Have Pay Equity Laws in Canada Helped Women? A Synthetic-Control Approach

ABSTRACT

Ontario passed a very aggressive Pay Equity Act in 1988, and in 1996 Quebec passed a similar Pay Equity Act. We use synthetic-control methods to examine what has happened to the gender pay gap (female–male earnings ratio) in Ontario since 2005 and to see whether Quebec’s Pay Equity Act has had any effect on its pay gap. Ontario and Quebec are chosen simply because they are the provinces with the most comprehensive and “aggressive” pay equity laws. We also use synthetic-control methods to investigate whether these acts may have had an adverse effect on the female–male employment ratios in Ontario and Quebec. We find that Ontario’s act has had a negligible effect on that province’s gender pay gap and employment ratio. However, Quebec’s act appears to have reduced the gender pay gap, although at a cost of a somewhat smaller employment ratio for women in that province.

KEYWORDS:

• Ontario’s Pay Equity Act
• Quebec’s Pay Equity Act
• gender pay gaps
• gender employment ratios
• labor-market discrimination

Acknowledgments

The authors are grateful to anonymous referees and participants at the 23rd ACSUS Biennial Conference for helpful suggestions.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. See, for example, Weiner and Gunderson (1990). The quotation is from Baker and Fortin (2001, 346). “Proactive” laws were developed in recognition of the fact that pay discrimination is systemic or system wide, rather than occurring only in specific jobs or particular workplaces.

2. There was a jump up to 74.4% in 2009, which Drolet (2011) attributes to several factors, including an increase of about 2 years in the average job tenure for women over the period 1978–2008, and an increase from 15.7% in 1990 to 29.3% in 2008 in the proportion of women with a university degree. Note that as of October 2015, the most recent female–male earnings ratio on the Statistics Canada website was for 2011; this series has since been terminated.

3. Federal equal-pay legislation, embedded in the Canadian Human Rights Act, means that all Canadians are covered by “complaint-based” pay equity laws, with female employees bearing the burden of proving that they were underpaid. Similar human rights provisions also cover the Yukon, the North West Territories, and Nunavut (Canadian Labour Congress 2008, 14–15).

4. See Ketterer (2015) for a strong criticism of California’s Fair Pay Act.

5. Employers are required to evaluate, without bias, jobs held mainly by women and men and to pay equal wages and benefits for jobs deemed comparable.

6. Maintaining pay equity means that once a pay equity comparison was complete, incumbents in every female job class must receive the same increases given to those in the male comparator job class. It also means that these comparisons must be made for new female job classes using the same comparison system that had been used for other job classes in the establishment (Pay Equity Commission 2013b, 8).

7. In March 2001, British Columbia’s New Democratic Party (NDP) government enacted complaints-based pay equity legislation, but during August of that same year the newly elected B.C. Liberal government repealed this legislation.

8. Rather they have developed frameworks for negotiating pay equity with some specific public-sector employees.

9. Synth is a package that is available in the computer language R (2014), available online at http://www.r-project.org

10. That is, all employers were required to: develop a pay equity plan or plans; determine male and female job classes; implement a “gender-neutral” job evaluation system; and compare the compensation of female and male job classes of similar value to determine the adjustments needed to attain pay equity. (For more details, see Pay Equity Commission 2013b).

11. Alberta and British Columbia (provinces without pay equity laws) both have larger pay gaps, 36.8% and 30.9% respectively, followed by Nova Scotia, with a gap of 30.2%, and Manitoba, with a gap of 28.3%. Throughout this article, we define the pay gap as 1 − (wages of females/wages of males), which is the wages of males minus the wages of females divided by the wages of males.

12. In ten of the twenty-seven interviews we conducted, examples of noncompliance or manipulation of some of the requirements of the act were clear. We discovered that (p. 195) many employers “manipulated and interpret[ed] the law so as to minimize their pay-equity payouts.” (McDonald and Thornton 1998)

13. In practice, finding the correct control (comparison) is not easy to do. Ideally such a control would be identical in every way to Ontario except that it did not pass any pay equity legislation; that is, the control would proxy for what would have happened in the “counterfactual Ontario,” the one without pay equity.

14. Admittedly there is no perfect comparison for Ontario; and Synth, just as with regression, cannot control for effects that cannot be measured.

15. Real GDP per capita is real GDP in 2002 constant dollars (from CANSIM Table 384–0002: Expenditure-based chained dollar real GDP, annual) divided by population (from CANSIM Table 051–0001: Both sexes, All ages).

16. The unemployment rate is obtained from CANSIM Table 282–0086: Labor force survey estimates, supplementary unemployment rates, official rates, both sexes, age group 25 years old and older.

17. The average female–male earnings ratio is obtained from CANSIM Table 202–0102: “Average female and male earnings, and female-to-male earnings ratio, by work activity, 2011 constant dollars, annual.” Data on full-year, full-time workers were used. As mentioned in note 2, this series has been discontinued.

18. Special predictors are chosen by the user and can be any variable used in the predictor list over any time period (i.e., that variable’s value from a single year or average over several years) as long as these years are during the pretreatment period. (http://www.insider.org/packages/cran/synth/docs/dataprep)

19. It should be noted that Synth is sensitive to which predictors and special predictors are used; that is, using different variables for predictors and special predictors than we do here will generate a different Synthetic Ontario. For example, when control provinces were restricted to those that did not pass pay equity laws−Newfoundland, Saskatchewan, Alberta, and British Columbia−the weights assigned to British Columbia and Alberta both increased (from 0.638 in unrestricted case to 0.693 for British Columbia; and from 0.194 to 0.282 for Alberta). Neither Newfoundland nor Saskatchewan was a control in the unrestricted case; however, here Newfoundland had a weight of 0.0011 and Saskatchewan’s was 0.024.

20. This illustrates a key idea behind the synthetic-control approach of Abadie, Diamond, and Hainmueller (2010)—that a combination of units often provides a better comparison for the unit exposed to the intervention than any single unit alone. Abadie et al. (2010, 3) note that both comparative case studies and studies using aggregate data can be useful in evaluating policy interventions at an aggregate (or macro) level. However, aggregate data-driven procedures like Synth reduce the amount of discretion used in the choice of the comparison control units and allow researchers to decide on study design without knowing how this will affect their conclusions. For more details, see Abadie and Gardeazabal (2003) and Abadie, Diamond, and Hainmueller (2010).

21. Figure 1 uses mean earnings data. We also calculated the Ontario versus synthetic-control earnings ratios using median earnings data, but there was very little difference between the two ratios. Consequently, we have used only mean data in subsequent figures.

22. The female–male employment ratios are calculated using CANSIM, Statistics Canada Table 202–0103: Number of full-year, full-time workers, by sex and work activity (total employment data were used).

23. See Kaufer and Brady (2014) and Cornish and Quito (2013) for more details about Quebec’s Pay Equity Act.

24. Baker and Fortin (2001) use Quebec as a control for examining whether Ontario’s Pay Equity Act had any effect on the wages of women in Ontario.

25. Weights using 1981–1995 as pretreatment years: Newfoundland (7.1%); Prince Edward Island (7.1%); Nova Scotia (14%); New Brunswick (13.2%); Ontario (7.4%); Manitoba (12.8%); Saskatchewan (3.6%); and British Columbia (34.9%). Weights using 1981–2000 as pretreatment years: Newfoundland (9.3%); Prince Edward Island (9.2%); Nova Scotia (11.3%); New Brunswick (13.25); Ontario (7.6%); Manitoba (8.5%); Saskatchewan (6.2%); Alberta (3.4%); and British Columbia (31.3%).

26. It is possible that changes in the birthrate in Quebec might also have affected the employment ratio in Quebec during part of this period. During the early 2000s, the fertility rate of women in Quebec rose as a result of the province’s generous parental leave program and inexpensive day care. There was also a temporary increase in the fertility rate in the 1980s as a result of the provincial government’s cash (“baby”) bonus for new births. Such increases in the birthrate would be expected to have at least a temporary negative effect on the employment ratio. We thank one of the anonymous referees for this suggestion.

27. Kervin (2007) also noted that men and women were not going into the same jobs in the same proportions in Ontario, nor were they doing so elsewhere, since many other countries have similar issues regarding “occupational segregation.” This segregation has been noted in many studies, including Fortin and Huberman (2002), who discuss the effect of occupational segregation on women’s wages in Canada, and the OECD (2008, 45), which points out that, although female students now outnumber male students at universities and colleges in many OECD countries (including Canada), men and women still choose different majors, with women largely preferring welfare and health subjects and men gravitating toward engineering and commerce. Consequently, women are more likely to end up in female-dominated fields, for example, nursing and teaching, typically characterized by lower-status and lower-paying jobs.

28. The eight additional steps according to Cornish (2014) are: treating closing the gap as a human-rights priority; increasing awareness of the gap through annual equal paydays and education; developing closing gender pay gap plans; enforcing and expanding pay equity laws; promoting access to collective bargaining; increasing the minimum wage; providing affordable and accessible childcare; and mainstreaming equity compliance not only into government laws and policies but also into work and business.

29. A recent Wall Street Journal Special Report “Women in the Workplace” (2015) also discusses the need for more flexibility in the workplace, including more paid leave and better childcare.

Additional information

Funding

The authors wish to thank the Government of Canada for financial assistance.

Notes on contributors

Judith Ann McDonald

Judith Ann McDonald is a professor of Economics at Lehigh University. Her primary areas of research are gender pay inequities and international trade and finance. She teaches International Finance, International Trade, Intermediate Macroeconomics, and Money and Banking at the undergraduate level. Her research has examined the effects of pay- equity laws on women’s pay and employment in Ontario and Quebec; she has also looked at starting salaries for new college graduates in the United States. A native of Canada, Professor McDonald received an undergraduate degree from the University of Western Ontario in 1979 and a doctorate from Princeton University in 1986.

Robert Thornton

Robert Thornton is MacFarlane Professor of Economics at Lehigh University. He received his PhD in economics from the University of Illinois and his BA degree in classical languages from Xavier University. He teaches courses in labor economics and statistics, and his research includes many articles on labor-market discrimination, unions and collective bargaining, occupational licensing, and forensic economics.