Chapter 2: Data justice

Concepts and practices

We live intersectional lives.

Powerful statements are made possible by disaggregated data.

By making systemic inequalities in our society visible, data can lead to positive change. The same data, used or collected poorly, can reinforce stigmatization of communities, leading to individual and community harm.

The task force heard strong and consistent support for the employment equity groups of workers to be referred to in a distinctions-based and intersectional manner. From the Public Service Commission of Canada to unions, employee networks, private sector employers, and civil society, alongside international experts and representatives from the International Labour Office in Geneva, we heard that it was particularly important to be able to see how sub-groups are affected. Aggregate data collected to reflect the four current employment equity groups prevented us from seeing what was happening to sub-groups in the workplaces. By failing to collect disaggregated data, employment equity has masked and at times perpetuated systemic inequality in the workplace.

A distinctions-based approach focuses on addressing equity groups with due attention to their specific histories and identities. This is particularly significant for Indigenous peoples. They are nations, who have treaty rights and Indigenous rights recognized in the Canadian constitution and at international law. As is discussed in Chapter 3, their inclusion in employment equity legislation needs to be rethought to reflect the specificity of their experience and honour and respect nation-to-nation/government-to-government relationships.^{Footnote 9}

An intersectional and distinctions-based approach shares much with the Canadian government’s championing of a Gender-Based Analysis Plus (GBA+) framework. GBA+ was broadly recognized as providing space for nuanced analyses, recognizing that people at once have multiple identities that interact; they cannot either be reduced to those identities. We heard from stakeholders, though, that they worried that the framework as characterized might inadvertently give the impression that it sought to subsume all other equity groups into the “plus” – meant to include identities such as race, ethnicity, religion, age and disability. This approach was considered to be a problem, potentially creating a hierarchy between gender and all the rest. As one person put it, groups do not want to be reduced to a “plus”. The Government of Canada should seriously consider renaming its framework so that its substantive equality framework for inclusion is evident in a name that all equity groups want to embrace.

Data disaggregation is widely accepted, including by the Public Service Commission of Canada itself. In a 2021 employment equity audit of representation in recruitment it found different experiences amongst external applicants to advertised positions and stressed “the importance of examining visible minority sub-group data.”^{Footnote 10}

Data disaggregation supports the distinctions-based approach. Instead of limiting employment equity data only to the composite group, for example “Indigenous peoples”, it turns attention to First Nations, Métis and Inuit workers. It is apparent that with time and with due regard to the considerations to be raised below, employment equity programs will need to be especially attentive to understanding the representation of specific nations and the precise barriers that may apply. Data disaggregation is recognized as one tool among others – including qualitative data - to be used in the context of the Employment Equity Act framework to achieve equality in the workplace.

[T]he [Employment] Equity Act will not be able to address intersectional barriers faced by women with disabilities if it does not require employers to collect disaggregated data on their workforce in order to determine the degree of underrepresentation of persons from multiple designated groups, including the experiences of those living at the intersection of one or more group.

A related concept is intersectionality.

The notion of intersectionality emerges in law from the work of Professor Kimberlé Crenshaw and in sociology from the work of Professor Patricia Hill Collins. Its profound roots should not be forgotten: they are embodied the life of the formerly enslaved African American woman, Harriet Tubman, who after freeing herself from slavery, kept going back at great risk, to bring others to freedom. She was not alone.^{Footnote 11} Intersectionality offers a way to understand and challenge systemic inequality. Keeping the focus on this point is crucial. Intersectionality is not an abstract adding of identities. Nor is it subdivision separated from social context into impossibly smaller and smaller subgroups.

Rather, the traffic intersection analogy used by Crenshaw in her leading 1989 law review article focused on employment discrimination law helps to explain the focus on grounds of discrimination as lived by actual people, in context:

Consider an analogy for traffic in an intersection, coming and going in all four directions. Discrimination, like traffic through an intersection, may flow in one direction, and it may flow in another. If an accident happened in an intersection, it can be caused by cars travelling from any number of directions and, sometimes, from all of them. Similarly, if a Black woman is harmed because she is in the intersection, her injury could result from sex discrimination or race discrimination.^{Footnote 12}

Intersectionality must focus on the actual people who seek to move from societal margins to the centre of policy change. In other words, intersectionality must remain centred on its critical “why”. Its “why” is to gain deep insight into how systemic discrimination is experienced by people who have been historically marginalized and continue to face exclusion. Its “why” is to ensure that context matters, in all its complexity. Its “why” is to bring lived experiences to the centre of policy and action for substantive equality.

Just as data disaggregation is increasingly used in data reporting, intersectionality is increasingly recognized in legislation: Section 3.1 of the Canadian Human Rights Act acknowledges that “a discriminatory practice includes one or more prohibited grounds of discrimination or on the effect of a combination of prohibited grounds”.

Some further clarifications are useful. Data disaggregation does not necessarily require intersectionality. For example, disaggregation of the racialized workers could mean reporting separately on the data collected specifically for Japanese workers and Filipino workers.

[The] Filipino Public Servants Network recently put together some stats showing the employment gaps for Filipinos in the public service at all levels. These types of realities would be difficult to notice if data about Filipinos is lumped in together with data about other Asian ethnicities under an "Asian" category.

Intersectionality does however require disaggregation, both within and across equity groups. For example, intersectionally disaggregated data allow us to report on all racialized men and all racialized women, or Inuit women with a disability and Inuit men with a disability, therefore within and across the equity groups constituted as visible minority, women and disabled workers.

Intersectionality also allows us to identify and remove barriers, based on qualitative systems reviews, for equity groups across multiple grounds of discrimination identified in the Canadian Human Rights Act. The task force heard from youth representatives, who called for an intersectional approach to young people, who are present in each of the employment equity groups.^{Footnote 13} The theme of intersectional barrier removal beyond the employment equity groups will be explored further in Chapters 3 and 4.

Finally, an intersectional and disaggregated approach makes space to address so-called “mixed-race” identities. Accurate Census data make accurate analysis possible. The key, throughout, will remain the purpose. For the Employment Equity Act framework, that purpose remains to achieve substantive equality, and equality is achieved by removing barriers to equitable employment across equity groups and for all.

A strong statistical base supports a deepened understanding of systemic discrimination – and in particular adverse impact discrimination. While the law on systemic discrimination has been clarified to confirm that statistics are not always required in court cases to establish adverse impact discrimination, for policy makers trying to make key decisions on representation in their workplaces, a thick, nuanced statistical base provides tools necessary to redress systemic discrimination.^{Footnote 14} This requires a true change in perspective, for “[m]ore than just extending the conventional use of legal sanctions to discriminatory acts, indirect discrimination requires the examination of all of the apparently neutral procedures and practices, and, above all, the active promotion of equality.”^{Footnote 15}

Collecting disaggregated data: The grandmother perspective

Rather than monitoring the lives of our citizens, collecting and using disaggregated data is about caring for our communities by informing law, policy and institutional practice that is in service of – and developed in collaboration with – those who are systemically discriminated against.

The grandmother principle applied by the BC Office of the Human Rights Commissioner, and referenced earlier in the chapter, underscores the need to align purpose, process and tool, in the context of the Anti-Racism Data Act that was subsequently adopted. In building this principle, the BC Human Rights Commissioner acknowledged notably that demographic data collected on Indigenous households by what was then known as the Department of Indian and Northern Affairs Canada supported the establishment and operation of so-called residential “schools”.^{Footnote 16} This history of trauma explains why it is so important to be extremely attentive to how disaggregated data are collected and used. The BC Office of the Human Rights Commissioner adds that “[i]n addition to the significant danger of disaggregated data leading to further stigma and discrimination, over-researching social inequalities with little follow-up action has also harmed communities …. In these cases, research is not grounded in respectful relationship.”^{Footnote 17}

Adapted to the employment equity context, the Grandmother principle yields the following:

• The purpose is achieving substantive equality in the workplace through employment equity. The purpose of disaggregated data is not individual or group stigmatization, or reinforcing deficit narratives. Rather, the focus is on redressing systemic workplace exclusion.
• The process must foster respectful relationship toward members of equity groups in the workplace context. Employers, unions and other workplace actors are integral to the process of achieving employment equity.
• An important tool to achieve this purpose is data collection, and in particular disaggregated data collection.

Emerging out of the COVID-19 pandemic, jurisdictions including British Columbia started collecting race-based and other disaggregated data to understand and address systemic discrimination. The human rights commissioner alongside the privacy commissioner urged the “grandmother perspective” be adopted in the collection of disaggregated data. Drawing on that important work, British Columbia’s 2022 Anti-Racism Data Act, co-developed with Indigenous peoples and racialized communities, is a significant example of leadership on dismantling systemic discrimination. With a focus on building trust with Indigenous Peoples and racialized communities, it retains the principle of voluntary self-reporting, safe collection, use and disclosure of demographic data. It builds in a committee-based structure for ongoing, robust and culturally safe consultations to establish a plan, and ensure implementation.

Finally, collecting disaggregated data is consistent with Canada’s international human rights treaty obligations. Most recently, the UN Committee on the Rights of Persons with Disabilities observed that “it is essential to monitor the barriers to employment for persons with disabilities on an equal basis with others”, adding that by appropriately disaggregating the data, it will be possible to identify who is working in the informal economy and what barriers may relate to self-employment and entrepreneurship.^{Footnote 18} The ILO’s Committee of Experts on the Applications of Conventions and Recommendations has also encouraged governments, in cooperation with workers’ and employers’ organizations and other interested bodies, to gather disaggregated data in a manner that is respectful of privacy through informed consent and voluntary self-identification, to take into account multiple forms of discrimination.^{Footnote 19}

Building trust

Critically, our task force was repeatedly told by a number of stakeholders from different constituencies that trust is not high.

Whether to disclose membership in a group that has faced historical marginalization may remain one of the most critical challenges that workers face. They might fear stigmatization or social isolation; they might risk harassment, career derailment or even job loss.^{Footnote 28} We heard workers have real fears about how information collected by individual employers can be used. This profoundly affects self-identification.

In its submission to the task force, the Office of the Privacy Commissioner emphasized that both employment equity legislation and privacy legislation trace their origins to the Canadian Charter of Rights and Freedoms, and to federal legislation. And under the Charter, as well as the Canadian Human Rights Act, everyone is entitled to exercise fundamental rights under conditions of equality and non-discrimination.

The Office of the Privacy Commissioner added that it recognizes that effective compliance with the requirements of the Employment Equity Act framework depends on the collection, analysis and reporting of employees’ personal information. Not only do privacy protections support data collection in fulfilment of the purposes of employment equity. Privacy protections can engender trust, through transparency and a high standard for safeguarding personal information. The Office of the Privacy Commissioner added that workers must:

• have confidence in their employers in order to accept to self-identify voluntarily
• trust that their personal information will lead to decision-making and implementation that achieves employment equity
• trust that their sensitive self-disclosed information will not be used for non-employment equity purposes, or improperly accessed or disclosed and used against them.

We would add that centering the purpose – substantive equality at work – and ensuring that the meaningful consultations pillar is strengthened, as discussed in Chapter 5, are crucial to engendering the trust necessary for the balancing of rights.

We must remember the ‘why’ of employment equity data collection. We urge a human rights, purpose-driven, trust-building approach to data collection – including disaggregated data collection –under the Employment Equity Act framework, with a view to achieving substantive equality.

In light of this focus, we make the following recommendation in keeping with our mandate to support equity groups, improve accountability and improve public reporting:

With that focus securely in mind, it is possible also to address data justice across all three pillars of the Employment Equity Act framework: implementation through barrier removal, meaningful consultations, and regulatory oversight.

The self-identification conundrum

Self-identification is the only permissible basis through which an employer may identify employees as members of one or more employment equity group for the purpose of implementing employment equity, according to Section 9(2) of the Employment Equity Act. Section 9(3) establishes strict rules governing the confidentiality of information and its designated use, in keeping with privacy legislation. The 1986 Employment Equity Regulations even included self-identification as part of the definition of the employment equity groups.

Although the definition includes persons who have been accommodated in their current job or workplace, Section 18(4) on self-identification states that only those who identify themselves to their employer or agree to be identified by their employer as a member of an employment equity group may be counted for the purposes of the report.

For strong employment equity data to become available, it is crucial that workers in employment equity groups trust that their participation, through self-identification and the various participatory methods stressed in this report, is a meaningful part of the process. Some employment equity groups have witnessed data collection and self-identification used against them – sources of disadvantage and stereotyping rather than proactive equality. It was disturbing to see that decades after the employment equity legislation has been in place, workers who can “pass” notably as white, or cisgender, or who may keep their disabilities hidden, may feel they are better off doing so.

While self-identification has problems, it is important not to overstate them. Some federally regulated employers report high rates of self-identification. For example, the Bank of Canada reports that 88% of employees participate in the self-identification program.^{Footnote 29} CBC/Radio Canada reports approximately 90% self-identification, and is working to improve the participation.^{Footnote 30}

It is also important to keep in mind that the problems with self-identification do not necessarily relate to all equity groups in the same manner.

Distinct challenges to self-identification for First Nations, Métis and Inuit peoples that emerge from taking nation to nation or government to government relationships seriously are addressed separately and in detail in Chapter 3.

Members of the Community of Federal Visible Minorities in the federal public service indicated that their members tend to self-identify.^{Footnote 31} In contrast, the Public Services Commission of Canada reported in 2014 that only one quarter of persons with disabilities report them within the first two years of public service; most report over the course of their careers in the federal public service.^{Footnote 32} Our task force recognizes that self-identification may be a particular challenge for persons with psychosocial or intellectual disabilities.

Special concerns were shared with our task force about the risks of declaring sexual identity or social expression:

In relation to Positive Space, the Act should address the risks LGBTQ2+ employees currently face when self-declaring (e.g., the risk of their manager being homophobic). It should also address the various social stigmas surrounding different groups (e.g., feminine-presenting employees may not be as readily seen as leaders; employees whose membership in the LGBTQ2+ community is more visible (by the way they present themselves, dress, speak, etc.) may be discriminated against and “othered” consciously or subconsciously).

The task force heard that for members of 2SLGBTQI+ communities, the risk of unintended consequences of gaining access to datasets collected for other reasons could be dire. People may not self-identify precisely because they do not want to be “discovered”. Indeed, researchers have reported that more than half (53%) of 2SLGBTQI+ workers may conceal their identity. They add that inclusive workplace environments can help: supportive policies are linked with an increased willingness to “come out” within the workplace context. Supportive workplace policies can foster inclusion for all workers.^{Footnote 33}

It will be important to anticipate that with the recommended inclusion of 2SLGBTQI+ communities as an employment equity group, additional self-identification questions will arise. Researchers from Toronto Metropolitan University’s Diversity Institute pointed out that some members might resist self-identification if it requires them to share information about their sex lives and they consider this to be “inappropriate or blatantly unfair.”^{Footnote 34}

The acknowledged challenges in establishing trust seem built into the regulations, in particular on employment equity in the Canadian Armed Forces. Section 16(1) directs that when exercising their oversight powers, agencies including the Canadian Human Rights Commission should take into account “the fact that Canadian Forces members who may be members of one or more of the employment equity groups of Aboriginal peoples, persons with disabilities and members of visible minorities, may choose not to identify themselves as such or not to agree to be identified as such.”^{Footnote 35} But the provision itself may indicate how deeply climate-related barriers persist, and that much more work needs to be done.

Public service self-identification modernization project

“In September 2020, the Treasury Board of Canada Secretariat launched the Self-Identification Modernization Project to increase the accuracy of data and reduce the perception of the risk associated with voluntary self-identification. With this project, there is hope for collecting more comprehensive data for persons with disabilities and for other equity seeking groups. For example, the Canadian Survey on Disability conducted by Statistics Canada in 2017 revealed that 15.6% of the Canadian workforce are persons with disabilities. However, in the 2017–18 fiscal year, only 5.3% of public service employees self-identified as persons with disabilities, and only 2.9% self-identified as persons with disabilities when applying for a position. Therefore, a new questionnaire is being designed to capture more relevant data.”

Finally, the Treasury Board Secretariat has issued a Directive on Employment Equity, Diversity and Inclusion effective as of 1 April 2020. It sets out mandatory procedures for self-identification:

We listened closely to the challenges of self-identification experienced both by workers who are members of employment equity groups and by covered employers. It is important to acknowledge right from the start that in the context of a nation reckoning with the need for truth and reconciliation in dealing with the impact of settler colonial laws and policies, Indigenous self-identification is laden and requires specific and particularly careful treatment. These issues are addressed with attention to their specificity and their complexity in Chapter 3.

Ultimately, one of the most promising ways out of the conundrum seems to be something that the Supreme Court of Canada recognized decades ago: achieving critical mass. When equity group members are well represented in workplaces that also include champions, allies, and others who actively support people from underrepresented groups, the workplace environment starts to change, too.^{Footnote 36}

Taking into account the various considerations, our task force considers that the Employment Equity Act framework should move toward making completion of the survey mandatory for all employees covered by the legislation, but self-identification should remain voluntary. In other word, the self-identification form should include the possibility to answer questions – indeed each question – with something like, “prefer not to state”. Currently, Section 3(7) of the Employment Equity Regulations permits this practice but does not require it. The forms should clearly indicate that self-identification itself is voluntary. This practice is already in place in some federal workplaces, including the Canadian Armed Forces.^{Footnote 37} For others, this would entail modifying the current self-identification form in the federal public service, both to include the third option AND to clarify that self-identification is voluntary.^{Footnote 38}

Within the federal public service, self-identification survey data should be centralized, so that employees do not need to complete the surveys each time that they change departments. A consultant engaged by the Treasury Board Secretariat also provisionally recommended in February 2023 that self-identification data should be centralized, noting that Treasury Board already maintains a central data bank of self-identification to facilitate transfers between departments. Centralization would reduce the time required to reconcile data collected locally, to avoid multiple counting of employees who might have been transferred. We recommend making the Treasury Board Secretariat the central record recipient and recorder of self-identification to support harmonization, improve efficiency and speed, with due regard for preserving privacy protections.

Self-identification and data on accommodations

Requesting a reasonable accommodation is not considered by the Labour Program to be self-identification for the purposes of the Employment Equity Act.

Some employers have specifically sought to be permitted to use the data on accommodation requests in their reports on the number of persons with disabilities in their workplaces to determine representation under the Employment Equity Act framework. After all, both concern human rights frameworks, and both require self-identification to the same employer.

The task force considered this proposal closely.

It is worth noting right from the start that this request by employers was raised only in the context of persons with disabilities. A person may request an accommodation on the basis of a disability without actually meeting the definition of “persons with disabilities” under the Employment Equity Act framework. We address definitional issues and offer recommendations for change in Chapter 3.

Our reflection focused on the privacy law dimensions. The task force noted that Section 8(2)(a) of the Privacy Act applicable in the public sector enables personal information to be compiled for a use that is consistent with the purpose for which it was collected. In addition to the 10 fair information principles that form the ground rules for the Personal Information Protection and Electronic Documents Act (PIPEDA) privacy framework in the private sector, would a reasonable employee not consider the use and disclosure appropriate in the circumstances?

The position of the Office of the Privacy Commissioner was that clear legal authority under Section 4 of the Privacy Act is needed. A recommendation would require a specific Employment Equity Act amendment that contained the necessary limitations. Schedule 1 of PIPEDA requires that personal information collected must be limited to what is necessary to the purposes identified by the organization. A new purpose must be required by law; if not, the consent of the individual is required before the information can be used for that purpose.

The Office of the Privacy Commissioner recommended that institutions collect personal information directly from workers, and that the purpose for the collection be made clear at the point of collection. Similarly, in our consultations drawing on experiences in other jurisdictions, the Quebec Commission des droits de la personne et des droits de la jeunesse confirmed that in the face of workers who are reluctant to self-identify, the Commission encourages greater sensitivity-training but considers that the exclusive responsibility for self-identification should rest with the person concerned.

We also received a powerful argument in favour of responding to this objection by being explicit about the multiple purposes of the data collection, including in the context of inter-departmental sharing of information out of which the submission below emerged:

While we understand that under the Privacy Act, data can only be used for the purpose it is collected, perhaps we should change the basis on which we collect data by explicitly stating all the purposes for which it could be used. Privacy is an important stewardship requirement, but stewardship in data also means collect it once and use it effectively. Every time we collect and store the data over again for another purpose we create further possibilities of data breaches. We should be aware of that and of the difficulty of harmonizing all of these processes.

We considered this suggestion closely. It raises concerns about the climate of trust in the workplace. We realized that disclosure is not an all or nothing proposition, but rather a continuum from full disclosure to non-disclosure.^{Footnote 39} Disclosure is linked to its purposes. Data on accommodation requests might be quite detailed because of the purpose, and the datasets might contain particularly intrusive and sensitive material. The risk of having that information more broadly available than necessary seemed to outweigh the benefits, as self-identification data are necessarily quite limited. We did not want to create a situation that could potentially serve as a break on workers coming forward to request accommodations when they need them.

We also wondered whether we might be overlooking something obvious. An employee who came forward to request an accommodation would have already disclosed at least the relevant part of their identity, and likely in greater detail than an employment equity self-identification form could legitimately require. It is possible that the extent of the contexts in which employment equity information could be used – including across the worker’s lifecycle in matters of promotion and leadership training and other aspects of their file – might well give rise to concerns. It was equally plausible that the matter was one of bureaucratic miscommunication, since self-identification has tended primarily to happen only once, when an employee is hired. Over the course of their working life with the employer, the worker might have acquired a disability and sought an accommodation but not thought to update self-identification data. We address the timing of self-identification later in this chapter.

In working through the options, the task force considered whether a method might be devised that would allow the worker to choose whether their accommodation data could be used for self-identification. But given the distinct data banks in which the information is required to be held, at least under the Privacy Act, the risks and complications outweigh the advantages. The Office of the Privacy Commissioner stressed that accommodation files may contain quite detailed and sensitive health or medical information that would not be required for self-identification purposes; indeed, its use could be inconsistent with federal privacy legislation.

In addition, if choice could be given on an individual basis, the risk that it could leave workers feeling compelled to self-identify for employment equity purposes in order to receive the accommodation would be a significant potential unexpected consequence. Given the need for trust for employment equity legislation to be effective, this seemed counter-productive.

Considering the depth of the concern expressed by employment equity groups and the importance of retaining strong privacy protections, our task force considered, instead, that a “soft” approach in this context would be most advisable. We consider that the approach most consistent with the spirit of both employment equity legislation and privacy legislation is to continue to encourage self-identification, but not to require it.

To be abundantly clear, this would under no circumstances permit the distinct datasets to be linked. Instead, an employer who receives a request for an accommodation should be permitted to ask the employee to complete a separate, confidential, and voluntary self-identification form for the purpose of employment equity, after the accommodation request has been addressed. The self-identification form should provide the option under each listed employment equity group to declare not only whether the worker is a member or is not a member, but whether the worker would prefer not to self-identify for the purposes of employment equity. The self-identification form would be kept separate from the accommodation request and remain confidential.

We are sensitive to the risk that some workers who have not already self-identified could feel implicit pressure to self-identify, even though the form would clearly indicate the option not to self-identify. However, we also heard that workers are in many cases self-identifying for the purpose of requesting some types of accommodations. The issue, in other words, is to ensure that distinct datasets are maintained. This recommendation would therefore ensure that the datasets remain distinct. The information in the employment equity dataset would therefore remain separate and strictly limited to the information necessary for the purposes of employment equity. It would only be requested once the accommodation process has been completed. In light of all of these considerations, we think the recommendation is fair and balanced, supporting the purpose of employment equity data collection and access to accommodations.

Data collection and disaggregation

The Act should stipulate that progress must be monitored by the employer … on an annual basis for the 10 ethno-cultural communities that are being tracked by Statistics Canada, and timely actions are taken when its employment equity plan is not on track.

It is important to share granular and nuanced information to achieving employment equity goals. This allows employment equity to be implemented in a sustained, supportive manner that is alive to those most in need of employment equity measures. It enables meaningful consultations and accountability.

It is also important to make sure that data used – and any datasets linked – sufficiently de-personalize individuals. This might be a particular challenge in relatively small workplaces.

For the Office of the Privacy Commissioners, even if this might limit the number of linkages that can be made or even prevent some datasets from being linked, the key is to be “as transparent as possible” about the reasons for the data collection, and to identify cut offs.^{Footnote 40}

The Office of Public Service Accessibility pointed out that it is not always possible to analyze disability by subtype, proposing instead that disaggregated data collection remain optional. A second option is to follow the Office of the Privacy Commissioner’s recommendation that if there are fewer than 10 employees in any given group or sub-group, reporting on those categories or sub-categories should be suppressed.^{Footnote 41} Moreover, we acknowledge that this could lead to challenges if sub-group data are systematically suppressed. A third option would be to aggregate subgroups, and in some cases employment equity groups, where necessary to permit granular reporting without sacrificing privacy. The advantage of the third option is that it reduces the risk of systematic data suppression in some of the smaller employment equity groups.

Given the importance of data disaggregation and the complementary concern for privacy protection, we are inclined to ensure that there is a requirement to disaggregate, but latitude to move between the second or third option. Those options should be exercised on the basis of meaningful consultations as discussed in Chapter 5.

We also agree with the Office of Public Service Accountability that future Employment Equity Act regulations should provide detailed guidance to organizations on how to collect and report on disaggregated data. We consider that this should include promising practices on how to enable a meaningful composite to emerge from the data available. Recent approaches adopted by Statistics Canada to ensure that consultations are undertaken with concerned communities are promising practices that should be further developed and shared. Promising practices should also be shared on how to avoid potentially misleading reporting when the same person is counted multiple times across a number of disaggregated or intersecting groups. Those promising practices exist: consider for example tables prepared by McMaster University to meet employment equity obligations under the Federal Contractors Program. The tables present employment equity groups and their overall composition in the initial columns, then add columns clearly identified to provide a portrait of each group’s intersectionality within the employment equity groups.^{Footnote 42}

Keeping it simple: Moving beyond workforce availability toward labour market availability and comprehensive barrier removal

FETCO members also feel that the dogged focus on Labour Market Availability (which is based on lagging Census figures that are often not detailed enough) is not working. Focusing on specific numbers can take away from successes. Could the focus be on ranges, or some other measure?

Our task force has looked closely at the available data in the ever changing, contemporary post-pandemic Canadian workplace. One message seems clear: we need to simplify.

We did not sense much love for Labour Market Availability and heard a rather fundamental rejection of Workforce Availability.

Section 5(b)(i) of the Employment Equity Act and Section 6(1)(b) of the Employment Equity Regulations set out the process for establishing the benchmark by which representativeness is to be established. It takes as a starting point the availability of each employment equity group in each occupational group, either in “the Canadian workforce as a whole,” or in “those segments of the Canadian workforce that are identifiable by qualification, eligibility or geography, and from which the employer may reasonably be expected to draw employees”.

This offers a fair amount of latitude, which in theory is helpful to setting the short-term goals (1 – 3 years) that the law requires. It can also get complicated quickly.

But our consultations and research confirm: we need to focus on three clear, simple things:

• We need to be able to rely on a more fluid set of data, including projections-data;
• We need to ensure that the benchmarks we choose do not embed the barriers we are trying to remove;
• We need to move ourselves away from excessively complicated calculations and toward a focus on removing barriers to reasonable progress that are sustained over time, so that employment equity is actually achieved.

Moving away from workforce availability in the federal public service

Workforce Availability in the Federal Public Service has exceeded its shelf life.

Introduced by the federal public service to take into account the specific requirements under the Public Service Employment Act in determining availability, WFA has become a primary example of how a measurement tool can embed the exclusionary practices that employment equity seeks to uproot. We heard repeatedly that the Canadian federal public service, of all employers in Canada, should have reached sufficient maturity to be able to use the same metric that is used by most federally regulated employers in Canada.

Below is the Treasury Board’s lay description of how workforce availability is calculated:

Workforce availability (WFA) calculation in the Public Service of Canada:

Workforce availability (WFA) estimates are used as a benchmark to assess the representativeness of employment equity designated groups within the CPA (organizations listed in Schedule I and IV of the Financial Administration Act) in accordance with the Employment Equity Act. Indeed, the dynamics of hiring depend on the availability of designated group members for public service employment. The WFA varies geographically (national or by province or territory) or by the specific qualifications that organizations have to fill. Four filters are taken into account:

1. citizenship: this filter was applied because the Public Service Employment Act gives preference to the hiring of Canadian citizens (section 39.1.c); this preference was extended to permanent residents as of June 29, 2021
2. classification: this filter narrows consideration to occupations that the government deems relevant to the public service
3. education: this filter is used for some classifications to include only persons who have an educational degree for scientific and professional occupations, taking into account the public service qualification standards for jobs (educational requirements)
4. geography: this filter assumes that most organizational hiring will be done locally for most occupational groups, rather than from wider geographic areas

WFA estimates are derived from the Labour Market Availability, which is derived from the Census and the Canadian Survey on Disability, which is performed every five years. The WFA [has in the past included] Canadian citizens who have been in the labour market for at least 15 years based on their presumed ability to work in occupations comparable to those in the public service.

Treasury Board of Canada Secretariat, Annual Report: Employment Equity in the Public Service of Canada 2020 – 2021 at 28-29.

Among many others, the Public Service Alliance of Canada considered WFA not to reflect the population in Canada and to be outdated. Repeatedly, our task force heard from employers, workers, equity groups, and experts that WFA does not reflect the population in Canada. We heard that it is outdated. We heard that the challenge is less the available pool than the policies and practices on hiring, promotion and retention.^{Footnote 46}

The most significant problems with WFA can be observed vividly in relation to the visible minority/ racialized workers category:

The manner TBS calculates WFA is the most significant systemic barrier and discriminates all Visible Minorities. According to the TBS report for 2020-21, Visible Minorities representation in the Federal Public Service is 18.9%, which appears favourable when compared with the VM’s WFA of 15.3%. However, it is worth noting that it is still less than the population of visible minorities which was 19.6% in 2011, their current Labour Market Availability is 21.3% and their population is 22.3%, based on 2016 census. Therefore, the reality is that the VMs representation is 2.4 points lower than their LMA (even though six years outdated), and 3.4 points lower than their representation in the Canadian population in 2016. For EX classification, VM representation is 12.4%, which is 1.8 points higher than their WFA but 9.9 points lower than their representation in the population of Canada. Keeping the benchmark for EX so low in fact institutionalises the assumption that visible minorities are less qualified to be public service executives. … It is also worth noting that the flawed benchmark used by TBS is based on census 2016, and used as a benchmark even in 2022, six years out of date. … instead of being a leader, the Federal Public [service] is a laggard, and is lagging behind the [federal] private sector.

As early as 2000, the task force report on the participation of visible minorities, Embracing Change in the Federal Public Service recommended that WFA be jettisoned. The Community of Federal Visible Minorities made the same recommendation.

Frankly, it was difficult to review a graph like the one submitted by the Public Service Commission, which clearly showed that applications from members of visible minorities significantly exceeded the WFA numbers, often by 10%, and hiring also tended to exceed WFA by between 4 – 6%, knowing the extent to which it underestimates racialized populations in Canada.

Text description of figure 2.1

Members of visible minorities: applicants to advertised processes and hires compared to workforce availability

Fiscal year	Percentage of applicants	Percentage of hires	Workforce availability
2015–16	21.3	17.3	13.0
2016–17	21.0	17.9	13.0
2017–18	22.2	17.7	13.0
2018–19	22.3	19.3	15.3
2019–20	24.3	21.3	15.3
2020–21	25.0	-	15.3

• Source: Public Service of Canada, Presentation to the EEART, 18 March 2022.

It is understood, too, that now that the Public Service Employment Act has been amended on 29 June 2021 and citizenship is no longer a requirement for employment, the biggest difference in WFA rates is likely to be experienced in the visible minority group. Permanent residents will be included in the availability rates. This was, however, one of the most important justifications for undertaking all the work of establishing a separate benchmark for the federal public service. That justification is gone.

The remaining differences – in particular, excluding temporary migrant workers and foreign students from LMA – may not justify the time and effort required to produce a separate standard for government. Although the rules are different, we should keep in mind that private sector employers cannot simply hire without regard to immigration status either. No matter, if LMA and WFA are substantially similar with the removal of the permanent residency restriction in the federal government, it is entirely reasonable to anticipate that the federal government will be able to base itself on LMA as characterizing the workforce from which it may reasonably be expected to draw. This would represent a more fluid approach and send the right signal to employers at large.

Some stakeholders went further, pointing to the significant disparity between WFA and population-based information. The outgoing Deputy Minister of Public Service Accessibility and Champion for Federal Employees with Disabilities, Yasmine Laroche, shared findings from consultations. The significant difference between WFA, situated at 9% for disabled workers, and actual population-based availability at 22% based on the 2017 Canadian Survey on Disability, was cited as a stark example of the representation challenge.^{Footnote 47}

It is important to note that 64% of employees surveyed on the draft Public Service Accessibility Strategy rated replacing the concept of workforce availability in setting employment targets for persons with disabilities with a concept that accounts for work potential as defined in the 2017 Canadian Survey on Disability to be either very important or somewhat important.^{Footnote 48} We acknowledge that work potential will change when there are fewer barriers. We also recognize the importance of supporting equitable inclusion for persons with disabilities within the workplace with basic income support measures at the societal level, as discussed in Chapter 4.

With the growing potential of remote work, some considered that the federal government should set an example by not relying on a benchmark that incorporates a geography filter to assume that hiring will necessarily be local.

The main argument against shelving WFA in the federal public service was the potential that departments and agencies might feel discouraged if they feel the targets are unrealistic. When we looked closely at the data and the significant differences in actual attainment rates between some governmental departments and agencies, we were told the key difference is that some were intentional about removing barriers and fostering equitable inclusion. Expending extensive resources to get a more precise WFA will not fix this cardinal difference. WFA has been a source of discouragement for too many employment equity groups. We were persuaded that our federal government, assuming its dual responsibility as a symbolically important federal employer as well as regulator, can do better.

The production of WFA is now understood to be a distraction, taking time and energy away from the fundamental work of barrier removal. Government should be the site for experimentation and creativity for sustainable change. It should be at the vanguard, not the rear guard.

To use WFA in 2023 was likened by one public service employee to trying to follow social media with a flip phone. We must do better.

During its consultations with the task force in Spring 2022, TBS-OCHRO informed the task force that it would be commissioning its own study of WFA. The Treasury Board’s president is a member of Cabinet and Treasury Board is a government employer. The Treasury Board exercised its capacity to seek recommendations for legislative change through a separate process, while the task force was ongoing. A consulting firm was hired in Fall 2022, and a draft report was shared with the task force Chair in February 2023. The consultant provisionally recommended retaining the use of WFA.

The consultant’s recommendation referred to a criterion that currently exists in the Employment Equity Regulations, that is the workforce from which the core public service may “reasonably be expected to draw” employees. The legislated standard itself – reasonableness – seems not to have been the one applied in the consultancy report. Instead, a different, higher standard - “accuracy” - was substituted.

Although the consultant’s report notes the benchmark lags up to seven years, the report seems not to have turned attention to the serious issue of the time and complexity involved in preparing a separate WFA. The task force learned that some departments faced with challenges in the WFA that limit the ability to measure representation when government priorities change and recruitment requires different NOC codes to be used, calculated their own WFA using Treasury Board’s methodologies. And the process is hardly transparent; calculation building blocks are not published or even well known, and interpretations may vary.^{Footnote 49}

The excessive focus on developing separate WFA benchmarks, potentially on an annual basis relying on Demosim despite real feasibility challenges discussed above, risks simply exaggerating the attention placed on representation numbers, to the detriment of actually achieving employment equity. This must stop.

Rather than building on Demosim while keeping WFA, our task force has recommended mandating the Employment Equity Data Steering Committee to consider how best to draw on existing and emerging projections capabilities, to redress the time lag in the calculation of LMA alone. We further recommend shelving WFA altogether.

Recommendation 2.25: The federal public service should cease producing and relying on workforce availability to meet its responsibilities under the Employment Equity Act framework.

Toward a fluid approach to representation and goal setting

Use representation by population rather that convoluted formula that [replicate] patterns of systemic discrimination.

So, what do we need to be able to represent our society the way it would be in the absence of systemic discrimination? How do we get there?

We heard loud and clear: the federal government should be aspiring to be the face of Canada.

Labour market availability is a good start for federally regulated workplaces. But our aspiration should be that our workplaces reflect our populations. A barrier-removal focus is best placed to get us there. This will require proactive initiatives by our government to address the barriers that intersect with the workplace discussed in Chapter 4. This cannot and should not be borne by employer action alone. It is an ongoing, interactive process and requires an all of society approach. It should be integrated into the Canadian government’s adherence to the United Nations’ Sustainable Development Goals, including on gender equality and decent work, and tackled holistically.

We cannot help noting, as well, that to calculate “availability” in the context of a tight labour market might in itself beg the question. We need everyone and the tight labour market reminds us of how important it is to be finding ways to draw people in and not just moving people around.

Because of the current restrictions on litigating employment equity, we may also have lost sight of the fact that in the two leading federal cases on employment equity, both the Supreme Court of Canada and the Canadian Human Rights Tribunal set targets that were above availability. The reason was clear: underrepresentation was a reflection of the discrimination experienced by employment equity groups historically. To wait a generation or two until they become “available” is to allow the effects of the past discrimination to persist.^{Footnote 50} Employment equity sought to correct the underrepresentation in a steadfast and sustainable manner.

Within the federal public service, the Canadian Human Rights Commission already advises public service employers that it is a good practice to exceed the employment equity attainment rates under workforce availability, knowing that the benchmarks are based on data from previous years.

The task force paid special attention when employers in both the public and private sector referenced basic Census or Canadian Survey on Disability workforce numbers, and rather than complicating the analysis, turned their focus on how they could remove barriers and transform their workplaces. Some other members of employment equity groups or subgroups have also called for representation numbers to be set on the basis of the general population, in particular the Black Class Action Secretariat.

Turning the efforts toward barrier removal made more sense than applying more and more complex analyses of the availability data. We discuss this in detail in Chapter 4.

For the federally regulated private sector and employers covered by the Federal Contractors Program, in addition to the federal public service, this means promoting a more fluid approach to representation.

A more fluid approach will help with some of the concerns raised by some private sector employers who have reported that NOC codes do not even fit their industries.^{Footnote 51}

For all, this means that the focus on Labour Market Availability, while important, should not become a death grip that locks workplaces into outdated data and sterile exercises that tell us little about what is actually happening to achieve and sustain substantive equality.

We have the labour market availability benchmark, of course, because the goals need to be realistic, and should not be arbitrary. And we want to foster reasonable progress over time to achieve and sustain employment equity.

Ultimately, labour market availability should be thought of as a floor, with a broad societal aspiration for population-based representation rates that show that it is possible to achieve and sustain substantive equality in a barrier-free workplace for all. Because it is 2023.

Our workplaces should increasingly look like Canada. The benchmarks help us to get there. They should not be hindrances. They should not get in the way of the real work of equity.

In our globalized, integrated societies, workplaces are able to recruit widely, and workers are increasingly mobile. Technology and our pandemic-era work-from-home experience have made this more apparent for many employers. Where there are labour market shortages, we are increasingly recruiting internationally. We are in a moment where we have shown creativity and experimentation on work-life issues. The creativity and experimentation should purposefully be used to foster greater accessibility and bring more people into the workforce. We should avoid artificial limits on employers’ ability to look for talent across the Canadian workforce, and to do so while fostering equitable inclusion. This is said with care, because we acknowledge the immense challenge of representation across a country as vast as Canada, with remote and rural communities alongside major metropolitan areas.^{Footnote 52}

So long as representation is lower than Census population levels, employers should be permitted to continue to work to correct underrepresentation.

We recommend that the “Canadian workforce” should be the default. Requests for derogations should be addressed to the Employment Equity Commissioner and assessed on a case-by-case basis for a time-limited period.

We also heard the frustration faced by employment equity groups whose labour market availability is consistently low and attainment rates rarely received.

Here, especially, we should take pressure off the narrow calculation and place efforts resolutely on fostering inclusion.

Recommendation 2.26: The “Canadian workforce” under Section 5(b)(i) of the Employment Equity Act should be the default benchmark in the Employment Equity Regulations.

Recommendation 2.27: Requests for derogations from the default benchmark should be addressed to the Employment Equity Commissioner on a case-by-case basis for a defined time period.

Recommendation 2.28: So long as representation is lower than Census population levels appropriate to the geographic context, employers should be permitted to continue to work to correct underrepresentation of employment equity groups, focusing on obtaining critical mass.

Keeping it simple: Reducing the reporting obligation

We were also wisely urged to keep reporting simple.^{Footnote 53} In Chapter 3, we recommend adding two employment equity groups. Challenges associated with certain subgroups require attention, so we enable and encourage data disaggregation. It is clear that the smaller the sub-group, the greater the risk of data suppression. We need to allow fluidity on assessing and calculating representation rates and attainment. We take up this suggestion with a specific recommendation to reduce the frequency of reporting by all employers in Chapter 4. We also offer recommendations on streamlining reporting in Chapter 6.

Public transparency: Democratizing access to data

The logic of the Employment Equity Act is that the Act’s effectiveness depends on scrutiny of employers’ results by the public as well as by government compliance authorities.

One of the reasons we have not advanced on employment equity is that the public has been required to rely on the symbolic as proxies for action on equity, rather than on the data. Workplace equity in a universe of EDI has tended to be measured by statements and communications strategies that have come to symbolize diversity.^{Footnote 54} But do they work? For example, do we have the information necessary to ensure that these workplaces have been promoting members of employment equity groups into managerial positions? How do we know whether they have revolving doors? How do we know whether sexual harassment claims are through the roof?

To meet our international human rights obligations, Canada has been encouraged to publish comprehensive, disaggregated data for the public service, for “[a]ll contractors to public service agencies”, and for private employers by the UN Committee on the Elimination of Racial Discrimination.^{Footnote 55}

Federal employers have themselves been calling for learning to be shared:

Suggest government share widely any learnings with the entire stakeholder community (perhaps look at best practices). Employers could then, for example, use this relevant information to improve their own EDI initiatives.

Data justice also includes democratizing access to data through greater disclosure obligations, to enable the public at large to understand whether employment equity is being achieved.

We were reminded that employers see a business case to equity, and that they want to be able to tell their positive narratives about success. They do not just want to be reporting their numbers outward like an academic exercise; they want a process of better data collection that allows them to have successes, and share them as promising practices.^{Footnote 56} If this is the ground we are rebuilding employment equity upon, it is pretty solid.

Currently, reports submitted on an annual basis to Parliament comprise the main approach to public disclosure by public oversight agencies.

Audits conducted by the Canadian Human Rights Commission of the federal public administration are not made public; they are not even shared with the Labour Program. They were redacted even in the case of review by the task force chair.

Employment equity results should not be hidden. Nor should the goal be public shaming. Rather, employment equity results should be the basis of deep learning. Success in achieving and sustaining representative workforces should become a sign of pride that we can all celebrate.

Currently the Employment Equity Act provides for reports to be made available, including from private sector employers:

The practical experience with this requirement leaves much to be desired. Consolidated versions of the report are included in the Annual Report to Parliament pursuant to Section 20 of the Employment Equity Act. Treasury Board also submits an annual report to Parliament. Although the Labour Program’s website normally allows reports to be searched, the search function was rendered inoperative at a critical period in the report drafting phase and the task force chair was not otherwise able to obtain access to the reports. The Labour Program must do better.

The Canadian Bar Association made the following comprehensive suggestions for rendering information publicly available, noting that none requires legislative change:

We agree. In practice, though, and in the task force’s own experience, the provisions in the Employment Equity Act on the availability of reports have not been interpreted to foster readily accessible, reasonable public access.

A range of stakeholders has called for reports to be made readily available and searchable online to ensure easy access to information. The DisAbled Women’s Network of Canada (DAWN) considered that having to proceed through an access to information request to obtain an employer request, with the associated fee, to be a barrier.^{Footnote 57} Several have also called for compliance and audit reports to be made publicly available.^{Footnote 58}

Reports filed by all those covered by the federal Employment Equity Act framework - including the federal public service, the federally regulated private sector, and FCP employers - should be accessible, both in location and format. They should be easy to find, and easy to read. They should respect current accessibility standards.

Summaries of audits, using promising practices on data sharing for comparability, should similarly be made available to the general public.

The task force was provided with a first glance of initiatives underway to render pay transparency data readily available to the public through the Workplace Equity Program of the Labour Department. With due regard to privacy protections, we would encourage the proposed Employment Equity Commissioner to develop tools that foster appropriate public sharing of employer reports. Protocols should be developed to ensure that proprietary information can be excepted from the information that is shared.

Guidelines should be put in place by the proposed Employment Equity Commissioner, who is encouraged to build an accessible, clear electronic form that can be completed with the expected data and that can then be made readily available and accessible online in a centralized and searchable site by the Commissioner.

A complementary initiative would be to create an open government site for employment equity reports, and make all reports filed under the employment equity framework available through that accessible, searchable database.

Data centralization in the federal public service

The task force heard loud and clear that data on compliance with the Employment Equity Act should be made available in a ready, accessible manner. There was frustration with how data from the federal public service was made available, when it was made available:

It would be great if all the data with regards to EE were clearly accessible to the public if you visit the page of any department (for example a quick graphic under the About Us Tab with a breakdown of just how diverse departments/regions are) If that information is available, it should be easy to find on the webpage and not buried in some long report somewhere that would be challenging to find.

Part of making the data publicly available and enabling regulatory oversight is to make sure that data are available in an accessible and organized format.

The task force was informed by the Public Service Commission that the slow release of data can be a problem. The current requirement dictates that a gap needs to be identified before action can be taken; this slows down proactive measures. Our task force also heard that in the public service, the collection, processing, reconciliation and quality control takes approximately 7 – 8 months. Since reports to Parliament are aligned with fiscal years, they are tabled by March 31. In light of this, the data are published on a yearly basis – a 12-month basis, even though they may be ready well in advance. Some entities wanted the data to be made available between the time that employment equity data is calculated and the time that it is tabled in Parliament. The Office of Public Service Accessibility put it this way:

Public service employees are often inhibited from accessing employment equity data for months after EE reports have been generated. This is a consequence of EE data having restricted access prior to being tabled in Parliament. Despite the understanding that using and integrating EE data in public service initiatives drives forward progress, accessing EE data during this ‘restricted’ phase puts analysts at risk of being in contempt of parliament. Delaying internal communication of EE data reduces accountability and strays from the spirit of the EEA: equitable access and pushing for improvements in the workplace for minority groups.

Some inspiration for the federal public service might be found from the United Kingdom. Since 2011, the United Kingdom’s Equality Act, 2010^{Footnote 59} has included a duty on public bodies to publish information annually that shows compliance with a duty to avoid discrimination, advances in equality of opportunity, as well as equality objectives, in addition to longstanding provisions permitting “positive action”. Regulatory oversight lies with the Equality and Human Rights Commission.

Of course, the recommendations in this Chapter move us toward greater access to data as well as streamlining and simplification of reporting. Further recommendations are formulated in Chapter 6. But we take the point that someone should have an eye to this specific problem.