5), 802–820. DOI: 10.1177/0950017015624402.
• Lewis, P. & Simpson, Ruth. (2017). Hakim Revisited: Preference, Choice and the Postfeminist Gender Regime, Gender, Work & Organization, 24(2), 115 – 133.
• Lin, Jing., Oxford, Rebecca L., & Culham, T. (2016). Toward a spiritual research paradigm: exploring new ways of knowing, researching and being.
• Majchrowska, A. & Strawinski, P. (2018). Impact of the minimum wage increase on the gender wage gap: Case of Poland, Economic Modelling, 70, 174-185.
• McAllister, Jenny. (2017). Gender segregation in the workplace and its impact on women’s economic equality.
• OECD (2017), Connecting People with Jobs: Key Issues for Raising Labour Market Participation in Australia, Connecting People with Jobs, OECD Publishing, Paris, https://doi.org/10.1787/9789264269637-en
• Perugini, C. & Vladisavljevic, M. (2019). Gender inequality and the gender-job satisfaction paradox in Europe, Labour Economics, 60, 129-147.
• Pocock, B. & Charlesworth, S. (2017). Multilevel Work-Family Interventions: Creating Good-Quality Employment Over the Life Course, Work and Occupations, 44(1), 23–46
• Raquel, F. (2007). Women, Work, and Culture, Journal of the European Economic Association, 5(2/3), 305-332.
• Roberson, Q. (2006). Disentangling the Meanings of Diversity and Inclusion in Organizations, Group & Organization Management, 31(2), 212-236.
• Stephanie, J. (2019, July 30). HILDA examines how Australians are balancing work and home. Retrieved from: https://about.unimelb.edu.au/newsroom/news/2019/july/hilda-examines-how-australians-are-balancing-work-and-home.
• T-F Lo, P-H, K., & W-J Tsay. (2018). Pairwise likelihood inference for the random effect’s probit model. Computational Statistics, 33(2), 837-861. doi:http://dx.doi.org.libraryproxy.griffith.edu.au/10.1007/s00180-017-0757-8.
• The Household, Income and Labour Dynamics in Australia Survey: Selected Findings from Waves 1 to 17 (2019), Melbourne Institute: Applied Economic & Social Research. Retrieved from https://melbourneinstitute.unimelb.edu.au/__data/assets/pdf_file/0011/3127664/HILDA-Statistical-Report-2019.pdf
• Towards 2025: An Australian Government Strategy to Boost Women’s Workforce Participation (2017). Retrieved from https://ministers.pmc.gov.au/cash/2017/towards-2025-boosting-australian-womens-workforce-participation.
• United Nations Development Fund for Women. (2003). Progress of the World’s Women 2000: UNIFEM Biennial Report.
• Veal, A. (2011). Research Methods for Leisure and Tourism: A Practical Guide, 4^th edition., Pearson Australian Group Pty Ltd, NSW.
• Verbeek, M. (2012). A Guide to Modern Econometrics, 4^th edition., John Wiley & Sons, LTD.
• Williams, R. (2009). Using Heterogeneous Choice Models to Compare Logit and Probit Coefficients Across Groups. Sociological Methods & Research, 37(4), 531–559. https://doi.org/10.1177/0049124109335735.
• Young, M., & Schieman, S. (2018). Scaling Back and Finding Flexibility: Gender Differences in Parents’ Strategies to Manage Work-Family Conflict, Journal of Marriage and Family, 80(1), 99-118.
• Zlatko, Spralja., Sebastien, Lucie., Alanna Sutcliffe, Ruel Abello., & Jill, C. (2010). The merits of a longitudinal survey: a methodologist’s view. Retrieved from: https://www.dss.gov.au/sites/default/files/files/settle/taking/following-migrants-forward_access.pdf

[1] The random effect model will be referred to as ‘REM’ for the remainder of this research proposal.

[2] The hourly wage rate is ‘current earnings per hour worked’. The HILDA survey does not ask respondents to report their hourly wage; rather, weekly earnings and weekly hours of work which obtained from everyone who is employed. Hourly wage rate calculated from this information.

[3] United Nations Development Fund for women (UNIFEM) refer unpaid care work as all unpaid services provided within a household for its members , including care of persons, housework and voluntary community work.

[4] Young women refer to women below 35 years.

[5] Educated group refer to women holding at least a bachelor’s degree.

[6] ANZSCO stands for the Australian and New Zealand Standard lassification of Occupations.  It is based on a conception of types of tasks and skill-level requirements. It has eight occupation groups.