‘A More Receptive Crowd than Before’: Explaining the World Bank’s Gender Turn in the 2000s

I Introduction

The comment was made by Danny Leipziger, one-time vice president for the Poverty Reduction and Economic Management network within the World Bank, in conversation in 2015. (The conversations took place during a two week visit to the World Bank in Washington DC in November 2014, via phone interviews in 2015 and with staff in a country office in July 2015.) He was reflecting back on the Gender Action Plan that ran from 2006 to 2010. The Gender Action Plan used ‘serious economists’ and worked in areas—land, labour, credit and agriculture—that had not been a focus of gender work previously. Leipziger emphasized that the Gender Action Plan came at a time when gender work was more legitimate and acceptable within the institution. The Plan was contrasted to the sort of mainstreaming work that had characterized gender work in the past: gender audits, gender scorecards and project oversight. Credibility was particularly focused on the sorts of analysis that was funded and subsequently published. I was told by another senior staff member that the papers that came out of the Gender Action Plan ‘gave credibility to gender as a research subject’ (Interview with staff member 11 December 2014, Washington DC).

What seemed to make the Gender Action Plan feel different and achieve a different sort of significance within the institution was the way it had a degree of intellectual currency. I was told by a senior economist that: ‘gender is not of immediate importance to economists but when it gets at something intuitive about how households work or other underlying causes of underdevelopment it becomes interesting’ (Interview with staff member 9 December 2014, Washington DC). The Plan came about at a time when there were developments in the field of economic and project analysis—impact evaluations, randomized control trials and household data disaggregation—which made gender work more credible and, at a more basic level, more doable for micro-economists working within the institution. The idea of gender itself had become part of orthodox micro-economic work on the household (e.g., Bardhan and Udry, 1999; Basu 2006; Duflo, 2003; Udry 1996). Randomized control trials had also shown the causal relationship between women’s roles and development outcomes (e.g., Beaman et al. 2009; Chattopadhyay and Duflo 2004). Part of the success of the Gender Action Plan was the way it funded a number of research projects and fed into publications in economics journals (e.g., Ali et al., 2014; Bandiera et al., 2014; Deininger et al., 2008). While this was not the only reason for its success of the Plan, the word ‘gender’ had begun to align itself with a set of techniques and technologies that were developing in the field of development economics.

The 2000s was also a time when the World Bank, like other organizations, had to make more sense of how its work looked to an outside audience. As Esser and Strömbäck (2014) mention the increasing role of media reflects a deep structural shift in the way large-scale organizations make sense of themselves. Alongside pressure from civil society groups and NGOs, there was the growth of the Internet, which meant the virtual publication of reports, the use of online videos and the possibility of reaching a much wider audience. The World Bank, like other development organizations, operated in a changing landscape. It was a time when development work was more likely to gather support if it worked in media terms, and projects such as Voices of the Poor gained support from the Office of the President in part because they worked in presentational terms. One way of conceptualizing this is in terms of a broad logic of mediatization where topics and themes take on new meanings because of how they work in media terms (Couldry, 2008; Mazzoleni and Schulz 1999). Women in development, which brings forward a set of images and associations, has a particular ‘media logic’ across the development sector (Altheide and Snow, 1979). The idea of women as responsible producers, or as the handmaiden’s of social transformation, which has a fairly long genealogy in the development sector, became entangled with an expanding catalogue of images, and of places - websites, reports, online videos, social media accounts - to put those images in the 2000s (Cornwall et al., 2008; Wilson, 2011).

As such, I argue that the meaning of gender was influenced by new techniques and technologies in the field of micro-economics and a growing concern with how things looked. Neither of these things was particularly linked to gender as a body of theory or as a normative agenda, and yet they helped shift what gender signified within the institution. The fact that gender had become a useful variable for economists interested in publishing papers and a set of images that appeared on websites and in PowerPoint presentations gave it greater resonance (Snow, 2004). Papers were published with the support of the Gender Action Plan on issues such as ‘gender gaps in entrepreneurial performance among firms’, ‘gender and labour market participation’, ‘gendered impacts of India’s National Rural Employment Guarantee Scheme’ and ‘impact evaluation of small enterprise development agency’s services in South Africa’. These would not have existed in the same form a decade earlier, because they would not have been technically possible. Similarly, online videos, micro-websites and social media technologies explaining the World Bank’s work to a global audience were products of the 2000s. These sorts of shift in an academic discipline, or the role of a media technology may not immediately link to a concern with gender equality, but they help explain the new interest in gender within the institution.

Before moving on, it is worth making some qualifications. This article looks at some of the reasons why gender became a less contentious term for people working within the World Bank. It does not examine the ‘on the ground’ effects of World Bank policies and programming. It looks instead at the way an idea—that had been difficult for many in the institution to engage with—became less contentious, and focuses on shifts in the field of development economics and the institutions greater use of media. A more complete analysis would pay attention to the role of individuals, what might be termed ‘norm entrepreneurs’, in the story of the Gender Action Plan. The actions of Danny Leipziger, for example, or other staff members and senior management at the institution would be part of a more comprehensive analysis of gender work at the World Bank. There would also be more discussion of the particular modes of governmentality that by the sorts of projects and policies, promoted by the Gender Action Plan, were produced with their focus on private sector development and market-based solutions (Chant, 2007). I should also add that the study draws on my own experience of working for the World Bank in the early 2000s, which gave a particular type of access to the institution during the data collection period in 2014 and 2015 (as discussed in the methods section).

II Norms and the World Bank

As noted in the introduction to this set of essays, the theoretical concern is to challenge both the ‘intentional’ and the ‘behind-the-back’ explanations of how norms spread (Cold-Ravnkilde et al., 2018). This means going beyond thinking of norms as behaviour or discourse-converging entities that travel and influence organizations without much diversity, by showing the ways in which the same norm comes to mean different things across different organizations (Acharya, 2004; Levitt and Merry, 2009; Nadelmann, 1990). The way ideas and practices influence the way norms play out has been described through concepts such as vernacularization, localization, domestication, adaptation and translation. Without wishing to add to this list, my concern is to show how things seemingly unrelated to a normative agenda—technical innovations in the field of micro-economics, the setting up of a website—helped explain why a particular normative agenda achieved more resonance.

As will become clear, ‘gender work’ was not something new to the World Bank in the mid-2000s (the question of gender and development has been part of the World Bank’s work since at least the late 1970s). Rather, changes somewhat at a right angle to norms of gender equality—new techniques in microeconomics, increasing use of media—affected its meaning. In making this point about changes disconnected from the norm in question, I also suggest the limits to research that focused narrowly on a particular agenda and do not consider changes in the wider environment.

This narrowness is apparent in existing work on norms at the World Bank. Goldman’s account of environmental policy and the World Bank, for example, argues that the World Bank actively sought to dominate global discourses on the environment. Goldman looks at the World Bank’s production of knowledge on the environment, and how that knowledge aligns to particular frameworks and interpretive schemes. The World Bank accounts ‘for them [people] and the qualities of their environments through new discourses of ecological improvement, which compel them to participate in the new neoliberal process of ecogovernment’ (Goldman, 2001: 499). Central to his thesis is the interlinking of scientific judgements, environmental knowledge and economic analysis in a way that ties the environment to the market and the private sector. Goldman emphasizes structure and discourse in his analysis, and draws on the earlier work of Escobar (1995) and to some extent Timothy Mitchell’s analysis of USAID engagement in Egypt (2002). (Mitchell shows the ways in which ‘experts’, many of them from international agencies, produced a version of post-war Egypt where a sort of ‘techno-politics’ came to define social problems.)What Goldman does not really reflect upon is how changes in the wider environment—new methodologies, ways of publishing and means of disseminating knowledge—came to structure the way people working at the World Bank were able to work.

Sarfaty’s work on human rights at the World Bank takes a slightly softer line. She puzzles over the way employees from a non-economics background—in this case lawyers in the legal department—had to frame human rights issues in ways that would make sense to economists (2009, 2012). (There is, perhaps, a less strongly felt difference between the norm of gender equality and the economizing logic of the World Bank. Whereas human rights has an absolute quality, and is embodied by the legal profession and international law; gender equality has long been picked apart in different ways by different professions and institutions.) Sarfaty points to the way staff working on human rights felt like outsiders and adopted a strategy of ‘economizing human rights’ (2012, Chapter 5). This meant, at a substantive level, a dilution of the more universalist claims of human rights having an intrinsic value, a dilution that was seen by those involved as a pragmatic decision (2012, Chapter 4). Those interviewed dissembled concerns about the dilution of the normative power of human rights, because they felt that once properly institutionalized inside the World Bank, they could push through ‘under the radar projects’ that would allow human rights to become radical again (2012: 129). Human rights was linked to economic growth, and staff members developed a series of indicators for ‘measuring justice’ (2009). The focus is very much on the human rights advocates themselves and the work they did to try to influence the institution as a whole.

Mosse makes a similar point in work on the social development division at the World Bank (2006, 2011). He found many occasions where social or political concerns had to be translated into what his respondents felt was the dominant operating logic to the institution: economics. Mosse discusses the work social scientists put into making things that achieve a particular value for a particular audience, and he sees the Bank as something of a marketplace where social scientists produce objects that can be packaged and sold. In terms of how agendas come to be established within the World Bank, Mosse points to the way policy ideas emerge from a set of ‘highly localized processes, significantly shaped by the positions and interests of particular professional groups’ (2011: 80). In some ways his study echoes John Harriss’s earlier study of social capital research at the World Bank, where social scientists decoupled the concept from the foundational work of Bourdieu, emphasising instead a genealogy that made social capital something more measurable and suited to policy and project analysis (2002). Mosse takes the example of community-driven development showing how it becomes something ‘strategic rather than conceptual or practical’ (Mosse 2006: 15). It operates as a sort of empty signifier around which different people and different agendas can invest meaning; a brand where ‘the sense of uncertainty or risk associated with innovation in the Bank is reduced’ (Ibid.).

If we turn to critical scholarship on the World Bank’s gender work, the main idea is that there has been an instrumentalist line of thinking that dates back to efficiency arguments from the 1970s and policy work around women and structural adjustment in the 1980s (Razavi and Miller 1995: 33–45). Partly the literature focuses on the resistance of the institution to gender concerns, partly the focus is on the way those concerns get economized. Zuckerman and Qing, for example, chart the failure of the distribution of documents such as the gender strategy or engendering development in the early 2000s to do much to change the mindset of ordinary staff (2005: 11–12). Chant notes that ‘one of the key problems with ‘smart economics’ has been the way gender advocates at the World Bank continue to deploy an instrumentalist use of gender, with women used explicitly, or more usually implicitly, as a ‘conduit for policy’ in the service of others, especially children’ (2012: 202; see also Molyneux, 2006, 2007). Bedford traces the connection between an economic model of the productive nuclear household with defined gender roles, and the promotion of neoliberal economics (Bedford 2009: 19; see also Wilson, 2011; Zuckerman, 2007).

What links these different sorts of studies is the way they begin from a very particular agenda—the environment, human rights, social development gender—and then move out to see the way that agenda gets framed and then marketed across the institution. The focus is on the environment division or human rights advocates working out how to ‘economize’ their agenda or social anthropologists marketing community-driven development for the consumption of World Bank employees. In academic work on gender at the World Bank, the focus has been on the extent to which gender advocates reframe gender equality around instrumentalist arguments as a way of gaining influence in an institution dominated by economists.

I make a different point. I suggest that elements, only loosely related to what we might think of as the agenda in question (in this case gender) changed the meaning of that agenda meant within the institution. It may be more useful to think of ‘loose coupling’, where effects and relationships between one field and another may not always be immediately apparent, obviously intended or easily related (Glassman, 1973). I look at changes in the field of development economics and changes in the use of media as part of the explanation as to why there was a ‘more receptive crowd than before’. These elements changed what gender meant to many people working in the World Bank. In the next section, I outline a history of seemingly unrelated developments in the World Bank’s work on gender equality, its innovations in the field of development economics and its increasing use of media. I also introduce the Gender Action Plan.

III Looking at the History of Gender at the World Bank

The 1980s and 1990s can be read as a period of the uneven institutionalization of gender into the work of the World Bank, much of it around the technologies of gender mainstreaming—gender audits, project oversight and trainings. In institutional terms, gender work can be dated back to the creation of a ‘women in development’ advisor post at the World Bank in 1977. The advisor was given a broad mandate though few resources, and her office played a monitoring and oversight role, and reviewed projects at the preparation and appraisal stages, providing criticisms and suggestions that would have to be addressed before projects could advance (Kardam, 1991). As Razavi and Miller argue, gender mainstreaming had little to say directly about economic development, and met with hostility and resistance in many parts of the institution. There was the ‘lack of an obvious organizational fit between the Bank’s mandate and ideology, on the one hand, and gender concerns, on the other’ (1995: iv). A Bank-wide reorganization in July 1987 resulted in the gender becoming a division within the Population and Human Resources department. This underlined the association between gender and what was seen as a ‘soft’ area of research and policy in the World Bank. Razavi and Miller attribute the growing space for gender work in the language and thinking of the institution to the increasing emphasis on poverty alleviation and human development, and found that gender advocates had ‘to mould gender issues into a shape that is acceptable to the Bank economists’ (Ibid.). In a comment that suggests a line through to the ‘smart economics’ agenda of the 2000s, they also point to the fact that the gender unit drew on World Bank research to show how ‘investing in women’ linked to the Bank’s main objectives: poverty reduction, increased productivity, more efficient use of resources and social returns’ (1995: v). This was the era of toolkits and ‘best practice’ and greater oversight over how policy and project documents were written. The early 1990s was also a time when more anthropologists and sociologists were brought into the organization (Cernea, 1994).

At about the same time, Chris Udry, an economist then based at Northwestern University, published ‘Gender, Agricultural Production, and the Theory of the Household’. The paper came out in the Journal of Political Economy, one of the four or five most influential journals in economics (the paper has been cited more than a thousand time). (At the time of writing the Journal of Political Economy ranked 4 out of 836 journals in the Scimago Institution Rankings database) Udry demonstrated that women-controlled plots were invested in less intensely than plots controlled by men from the same household. The data came from 150 households across three different agro-climatic zones in Burkina Faso. Udry’s paper showed that a gender-disaggregated analysis helped to improve standard work on the household economy by showing the errors that are produced when the analysis assumes households allocate factors efficiently. Later in the paper, he suggests the need for a ‘new approach for modelling intrahousehold allocation’ (1996: 1010). The study became a benchmark in research on gender in the field of economics and helped make gender a non-controversial analytical category for orthodox economists. The type of ‘gender work’ in the field of micro-economics that developed in the World Bank through the mid-2000s was beginning to emerge in the late 1990s (cf. Bardhan and Udry 1999; Duflo, 2003; Udry, 1996). (In looking at these publishing dates, it is worth noting the long gestation period for publication. A version of the Chattopadhyay and Duflo paper, for example, appeared as a National Bureau of Economic Research (NBER) working paper in 2000).

A few years after the publication of Udry’s paper, the gender unit was moved from the social development grouping to a newly created Poverty Reduction and Economic Management network. This placed gender in a ‘hard’ part of the bank, a network that also included the poverty group, public sector governance, economic policy and debt, international trade. (I asked a number of interviewees why gender had been moved into an economics division. A couple of respondents suggested it was partly a desire for the new vice president for Poverty Reduction and Economic Management to have a ‘larger empire’. A very human take on norm entrepreneurship.)This reorganisation was part of the ‘strategic compact’ that set up the hub-and-spokes, or ‘matrix management’ model of the late 1990s and 2000s (with different technical units at the centre supporting the regional divisions and country offices) (Zuckerman and Qing 2004: 9). The early 2000s also saw the decentralization of operational work to country offices. These were the years of the Wolfensohn presidency, a time when the World Bank brought in critics such as the economists Joseph Stieglitz and Ravi Kanbur, and embarked on agendas around participation, empowerment and social capital, and expanded the work of its social development division (Bedford 2009: 3). At the time, the era was described as a period of ‘mission creep’ and of passing fashions rather than a thought out response to criticisms of policies of structural adjustment or the Washington Consensus (Naim, 1999). In terms of gender work scholars remained critical of the gap between the social and economic policies of the World Bank (Schech and Dev 2007).

In 2004, Raghabendra Chattopadhyay and Esther Duflo published ‘Women as Policy Makers: Evidence from a Randomized Policy Experiment in India’. The paper came out in Econometrica (another one of the top economics journals). (In the Scimago Institution Rankings database Econometrica ranked 3 out of 836.) The authors showed a causal link between women’s representation on councils or gram panchayats and the types of investments councils made. Specifically, they found leaders invested more in infrastructure directly relevant to the needs of their own genders on councils with higher levels of female representation. It was presented as a ‘natural experiment’ in that the Indian government had, since 1993, reserved one-third of the positions of village chief for women; these reserved positions were done at random. Their analysis showed that:

In West Bengal, women complain more often than men about drinking water and roads, and there are more investments in drinking water and roads in GPs [Gram Panchayats] reserved for women. In Rajasthan, women complain more often than men about drinking water but less often about roads, and there are more investments in water and less investment in roads in GPs reserved for women (2004: 1411).

The causal element in the study made this sort of work appealing to policy makers. At the time ‘very little’ was ‘known about the causal effect of women’s representation on policy decisions’ (2004: 1410). And as with the Udry paper, the study by Chattopadhyay and Duflo made gender work a non-controversial subject for economic analysis.

The 1990s and early 2000s were also a time when the World Bank began making claims about becoming a ‘knowledge Bank’ (Goldman, 2005: 103; cf. Talero and Gaudette 1996). This was made possible by the spread of the Internet and new information communication technologies. The World Bank invested in the building and maintenance of a website, which has, over time, become the main site for data on development and the main avenue through which World Bank knowledge is transmitted. (An online search of the words ‘Burkina Faso poverty’, for example, produces the World Bank as the first suggested link. The link is: http://data.worldbank.org/country/burkina-faso last accessed on 1 June 2016.) Goldman describes the ‘assembly line of production’ of data and statistical information (2005: 149); a territorialization of knowledge that has made the World Bank ‘authoritative’ in areas where it had limited past experience. Goldman suggests that it was through producing and publishing this knowledge that the World Bank has defined the way development problems are measured, empiricized and analysed (2005: Chapter 4). (Though Goldman, like other authors writing about the World Bank, do not really reflect on the way the Internet changed the work of the institution). The sort of authority Goldman is interested in was achieved through the production of ‘benchmark’ publications as well as a raft of journal articles and in-house publications, (whose influence has been magnified by the rise of information communication technologies). Goldman estimates that the World Bank spends somewhere around 5 per cent of its operating budget on research work making it, in his view, something close to the ‘Knowledge Bank’ it aspired to be (Goldman 2005; also Mehta, 2001, see also Amsden 1994).

The World Bank also has an external affairs division, responsible for managing press relations and for promoting research and development work across different media. The World Bank runs its own internal intranet site, providing staff with news updates, links to projects, speeches by the president and other online content. The Internet has also promoted a number of ‘watchdog’ websites that monitor the work of the World Bank and other developing institutions, most obviously http://www.brettonwoodsproject.org and, in terms of the World Bank’s work on gender http://www.genderaction.org. And there is a sizeable literature looking at the role of media in campaigns to challenge the work of the World Bank (an early example being van Aelst and Walgrave, 2002). Though again what is striking is the lack of any comparable analysis of the changing way the World Bank uses media and how this reshapes the workings of the institution itself.

The Gender Action Plan was introduced to the executive board of the institution in 2006 at the end of Wolfensohn’s presidency. The Plan was led by Mayra Buvinic from within the gender unit and Danny Leipziger as vice president for Poverty Reduction and Economic Management. The particular entrepreneurial skills of Mayra Buvinic were a focus of discussion in a number of interviews. Buvinic published on the subject of ‘smart economics’ with the economist Elizabeth King (Buvinic and King 2007) and was seen as a capable norm entrepreneur, ‘someone who was more credible and savvy’ about the workings of the World Bank. (From an interview 12 December 2014, Washington DC.) It commenced operations a year later after receiving support from World Bank/IMF Development Committee. Of particular importance from the donor side was the support of Norwegian and German governments and the role of five women directors of the Executive Board, often referred to at the time as the ‘Gang of Five’. The term ‘Gang of Five’ cropped up in a number of interviews, including with a former employee who was central to gender work in the late 2000s and early 2010s. The employee referred to five women members of the World Bank’s board ‘who were unpopular, but who kept pushing the gender agenda, asking questions on the subject’. (From an interview 11 December 2014). The stated objective of the Gender Action Plan was:

… to advance women’s economic empowerment in the World Bank Group’s client countries in order to promote shared growth and accelerate the implementation of Millennium Development Goal 3 (promoting gender equality and women’s empowerment) (World Bank, 2006: iv).

The plan ran for four years, mostly during the Presidency of Robert Zoellick, and was designed to ‘intensify and scale up gender mainstreaming in the economic sectors’ which, in the language of the document, were in four markets areas: products, labour, land and finance (World Bank 2006, 2008, 2011a). These markets were ‘chosen because of their potential to produce rapid and sustained increases in women’s productivity and incomes’. The original proposal focused on ‘results’; on the understanding that ‘observable results’ would create a demonstration effect; one of the guiding principles was ‘oriented to results—processes, outcomes and impacts’. Projects and programmes had to apply to the fund for support for gender work. In the original document, economic empowerment was described as about ‘making markets work for women and empowering women to compete in markets’. Absent from the report was any language of women’s rights (Chant & Sweetman, 2012). Of the ‘four action’ areas, two are of particular relevance to our discussion later on: improved research and statistics and a targeted communications campaign. The final progress report concluded that ‘externally, it contributed to placing women’s economic empowerment squarely on the international development agenda, galvanizing donor and member countries to action, as reflected by the GAP’s capacity to mobilize resources’. The Gender Action Plan allocated more than US$70 million over four years and was used as seed money to fund gender work as part of existing projects and programmes (the initial budget was US$24.5 million). The US$70 million can be contrasted with the US$600,000 incentive fund allocated in 2002 for gender mainstreaming (Zuckerman and Qing, 2004: 4). The Gender Action Plan also received funding from the Nike Foundation and a number of other ‘non-traditional’ donors.

This slightly fragmented analysis looks at the work of gender advocates, the growing body of work on gender within development economics, and the changing demands on organizations with the growth of information communication technologies. Its fragmented nature suggests a way of thinking about the particular coalescence of different fields in the mid-2000s, which acted on gender: development economics; new information communication technologies; a concern with becoming a ‘knowledge bank’.

IV Methods and an Outline of the Remaining Sections

This article is a reflection on a series of conversations with World Bank employees in 2014 and 2015. A total of 21 interviews were conducted with current and former World Bank staff, most of them senior and all of them involved in gender work in one way or another. Interviewees included senior staff in the gender unit, economists in the Africa and East Asia region and employees who had played a key role in gender work in the World Bank in the 1990s and 2000s and had since moved on. To my surprise, those I interviewed spoke more about the Gender Action Plan than the more recent World Development Report on Gender from 2012 (staff also spent quite a lot of time discussing the future of the institution and an ongoing reorganization) (World Bank, 2011b). When I was referred on to other colleagues, it was typically with regard to what they could tell me about the Gender Action Plan. I was told that the Plan promoted work in areas such as labour reform, care work, land and agricultural interventions rather than fields of education or health, where the World Bank had concentrated gender work in the past. Though even here, the extent to which gender concerns play a role in the World Bank’s work has been questioned. The NGO, Gender Action observed, for example, that in few of the actual projects in Uganda and Cameroon, they studied facilitated men and women’s participation in project planning and design, or promoted gender-equitable access to project benefits (Gender Action, 2012; Arend, 2010). Staff members also pointed to the way the Gender Action Plan funded research work that drew on new technologies of impact evaluation and randomized control trials. The Gender Action Plan was something people found easy to talk about and around which there was a clear narrative for those involved.

As noted in the introduction, these conversations were also shaped by the fact I worked for the World Bank for two years from 1998 to 2000. I worked in the PREM network where I understood gender to be marginal, even within the network to which it belonged. My time at the World Bank also gave me some practical knowledge of the institution—acronyms, job titles, hierarchies, politics—that I could not have acquired so easily as a researcher coming entirely from the outside. Maia Green similarly reflects on her experience as a consultant for the UK’s Department for International Development in discussing policy making within the bureaucracy (Green 2011: 38-39). There was a palpable difference when interviewing those who remembered me as an ex-World Bank employee when compared to those who saw me simply as an outside researcher. Ex-colleagues sometimes joined in interviews with those new to me, producing a more relaxed conversation. This was a reflection of the way the World Bank works as a sort of society that has its own codes and rules where knowledge from within is valued. My earlier experience also shaped my sense that economics had moved on, and that gender had become less easy to disregard than it had been in the past.

V Gender as Regression

Economist 1: Where they had to conjecture ‘no evidence’ before, we now have more robust empirical work. Theoretical work on intra-household work is a lot more than before.

Economist 2: When you look at Young Professional class, there are many more micro people now; a more receptive crowd than before.

Economist 1: It is also reflective of how economics in the Bank has changed. Probably, there is more sway from the other social sciences as well …. But it is still a place where economic arguments, where [the] ‘empirical robust’ holds sway (Interview with staff members on 10 December 2014, Washington DC).

The above was taken from a conversation with two economists on the subject of how their discipline had changed over the past decade. They mentioned a number of examples, including the growing ranks of micro-economists in the Young Professionals Program as well as the fact that a World Development Report was, for the first time, focused on gender equality. Much of the conversation focused on the changing nature of the audience within the institution—that people were more receptive to gender work—and that this work appealed because it was to be found in new sorts of analytical approaches, both within micro-economics and in the ‘evidence’ of randomized control trials and impact evaluation; what Economist 1 terms the ‘empirical robust’.

It is a commonplace in the literature on the World Bank to observe that the dominant language in the Bank is economics (Mosse, 2011; Sarfaty 2012). Sarfaty emphasizes the dominance of economic knowledge and the prestige of economics as a field: ‘[their] expertise ranks as the most valuable … and whose language is the dominant mode of communication and rationality’ (2012: 96). The dominance of economists can be seen in the management of the organization—senior staff and country directors almost always claim some sort of economics background. A large proportion of staff are economists and those who are not economists generally acquire a way of framing things that would make sense to economists (ibid.). (As already noted, in the earlier discussion of agendas within the World Bank—environment, human rights, social development—advocates of those agendas talk about economic viability, growth and productivity. In promoting the ‘participation’ .)

While gender had been connected to efficiency arguments in the past, the rise of methodologies on intra-household dynamics, randomized control trials and impact evaluation meant there was a particular intersection between gender work and innovations in the field of economics. If we go back to the conversation:

Economist 1: Impact evaluation works because it produces robust analysis and persuades colleagues. Randomized control trials work because they are strongest in health, education and social protection, areas where work on gender is easy to do. Gender is more exciting (Interview with staff member 10 December 2014, Washington DC).

In this way, gender was less tied to an agenda of gender advocates, and had become more of a part of the language of economics that provoked intellectual curiosity as we as the ‘empirical robust’. Indeed part of the frustration of an older generation of gender advocates was the sense that many of the people using the word ‘gender’ within the institution imagined it as a way of managing data, getting economics papers published or loans through. Those who supported the Gender Action Plan spoke in a language that was focused more on the technologies involved that the subject of gender itself:

… the really interesting perspective that behavioural economics plays in terms of the design of the intervention should move the needle more. And the whole approach of RCTs [randomized control trials] and whatever encourages disaggregation and looking at games. Doing trials with men and women. The other thing that really should move the needle forward a lot is the focus on impact evaluation and the focus on results; that should move things along a lot. (Interview with staff member 12 December 2014, Washington DC)

A simple way of showing the growing legitimacy of gender in the field of economics in the 2000s is to look at changes in the number of articles published. A search of economics journals via a meta-database showed that in 1998, 35 articles were published with ‘gender’ in the keyword, abstract or title. In 2015, the figure was 219, a more than six-fold increase. (The economics database of ScienceDirect contains 20 economics journals, including World Development and the Journal of Development Economics, Economics Letters, the Journal of Economic Behaviour and Organization, Journal of Public Economics and the Journal of Labour Economics.) This may partly be a consequence of the increasing number of journals and journal issues, though comparable data for more established economics keywords suggests a smaller increase in the number of articles. The number of journal articles in the database as a whole increased just under three-fold from 3,417 in 1998 to 9,657 in 2015. And if we look at articles by subject those with ‘labour’ in the keyword, abstract or title in the 1998 figure was 193, while in the 2015 figure was 653 (a three-fold increase). For ‘environment’, another increasingly popular area in the field of economics, the 1998 figure was 150, while the 2015 figure was 773 (a five-fold increase). For ‘land’, the 1998 figure was 65, while the 2015 figure was 280 (a four-fold increase). More relevant to the argument of this article, perhaps, is the late bunching of gender papers in one the go-to journal for World Bank economists. Of the 51 articles published with gender in the title, keyword or abstract, in the Journal of Development Economics, 34 of the articles were published after 2009 (the journal started publishing in 1974). (The first article with the word ‘gender’ in the abstract appeared in 1986.)

This growing body of scholarship helped create a ‘knowledge community’ where work on gender and economics was uncontroversial. To take an example of the type of studies that had become possible, the Journal of Development Economics published a paper ‘Environmental and Gender Impacts of Land Tenure Regularization in Africa: Pilot Evidence from Rwanda’ in 2014 (it was received by the journal in November 2011). The paper looks at the effects of a pilot programme of land tenure regulation and shows considerable congruence with the World Bank’s land reform agenda. The analysis suggests that married women had improved access to land as a result of the reform, that there was an increase in investment and maintenance of soil conversation, and that land sales declined. On this last point, the authors reject claims that such programmes resulted in distress sales or more landlessness. The land tenure regularization programme in Rwanda, as in other developing countries, has been the subject of concern for scholars who argue that the benefits tend to favour the wealthy and those with access to social and political networks. What is more notable, from the point of view of the argument made here, perhaps, is the way a gender finding is placed alongside an environmental finding, it is not something exotic or unusual but is rather one more variable to be regressed.

Funding for the research came from DFID, the Bank–Netherlands Partnership Program, the Knowledge for Change Program and the Global Land Tools Network as well as the Gender Action Plan. The epigraph in the ‘Introduction’ is itself a descriptor of the knowledge community from which such work was possible:

We thank Rodney Dyer, Sion McGeevr and Cyriaque Harelimana for initiating and supporting this evaluation, Clive English for insights on the modality of pilot implementation, Peter Berck, Stein Holden, Niels Kemper, Gunnar Kohlin, Jeremy Magruder, Lorraine Ronchi, Chris Udry and two anonymous referees.

Thanks are also given to participants at conferences and workshops where drafts of the paper were presented, including the African Studies Centre at Oxford, the University of California, Berkeley, the World Bank’s annual conference on land and poverty in 2011, and the National Land Centre in Kigali. This gives a sense of what could be termed an ‘epistemic community’ that had built up around gender work in the field of development economics and fits well with Haas’s definition of an epistemic community as a network with ‘a shared set of causal principles (analytical and normative) and beliefs, a consensual knowledge base and a common policy enterprise (common interests) that distinguishes epistemic communities from other groups’ (1992: 18). In other words, the shifting meaning of gender within the institution resists a simple story of the instrumentalization of gender by a number of gender advocates. Instead work on gender achieved more ‘resonance’ because of methodological innovations that were not particularly interested in ‘gender’ (Snow, 2004). The shifting dynamics within development economics helps explain how a particular normative agenda—‘gender equality as smart economics’—was able to make meaning within an organization, and why it may gain purchase at one time rather than another. As such, it may be less useful in this instance to speak of vernacularization (Levitt and Merry, 2009) or even localization or domestication (Acharya, 2004; Nadelmann, 1990) and more helpful to focus on the reasons a particular epistemic community developed around a particular norm.

VI Gender as Image

In a video explaining the logic of the World Bank’s ‘gender equality as smart economics’, Danny Leipziger speaks of ‘the link between women’s economic empowerment and development and between gender equality and poverty reduction’. The video uploaded in 2008 to the World Bank’s YouTube Channel ‘Growth Commission’ explains the logic of the Gender Action Plan (https://www.youtube.com/watch?v=av1FFB9M5uU). Leipziger goes on to argue that while education and health had received considerable attention:

unfortunately, however, equal attention has not been given to women’s economic livelihoods, to giving women the opportunity to generate income themselves and for their families. We are missing this opportunity to invest in women.

Leipziger then makes the link between the World Bank’s Gender Action Plan (2006–09) and ‘gender equality as smart economics’. The video is about nine-minutes long and is a mixture of graphs, visuals and film, and shows the World Bank was directing funds towards work in areas such as the labour market, land titling, credit and agricultural production. Cases are drawn from Vietnam and Mexico. Other countries mentioned include Egypt, Uganda, Nigeria, Kenya, Bangladesh and Brazil. Women offer testimonies linking programme work to economic change. As one woman from Mexico testifies, through a translator, ‘I got my house through the gender equality programme; I was able to qualify for a loan because of my salary level’. Links are made between entrepreneurship and poverty reduction, and also between access to credit and family health. In other words women are presented as effective instruments of economic development.

On the Gender Action Plan’s website, there is a brochure made up mostly of pictures rather than text. (This was a different iteration of the World Bank to the one I had known in 1998–2000, where there was a degree of equivocation about the use of images and where the public face of the institution was less celebratory in tone). In the brochure, a group of women in an unidentified African location carry trays of eggs. Across the middle of the page is the strapline ‘when given the chance in the form of economic opportunity, women can thrive’. Another page has women in a rice field in Asia their heads bent over their work harvesting the crop; the text running through the image exclaims ‘a simple change in land title certificate can make all the difference!’ Germany and Norway are singled out for their support for the initiative, and there is a link to the World Bank’s gender webpage. The project is rendered into a series of graphs and charts. The logic of the plan is explained as a way of addressing women’s lack of progress in the ‘world of work’. The homepage of the Gender Action Plan includes a mixture of images of women and girls. The understanding that gender mostly concerns women is apparent across the World Bank’s work on the subject. While a few of those I spoke to pointed to work on masculinities, identity, intersectionality or relations between men and women, most of those referenced an older idea of ‘women in development’. As Bedford points out, this is not because the World Bank is un-gendered in how it thinks of things rather that the word gender signifies women, typically married women (2009, Chapter 1).

This shift towards the use of media can also be observed in the funds allocated to media work, and the quantifiable ways in which public relations has become more significant for the World Bank. As Goldman observes: ‘improving its public image has become such a priority that the World Bank’s public relations expenditures have surpassed its research budget in recent years’ (2005: 299–30). The 2016 the expenditure for external and corporate relations was on a par with its total spending on development economics (US$44 and US$49 million, respectively), while ‘information and technology solutions’ was budgeted at US$233 million. Beyond these crude budget lines in the annual review, there is also increased expenditure on dissemination and communications work within projects. The argument is less that the World Bank sought to use public relations to create some sort of hegemony of development knowledge, as Goldman argues, rather that the World Bank, like other institutions, was more concerned with the use of media, which was itself more widespread and varied (Esser and Strömbäck, 2014). This is a structural shift in the way large organizations make sense of themselves, and one where gender has a particular resonance as a set of images. Interestingly, the proposal felt the need to dismiss the role of the Internet on its opening page with the following quote from The Economist magazine: ‘Forget China, India and the Internet: Economic Growth is Driven by Women’ (2006: 1).

The original mandate for the Gender Action Plan included a section on ‘targeted communications campaign’ (2006: 16–18). The Plan, like any other activity at the institution, required a level of dissemination and public engagement. Throughout Action Area 4 of the original proposal, particular reference was made to ‘[p]repare and disseminate a collection of lessons learned to donors, client country ministers, Bank country directors and managers, UN agencies, and NGOs’. The report also stated a commitment to ‘regularly brief international economic press on women’s economic participation issues and Action Plan results’, to ‘brief media regularly in RBI countries on issues and results of Action Plan’ and to ‘[p]ublish and disseminate brochures on RBI lessons learned’ (Ibid.). This points to another element of change that helps explain the shifting significance of gender: the fact that the World Bank, like other development organizations, had become more concerned with presenting its work to internal and outside audiences. As noted earlier in the article, the Gender Action Plan received funding from the Nike Foundation and gender work has, across the institution, increasingly been supported by publicity hungry philanthropic foundations. In conversation with a consultant working in the World Bank’s extractive industries division, I was told that ‘it [gender] looks good … you could see pictures everywhere’ (Interview with staff member 11 December 2014, Washington DC). Those I spoke to pointed to the ways gender produces certain sorts of images at a time when representations of development have increasingly come to matter. A young female economist noted that it was ‘sexy’ as a subject of research (Interview with staff member 10 December 2014, Washington DC). The role of media and imagery could also be read into the way certain tropes and stories found themselves repeated in conversations.

A number of scholars have emphasized the way ‘cultural and social processes are increasingly constrained to take on a form suitable for media representation’, a logic of mediatization (Couldry, 2010; Hjarvard, 2007). In other words, activities and agendas that can be easily communicated—as images, data, on websites, in glossy publications—are increasingly valued and influential. Esser and Strömbäck, for example, argue that in the realm of policymaking, the question of what works in media terms should be read into the way certain programmes get picked up while others do not (2014). Such that business loans or education reforms may take on a gendered quality because they produce the right sorts of images: the capable woman in front of her new shop, the girl child going to school.

Those I spoke to about the Gender Action Plan remembered the wider environment within which gender was gaining momentum within the organization. In many instances, images or moments that reflected the organizations public profile were mentioned: ‘the push of Hilary Clinton that has made this a legitimate thing’. I was also told that the Bank’s work on impact evaluation made it interesting to organizations such as the Nike and Gates Foundations (the Nike Foundation being responsible for perhaps the most media-savvy campaign in the field of gender and development: ‘the girl effect’). At the same time, the extent to which images increasingly matter in the field of development organizations helps explain the growing commitment to gender work on the part of organizations that might not otherwise take an interest. Women have a particularly strong iconography in the world of development institutions. They are presented as mothers, carers, daughters. As Cornwall observes, women tend to be subsumed into:

an image of the protective mother who will translate any gains from the market into the means for household survival, and will be prepared to make unlimited personal sacrifices to provide the household.

One of the arguments to make about the ‘gender equality as smart economics’ agenda was that it came at a time when the World Bank, like other organizations, was more conscious of how it used images and media technologies. Just as changes in the field of micro-economics acted on gender within the institution, the question on how things looked on the outside also mattered more.

VII Conclusions

In the introduction to this set of essays, we refer to the idea of gender equality as an ‘overloaded signifier’ involving a range of different and sometimes even contradictory understandings (Cold-Ravnkilde et al. 2018, cf. Verloo and Lombardo 2007). The role of economics and media offer two ways in which the word ‘gender’ was signified in the World Bank in the 2000s. These new meanings did not come about simply because actors worked to put them there. Instead they came about because gender had become a new sort of variable to regress and because staff were under pressure to present their work to audiences in the digital age. These shifts, not particularly about gender in and of themselves, nonetheless go some way to explaining the changing significance of gender within the institution. They also suggest why ‘gender equality is smart economics’ was a happier moment for an institution that had been fairly uncomfortable with work on gender in the past.

In making this argument, it is possible to move away from both the ‘intentional’ and the ‘behind the back’ explanations of norm translation (where norms change either because of public advocacy, or because of private lobbying). Instead by linking the changing meaning of gender equality to new techniques in development economics, or the increasing role of media in the work of the World Bank, there is a more open-ended, more loosely coupled understanding of the ways a word like gender takes on meaning.

The life of the norm should not be seen as a chronological, originary story of diffusion or a story of translation and adoption but rather as an anachronic set of shifts in different, sometimes unexpected, fields (in this case the fields of economics and media). It is important then to look at what might exist at a right angle to the norm concerned, and to developments that fall outside the particular agenda being discussed. Making sense of the ‘gender equality as smart economics’ agenda also means thinking about the ranking of economics journals or the need for images to be uploaded onto a website.

I have argued that gender work gained legitimacy in the World Bank ‘because’ the field of development economics shifted and as the World Bank became more concerned with the use of media. The rise of impact evaluation and, more generally, the tilt towards microeconomics in the 2000s helped make new sorts of gender-based analysis possible. In this way, gender achieved a signification within the World Bank because it was partly uncoupled from its usual association with gender audits, gender mainstreaming and gender monitoring. The word ‘gender’ also attached itself to a set of images of women and girls that became more pervasive as the use of media became more central to the work of the World Bank. As such, it seems less useful to think of norms in relation to ‘discourse convergence’ or a single direction of travel than to think of norms as they are read against a churn of different epistemes, fashions and evolving technologies.

As a final example: in putting together this article, I met with staff in a country office of the World Bank (Interview with a staff on 13 August 2015, World Bank Country Office). The consultant was working on a gender and land study in a neighbouring African country. He had moved from Washington DC during a bout of staff restructuring, preferring the relative isolation of the country office. In the air-conditioned country office that was a copy of the hushed environment of the World Bank’s headquarters in Washington DC, we talked about the particulars of the study. What was clear from our conversation was the fact that the gender dimension to the research mattered because it presented an intellectual and methodological challenge, and because it belonged to a genealogy of work that was published in top economics journals. When pressed on the subject of gender as something about the role of women in society or about the relations between men and women, the conversation seemed to lose focus.