Evidence and equity: struggles over federal employment equity policy in Canada, 1984-95.

Grundy, John ; Smith, Miriam

John Grundy is a doctoral candidate, Department of Political Science, York University. Miriam Smith is professor, Department of Social Science, York University. This research was supported by a Social Sciences and Humanities Research Council of Canada grant to Professor Smith. An earlier version of this article was presented at the annual meeting of the Canadian Political Science Association, Concordia University, Montreal, 1-3 June 2010. The authors thank Barbara Cameron and Leah Vosko and the journal's anonymous reviewers for helpful comments. Evidence-based policy-making is central to government modernization initiatives across a range of jurisdictions, including Canada. Based in part on the ascendance of evidence-based medicine in the 1990s (Mykhalovskiy and Weir 2004), this model of policy-making emphasizes the need for rigorous appraisal of scientific evidence in the formulation of policy. Proponents of evidence-based practice argue that incorporating the best available social scientific evidence into policy deliberation results in public policies that are more effective and rational than those forged on the basis of consensus, partisanship or ideology (Packwood 2002). The influence of the evidence-based model is further evident in the field of public administration as public-sector organizations increasingly adopt quasi-scientific performance measurement strategies. In this way, the evidence-based movement is closely aligned with the principles and techniques of "new public management" (Clarke 1998; Saint-Martin 2000; Solesbury 2001).

A growing body of scholarship calls into question the tenets of the evidence-based policy paradigm (see Laforest and Orsini 2005; Parsons 2002). According to a number of researchers, the enthusiasm shown by policy-makers for evidence-based practice reflects an increasingly technocratic approach to policy-making and implementation that is associated with neoliberal governmentality (Hall 2005; Peck and Theodore 2010). In this light, the evidence-based model serves not simply to augment the knowledge base of policy-makers but to contain fundamental disputes over the direction of public policy by imposing new technocratic policy discourses that further limit the kinds of knowledge and actors that count in decision-making processes (Grundy and Smith 2007). Evidence-based policy-making is cast from this perspective as a mode of anti-politics, usefully defined in recent work by the anthropologist Tania Murray Li as the practice of "reposing political questions as matters of technique; closing down debate about how and what to govern and the distributive effects of particular arrangements by reference to expertise; [and] encouraging citizens to engage in debate while limiting the agenda" (2007: 265).

This article examines federal employment equity policy in Canada through the lens of evidence-based policy-making. While developments in this policy field are well documented (Abu-Laban and Gabriel 2002; Agocs 2002; Lum 1995; Timpson 2001), revisiting them yields a range of insights that can inform and advance recent critical scholarship on evidence-based policy-making. The Employment Equity Act of 1986 represented the Mulroney government's response to growing political pressure to address the problem of systemic employment discrimination in the federally regulated sector. The act committed extensive resources to the development of an administrative machinery for the collection, analysis and dissemination of statistical data on the representation of four designated equity groups (women, aboriginal people, visible minorities, and people with disabilities) employed in this sector. Rather than imposing measures that would directly interfere in employer practices, the only enforcement measure contained in the act was a $50,000 fine that could be levied on employers who failed to submit annual statistical reports on the representativeness of their workforce to the federal government. The data-centric employment equity framework promoted by policy-makers was premised on a depoliticized depiction of workplace inequality as neither intentional nor ingrained but rather as rooted in a lack of knowledge on the part of employers about the representativeness of their workforce (Lum 1995: 48). The federal government claimed that mandatory statistical reporting of workforce composition would improve federally regulated employers' awareness of workplace inequity and, in turn, that such knowledge would result in more equitable employer practices. Employment equity policy thus exemplified certain tenets that scholars suggest underlie the evidence-based model--specifically, the belief that technical solutions may be found for political problems and the belief that policy deliberation organized around acquiring more and better knowledge can lead to consensus between conflicting policy stakeholders (in this case, equity-seeking groups fighting for workplace equality and employers defending their autonomy in workplace decision-making).

The objective of this article is to highlight the federal government's failure to depoliticize the problem of workplace inequality with the 1986 Employment Equity Act. The article traces how policy-makers were unable to redefine the problem of workplace inequality in terms of the technical, data-centric solution offered by the act. It demonstrates that, rather than unifying conflicting stakeholders in a field of technocratic policy-making, employment equity policy gave rise to extensive struggles over the nature and role of evidence in policy-making and implementation. From the passage of the Employment Equity Act in 1986 until its reform in 1995, employers and equity groups battled over the efficacy and legitimacy of the statistical practices that would be used to assess the representation of each group. Struggle between stakeholders over the meaning and use of statistical evidence unfolded in three phases. The first phase of debate, preceding the passage of the initial legislation (1984--86), was characterized by growing conflict over the role statistical data would play in implementation and enforcement. The government argued that numbers alone would elicit change from employers, while equity-seeking groups argued for strong enforcement of workplace remedies. The second phase of debate (1988--92), prompted by the release of the first set of employment equity statistics, involved deepening conflict between policy stakeholders over the enforcement of employment equity. During this period, the government instituted a weak enforcement process designed to forestall opposition from employers while placating equity-seeking groups. The government's attempt to manage deepening conflict was unsuccessful, as it became the target of litigation by both equity groups and employers, centring on the legal actionability of the newly generated data. During the third phase of policy debate (1991--95), which was occasioned by a mandatory review of the original act, equity groups and employers mobilized methodological critiques of employment equity statistics from different perspectives and to different ends, battling over the shape of proposed reforms. The very categories created under employment equity for some of the designated groups (e.g., disability) and the methods used to assess progress for each group (e.g., how to count members of each group in the workplace) proved to be unstable and contested.

The fourth section of the article considers how the Chretien government revised the Employment Equity Act (in 1995) to eliminate conflicts over the nature of evidence and to restore the possibility of a technocratic space in which workplace discrimination would be managed without disruptive political conflict. To this end, the enforcement regime introduced under the 1995 act emphasizes the use of "soft law," negotiated remedial measures for non-compliant employers, and severely restricts the capacity of equality advocates to participate in enforcement. Overall, this study sheds new light on the case of employment equity policy by demonstrating how policy battles of the 1980s and 1990s were driven not only by conflicts between equity groups and employers but also by policy-makers under governments of both political parties, who continually sought to impose an anti-political solution that would relieve them of the political pressure.

The project of anti-politics: statistical knowledge as a policy solution to systemic discrimination, 1984-86

In Canada, as elsewhere, legislative action against employment discrimination took off in the 1960s and 1970s. By the late 1970s, human rights legislation that prohibited discrimination based on sex, race, national origin and other grounds had been established in federal and provincial jurisdictions. Mechanisms of redress were put into place based on a system of individual complaints. By the early 1970s, the concept of discrimination was in transition in Canada, as in the U.S. and elsewhere. The system of individual complaints was criticized as a cumbersome and reactive system that created backlogs and placed an inordinate burden of time and money on the individual complainant. The newly coined concepts of "affirmative action" and "systemic discrimination" were based on the assumption that equality in the workplace was not simply a question of individual conduct; rather, it was a question of structural barriers as well as organizational and workplace cultures that defined certain types of people and behaviour as normative (Black 1985: 15; see also Agocs and Jain 2001).

Civil rights reform in the U.S. in the 1960s provided some of the conceptual parameters for Canadian debates on the use of statistics to assess systemic discrimination. In 1961, President Kennedy issued an executive order prohibiting discrimination among federally regulated contractors, requiring them to assess and report on the representativeness of their employee base and to develop affirmative action plans, including goals and timetables to correct the underrepresentation of minorities (Leonard 1990: 48). Where statistical data showed underrepresentation, employers had to develop affirmative action goals and timetables to ensure workforce equity, regardless of the cause and without any further investigation. The sanctions for non-compliance ranged from warnings to contract debarment (Geller 1985: 20). This approach was confirmed in a landmark U.S. Supreme Court decision in Griggs v. Duke Power Company (401 U.S. 424 [1971]), in which the court accepted the concept of systemic discrimination defined by numerical imbalances and confirmed that workforce statistics constituted prima facie evidence of discrimination, thus placing a heavier burden on all employers to avoid the appearance of discrimination (Leonard 1990: 60). According to Peter Robertson, this was "a totally new way of thinking about the workplace" (1987: 8), based on what Patrick Simon later called an "arithmetical interpretation of equity" (2005: 13).

The focus on systemic discrimination was also reflected in Canada's federal human rights legislation and in the Charter of Rights and Freedoms (1982), which recognized the validity of affirmative action and specifically permitted such action under Section 15. In 1976, the federal government implemented a voluntary federal contracting program, which encouraged companies contracting with the federal government to establish affirmative action programs (Agocs 1986). In 1979, the federal government also introduced a voluntary affirmative action program. Under this program, staff in the Canada Employment and Immigration Commission sought to convince employers to develop data on their employment systems, to identify potential forms of underrepresentation and discriminatory practice, and to develop targets for improving equity. However, this voluntary program had a very low take-up rate among employers. Out of 1,400 employers contacted, only 71 agreed to participate (Canada, Royal Commission on Equality in Employment 1984: 195). Throughout this period, the Canadian Human Rights Commission and the Canada Employment and Immigration Commission were key players in pushing for a focus on systemic discrimination and affirmative action programs (see the detailed discussion in Timpson 2001:70-88).

In light of the federal government's failed voluntary program, a number of stakeholder groups representing labour, the women's movement, aboriginal people and others recommended that the federal government undertake a mandatory affirmative action program. This recommendation was supported by the Canadian Bar Association, as well as in federally commissioned studies of the labour market (Agocs 1986: 155--56; Hunter 1993). In response to these calls for action, in 1983, the federal government appointed the Royal Commission on Equality in Employment (the Abella Commission) to explore problems of systemic discrimination in the workplace.

The Abella Commission's formulation of employment equity laid the groundwork for a made-in-Canada solution to the problem of systemic discrimination, one that sought to avoid the political backlash over affirmative action that had occurred in the U.S. (Bakan and Kobayashi 2007). The report advanced a central role for data in rendering visible barriers flowing from seemingly neutral structures, which had the effect of discriminating against specific groups, as well as barriers based on stereotyping of specific groups based on putative characteristics (Canada, Royal Commission on Equality in Employment 1984: 9--13). It affirmed the importance of statistical data in providing a way around the complaints system that focused on individual complaints and that, in some cases, might require evidence of intent to discriminate (8).

Yet, the Abella Commission report invoked a critical caveat regarding the role of statistical data in redressing systemic discrimination. Departing from an important principle of affirmative action in the United States, it explicitly disavowed the idea that data could serve as prima facie evidence of discrimination. According to the report, "[d]ata are indicators. They can signal the possibility that a group is being adversely affected by discriminatory practices .... But data are by no means determinative. They simply indicate whether further investigation is necessary (Canada, Royal Commission on Equality in Employment 1984: 206). According to Justice Rosalie Abella, where the data indicated underrepresentation, an enforcement agency would investigate. If the investigation revealed evidence of discrimination, the enforcement agency would suggest that the employer alter his or her employment practices (203--206). Numbers alone could not serve as evidence of discrimination.

In response to the political pressure from equity stakeholders and the release of the Abella Commission's report, the government enacted the Employment Equity Act in 1986. The act applied to employers in federal jurisdiction with more than 100 employees, with the exception of the federal public service, the military and the Royal Canadian Mounted Police. It required employers to submit reports, beginning in 1988, to the Canada Employment and Immigration Commission on the representativeness of each of the four designated groups (women, aboriginal people, visible minorities, and people with disabilities), indicating their hiring, promotion and pay levels. Reflecting the Abella report's rejection of affirmative action, the legislation did not mandate the hiring and promotion of employees from designated equity groups based on the data. The only enforcement mechanism in the act was a $50,000 fine for failing to submit a statistical report to the Canada Employment and Immigration Commission. Enforcement was left to the Canadian Human Rights Commission under its then-existing legislative mandate to initiate a complaint if it had reasonable grounds to believe that a discriminatory practice had occurred (Hucker 1997: 565). The government expected that data collected under employment equity could furnish such reasonable grounds.

The government's employment equity framework downplayed conflicts between employers, the Canadian Human Rights Commission and equity groups. As Janet Lum argues, the rationale underlying the government's approach to employment equity was that there existed a shared concern with fairness among all parties in the policy field. Employers were not intentionally discriminating nor was discrimination ingrained in the workplace; rather, the underrepresentation of designated groups reflected the "thoughtlessness of well-meaning people" (1995: 48). Once employers had more and better statistical knowledge about the extent of underrepresentation in their own backyard, they would be prompted to review human resources management systems and design programs to improve representation. The government stated that, through the power of evidence, the objective of workplace equality could be squared with the desire of employers for control over hiring and promotion (Canada, Parliament, House of Commons, Legislative Committee on Bill C-62, Employment Equity Act 1985a: 2:7).

Rather than providing a technical solution to the conflict among stakeholders, the proposed employment equity enforcement measure was a flashpoint for disagreement. Equity groups pointed to the lack of mandatory compliance mechanisms. A representative of the Federation des femmes du Quebec and the National Association of Women and the Law challenged government officials to produce an example of a law mandating compliance without sanctions for non-compliance (Suzanne Boivin, cited in Canada, Parliament, House of Commons, Legislative Committee on Bill C-62, Employment Equity Act 1985b: 3:49). Akua Benjamin, of the Urban Alliance on Race Relations, summed up the view held by many equity advocates by stating that "[t]he bill is empty, as far as we are concerned. The bill actually is a data-collecting instrument" (Canada, Parliament, House of Commons, Legislative Committee on Bill C-62, Employment Equity Act 1985d: 7:115). Equity groups also criticized the proposed use of employment equity data to initiate complaints under the Canadian Human Rights Act as conflating incommensurate approaches to discrimination. While the equity stakeholders were not homogeneous in their views, many of them did point to the contradiction of committing extensive resources to the creation of a new system of data collection and reporting to make visible systemic modes of inequality while preventing redress for statistical underrepresentation through means other than the traditional human rights tribunals with their reactive, individualized and fault-finding focus. Under the proposed measures, rather than proactive management through state policy, individuals or groups (namely, organizations representing designated groups) would have to analyse statistical data and bring cases of non-compliance to the attention of the Human Rights Commission. Equity groups considered this arrangement as undermining the proactive systemic approach to discrimination that was made possible through statistical reasoning (Canada, Parliament, House of Commons, Legislative Committee on Bill C-62, Employment Equity Act 1985d: 7:102, 106). Moreover, some of the groups further anticipated extensive difficulty in operationalizing the new enforcement scheme. They pointed out that there was no experience in Canada with using statistical data as evidence of discrimination under existing federal human rights legislation and that, in the absence of a strong mandate for such enforcement, "[i]t will be a long and arduous process to develop the necessary jurisprudence required in the area" (Canada, Parliament, House of Commons, Legislative Committee on Bill C-62, Employment Equity Act 1985c: 5:47). Indeed, there was hesitation in Canadian courts in using statistics as evidence of discrimination. In the key case of Action Travail des Femmes (CN Rail) v. Canada [1987] 1 S.C.R. 1114, the courts relied on evidence of discriminatory practices and attitudes and not solely on data on the underrepresentation of women in the workplace (see Charbursky 1992: 342; Timpson 2001: 77--80). In sum, equity groups contested the government's technocratic solution and doubted that the legislation could be more than a "data-collecting instrument," which would not lead to the workplace change for which these groups were battling. In doing so, equity groups continued to inject politics into the government's anti-political policy template.

In contrast, employers played a direct role in pressuring government to avoid strong affirmative action measures (Agocs and Burr 1996). They were disarmed by the Abella report and the government's proposal, given that Abella's disavowal of mandatory affirmative action was carried forward in the legislation. A prominent employer organization, Federally Regulated Employers--Transportation and Communications, defended the proposed enforcement mechanisms, emphasizing that statistical data would play an important informal role in enforcing equity measures. In place of a plan to use the data generated by employment equity reporting to craft mandatory enforcement measures, government and business repeatedly asserted that the data would be used as moral leverage to encourage compliance among employers (Canada, Parliament, House of Commons, Legislative Committee on Bill C-62, Employment Equity Act 1985a: 2:7). In other words, employers favoured an evidence-based approach, as long as the evidence was not used by government to force them to action in the workplace. Equity groups doubted that evidence alone would serve as moral leverage in the absence of strong enforcement (Canadian Human Rights Advocate 1986: 13). Thus, the government's search for a technocratic solution failed to secure agreement among stakeholders.

In response to these debates, however, a few additions were made to the legislation, none of which served to embed the strong enforcement that had been favoured by equity stakeholders. At the recommendation of the Canadian Human Rights Commission, a clause was added requiring employers to prepare employment equity goals and timetables, although no sanction was included for failure to meet them. In addition, as a means to confirm the commission's enforcement role, a requirement was added stipulating that the reports had to be submitted to the commission (Hucker 1997: 564--65). At the same time, the Federal Contractor Program was established, which required companies of a certain size bidding on federal government contracts over a certain amount to submit equity plans and to provide reports on the numbers of employees in each of the four designated groups. Employers who failed to comply could be excluded from future bidding on government contracts.

To summarize the first phase of debate, the federal government advanced employment equity as its response to the increasingly high-profile problem of systemic discrimination. In place of measures that would interfere in the human resources' decisions of private companies, the legislation required the implementation of a data-reporting regime, putatively furnishing the statistical knowledge base for more equitable human resource management practices for federally regulated employers. Yet, the federal government was unable to reconcile its technical solution with the deeply politicized issue of workplace discrimination. Equity groups disrupted the anti-politics of employment equity and contested it as a near complete capitulation to employers and an empty gesture in addressing systemic discrimination. The federal government's attempt to mitigate political conflict, in Tania Murray Li's terms (2007: 265), and to render questions of justice and inequality in the workplace as technical issues manageable through the tools of data collection and statistical measurement, had already run aground before the legislation was enacted.

The failure of anti-politics: implementing equity, 1986-92

Struggles over the role of evidence intensified during the implementation phase of the policy. Despite the work undertaken by policy-makers to create alliances with employers, some employers resisted and refused to cooperate in implementation, going so far as to launch legal challenges to the Employment Equity Act. The Canadian Human Rights Commission, as the main state agency charged with enforcing the employment equity legislation, sought to defend the policy and its own legitimacy in the face of conflicting challenges from equity-seeking groups and employers. In this phase, evidence-based policy broke down entirely and became engulfed in political conflict that led to litigation against the government by both employers and equity groups.

As the new legislation came into effect, the federal government set about to consolidate an employment equity data-reporting regime, while, simultaneously, seeking to limit the role such data would play in employer decisions. The first step in policy implementation was the development of numerical benchmarks. Employment and Immigration Canada established a branch responsible for employment equity, which gathered data from employers on the representation of designated groups. Initially, this data was difficult to obtain: some employers had subsidiaries, which were sometimes counted as separate firms; some companies did not have a clear count of their employees; and the government did not have a list of employers with more than 100 workers (Cardillo 1993). Once this list was established, the participation rate for each particular employer was then compared with Statistics Canada labour force availability data for the four designated equity groups, organized by region and sector. In other words, the Statistics Canada data on availability provided the benchmark for the assessment of equity for particular employers, as required under the Employment Equity Act (Hucker 1997: 566). Employment and Immigration Canada provided the availability data to employers so that employers could assess the representativeness of their workforces and develop action plans (Canadian Human Rights Commission 1991: 34). The participation and availability data was also disseminated to the public so that equity-seeking groups would have access to it, although, in practice, the data was very hard for the equity groups to obtain (Rauhala 1988a: A2; 1988b: A12). The data collected from employers (participation data) and the data provided by Statistics Canada (availability data) were forwarded to the Canadian Human Rights Commission. This was intended to provide the basis on which the commission would initiate complaints against employers (Robertson 1987: 19-20).

The role of the Canadian Human Rights Commission in enforcing employment equity was highly contentious for employers, as they had opposed the addition of this role for the commission. Over the summer of 1988, some employers brought pressure to stop the implementation of the Employment Equity Act through the commission. Employers contacted the commission to voice concerns about implementation and to seek information on exactly how employment equity data would be used by the agency (Canadian Human Rights Advocate 1988a: 1, 12). Indeed, in late 1988, in response to demands for administrative transparency from employers, and without consulting the equity-seeking groups, the Canadian Human Rights Commission released an operational procedures manual for the enforcement of the act.

The manual indicated a crucial shift in policy for the commission. It indicated that the statistical data collected under the employment equity legislation would be used to initiate contact with employers, and not, as it had been previously understood, to initiate complaints (Canadian Human Rights Commission 1988: 4--7). According to the operational procedures manual, the purpose of contact was to inform the employer that its data showed "what appears to be an underutilization of designated groups" (5) and that, although a complaint was not being initiated, the commission "would like to undertake a review with the employer to determine if there are problems which must be corrected" (6). In the review, all of the employment systems would be assessed to see if they posed barriers; in other words, the review would seek to find the causes of the underrepresentation of the designated groups (7). During the review, the commission would encourage employers with problematic numbers to undertake a voluntary equity program, with commission monitoring. The manual was clear that, whether on a voluntary review or on a formal complaint, "the numbers resulting from a comparison of data will not constitute proof that an employer is discriminating but merely indicate where there may be problems" (ii). The manual also stated that the commission would not initiate a complaint if the employer complied voluntarily and if the employer "[was] prepared to correct discriminatory problems discovered and remedy their effect" (iv).

The American consultant to the Canadian Human Rights Commission, former Equal Employment Opportunity Commission official Peter C. Robertson, recommended that the Canadian Human Rights Commission initiate complaints based on the numbers, the procedure followed in the U.S. after the Supreme Court had upheld the validity of "the numbers game" in the Griggs decision in 1972 (Robertson 1987). On this basis, the Equal Employment Opportunity Commission had gone after a number of employers in large firms. The Canadian Human Rights Commission explicitly rejected this advice and publicly stated to the equity groups that the voluntary approach was preferred (Canadian Human Rights Advocate 1988a: 12; see also Charbursky 1992: 347). While the commission moved away from the idea of initiating complaints based on employment equity data alone, it maintained that it would still investigate complaints brought to it by third parties on the basis of employment equity data (Canadian Human Rights Advocate 1988b: 2).

The release of the first set of employment equity data in 1988 was a politicizing event for policy stakeholders. The initial round of reporting showed that there had been little improvement in the representation of designated groups, especially for workers with disabilities and aboriginal people (Hucker 1997: 565). In response, the commission initiated joint reviews with five companies and five federal departments. Although the act did not apply to federal government departments, they were included in order to send a message to private-sector employers and to provide representativeness and fairness (Rauhala 1988d: A1). There was also concerted activity on the part of equity groups in response to the initial release of data. The National Employment Equity Network was formed in 1989 following the publication of the first employment equity reports in order to coordinate a response (Rebick 1989). Two of the equity groups filed complaints with the Canadian Human Rights Commission on the basis of reported data. A group representing disabled Canadians, Disabled People for Employment Equity, initiated action against nine federally regulated employers, including Bell Canada, Canada Post, some of the big banks, the Canadian Broadcasting Corporation (CBC) and Canadian National Railways, in November 1988, as soon as the first data was released, because it showed very low levels of representation of the disabled among federally regulated employers (Canadian Human Rights Advocate 1988c: 1; Rauhala 1988b: A12; 1988d: A1). The Assembly of Manitoba Chiefs lodged complaints against seventeen private-sector employers and thirty-four government departments after receiving data from the employment equity reports and from the Treasury Board showing the underrepresentation of aboriginal people (Canada, Parliament, House of Commons, Special Committee on the Review of the Employment Equity Act, 24 February 1992a: 9:20). In addition to complaints from aboriginal people and the disabled, the Canadian Ethnocultural Council used the data to directly contact sixteen employers where data showed that their representation of visible minorities was far below their representation in the workforce (Armstrong 1988: A18; Canadian Human Rights Advocate 1989: 9).

Employers had various reactions to the reviews. Some employers agreed to cooperate with the Canadian Human Rights Commission and reviewed their employment policies to bring them into line. Others challenged the commission's right to bring complaints based on the numbers, believing that it was encouraging litigation (Benimadhu and Wright 1992). The commission also initiated reviews with the employers who had been pinioned by disability advocates and by the Association of Manitoba Chiefs (Canadian Human Rights Commission 1991: 46--47). As of summer of 1989, Bell and the CBC were facing complaints; Bell refused to allow materials it had given to the commission to be used in a human rights complaint by Disabled People for Employment Equity, while the CBC had simply not responded sufficiently to the inquiry. Both Bell and the CBC challenged the commission in Federal Court, arguing that the commission could not initiate action against them based on numbers alone (Canadian Human Rights Commission 1991: 46--47). Air Canada, Canadian Airlines, CP Rail, CP Express and Greyhound all challenged the commission's already limited capacity to enforce the Employment Equity Act. Even the Treasury Board questioned the commission's role in reviewing and formulating equity plans in government departments (Canadian Human Rights Commission 1993: 68).

The commission was caught in a dilemma. Presupposing the willingness of employers to act on the basis of statistical knowledge, the federal government devised employment equity without an enforcement mechanism. The threat of a complaint under the Canadian Human Rights Act based on employment equity data provided the only redress for the discriminatory practices of recalcitrant employers. Yet, as equity groups had pointed out prior to the passage of the legislation, bringing individual complaints under human rights legislation would not provide an effective means of securing employer compliance with employment equity. The groups asserted that the statistical data pointed to forms of systemic discrimination that could not be equated with individual acts or practices. These systematic cases could not be easily taken up by the commission, which required evidence of specific discriminatory acts. Therefore, the commission encountered great difficulty in its attempts to enforce the compliance of employers with employment equity through the Canadian Human Rights Act, stemming from complaints it initiated or that were initiated by a third party (see Canadian Human Rights Commission 1993: 68).

The commission's compromised position was made further evident in an arrangement struck between it and the CBC in June 1991, allowing for a voluntary review of employment equity in the CBC while the CBC continued its court challenge seeking to strike down the complaint brought against it by Disabled People for Employment Equity. The agreement, established between the commission and the CBC, in the face of objections of Disabled People for Employment Equity, specified that whatever was determined about the state of employment equity in the CBC would remain confidential. This directly contravened the original emphasis of the Employment Equity Act policy on disclosure of employer practices. Rather than shaming employers by publicizing discriminatory conduct, the commission was forced to close the file from public scrutiny. As a representative of Disabled People for Employment Equity stated, "If these large and powerful respondents want to say, 'no, we don't admit we've discriminated' and 'no, we don't want your investigators here,' they are basically able to dictate the terms and conditions on which the investigation will take place, if it takes place at all" (Canada, Parliament, House of Commons, Special Committee on the Review of the Employment Equity Act 1992b: 9:43--44). The sheer length of time it took to proceed with the complaints against employers indicated that the commission was overwhelmed with its mandate. Of the nine complaints filed by Disabled People for Employment Equity in 1988, by 1995 only one had been settled (Canada, Parliament, House of Commons, Standing Committee on Human Rights and the Status of Disabled Persons 1995a: 38:9).

The failure of anti-politics: conflicts over calculation and categorization, 1991-95

By 1991, when a parliamentary committee was established to hold hearings to review the legislation at its five-year mark, even more serious problems with the employment equity framework had emerged. The methods used to categorize designated groups and calculate their prevalence in both the workplace and broader population encountered a series of technical and political challenges. The question of who was or was not a member of a designated group and who would have the last word in making such a determination proved highly controversial. Conflicts between equity-seeking groups and employers over the definition of concepts such as "disability," as well as problems in securing comparable data on workforce availability and representation of the designated groups within particular companies, posed obstacles to the implementation of employment equity policy (for an overview, see Timpson 2001: 161--68). Rather than evidence and data providing the facts that could be used to guide implementation, the data proved to be unstable, difficult to identify, problematic to collect, and politically contested.

The collection of availability data was a complex undertaking, as an assistant chief statistician at Statistics Canada testified to the parliamentary review committee. According to him, of the four designated groups, the only one that was not problematic was "women" (Canada, Parliament, House of Commons, Special Committee on the Review of the Employment Equity Act 1992a: 6:6). For other designated groups, the procedures used to establish availability benchmarks were indirect and problematic. Aboriginal organizations pointed to the artificially low availability benchmark for aboriginal people. This was because about ninety of six hundred bands (approximately 45,000 aboriginal people) did not participate in the 1986 census. In the 1991 census, the level of non-participation decreased but remained substantial, estimated at 25,000 (Canada, Parliament, House of Commons, Special Committee on the Review of the Employment Equity Act 1992a: 6:8).

Statistics Canada's methods for calculating the availability of visible minorities were also problematic. Neither the 1986 nor the 1991 censuses included a direct question related to visible minority status. Instead, the method derived by the Employment and Immigration Canada-Statistics Canada interdepartmental working committee calculated the prevalence of visible minorities from census information recorded on ethnic origin, language, place of birth and religion. This method excluded from the visible minority category anyone who recorded their ethnic origin as Canadian. As with attempts to capture statistical portraits of aboriginal labour force availability, then, visible minorities were underrepresented in availability data. Statistics Canada tested ten different methods for measuring the category of visible minorities, thus showing the many challenges in stabilizing this category (Potvin 2005: 38). In the 1996 survey, Statistics Canada included for the first time a question directly asking about visible minority status (Potvin 2005; see also Lum 1995).

Struggles also occurred around the definition and calculation of designated groups in internal workplace surveys. One of the most pronounced conflicts took place over who would define disability and whether a social or medical model of disability would prevail in definitional practice. The Abella Commission report had advanced the principle of self-identification for all designated groups. Accordingly, in the employment equity legislation, people with disabilities were defined as those having "(i) ... any persistent physical, mental, psychiatric, sensory or learning impairment; (ii) consider themselves to be, or believe that an employer or a potential employer would be likely to consider them to be, disadvantaged in employment by reason of an impairment referred to in (i); and (iii) ... identify themselves to an employer or agree to be identified by an employer, as persons with disabilities" (Employment Equity Act regulations, cited in Cardillo 1993: 52).

Employers vigorously contested this definition. They maintained that the option for voluntary self-identification and the weight given to perception in employer surveys resulted in significant underrepresentation of the number of disabled people who were employed. They pointed to cases where individuals chose not to report a disability after accommodation had been made. Employers sought the ability to record the presence of disabled individuals themselves, a capacity premised on the medical model in which disability is intrinsic to the bodies of the disabled rather than barriers posed by the social and physical environment (Tyler 1991: A13).

Disability advocates argued that, if employers had the ability to define and record disability, the extent of underrepresentation would be underreported. Disabled People for Employment Equity argued that banks were already taking license with the definition of disability. They showed that the banks were counting people with gout or eyeglasses as disabled. The organization initiated a court challenge against the government, demanding that it enforce the legislative definition of disability. The government subsequently issued a directive ordering employers to comply with the definition (McGregor 1991). In turn, the Bank of Nova Scotia and the Toronto Dominion Bank filed a counter challenge seeking to have the definition of disability struck from the legislation (Tyler 1991: A13).

Therefore, the categorizations underlying employment equity were unstable. This instability was cited by employers as evidence of the need to further reduce the role of quantitative measures in implementing the policy. Employers argued that, given the various difficulties faced by the government in collecting availability data and given what they viewed as inaccuracies generated by provisions that allowed only for the self-identification of individuals belonging to designated groups, a re-weighting of employment equity reporting was necessary to place more emphasis on the qualitative and process-oriented measures undertaken by employers. Equity groups viewed this as reflecting employers' reluctance to take action to rectify the patent, obvious and evident underrepresentation of the designated groups.

Recourse to facts, it seemed, could not resolve the conflict. The seemingly simple technique of counting the members of the designated groups who were employed and comparing this to labour force availability proved to be fraught. While the government preferred to render employment equity as a problem to be solved through the use of technique and expertise (i.e., data collection and reporting), the first iteration of federal employment equity policy failed to gel into a stable technical solution that would secure the allegiance of policy stakeholders. The Canadian Human Rights Commission itself was forced to defend its legitimacy as an agency in the face of conflicting court challenges from equity-seeking groups and employers.

Anti-politics renewed: the Employment Equity Act of 1995

The House of Commons' Special Committee on the Review of the Employment Equity Act reviewing the 1986 act recommended a number of changes to improve the legislation in its majority report. It called for employment equity to be extended to the federal public service. Further, the report recommended stepped-up enforcement mechanisms, including sanctions for non-compliant employers, and the implementation of an auditing system to verify employer compliance with the act (Canada, Parliament, House of Commons, Special Committee on the Review of the Employment Equity Act 1992c). In June 1993, in response to the committee's report, employment equity was made mandatory in the public service (Canadian Human Rights Commission 1994: 64). Further measures responding to the report had to await the election of a Liberal government, in October 1993.

In 1995, the Liberal government introduced a new Employment Equity Act as a way out of the impasse that had come to paralyse the policy. To remedy the most serious deficiency of the 1986 act, the new legislation specified a clear enforcement role for the Canadian Human Rights Commission. The commission became responsible for conducting compliance audits with employers, a new mechanism that became the core element of the federal policy. These audits assessed the extent to which employers complied with the statutory requirements. If the audit determined that the employer was lacking in his or her obligations under the act, the employer and the commission negotiated measures to bring the employer into compliance. However, the commission also gained power to issue a direction to the employer outlining mandatory compliance measures. It could take complaints to an Employment Equity Tribunal. An employer could also contest the commission's direction before a tribunal. The orders of Employment Equity Tribunals had the force of an order of the Federal Court of Canada (Canadian Human Rights Commission 2000).

It is important to also note that employer compliance under the 1995 act did not mean equitable representation of designated groups. Like its predecessor, the 1995 act did not mandate equal representation. Compliance under the revised Employment Equity Act referred most directly to the existence of an employer-devised plan to redress workplace underrepresentation. The act required that employers assess their workforce, determine barriers to equity in their organization, and develop employment equity plans that specify short-term numerical goals (and long-term non-numerical goals), as well as the measures and timetable that would be put into place to achieve these goals. The act required that measures undertaken by employers constitute "reasonable progress toward implementing employment equity" (Employment Equity Act, S.C. 1995, c. 44, s. 11).

Two features of the post-1995 employment equity enforcement regime were indicative of an attempt to restore the depoliticized approach to the problem of workplace discrimination that eluded policy-makers under the first iteration of the act. The first was the government's preference, despite its newly expanded authority, for "soft-law" regulatory strategies that contained and minimized overt conflict between employers and the human rights commission. In testimony to the parliamentary committee reviewing the draft legislation, Lloyd Axworthy, minister of human resources development, promoted the new enforcement system on the basis of its "pro-active, non-confrontational manner," and he noted the new emphasis of the act would be to "negotiate mutually acceptable plans with employers" (Canada, Parliament, House of Commons, Standing Committee on Human Rights and the Status of Disabled Persons 1995a: 23:12; for an overall discussion of the 1995 act, see Timpson 2001: 174ff and Hart 2010).

The revised Employment Equity Act's emphasis on negotiation enabled the Canadian Human Rights Commission to manage persistent employer recalcitrance while avoiding policy breakdown. During the first two years of the new enforcement system, the commission completed 111 audits and found that four employers were compliant with the terms of the act (Canadian Human Rights Commission 2000: 65). It then entered into negotiations with 103 employers to move them towards compliance. This soft-law strategy resulted in limited success from the perspective of compliance. Follow-up audits demonstrated that only eight of the 103 employers engaged in negotiated undertakings had achieved compliance (65). Nonetheless, the emphasis on negotiations minimized the commission's use of "hard" enforcement measures. Despite the high level of non-compliance, between 1997 and 2002, the commission issued a mere twenty-three directions and held only eight employment equity tribunals (Canadian Human Rights Commission 2003).

The commission's minimal recourse to "hard" enforcement measures has come to be seen, not as a sign of regulatory failure, but, through the prism of anti-politics, as one of regulatory success. The report of the Standing Committee on Human Resources Development and the Status of Disabled Persons, following its 2002 review of the Employment Equity Act, commented that "members of the Committee are heartened by the fact that the [Canadian Human Rights Commission] has had to issue so few Directions since 1998. We believe that most employers genuinely want to comply with the Act, bur many need help to do so. We also believe that the CHRC may need additional resources to provide this guidance" (Canada, House of Commons, Parliament, Standing Committee on Human Resources Development and the Status of Disabled Persons 2002: n.p.). Such logic reflected policy-makers' attempts to reinvoke a depoliticized depiction of workplace discrimination as an unintentional problem awaiting a technical solution. This conception of workplace inequality underlay the original employment equity legislation but proved untenable amidst deepening political conflict between policy stakeholders and evidence of widespread non-compliance on the part of employers.

The second feature of the new enforcement regime that shut down conflict was a new rule preventing the Canadian Human Rights Commission or third parties from instigating complaints under the Canadian Human Rights Act based on the statistical data generated under the new legislation. This measure was imposed despite complaints from the commission's leadership and from equity groups (Canada, Parliament, House of Commons, Standing Committee on Human Rights and the Status of Disabled Persons 1995a: 25:8, 31:5). The government and employers stressed that the ban was necessary because the threat of commission or third party--initiated complaints would open up conflict and undermine the willingness of businesses to cooperate with the new audit process. The ban would also prevent employers from being liable to employment equity orders issued under both the Employment Equity Act and the Canadian Human Rights Act.

The parliamentary committee reviewing the 1995 bill similarly viewed third-party access to complaint-based mechanisms as inconsistent with the new emphasis on non-adversarial enforcement. In the committee's view, third parties had no legitimate role in the new enforcement system. The committee's majority report on the 1995 bill cast employment equity enforcement as a matter between employers and the government alone: "[e]mployers are responsible for achieving employment equity in the workplaces and the [Canadian Human Rights Commission] would be responsible for ensuring that employers met their obligations" (Canada, Parliament, House of Commons, Standing Committee on Human Rights and the Status of Disabled Persons 1995b: 59). Nonetheless, the committee proposed an amendment mandating that employee representatives, including certified bargaining agents, have input into the development of equity plans and measures. As the committee asserted in its final report, "we see this type of positive and collaborative involvement of third parties as more beneficial than an adversarial route ..." (Canada, Parliament, House of Commons, Standing Committee on Human Rights and the Status of Disabled Persons 1995b: 59). The committee's decision to uphold the ban on third-party complaints while encouraging their participation in employment equity plan development reflects one of the strategies of anti-politics as described by Li (2007), namely, the attempt to delineate and narrow the scope for actors to intervene in policy debate while encouraging their participation. The injunction against third-party complaints took place at a time when the institutional capacity of the Canadian Human Right Commission was eroding. Over the course of the mid- to late 1990s, deficit-fighting was high on the federal policy agenda. The commission saw its resources diminish, undercutting its ability to enforce employment equity (Agocs 2002: 261; on other aspects of the 1995 restructuring, see Timpson 2001: 181ff).

Conclusions

Evidence-based policy mechanisms were central to the adoption of federal employment equity policy and profoundly influenced policy debates during the inception of the policy and its initial implementation from 1986 to 1996. Drawing momentum from affirmative action policy in the United States, equity advocates in Canada pushed for the adoption of statistical reasoning as part of their collective claim for workplace equality. They viewed statistical analysis as a powerful method for bringing the problem of systemic employment discrimination to light and for enforcing proactive workplace remedies. Employers, unsurprisingly, vehemently resisted any policy proposal that would circumscribe their authority to hire, promote and fire. Caught between the conflicting demands of equity advocates and employers, policy-makers' response bore the governmental logic of anti-politics. The policy framework of employment equity promoted by policy-makers was premised on a particular construction of the problem of workplace inequality as neither intentional nor ingrained but, rather, as rooted in a lack of knowledge on the part of employers of the representativeness of their workforces. This depoliticized account of workplace inequality corresponded with the technical, evidence-based solution on offer--that of data-collection and dissemination. By rendering employment inequality as a lack of technical expertise amenable to solution through a data-gathering exercise (as it was characterized by one of its early critics), the federal government appeared to take action to bring fairness to the workplace while respecting employer autonomy.

As we have shown, however, federal employment equity policy failed to gel into a stable technical solution that would defuse conflict between policy stakeholders. The development and implementation of the original Employment Equity Act was characterized by ongoing struggle over the meaning, legitimacy and use of statistical knowledge generated under its provisions. Even the most seemingly technical aspects of measurement and calculation became the basis of intense political controversy and litigation. Only with the introduction of a revised Employment Equity Act by the Liberal government in 1995 could conflict in this policy field be effectively mitigated. The 1995 act managed employer recalcitrance through its "soft-law" emphasis on negotiated workplace remedial measures and by narrowing the bases on which employers may be deemed non-compliant with its terms. The 1995 act further displaced conflict on the part of equity groups by removing their capacity to participate in the enforcement process. The power and visibility of equality advocates under the post-1995 framework is severely weakened.

Finally, our analysis has implications for students of evidence-based policy-making. While several observers emphasize how the evidence-based policy model's promise of more effective policies may not be realized because of lack of capacity among governmental and nongovernmental actors to acquire and effectively use policy knowledge (Laforest and Orsini 2005; Howlett 2009), our analysis of employment equity points towards another dilemma of the evidence-based paradigm. It demonstrates how extensive investment in the development of analytical capacity in a given policy field may do nothing to prevent struggle over the meaning and use of policy knowledge. The adoption and deployment of evidence-based techniques in the policy-making process may provide a means of defusing political conflict and closing off the space for social contestation and citizen engagement, but its success as anti-politics is by no means inevitable. Contestation over the nature of data and evidence may be central to policy-making and to the process by which social and political stakeholders accept and legitimate policy outcomes.

References

Abu-Laban, Yasmeen, and Christina Gabriel. 2002. Selling Diversity: Immigration, Multiculturalism, Employment Equity, and Globalization. Peterborough, Ont.: Broadview Press.

Agocs, Carol. 1986. "Affirmative action Canadian style: A reconnaissance." Canadian Public Policy 12 (1) March: 148-62.

--. 2002. "Canada's employment equity legislation and policy, 1987-2000: The gap between policy and practice." International Journal of Manpower 23 (3): 256-76.

Agocs, Carol, and Catherine Burr. 1996. "Employment equity, affirmative action and managing diversity: Assessing the differences." International Journal of Manpower 17 (4/5): 30-45.

Agocs, Carol, and Harish C. Jain. 2001. Systemic Racism in Employment in Canada: Diagnosing Systemic Racism in Organizational Culture. Ottawa: Canadian Race Relations Foundation.

Armstrong, Jane. 1988. "Minority hiring poor at 'federal' bodies." The Toronto Star, 12 December: A18.

Bakan, Abigail B., and Audrey Kobayashi. 2007. "Affirmative action and employment equity: Policy, ideology, and backlash in Canadian context." Studies in Political Economy 79 Spring: 145-66.

Benimadhu, Prem, and Ruth Wright. 1992. Implementing Employment Equity: A Canadian Experience. Ottawa: Conference Board of Canada.

Black, Bill. 1985. "Systemic discrimination." Canadian Human Rights Advocate February: 7, 15.

Canada Parliament. House of Commons. Legislative Committee on Bill C-62, Employment Equity Act. 1985a. Minutes of Proceedings and Evidence. 33rd Parliament, 1st Session, 4 December. Ottawa: Supply and Services Canada.

--. --. --. --. 1985b. Minutes of Proceedings and Evidence. 33rd Parliament, 1st Session, 5 December. Ottawa: Supply and Services Canada.

--. --. --. --. 1985c. Minutes of Proceedings and Evidence. 33rd Parliament, 1st Session, 12 December. Ottawa: Supply and Services Canada.

--. --. --. --. 1985d. Minutes of Proceedings and Evidence. 33rd Parliament, 1st Session, 17 December. Ottawa: Supply and Services Canada.

--. --. --. Special Committee on the Review of the Employment Equity Act. 1992a. Minutes of Proceedings and Evidence. 34th Parliament, 3rd Session, 12 February. Ottawa: Supply and Services Canada.

--. --. --. --. 1992b. Minutes of Proceedings and Evidence. 34th Parliament, 3rd Session, 24 February. Ottawa: Supply and Services Canada.

--. --. --. --. 1992c. Report: A Matter of Fairness. 34th Parliament, 3rd Session. Ottawa: Supply and Services Canada.

--. --. --. Standing Committee on Human Rights and the Status of Disabled Persons.

--. --. --. --. 1995a. Minutes of Proceedings and Evidence. 35th Parliament, 1st Session, 31 January. Ottawa: Supply and Services Canada.

--. --. --. 1995b. Employment Equity--A Commitment to Merit: Report. 35th Parliament, 1st Session. Ottawa: Supply and Services Canada.

--. --. --. Standing Committee on Human Resources Development and the Status of Disabled Persons. 2002. Promoting Equality in the Federal Jurisdiction: Review of the Employment Equity Act. 37th Parliament, 1 st Session. Ottawa: Public Works and Government Services.

--. Royal Commission on Equality in Employment [the Abella Commission]. 1984. Equity in Employment: Report. Chaired by Justice Rosalie Silberman Abella. Ottawa: Supply and Services Canada.

Canadian Human Rights Commission. 1988. Operational Procedures for Ensuring Compliance with Employment Equity. Ottawa: The Commission.

--. 1991. Annual Report 1990. Ottawa: The Commission.

--. 1993. Annual Report 1992. Ottawa: The Commission.

--. 1994. Annual Report 1993. Ottawa: The Commission.

--. 2000. Annual Report 1999. Ottawa: The Commission.

--. 2003. Annual Report 2002. Ottawa: The Commission.

Canadian Human Rights Advocate. 1986. "CN can't win, MacDonald says," February: 13.

--. 1988a. "Employers gut job equity act," August: 1, 12.

--. 1988b. "Employment Equity Act intact, director writes," September: 2.

--. 1988c. "Job equity 'action' promised by November," September: 16.

--. 1989. "Minority group targets 16 employers," March/April: 9.

Cardillo, Brenda. 1993. "Defining and measuring employment equity." Perspectives on Labour and Income 5 (4) Winter: 43-56.

Charbursky, Lubomyr. 1992. "'The Employment Equity Act: An examination of its development and direction." Ottawa Law Review 24 (2): 305-61.

Clarke, John. 1998. "Thriving on chaos? Managerialism and social welfare." In Post-modernity and the Fragmentation of Welfare, edited by John Clark. London: Routledge.

Geller, Carol. 1985. "A critique of the Abella Report." Canadian Women's Studies 6 (4): 20-22.

Grundy, John, and Smith Miriam. 2007. "Activist knowledge in queer politics." Economy and Society 36 (2) May: 294-317.

Hall, Kathleen. 2005. "Science, globalization, and educational governance: The political rationalities of the new managerialism." Indiana Journal of Global Legal Studies 12 (1) Winter: 153-82.

Hart, Susan Margaret. 2010. "Self-regulation, corporate social responsibility, and the business case: Do they work in achieving workplace equality and safety?" Journal of Business Ethics 92 (4) April: 585-600.

Howlett, Michael. 2009. "Policy analytical capacity and evidence-based policy-making: Lessons from Canada.'" Canadian Public Administration 52 (2) June: 153-75.

Hucker, John. 1997. "Antidiscrimination law in Canada: Human Rights Commissions and the search for equality." Human Rights Quarterly 19 (3) August: 547-71.

Hunter, J. 1993. The Employment Challenge: Federal Employment Policies and Programs 1900-1990. Ottawa: Government of Canada.

Laforest, Rachel, and Michael Orsini. 2005. "Evidence-based engagement in the voluntary sector: Lessons from Canada." Social Policy and Administration 39 (5) October: 481-97.

Leonard, John S. 1990. "The impact of affirmative action regulation and equal employment law on black employment.'" Journal of Economic Perspectives 4 (1) Autumn: 47-63.

Li, Tania Murray. 2007. "Practices of assemblage and community forest management." Economy and Society 36 (2): 263-93.

Lum, Janet. 1995. "The federal Employment Equity Act: Goals and implementation." Canadian Public Administration 38 (1) Fall: 45-76.

McGregor, Carol. 1991. "Perspectives on employment equity: Disabled people for employment equity." Canadian Business Review 18 (2): 30-32.

Mykhalovskiy, Eric, and Lorna Weir. 2004. "The problem of evidence-based medicine: Directions for social science." Social Science and Medicine 59 (5) September: 1059-1069.

Packwood, Angela. 2002. "Evidence-based policy: Rhetoric and reality." Social Policy and Society 1 (3): 267-72.

Parsons, Wayne. 2002. "From muddling through to muddling up: Evidence-based policy-making and the modernisation of British government." Public Policy and Administration 17 (3) July: 43-60.

Peck, Jamie, and Nik Theodore. 2010. "Mobilizing policy: Models, methods, and mutations." Geoforum 41 (2) March: 169-74.

Potvin, Maryse. 2005. "The role of statistics on ethnic origin and 'race' in Canadian antidiscrimination policy." international Social Science Journa1 57 (183) March: 27-42.

Rauhala, Ann. 1988a. "Study finds women remain stuck on lower work rungs." Globe and Mail (Toronto), 22 October: A2.

--. 1988b. "Tougher affirmative-action legislation sought." Globe and Mail (Toronto), 25 October: A12.

--. 1988c. "9 companies told to review hiring habits." Globe and Mail (Toronto), 21 December: A13.

--. 1988d. "9 firms called biased against hiring disabled." Globe and Mail (Toronto), 17 November: A1.

Rebick, Judy. 1989. "National employment equity network." Women's Education 7 (3) Summer: http://www.nald.ca/litweb/other/cclow/newslet/1989/Sep_v7/contents.htm.

Robertson, Peter C. 1987. The Canadian Human Rights Commission as the Enforcement Mechanism under the Employment Equity Act: Recommendations based on the U.S. Experience. Ottawa: Canadian Human Rights Commission.

Saint-Martin, Denis. 2000. Building the New Managerialist State: Consultants and the Politics of Public Sector Reform in Comparative Perspective. Oxford: Oxford University Press.

Simon, Patrick. 2005. "The measurement of racial discrimination: The policy use of statistics." International Social Science Journal 57 (183) March: 9-25.

Solesbury, William. 2001. "The ascendancy of evidence." Planning Theory and Practice 3 (1): 90-96.

Timpson, Annis May. 2001. Driven Apart: Women's Employment Equality and Child Care in Canadian Public Policy. Vancouver: University of British Columbia Press.

Tyler, Tracey. 1991. "Big banks challenge definition of disabled." Globe and Mail (Toronto), 16 August: A13.