What's the problem with gender-based analysis? Gender mainstreaming policy and practice in Canada.
"Gender mainstreaming" is an approach to policy-making; it provides a blueprint for how analysis is to be conducted and on what basis, as well as how decisions are to be made and by whom. As a result, it is a critical but contested concept in feminist politics. It is critical because it attempts to apply a gender lens to the decision-making apparatus of the state. It thus offers the potential to engage in a "politics of process'" that encourages women's input into decision-making (Squires 2007). At the same time, however, it is a contested concept. Scholars debate not only what it means and what it requires of policy analysts but also the ability of gender mainstreaming to transform either the decision-making apparatus of the state or gender relations. While there seems to be general agreement about its transformative potential, exactly how such potential can or will be realized remains a matter of debate. (1)
In Canada, debates about the limits of gender mainstreaming have focused on its peripheral location in the bureaucracy (Burt and Hardman 2001; Grace 1997); the impact of neo-liberalism, which has constrained both the political will and resource allocation for gender analysis (Burt and Hardman 2001; Hankivsky 2009; Teghtsoonian 2000, 2003, 2004, 2005; Teghtsoonian and Chappell 2008); the expert-bureaucratic model employed, which marginalizes women's groups (Rankin and Wilcox 2004); and its privileging of gender as the primary axis of oppression, which obscures and potentially exacerbates differences between women (Hankivsky 2005, 2007). While analyses of these issues have enhanced our understanding of the limited outcomes of gender mainstreaming, they also explore gender mainstreaming as an "instrument," that is, as a tool with which to alleviate gender inequality. While gender mainstreaming is indeed an instrument, it is also a creative or productive force, bringing into being particular social identities and realities. It is this creative dimension of current gender mainstreaming practice in Canada, I suggest, that ultimately limits its transformative potential.
After providing a brief overview of gender mainstreaming in Canada, I apply the approach of problem representation to gender mainstreaming to critically explore the ways in which current policy and practice discursively construct the policy space. I argue that gender mainstreaming constructs a new form of worker: the "gender expert" who is then given authority to analyse, monitor and suggest interventions based on "expert analysis" (see also Bacchi et al. 2005: 63; Carney 2005). Thus, gender analysis becomes a "technology of rule," constructing gender experts whose power ultimately goes unscrutinized in the context of the organization, thereby obscuring the ways in which gender systems are reproduced or fractured by gender mainstreaming itself. I call for a reorientation of gender mainstreaming, away from an approach that considers only the instrumental effects of policy and towards one that considers both instrumental and creative effects of policy.
Gender mainstreaming in Canada
Adopted in 1995 at the Beijing Conference, the concept of gender mainstreaming in the Canadian federal government in many ways came to justify the dismantling of the "women's state," erected during the post-war era by largely Liberal governments (Burt and Hardman 2001; Rankin and Wilcox 2004). (2) Rather than being present only within specialized policy units, gender analysis would take centre stage within policy planning and development within the federal government. Responsibility for gender mainstreaming, "gender-based analysis," as it is known in Canada, was placed within the Status of Women Canada, a horizontal department charged with coordinating the government's gender priorities across all departments and agencies (Doern 1974; Grace 1997). The department is responsible for implementing gender-based analysis, including its development and training, as well as monitoring and enforcement. To date, ten departments and agencies have initiated their own gender-based analysis units. (3)
The Canadian models of gender mainstreaming are "expert-bureaucratic" as opposed to "participatory-democratic." Participatory-democratic models seek input from community groups, while expert-bureaucratic models require gender experts to conduct policy analysis (Beveridge, Nott, and Stephen 2000). In general, expert-bureaucratic models are considered to be more integrationist, attempting to "introduce a gender perspective into existing policy paradigms without questioning them" (Lombardo 2005: 415), while participatory--democratic models are considered to be agenda-setting, which attempt to destabilize policy-making structures (Jahan 1995; Squires 2005: 372). (4) It is important to note, however, that the expert-bureaucratic/ participatory-democratic affinity is perhaps more a matter of degree than of distinctness (Osborne, Bacchi, and McKenzie 2008; Squires 2005). For example, Scotland has adopted an expert-bureaucratic model but relies heavily on community consultation (Beveridge, Nott, and Stephen 2000). Similarly, where impact analysis--a tool of the technocratic state--is formally executed, the degree to which participatory-democratic models are truly agenda-setting is potentially limited. Using an expert-bureaucratic approach, gender-based analysis in Canada relies on bureaucrats to conduct gender-impact analysis--whether or not a proposed policy will result in differential outcomes for men and women--of policy and program proposals. It begins with the assumption, first explicated in Canadian political circles by the Abella Commission, that equality does not necessarily mean sameness (Canada, Royal Commission on Equality in Employment [Abella Commission] 1984; Rankin and Vickers 2001). More specifically, it requires analysts to follow several steps in the analysis of impacts on gender, outlined below.
Despite enthusiasm for gender-based analysis, it has been met with considerable criticism since its adoption. As noted above, for example, the position of Status of Women Canada within the bureaucracy has been flagged as a key concern (Burt and Hardman 2001; Grace 1997). Not only does the department have limited political influence but its limited resources have also meant implementation problems for gender-based analysis. Indeed, the implementation and application of this analytical tool throughout the federal government has been uneven at best, even in those departments with specialized units (Canada, Office of the Auditor General 2009; Canada, Parliament, House of Commons, Standing Committee on the Status of Women 2005). Furthermore, with no compliance mechanism, very few of these departments have actually acted on their results. In response to these concerns, as of December 2007, the federal government requires gender-based analysis to be conducted for all Treasury Board submissions, thus elevating gender analysis to the level of official languages and the environment. The effect of this change remains to be seen.
Similarly, the impact of neoliberalism on the outcomes of gender mainstreaming has been illuminated. For example, Katherine Teghtsoonian demonstrates the ways in which neoliberal discourse has "reduced the institutional spaces within which a gendered understanding of public policy and its impacts might be articulated and made visible" (2003: 28). Similarly, S. Burt and S.L. Hardman posit that gender mainstreaming in Canada has yet to question the ideological contexts of policy proposals, therefore obscuring the ways in which neoliberalism exacerbates gender inequality (2001: 210). Moreover, by obscuring the impact of policy proposals on gender, neoliberalism has limited political will and resource allocation towards alleviating gender inequality (Teghtsoonian 2000, 2003, 2005). These observations reveal the ways in which neoliberal discourse shapes gender analysis, thereby limiting its transformative potential.
In addition to contextual problems, several conceptual limits that direct attention to the degree to which gender-based analysis is, or can be, transformative in Canadian context have been acknowledged. For example, a number of scholars have criticized the expert-bureaucratic model employed by the federal government (Grace 1997; Rankin and Wilcox 2004). Pauline Rankin and Krista Wilcox argue that the model works to the detriment of state--civil society relations by privileging expert voices over women's groups (2004), which in turn have limited access to the state (for a more general discussion of this phenomenon, see Judith Squires 2007). One implication of this criticism is that participatory-democratic approaches would better facilitate strong linkages between state actors and women's groups, offering more avenues in which to voice concerns. (5) In turn, with greater voice in policy issues, women's groups could ensure that gender-based analysis becomes increasingly transformative.
Another key issue with gender-based analysis is its tendency to privilege gender as the primary axis of oppression, neglecting the complexities of social location, understood as the intersections of gender, race, class, ethnicity, sexuality, nation, ability and so on (Bacchi et al. 2005; Bacchi and Eveline 2003; Hankivsky 2005, 2007; Teghtsoonian 2000; for discussions of "intersectionality" and policy studies, see Manuel 2006 and Verloo 2006). Indeed, despite its potential (and explicit commitment) to consider multiple axes of difference (Rankin and Vickers 2001), gender-based analysis has yet to grapple with the complexities of identity, including the ways in which gender mainstreaming itself produces and reproduces gender, thereby reducing its potential to transform social relations. To remedy this, scholars have pointed to the need to integrate more sophisticated analyses of gender and diversity into mainstreaming frameworks. For example, Olena Hankivsky advocates a diversity mainstreaming approach in order to tap into the transformative potential of gender mainstreaming (2005, 2007). Such an approach would require analysts to assess policy impacts along the lines of difference, using intersectional analysis, exploring the multiplicative effects of difference rather than simply of gender.
All of these studies effectively point to the limits of gender mainstreaming in Canada. Focusing on the impacts of gender mainstreaming, these studies expose the ways in which policies intersect existing social relations, with the potential to exacerbate or alleviate inequalities. However, as explained by Carol Bacchi and Joan Eveline, "Policies do not simply 'impact' on people; they 'create' people" (2003: 110), intersecting, creating and contesting existing social relations. We must then consider the ways in which policies bring certain realities into being. That is, we must explore the degree to which gender mainstreaming creates gender identities and issues, rather than simply responding to them. Certainly this is implicit in the issues discussed above. Underlying many of the critiques is the assumption that policies are discursive, constructing particular realities. For example, the literature examining neoliberalism explores the ways in which neoliberal discourse shapes political spaces, illuminating the role of discursive politics in gender analysis. Similarly, for scholars examining the impact of the expert-bureaucratic model, gender mainstreaming constructs and reinforces gender "experts" while reducing the legitimacy of experiential knowledge. Finally, for those exposing the primacy of gender, gender mainstreaming uses a particular construction of women's reality as referent, while neglecting key differences between women, thus reproducing inequalities based on class, race, ability, sexuality and so on.
Inherent in each of these claims is that policy is both instrumental and creative (Bacchi and Eveline 2003; Forester 1993). On the one hand, policies are instrumental, that is, they allow social actors to get things done. However, in "getting things done," they also shape and constrain social interaction, constructing subjects, and opening or closing opportunities for resistance to dominant modes of social organization (see, for example, Bacchi 1999; Bacchi and Eveline 2003; Ball 1993; Forester 1993; Schneider and Ingram 2007; Schram 1993). By implication, policies have a creative or productive force; that is, they also play a role in ensuring the reproduction of the necessary social conditions, norms, values, relations of ruling, etc., that allow for (present and future) action. John Forester writes, "Political organizers must pay attention simultaneously to the goals they wish to achieve and to the ways in which their actions - their speech, gestures, pronouncements, expressions - reconstitute their own identities, shape their reputations, and thus further enable or undercut their future abilities to act" (1993: 9, emphasis in the original). Similarly, public policies are "active texts," serving to position (and shape) identities and social locations of non-state actors (Smith 1993). From this perspective, it is possible that the transformative potential, or lack thereof, of gender mainstreaming resides in the design of the policy itself. It is this dimension of gender mainstreaming that I suggest deserves much more attention.
Gender mainstreaming in Canada: What's the problem?
Since the early 1990s, students of critical policy studies have emphasized the "argumentative turn" of public policy (Fischer and Forester 1993; for an overview of this literature, see Hajer 2002). Within this broad family of approaches, including hermeneutics, critical theory and deconstructionism, it is acknowledged that policies are arguments, discursive constructs that bring into being particular social realities. Among these approaches is the "problem representation" framework, developed by Carol Bacchi (1999, 2000). Problem representation integrates insights from Foucauldian, neo-Marxian and feminist theory to investigate the ways in which power works through public policy, reinforcing or challenging existing modes of social organization. Underlying the problem representation approach are several key assumptions. First, the state is a space through which social forces work, where actors engage in discursive struggle. As a result, state actors and institutions are not neutral, a feature that serves to distinguish this approach from problem definition approaches (see Rochefort and Cobb 1993 for a classic articulation of this approach). Policies are the result of those struggles and thus reflect (and obscure) the power relations that structure public discourse. Second, all discourses are gendered. Policies shape and are shaped by the particular gender order of any polity. Thus, problem representation is inherently a feminist approach to public policy. Third, as noted above, it is assumed that public policies are "active texts" constructing the very problems they seek to solve in ways that obscure and, consequently, reproduce power relations. In other words, policy is discourse.
As a mode of analysis, problem representation begins with the policy text, requiring analysts to consider the ways in which the "problem" is represented by (and in turn shapes) policy discourse and language, what is problematized and what is not, what are the underlying assumptions and presuppositions of the problematization, and who is constructed as a policy constituent and who, by implication, is defined away from the "problem." In addition, it requires analysts to consider the effects of such representations, including the ways in which they restrict or close the policy space and their "lived" or "material effects" on both policy constituents and those who have been defined away from the policy framework. Bacchi describes this approach in the following way:
As a procedure, I suggest starting with the policy and working backwards. Start with what it recommends and see how this reveals what it assumes needs to change - this is what it represents as the problem. This is only the beginning of the exercise, however. We have to interrogate the proposal to see the underlying presuppositions which ground this representation of the problem. We have to uncover what is considered to be unchallengeable and unchangeable. We have to consider what will follow from this representation of the problem. We have to consider who is identified as the target of change and with what effects for that group. We have to ask whose behaviour remains unscrutinised (2000: 18).
Finally, organizational practices are also scrutinized "on the assumption that these are intimately related to the ways of thinking about the issues under construction" (Bacchi and Eveline 2003: 111). Here, the emphasis is on policy as "creator" rather than simply "response." Policy texts represent problems in ways that precondition a particular response, shaping both how we think about and respond to "problems." By implication, policies are argumentative claims about what the problem is and where subjects are positioned therein.
Applying this framework to gender-based analysis in Canada, we see a number of factors that might limit its transformative potential. As noted earlier, gender-based analysis requires an application of a gender lens to all policy proposals in order to reveal the potentially disparate impact between men and women. The training modules in this approach outline a number of stages, including
--preliminary assessment of impacts of an issue on gender equality;
--definition of outcomes, goals, objectives and indicators;
--development and analysis of policy options;
--communication of initiative;
--program/service delivery; and
--evaluation (Canada, Status of Women 2007a: 1-3).
From this, we see that gender-based analysis represents the "problem" as an androcentric bias in policy and decision-making, where policy has historically been assumed to be gender neutral. Thus, unlike targeted equity initiatives, such as Employment Equity or Affirmative Action legislation, which serve to construct "target populations" as "disadvantaged," gender mainstreaming problematizes public decision-making itself, not women or other marginalized groups (Bacchi et al. 2005; Bacchi and Eveline 2003; Verloo 2005). However, the cause of the "problem" is not patriarchal structures or institutions, or even analysts and their frameworks, as we can see with the rigid adherence to a rational analytic framework, discussed in more detail below. Rather, the cause of the "problem" is limited information, as noted in the Australian context by Bacchi and Eveline (2003). This is obvious in the following passage from a training document developed by Status of Women Canada:
The first step for any [gender-based analysis] is to collect information that is sex-disaggregated. The most basic [gender-based analysis] you can do is to count the number of women and men involved in any activity. If there is a significant difference, then you can ask why. The answers that the women and men involved give you will often also provide the basic information you need to figure out the changes that you need to make. You may also want to ask in what ways and at what levels do women and men participate in a particular program. For example, are most of the women working as volunteers while most of the men are being paid for their work? A slightly more in-depth analysis would lead you to the reasons for this situation and thus to ways to change it (2007b: 2).
With "better" information - information that is "sex-disaggregated"--analysts will be better equipped to make informed decisions to minimize differential impacts.
Upon closer scrutiny, several assumptions and presuppositions underlie this representation. First, it is assumed that "better" information leads to better policy. Rather than interrogate the implications of structures, institutions and processes under which gender analysis takes place, it is assumed that gender differences will be ameliorated with more information. This is related to the second assumption that policies are solely instrumental, neglecting the "creative" dimensions of public policy, which is captured by the following statement: "Most actions, policies, programs, projects, or socioeconomic trends do not have the same impact on women and men" (Canada, Status of Women 2007b: 2, emphasis added)." This reflects what Bacchi (1999: 19-22) calls a "problem-solution model," where the emphasis is on finding better solutions, assuming that "solutions" and "problems" exist independently.
Third, it is assumed that "gender" is a binary relation. This is most pronounced in training guides and information documents provided by Status of Women Canada, where gender and sex are used interchangeably and where references to "gender" include "women" and "men." Perhaps one of the most obvious examples is the following passage:
Measuring the differences between women and men gives us the information we need to figure out why these differences exist. We can then work on developing ways to remove or reduce these differences. We have also learned from past experience that:
--You can never assume that women and men benefit equally from any given action, policy, program, or socio-economic trend;
--Most actions, policies, programs, projects, or socio-economic trends do not have the same impact on women and men;
--Women and men often have different priorities and perspectives;
--To achieve gender equality, we need to integrate gender equality measures into all policies and programs from the design stage onwards. Otherwise, our policies and programs may inadvertently increase inequality between women and men and make their socio-economic situations worse (2007b: 2, emphasis added).
Finally, in relation to the previous assumption, it is assumed that expertise within the bureaucracy is the best means through which to achieve change.
The assumptions undergirding this representation of the issue are problematic from the perspective of transformative politics. Since the "problem" is information, where no one is asked to change and where organizational practice remains unscrutinized, gender-based analysis necessarily complements, or coheres with, bureaucratic discourse. Gender-based analysis is a technocratic exercise that requires objective, neutral, and efficient analysis and decision-making. More specifically, it is premised on the comprehensive rationality model of policy-making, which assumes (indeed, imposes) a linear decision-making process, beginning with the "recognition'" of "problems," the assessment of alternatives, and the implementation and evaluation of the "solution." As depicted in Table 1, the parallels between the comprehensive rationality model (adapted from Howlett and Ramesh 2003: 13) and gender-based analysis are striking.
This "decision process," as described by Harold Lasswell (1956, cited in P. deLeon 1999) during the 1950s, has been criticized extensively. Perhaps the most potent criticism from the perspective of this article is that in conflating, or perhaps confusing, the positive and normative dimensions of the model, scholars often overlook the creative dimensions of decision-making and policies, obscuring the ways in which such an approach serves to construct the identity of the analyst and other social actors (Bacchi 1999, 2000; Ball 1993; Schram 1993).
Purposively rational models, such as the comprehensive rationality model on which gender-based analysis is premised, construct the analyst as "expert," privileging expertise at the expense of experience and thus delegitimizing the voice of "non-experts" (Fischer 1990, 2009). As Bacchi et al. observe, "Just as problems are not simply there to be found but are actively assembled through particular assumptions and responses there is no predetermined subject of policy-making but rather people engaged in acts of incomplete becoming.... Hence, in the case of gender analysis, calling for new practices in policy formulation, overcoming the divide between development and implementation, stands to create new types of policy workers" (2005: 63). Gender-based analysis, then, functions as a "technology of rule," serving to construct a new form of policy-worker: the "gender expert." As noted by Teghtsoonian, "technologies of rule" operate within government and exert "a significant constraining influence on the decisions and self-understandings" (2003: 31) of organizations and individuals within them clearly exemplified by gender-based analyses. For example, Status of Women Canada writes, "Including gender expertise in the policy process helps policy makers become more gender-aware and encourages them to incorporate that awareness into their work" (20071o: 1).
As a "technology of rule," gender-based analysis works in conjunction with other "technologies of rule" that have accumulated over rime within government organizations (Teghtsoonian 2003, 2005). As a result, "gender experts" are required to work gender into accepted modes of analysis, such as the comprehensive rationality model outlined above, thus calling into question the degree to which mainstreaming can interrogate already accepted analytical starting points, such as "accountability,.... ," "cost-effectiveness" and so on. Indeed, in the Canadian case, gender-based analysis has been justified in technocratic terms, suggesting that it shares "discursive affinity" with bureaucratic discourse, as reflected in the following statement from Status of Women Canada: "[Gender-based analysis] makes gender equality issues visible in the mainstream of society, taking into account the full diversity of women and men. [It] also contributes to the redress of past and systemic discrimination and contributes to more efficient policies" (2007b: 1). In this example, gender-based analysis is justified on efficiency grounds.
Similarly, in the Treasury Board Secretariat's Management Accountability Framework, gender analysis is identified as one "indicator" of the "quality, adequacy, and soundness of analyses" in Treasury Board submissions, thereby aligning gender-based analysis with accountability (see Canada, Office of the Auditor General 2009: 29). Thus, as observed by Teghtsoonian, the space in which to identify, assess and remedy the gendered dimensions of public policy is narrowed, not only by neoliberal discourses but also by the particular ways in which gender-based analysis as organizational practice intersects these discourses (2003).6 Within this context, the basis on which these new policy-workers--"gender experts"--is determined remains uninterrogated and unchallenged. Thus, the ways in which power works through bureaucracies, policy analytic frameworks, and policies is obscured from view. Bureaucratic norms, values and practices are defined out of sight and are therefore not interrogated for the ways in which these factors might in fact contribute to the current gender order (Bacchi and Eveline 2003; Benschop and Verloo 2006).
Rather than requiring "gender experts" to understand the complexities of social location, gender-based analysis assumes that knowledge is apolitical, something to be acquired through the separation of facts and values, and assumes the analyst can step outside of discursive spaces to conduct objective, neutral and value-free analysis. Neglected are the normative underpinnings of theoretic and analytic frameworks and the ways in which such constructs serve to "organize attention" to some dimensions of an issue, while organizing away from others (Forester 1993; see also Torgerson 1986 for a discussion of the "political" dimensions of this methodology). Thus, in constructing the "gender expert," gender-based analysis obscures from view the ways in which the analyst him--or herself becomes a constructor of policy problems and social identities, which is necessitated by the mode of analysis. Consider, for example, Forester's critique of the comprehensive rationality model, in which he observes that rational analysis requires the following:
--a well-defined problem;
--a full array of alternatives for consideration;
--full baseline information;
--full information about the consequences of each alternative;
--full information about the values and preferences of citizens; and
--fully adequate time, skill and resources (1984: 23).
At issue, however, is that not only are problems very rarely well defined, or objective for that matter, but also that information and resources are always limited. The result is that the comprehensive rationality model reduces "problems" to what is manageable and therefore disarticulates them from the broader socio-political context. In turn, the power of the analyst in constructing those problems is rendered invisible and remains unchallenged.
Key here are the ways in which discursive politics are obscured by the mode of analysis. Quite simply, gender-based analysis assumes that "problems" are "out there" awaiting discovery through objective analysis of "the facts." However, as students of critical and postmodernist policy perspectives have convincingly argued, problems do not appear out of nowhere; they are given meaning through discursive politics (for examples, see Bacchi 1999; Danziger 1995; Schram 1993). Sanford Schram writes of modernist policy analytic models, "[S]uch a perspective obscures the extent to which it [e.g., the comprehensive rationality model] itself fails to address how the discursive practices involved in the analysis and making of public policy themselves contribute to the reality being addressed" (1993: 253).
Instead, the model assumes away "systematic distortions," which shape the analyst's attention in key ways (Forester 1980, 1982, 1984, 1993). Systematic or structural, distortions, in Habermasian terms, can be defined as the communicative blockages resulting from the particular constellation of power and authority within social organizations. That is, systematic distortions "organize attention" towards some issues while rendering others illegitimate or invisible. To consider a classic example, framing only economically valued activity as "productive" renders invisible the value of unpaid labour, labour that is predominantly performed by women (see Waring 1999 for a discussion of this classic example). Policies then that fail to acknowledge the value of unpaid labour and the sexual division of labour reproduce inequalities based on gender. Moreover, policies that take as given the role of capitalist systems in assigning value to activities reproduce inequalities based on class, pointing to important intersections of class and gender.
Systematic distortions are therefore historically and socially contingent and are both structural and discursive. They are often "referentially transparent" in that we do not often acknowledge them (Hutchins 1980, cited in Reskin and Hartmann 1986; see also Torgerson 1996: 275). It is this "taken for grantedness," however, that makes them so dangerous (Fischer 1990). For example, in a study conducted by Emanuela Lombardo and Petra Meier, it was discovered that gendered policy analysis does not necessarily imply feminist policy analysis, highlighting the myriad ways in which analysts interpret and apply analytical frameworks to policy "problems" (2006). Similarly, as observed by Yvonne Benschop and Mieke Verloo in their analysis of a gender mainstreaming project in Belgium, the "genderedness" of organizations often makes distortions difficult to remedy, even where analysts acknowledge them (2006). Without attending to the ways in which social relations are patterned, which inevitably filter into the analyst's decision and, indeed, the context in which those decisions are made, there is the danger that such relations will simply get reproduced by technocratic activity. Gender-based analysis as currently practiced, however, makes it impossible to detect and remedy systematic distortions.
These issues limit the degree to which gender-based analysis can transform social relations. Without an understanding of the ways in which problems come to be and how they are given particular shape by discursive activity, "solutions" are inevitably limited (Bacchi 1999). For example, as noted by Olena Hankivsky in the Canadian context (2009) and Bacchi and Eveline in the Australian context (2003), with gender-based analysis, "gender" is stripped of its socio-political content and instead reduced to (and reproduced as) a physical, individual attribute. In effect, "gender" becomes "sex." This not only reinforces the notion of a gender binary, a claim that has been convincingly problematized by post-structuralist feminists, but also assumes that women share a common experience on the basis of their sex/gender (see, most notably, Judith Butler 1990). This claim has also been problematized by difference feminists (see, for example, Crenshaw 1991; Hooks 2000; and in the context of mainstreaming, see Bacchi and Eveline 2003; Bacchi et al. 2005; Hankivsky 2005, 2007) and neglects the ways in which "gender" is mediated by social location. Moreover, in adopting ex poste analysis, gender-based analysis takes differences based on sex as given rather than interrogating how such differences are given meaning through discursive politics (Bacchi and Eveline 2003). Failing to challenge the gender dichotomy, this methodology not only obscures and potentially exacerbates differences between women (as well as differences between men) but also renders trans-groups invisible. Policies resulting from this type of analysis reproduce rather than resist the gender order and its intersection with other axes of difference.
In sum, gender-based analysis constructs the "problem" of gender inequality as androcentric bias in decision-making as a result of limited "information." The solution, then, is to gather better (read: sex-disaggregated) information on which to analyse policy proposals and make decisions. In so doing, however, gender-based analysis operates as a technology of rule, constructing the "gender expert," whose power and position within the unchallenged bureaucracy remains unscrutinized. Using a technocratic mode of analysis, which assumes that "problems" and "solutions" exist independently, the power to represent those "problems" and to propose "solutions" resides with the "gender expert" and is obscured from view. We therefore have no sense of how systematic distortions filter into the analysis, reinforcing rather than resisting axes of oppression. Thus, gender-based analysis becomes a mode of "othering" (see also Bacchi et al. 2005; Carney 2005).
Transforming gender-based analysis? Concluding remarks
Assessing the transformative potential of mainstreaming would require us to explore, on the one hand, the degree to which analysts conform to or challenge organizational practice and, on the other hand, the ways in which patterns of social relations get reproduced or challenged by the analyst. In the case of the former, we need to critically explore analytical frameworks for policy analysis and the degree of freedom they offer analysts in transforming organizations. In the case of the latter, we need to critically examine the ways in which the outcomes of analysis and policies intersect existing social relations, reinforcing or offering modes of resistance to power relations. Both of these dimensions are of crucial importance to gender mainstreaming, since they call into question the transformative potential of such initiatives for both decision-making processes and social relations more generally.
Although feminist "models" of policy-making are scarce, a number of approaches have emerged during the past decade in an attempt to scrutinize policies from the perspective of both "difference" and "transformation" (see, for examples, Marshall 1997, 1998, 1999; McPhail 2003). Often these approaches are "analytical frameworks" that present open lists of questions analysts should ask to identify disparate impacts between men and women and between groups of women. These are important contributions to feminist policy studies, pushing the boundaries of analysis by advocating explicitly gendered analyses and challenging the supposed gender "neutrality" of traditional policy analytic frameworks. However, again, emphasis is on the "impact" of policies rather than on its "creative" dimensions.
What is required is an approach that enables analysts to focus on both instrumental and creative dimensions of policy. Bacchi's work on feminist policy studies led her to propose a "dual-focus agenda" to "identify both the ways in which interpretive and conceptual schemas delimit understandings, and the politics involved in the intentional deployment of concepts and categories to achieve political goals" (2005: 207, 2009). Such an approach would require analysts to critically evaluate the assumptions and presuppositions grounding "feminist" policies, locating initiatives, including proposed reforms, in both discursive and material contexts, a concept Bacchi (2005: 208, 2009) refers to as "reflexive framing."
Reflexive framing requires the analyst to turn inward, to reflect on the concepts guiding analysis and suggested outcomes. Such an approach would entail a dynamic "policy design" evaluation of policy proposals, interrogating and scrutinizing the underlying logic, assumptions and presumptions informing policies prior to implementation (see Bustelo and Verloo 2009 for a discussion). Through this process, reflexive framing would attempt to detect--and remedy--systematic distortions, offering potential to dissect and challenge broader discourses, such as neoliberalism, as discussed by Teghtsoonian (2003, 2005, 2007). In so doing, it brings into focus the ways in which policies marginalize and/or silence some groups over others. In this sense, reflexive framing would address intersectionality and diversity and attempt to understand how "difference" comes to take on political and social meaning and gets reproduced by policy texts and discourses. By interrogating and scrutinizing "difference" rather than accepting it as given, reflexive framing addresses the concern with diversity identified by Hankivsky (2005, 2007) but avoids the "reification problem" identified by Squires (2005, 2007). Moreover, since it acknowledges that the analyst is only part of the analysis, central to understanding policies is dialogue with various groups to explore the various understandings and potential "lived effects" of proposals, thus attending to the concern with the expertise and participation described by Rankin and Wilcox (2004). It is here also where reflexive framing points to a key role of women's policy units in facilitating public dialogue (see also Bacchi et al. 2005 and Squires 2007).
In short, reflexive framing would require an analytical reorientation for gender-based analysis away from an instrumental model towards an approach that enables the analyst to identify both instrumental and creative dimensions of policies. I suggest that "critical frame analysis" is useful in this respect. Critical frame analysis, identified and developed by Mieke Verloo and Emanuela Lombardo (2007), refers to an approach that seeks to understand policies as discursive constructs, interrogating the ways in which frames both intentionally and unintentionally open or close discursive space (see also Lombardo, Meier, and Verloo 2009; Roggeband and Verloo 2007). In this context, critical frame analysis is more than simple "frame analysis," which often overlooks the ways in which power works through frames; rather, it explores the ways in which frames are grounded in discourses and identifies their constitutive effects, including the scope for empowerment or marginalization. The purpose of this methodology is pragmatic, that is, to detect inconsistencies in policy texts and to facilitate public debate (Verloo and Lombardo 2007: 30).
Underlying critical frame analysis is the "problem representation approach," the analytical framework deployed in this article, which has served as the basis of what Bacchi and Eveline (2003) have entitled "deep evaluation," developed for a pilot study for gender mainstreaming in Australia (see also Hankivsky 2009 for a discussion of this approach). The problem representation approach, and, indeed, critical frame analysis more generally, focuses as much on what is not present in the text as what is present, and it emphasizes not just policy frames but also hidden inscriptions and assumptions that limit or open space for change. It thus "organizes attention" to the ways in which our attention has been constructed, exposing the hidden biases, presumptions and assumptions, and illuminates the constructions and "lived effects" generated by policy texts. From this perspective, problem representation provides analysts with a tool with which to assess who's in, who's out, and to what effect, as well as to expose the systematic distortions shaping analysis. In short, the approach embraces reflexive framing as a normative standard for policy analysis, seeking better understanding of the discursive contexts in which policies are proposed, implemented and evaluated. It thus offers the potential for the transformation of social relations initially promised by gender mainstreaming.
In closing, gender mainstreaming is an important concept in feminist politics. Attempting to integrate a gender lens to all policies and programs is one way to ensure that women's voices are addressed within the state. At the same time, however, current practice represents the "problem" of gender inequality as information, thereby constructing "gender experts" whose power to construct policy "problems" and social identities remains unchecked. As a result, bureaucratic discourses shaping norms, values and organizational practice get reproduced rather than resisted. In effect, gender mainstreaming becomes a technology of rule. From this perspective, gender and other axes of "difference" are taken as given rather than questioned. Thus, gender mainstreaming as currently practiced in Canada cannot provide us with insight about the ways in which "difference" comes to be and how or why this is politically important. The danger, of course, is that inequalities are reproduced rather than challenged. It is only through analysis of the creative dimensions of policy that gender mainstreaming will transform social relations.
Bacchi, Carol Lee. 2001. "Gender mainstreaming: A new vision, more of the same, or backlash?" Dialogue: Academy of the Social Sciences in Australia 20 (2): 16-20.
--. 2000. "Policy as discourse: What does it mean? Where does it get us?" Discourse 21 (1) Winter: 45-57.
--. 1999. Women, Policy and Politics: The Construction of Policy Problems. London: Sage Publications.
Bacchi, Carol Lee, and Joan Eveline. 2003. "Mainstreaming and neoliberalism: A contested relationship." Policy & Society 22 (2): 98-118.
Bacchi, Carol Lee, Joan Eveline, Jennifer Binns, Catherine Mackenzie, and Susan Harwood. 2005. "Gender analysis and social change: Testing the water." Policy & Society 24 (4): 45-68.
Ball, S. 1993. "What is policy? Texts, trajectories and toolboxes." Discourse 13 (2) Spring: 10-17.
Benschop, Yvonne, and Mieke Verloo. 2006. "Sisyphus' sister: Can gender mainstreaming escape the genderedness of organizations?" Journal of Gender Studies 15 (1) March: 19-33.
Beveridge, Fiona, Sue Nott, and Kylie Stephen. 2000. "Mainstreaming and the engendering of policy-making: A means to an end?" Journal of European Public Policy 7 (3): 385-405.
Burt, S., and S.L. Hardman. 2001. "The case of the disappearing targets: The Liberals and gender equality." In How Ottawa Spends, 2001-2002: Power in Transition, edited by Leslie A. Pal. Don Mills, Ont.: Oxford University Press Canada
Bustelo, M., and Mieke Verloo. 2009. "Grounding policy evaluation in a discursive understanding of politics." In The Discursive Politics of Gender Equality: Stretching, Bending and Policymaking, edited by Emanuela Lombardo, Petra Meier, and Mieke Verloo. London: Routledge.
Butler, Judith. 1990. Gender Trouble: Feminism and the Subversion of Identity. London: Routledge. Canada. Office of the Auditor General. 2009. Report (Spring). Ottawa: Public Works and Government Services Canada
--. Parliament. House of Commons. Standing Committee on the Status of Women. 2005. Gender-Based Analysis: Building Blocks for Success. Ottawa: Public Works and Government Services Canada
--. Royal Commission on Equality in Employment (Abella Commission). 1984-85. Report. Ottawa: Supply and Services Canada Chaired by Rosalie Silberman Abella.
--. Status of Women. 2007a. An Integrated Approach to Gender-Based Analysis--Gender-based Analysis (GBA) Training. Ottawa: Public Works and Government Services Canada. Available at http://www.swc-cfc.gc.ca/pol/gba-acs/guide/2007/train-forma-eng.pdf.
--. --. 2007b. An Integrated Approach to Gender-Based Analysis--What is GBA? Ottawa: Public Works and Government Services Canada Available at http://www.swc-cfc.gc.ca/pol/gba-acs/guide/2007/gba-acs-eng.pdf.
Carney, G. 2005. "Data, anecdote and metaphor in gender equality policy-making: Merging 'intellectual and real world mainstreaming.'" Paper presented at the ECPR Workshop: Metaphor in Political Science, 14-19 April, Granada
Crenshaw, Kimberle. 1991. "Mapping the margins: Intersectionality, identity politics, and violence against women of colour." Stanford Law Review 43 (6) July: 1241-299
Danziger, Maire. 1995. "Policy analysis postmodernized: Some political and pedagogical ramifications." Policy Studies Journal 23 (3) Fall: 435-50.
deLeon, P. 1999. "The stages approach to the policy process: What has it done? Where is it going?" In Theories of the Policy Process, edited by Paul A. Sabatier. Boulder, Colo.: Westview Press.
Doern, G. Bruce. 1974. "Horizontal and vertical portfolios in government." In Issues in Canadian Public Policy, edited by G. Bruce Doern and Seymour Wilson. Toronto: Macmillan.
Donaghy, Tahnya Barnett. 2004. "Applications of mainstreaming in Australia and Northern Ireland." International Political Science Review 25 (4) October: 393-410.
Fischer, Frank. 2009. Democracy and Expertise: Reorienting Policy Inquiry. New York: Oxford University Press.
--. 1990. Technocracy and the Politics of Expertise. New York: Sage Publications.
Fischer, Frank, and John Forester (eds.). 1993. The Argumentative Turn in Policy Analysis and Planning. Durham, N.C.: Duke University Press.
Forester, John. 1984. "Bounded rationality and the politics of muddling through." Public Administration Review 44 (1) January/February: 23-31.
--. 1980. "Critical theory and planning practice." Journal of the American Planning Association 46 (3) July: 275-86.
--. 1993. Critical Theory, Public Policy, and Planning Practice: Toward a Critical Pragmatism. New York: State University of New York Press.
--. 1982. "Planning in the face of power." Journal of the American Planning Association 48 (1) Winter: 67-80
Grace, J. 1997. "Sending mixed messages: Gender-based analysis and the 'status of women'." Canadian Public Administration 40 (4) Winter: 582-98.
Hajer, Maarten. 2002. "Discourse analysis and the study of policy making." European Political Science 2 (1): 61-65.
--. 1993. "Discourse coalitions and the institutionalization of practice: The case of acid rain in Great Britain." In The Argumentative Turn in Policy Analysis and Planning, edited by Frank Fischer and John Forester. Durham, N.C.: Duke University Press.
Hankivsky, Olena. 2007. "Gender mainstreaming in the Canadian context: One step forward and two steps back." In Critical Policy Studies, edited by Michael Orsini and Miriam Smith. Vancouver: University of British Columbia Press.
--. 2009. "Gender mainstreaming in neoliberal times: The potential of 'deep evaluation'." In Public Policy for Women: The State, Income Security, and Labour Market Issues, edited by Marjorie Griffen Cohen and Jane Pulkingham. Toronto: University of Toronto Press.
--. 2005. "Gender vs. diversity mainstreaming: A preliminary examination of the role and transformative potential of feminist theory." Canadian Journal of Political Science 38 (4) December: 977-1001.
Hooks, Bell. 2000. Feminist Theory: From Margin to Center, 2nd edition. Cambridge, Mass.: South End Press.
Howlett, Michael, and H. Ramesh. 2003. Studying Public Policy: Policy Cycles and Policy Subsystems, 2nd edition. Don Mills, Ont.: Oxford University Press Canada.
Jahan, Rounaq. 1995. The Elusive Agenda. Mainstreaming Women in Development. Atlantic Highlands, N.J.: Zed Books.
Lombardo, Emamuela. 2005. "Integrating or setting the agenda: Gender mainstreaming in the European constitution-making process." Social Politics 12 (3) Fall: 412-32.
Lombardo, Emanuela, and Petra Meier. 2006. "Gender mainstreaming in the EU: Incorporating a feminist reading?" European Journal of Women's Studies 13 (2) May: 151-66.
Lombardo, Emanuela, Petra Meier, and Mieke Verloo (eds.). 2009. The Discursive Politics of Gender Equality Stretching, Bending and Policymaking. London: Routledge.
Manuel, Tiffany. 2006. "'Envisioning the possibilities for a good life: Exploring the public policy implications of intersectionality theory." Women, Politics and Policy 28 (3): 173-203.
Marshall, Catherine. 1997. "Dismantling and reconstructing policy analysis." In Feminist Critical Policy Analysis: A Perspective from Secondary Schooling, edited by C. Marshall. London: Falmer Press.
--. 1998. "Feminist critical policy analysis: Toward demanding and disrupting policy analyses." Paper presented at the annual meeting of the American Educational Research Association, 13-17 April, San Diego, California.
--. 1999. "Researching the margins: Feminist critical policy analysis." Educational Policy 13 (1/2) January-March: 59-76.
Mazur, Amy. 2002. Theorizing Feminist Policy. Oxford: Oxford University Press.
McPhail, Beverly. 2003. "A feminist policy analysis framework: Through a gendered lens." Social Policy Journal 2 (2/3): 39-61.
Osborne, Katy, Carol Bacchi, and C. Catherine McKenzie. 2008. "Gender analysis and community consultation: The role of women's policy units." Australian Journal of Public Administration 67 (2) June: 140-60.
Rankin, Pauline, and Jill Vickers. 2001. Women's Movements and State Feminism: Integrating Diversity into Public Policy. Written for Status of Women Canada. Ottawa: Public Works and Government Services Canada.
Rankin, Pauline L., and Krista D. Wilcox. 2004. "De-gendering engagements? Gender mainstreaming, women's movements and the Canadian federal state." Atlantis 29 (1) Fall: 52-60.
Rees, T. 1999. "Mainstreaming equality." In Engendering Social Policy, edited by Sophie Watson and Leslie Doyal. Buckingham, U.K.: Open University Press.
Reskin, Barbara, and Heidi Hartmann. 1986. Women's Work, Men's Work: Sex Segregation on the Job. Washington, D.C.: National Academy Press.
Rochefort, David A., and Roger W. Cobb. 1993. "Problem definition, agenda access, and policy choice." Policy Studies Journal 21 (1) Spring: 56-71.
Roggeband, Conny, and Mieke Verloo. 2007. "Dutch women are liberated; Migrant women are a problem." Social Policy and Administration 41 (3): 271-88.
Schneider, A., and H. Ingram. 2007. "Social construction and policy design." In Theories of the Policy Process, 2nd edition, edited by Paul A. Sabatier. Cambridge, Mass.: Westview Press.
Schram, Sanford. 1993. "Postmodern policy analysis: Discourse and identity in welfare policy." Policy Sciences 26 (3): 249-70.
Smith, Dorothy E.. 1993. Texts, Facts, and Femininity: Exploring the Relations of Ruling. New York: Routledge.
Spalter-Roth, R., and R. Schreiber. 1995. "Outsider issues and insider tactics: Strategic tensions in the women's policy network during the 1980's." In Feminist Organizations: Harvest of the New Women's Movement, edited by M. Marx Ferree and P. Yancy Martin. Philadelphia: Temple University Press.
Squires, Judith. 2005. "Is mainstreaming transformative? Theorizing mainstreaming in the context of diversity and deliberation." Social Politics 12 (3) Fall: 366-88.
Stratigaki, Maria. 2004. "The cooptation of gender concepts in EU policies: The case of 'reconciliation of work and family'." Social Politics 11 (1) Spring: 30-56.
--. 2007. The New Politics of Gender Equality. Houndmills, U.K.: Palgrave.
Teghtsoonian, Katherine. 2005. "Disparate fates in challenging times: Women's policy agencies and neoliberalism in Aoteara/New Zealand and British Columbia." Canadian Journal of Political Science 38 (2) June: 307-33.
--. 2000. "Gendering policy analysis in the Government of British Columbia: Strategies, possibilities and constraints." Studies in Political Economy 61 Spring: 105-27.
--. 2004. "Neoliberalism and gender analysis mainstreaming in Aotearoa/New Zealand." Australian Journal of Political Science 39 (2) July: 267-84.
--. 2003. "W(h)ither women's equality? Neoliberalism, institutional change and public policy in British Columbia." Women, Politics and Policy 22 (1): 26-47.
Teghtsoonian, Katherine, and Louise Chappell. 2008. "The rise and decline of women's policy machinery in British Columbia and New South Wales: A cautionary tale." International Political Science Review 29 (1) January: 29-51.
Torgerson, Douglas. 1986. "Between knowledge and politics: Three faces of policy analysis." Policy Sciences 19 (1) February: 33-59.
--. 1980. Industrialization and Assessment: Social Impact Assessment as Social Phenomenon. North York: President's Advisory Committee on Northern Studies, York University.
--. 1996. "Power and insight in policy discourse: Post-positivism and problem definition." In Policy Studies in Canada: The State of the Art, edited by Laurent Dobuzinskis, Michael Howlett, and David Laycock. Toronto: University of Toronto Press.
True, Jacqui. 2003. "Mainstreaming gender in global public policy." International Feminist Journal of Politics 5 (3): 368-96.
True, Jacqui, and Michael Mintrom. 2001. "Transnational networks and policy diffusion: The case of gender mainstreaming." International Studies Quarterly 45 (1) March: 27-57.
Verloo, Mieke. 2001. "Another Velvet Revolution? Gender mainstreaming and the politics of implementation.'" Vienna: IWM Working Paper No. 5/2001. Available at http://www.univie.ac.at/iwm/p-iwmwp.htm.
--. 2005. "Displacement and empowerment: Reflections on the concept and practice of the council of Europe approach to gender mainstreaming and gender equity." Social Politics 12 (3) Fall: 344--65.
--. 2006. "Multiple inequalities, intersectionality and the European Union." European Journal of Women's Studies 13 (3): 211-28.
-- (ed.). 2007. Multiple Meanings of Gender Equality. A Critical Frame Analysis of Gender Policies in Europe. Budapest: CEU Press.
Verloo, Mieke, and Emanuela Lombardo. 2007. "Contested gender equality and policy variety in Europe: Introducing a critical frame analysis approach." In Multiple Meanings of Gender Equality: A Critical Frame Analysis on Gender Policies in Europe, edited by Mieke Verloo. Budapest: Central European University.
Waring, Marilyn. 1999. Counting for Nothing: What Men Value and What Women are Worth. Toronto: University of Toronto Press.
Woodward, A. 2001. "Gender mainstreaming in European policy: Innovation or deception?" Manuscript based on a lecture given at the Wissenschaftszentrum Berlin fur Sozialforschung in June 2001.
--. 1999. "Women acting rationally: Speaking the language of power to test the state." In Rationalisation, Organisation, Gender, edited by M Goldman. Proceedings of the International Conference, October 1999, Dortmund.
(1) The literature is quite extensive. For examples of those who view gender mainstreaming as offering potential to transform organizational life, see T. Rees (1999), Jacqui True (2003), Jacqui True and Michael Mintrom (2001). In contrast, for examples of those who are more critical, see Caroll Bacchi (2001), Carol Bacchi et al. (2005), Carol Bacchi and Joan Eveline (2003), Mieke Verloo (2005, 2001), and A. Woodward (1999, 2001). In addition, see the special issues of Social Politics 12 (3), the International Feminist Journal of Politics 5 (3), and Gender and Development 13 (2), all of which contain articles exploring practical and theoretical dimensions of mainstreaming.
(2) Note that this has been flagged as problematic. Olena Hankivsky (2007) and Katherine Teghtsoonian (2000, 2003, 2005) in Canada, as well as a host of international literature (for an example, see Judith Squires 2007), point to the importance of the relationship between women's policy units, strong women's movements, and elected officials for achieving gender equality. This is also known as the "strategic partnership" or the "triangle of empowerment." See Amy Mazur (2002) for an overview of this literature.
(3) These include Canadian International Development Agency, the departments of Agriculture and Agri-Food, Citizenship and Immigration, Foreign Affairs and International Trade, Health, Human Resources and Skills Development, Indian and Northern Affairs, Justice, the Office of the Solicitor General, and Statistics Canada. It should also be noted that the Department of Finance has recently committed itself to conducting gender analysis.
(4) Judith Squires adds "transformational" approaches (2005) to the typology developed by Rounaq Jahan (1995), which includes integrationist and agenda-setting approaches.
(5) It should be noted, however, that evidence suggests there is little difference in terms of results between expert-bureaucratic and participatory-democratic models. See Tahnya Donaghy (2004).
(6) This speaks to the organizational and normative constraints posed by bureaucratic practice, for which there is rich debate among students of gender mainstreaming. For example, a number of scholars warn of the technocratic nature of mainstreaming, leading to a paradox whereby gender mainstreaming advocates must use the master's tools to dismantle the master's house. See Yvonne Benschop and Mieke Verloo (2006), Mieke Verloo (2001, 2005), and A. Woodward (1999, 2001). More specifically, supporters of gender mainstreaming must use rationalist, often gender neutral, rhetoric to advance their goals. However, in doing so, they risk becoming entrapped by that rhetoric, using tools that prohibit the transformative potential of the mainstreaming project. In becoming entrapped by standard administrative practice, feminists risk losing their outsider perspective or opening themselves up to cooptation. See also R. Spalter-Roth and R. Schreiber (1995), Judith Squires (2007), and Maria Stratigaki (2004) for relatively recent articulations of this argument.
The author is assistant professor, Department of Political Science, Concordia University. An earlier version of this article was presented at the annual meeting of the Canadian Political Science Association, Vancouver, June 2008. The author wishes to thank Cherry Marshall, for invaluable research assistance, as well as the editors of the journal and the three anonymous reviewers for their comments and suggestions.
Table 1. Gender-Based Analysis as Rational Policy Analysis Comprehensive rationality model Gender-based analysis Problem recognition Preliminary assessment of impacts of an issue on gender equality Definition of outcomes, goals, objectives and indicators Proposal of solution Research Consultation Development and analysis of policy options Consideration of solution Making recommendations Communication of initiative Putting solution into effect Program/service design Program/service delivery Monitoring results Evaluation