Bureaucratic Representation in the IMF and the World Bank

In the past few decades, the two major international financial institutions (IFIs)—the International Monetary Fund (IMF) and the World Bank—have faced challenges to their perceived relevance, legitimacy, and effectiveness in a rapidly changing global governance landscape. These challenges are linked to concerns regarding member states’ representation and the policies the institutions prescribe, as well as widely held expectations that modern international organizations (IOs) ought to be as democratic, diverse, and inclusive as possible (Woods 2000).

Over the past decades, numerous high-level governance reform efforts have focused on voice and vote reforms on the institutions’ executive boards (Zedillo Comission 2009). However, far less attention has been focused on representation within the bureaucratic ranks of management and staff. Yet global governance scholars from both rationalist principal-agent and sociological constructivist traditions have increasingly highlighted how much staff really matter and have taken seriously the empirical challenge of measuring and explaining the relative autonomy and influence of staff within these IFIs (Chwieroth 2015; Eckhard and Ege 2016; Fleischer and Reiners 2021; Hawkins et al. 2006; Heinzel and Liese 2021).

If bureaucratic agents indeed matter in shaping organizational ideas, policies, and outputs, then how well these agents represent their broader stakeholders and their development interests—especially citizens in low- and middle-income countries—is an important normative and empirical question (Sebastian, Gulrajan, and Weinlich 2022; Ferreira 2022). Normatively speaking, representation that may be perceived as skewed or unfair in any particular way can weaken the trust of key constituencies and undermine the degree to which stakeholders see IMF and World Bank decisions as legitimate. At the same time, politically driven goals around “balanced” or “equitable” representation, at either the level of the boards or within management and staff, may undermine IOs’ ability to mobilize the staff technical expertise necessary to fulfill operational mandates and maintain organizational stability (Cottarelli 2005). At its heart, the debate over representation is deeply entangled with the challenges of democratizing development (Zoellick 2010) and expanding the political and intellectual space for new voices and ideas that coincide with the expansion of IO membership and the global development agenda. Defining “optimal representation” is a subjective, value-laden task. With this in mind, how can we examine representation at the World Bank and the IMF?

In this article, we begin our inquiry into representation by explicitly sticking to the basic yet fundamental tasks. We do not seek to define what optimal representation should be for the World Bank and the IMF. Instead, we aim to fully understand what actual representation is in the World Bank and the IMF and how it might be changing over time in response to growing attention to questions regarding representation. We do so for two crucial reasons. First, we seek to fill an empirical gap in data on World Bank and IMF staff representation. Unlike peer institutions such as the European Union and specialized agencies within the United Nations system, very little disaggregated data is available on World Bank and IMF staff in a way that allows for even the most basic descriptive statistical analysis. We address this lacuna by utilizing a novel dataset that includes detailed individual-level data on staff nationality, education, and gender between 2000 and 2020. We focus on these three dimensions of representation, as they are amenable to empirical observation and measurement through methods described below, unlike more difficult to measure dimensions such as self-reported race, ethnicity, and ideology. Second, we analyze this data to discern basic patterns and trends in staffing, to document reform efforts over the last twenty years, and to discuss to what extent they have effectively altered the representativeness of World Bank and IMF staff between 2000 and 2020. We specifically focus on key bureaucratic positions where authority and decision-making power are most concentrated, including vice presidents and country directors (World Bank) as well as regional directors and resident representatives (IMF). This analysis lays the groundwork for further research studying the impact of staff representation on important outcomes such as loan conditionality, project selection, resource allocation, and operational safeguards, as well as research and data production. Ultimately, the cumulative research will inform a more substantive dialogue on bureaucratic representation.

Our results reveal considerable but uneven change. First, individuals from high-income countries continue to hold, in relative terms, more positions of authority than might be predicted from simple population-based estimates. Second, the World Bank and (especially) the IMF staff remain dominated by Anglo-American trained economists, even as the development space (as manifested in the Millennium and Sustainable Development Goals) has become much more interdisciplinary and inclusive of noneconomic development objectives. Third, while both IFIs have attempted to increase the representation of women on their staff, these efforts have not always translated into more women in critical positions of authority. Particularly in the IMF, the vast majority of these positions remain held by men. Finally, there is a compounding effect for staff who align with multiple dimensions of underrepresentation (for example, the intersection of low-income country nationality, women, and non-economists). While we do not go so far as to make normative claims about these results, we believe these descriptive findings provide a basis for further critical research and discussion.

In section 1, we introduce the general arguments about why and how representation matters in global institutions. Section 2 further breaks down the concept of representation and its potential consequences along the three dimensions we focused on in our data collection: nationality, education, and gender. In section 3, we discuss the IMF’s and the World Bank’s strategies around representation, diversity, and inclusion over the past twenty years. Section 4 discusses how we collected our novel data on staff composition in key staff positions of authority. We subsequently use this data in section 5 to assess the degree to which the IFIs have made progress in meeting their own goals around representation. With respect to these espoused goals, we conclude with recommendations, first on how the IFIs could improve their diversity and inclusion policies to enhance representation broadly defined, and second on how future research can help us understand what representation should look like given its real-world effects on IO behavior and outcomes.

The literature on bureaucratic representation in IOs has increasingly focused on the determinants and effects of workforce composition. Studies have highlighted the benefits of equitable representation within IOs, defined as increasing diversity in nationality, educational background, and gender (Abels and Mushaben 2020; Badache 2020; Christensen 2020; Eckhard and Steinebach 2021; Haack 2014; Momani 2005; Novosad and Werker 2019). Scholars argue that IOs can reap three broad benefits from addressing the representativeness of their staff: increasing legitimacy, responsiveness, and performance. We discuss these three general benefits before breaking down the dimensions of representation.

First, the perceived representativeness of staff can be one principal source of legitimacy from IOs’ critical external stakeholders (McDowell et al. 2020; Rapkin, Strand, and Trevathan 2016; Woods 2022). Legitimacy is fundamental to IOs; it is an essential factor for securing material resources, capturing autonomy, enabling IOs to develop new policy norms, carrying out operations in countries, and attaining states’ compliance with international agreements (Tallberg and Zürn 2019). Legitimacy can be derived from democratic principles such as equitable (versus weighted) voting systems or staff representation rules that are derived from a transparent set of proportionality principles such as population (Tallberg, Bäckstrand, and Scholte 2018; Buchanan and Keohane 2006; Christensen 2020; Dellmuth, Scholte, and Tallberg 2019; Rapkin, Strand, and Trevathan 2016). For example, Johnson (2011) argues that IOs may suffer “guilt by association” if they are seen as overly influenced by any particular state or group of states. Governments may be less likely to implement IO policies if citizens and public officials do not view the IO as a legitimate authority that will treat each member state fairly (Heinzel et al. 2021; McDowell et al. 2020). Thus, IOs have recently dramatically increased the size of external affairs offices and engaged in extensive public communication strategies to shape public awareness and activism directed at IOs (Ecker-Ehrhardt 2018; Gronau and Schmidtke 2016)—tasks that are increasingly complicated by shifts in the balance of power and influence of member states within IOs. While the vast majority of studies on skewed representation emphasize higher levels of governance, such as executive board voting rights and organizational leadership selection, increasingly studies of IOs have focused on representation at the staff level (Badache 2020; Eckhard 2021; Haack, Karns, and Murray 2020; Parízek 2017).¹

It should be noted that legitimacy based upon principles of staffing representation can easily be seen as a trade-off with other staffing principles, such as meritocracy. From a meritocratic perspective, especially in more technically oriented service IOs such as the IMF and the World Bank, staffing is driven in large part by the perceived need for the most qualified personnel who have the training and professional experience to carry out the organizational mission (Abbott and Snidal 1998; Barnett and Finnemore 2004). In this instance, legitimacy (and influence) is derived from an IO’s ability to mobilize and exercise expertise. Yet even here, the legitimacy of certain expertise is often contested, especially in IOs working on complex and increasing interdisciplinary issues such as international development and global financial stability and growth (Rao and Woolcock 2007; Chwieroth 2010; Nelson 2017). Organizations that highly prize certain expertise may prove less adaptive, given the ideological challenges of abandoning deeply held disciplinary tenets even in the face of contending evidence (Kentikelenis and Babb 2019).

Second, in the absence of more substantive reform in the formal governance structures of multilateral institutions, such as the executive boards, more balanced representation among organizational management and staff may help boost the degree to which crucial stakeholders feel represented by, and thereby invested in, IOs. States consider their presence within IO staff (in addition to formal power) as a fundamental factor in deciding to continue participating in IOs (Morse and Keohane 2014; Parizek and Stephen 2021). This has become especially evident in recent public comments by BRICS countries calling for a greater “voice” in IOs to continue their support of Western-led institutions. In response, status quo powers have supported governance reforms, albeit reluctantly in some cases. Both the IMF and the World Bank have acknowledged the threat of “exit” by emerging powers due to perceived underrepresentation as a fundamental reason for initiating voice and vote reforms in 2008 and 2010 (Kaya 2015; Moschella and Weaver 2017; Vestergaard 2011; Vestergaard and Wade 2013; Felicity and von Borzyskowski 2022).

Third, more balanced representation across dimensions such as nationality, education, and gender can affect organizational performance by closing bureaucratic blind spots. Representation is linked here not just to expertise but to professional and lived experiences, including contextual knowledge of countries or knowledge about particular policy areas (Evans and Finnemore 2001; Heinzel 2022; on the other effects of enhanced representation of the Global South, see Roy 2022). In turn, this diversity can affect stakeholders’ perceptions of organizational performance, as demonstrated by recent research on perceptions of IOs among stakeholders (Liese et al. 2021) and the public (Dellmuth and Tallberg 2015; Dumdum 2022). By including more diverse staff with respect to professional education (both discipline and institutions of training), IOs can incorporate broader skills and knowledge into their operations, mitigating groupthink and organizational myopia (Chwieroth 2010; Rao and Woolcock 2007), and countering external perceptions of professional biases linked to the hegemony of Anglo-American–trained mainstream economists (Kentikelenis and Babb 2019; Woods 2006).

This article focuses on representation along three dimensions: nationality, education, and gender. As briefly discussed above, these dimensions were chosen for two reasons. First, the three dimensions have been discussed as crucial by various authors studying descriptive and substantive representation in IO staff (Badache 2020; Chwieroth 2013; Haack, Karns, and Murray 2020; Karim and Beardsley 2013; Parízek 2017). More critically, in our cases of the IMF and the World Bank, our desire to include other important attributes—such as race and ethnicity, socioeconomic background, or ideological outlook—is limited by our ability to empirically observe and collect data. As we discuss in our methods section, the lack of publicly available human resources data precludes our ability to capture this information. Trying to “guess” ethnicity, race, socioeconomic status, or ideological views from our data sources (CVs, dissertation databases, and professional websites) would exacerbate biases in our dataset. Thus, our analysis is more narrowly focused and hopefully will help incite the IOs to be more proactively transparent about staffing data in the future.

The IMF and the World Bank have historically hired staff who overrepresent individuals from English-speaking, advanced industrialized countries relative to the broader population of their member states (Evans and Finnemore 2001, 11). These patterns likely result from powerful states’ interests. Member states may seek to place more of their nationals in important staff positions to solidify influence within IOs (Dijkstra 2017; Kleine 2013; Novosad and Werker 2019). For example, Kleine (2013) argues that states can trade control over certain policy portfolios by ensuring that their nationals control these portfolios. Similarly, McKeown (2009) uses internal studies by the US State Department to show that the United States sees its overrepresentation in IFI staff as an essential means to lobby for policy changes informally. More generally, member states that hold disproportionate power over IOs due to their financial contributions or influence over governance (such as leadership selection or voting shares) may also exert influence over selecting and screening staff from preferred institutional or professional networks.² For these reasons, the number of nationals a country has within IO staff is also a “measure of institutional power” (Novosad and Werker 2019, 26). Some countries from the Global South have contested the perceived imbalance of Western versus developing country staff and have called for broader reforms (Evans and Finnemore 2001; Kaya 2015). While it is unclear to what extent IO staff’s interests and behaviors are driven by their national identities (versus their professional expertise or other personal attributes and experiences), many states nonetheless consider their representational presence within IO staff as one important factor in evaluating their participation and trust in IOs (Evans and Finnemore 2001; Parizek and Stephen 2021).

At the same time, national representation can be linked to the previously discussed notion of expertise, especially contextual knowledge of countries’ political economies, histories, and languages (Heinzel 2022). Recent interviews with World Bank staff suggest that this has motivated bank management to increase the representation of staff from the low- and middle-income countries where the bank works. These interviews also support the findings of comparative studies of IOs that point to the relevance of national expertise at several levels. For example, each loan or project is based on an extensive analysis of the country context where a project is supposed to operate. Thus, local staff are highly valued for their contextual knowledge and their recipient country networks during project negotiations. A narrow reliance on (primarily Western) ex-pat staff who possess more policy-specific rather than regional expertise can undermine the extent to which such analysis can capture recipient countries’ specific problems and implementation challenges (Eckhard 2021; Evans and Finnemore 2001, 1). For instance, UN staff perceive local knowledge as a particular strength of staff from recipient countries (Eckhard and Parizek 2020). Incorporating such knowledge has also been shown to increase the effectiveness of IFI interventions and the support of local stakeholders (Christensen 2020; Heinzel 2021). The need to integrate local knowledge may help explain the distribution of nationality in the UN system, which shows a U-shaped pattern, with the overrepresentation of powerful states but also an overrepresentation of nationals from lower-income states in lower-level operational roles (Parízek 2017).

Overall, the IMF and the World Bank have signaled a commitment to greater national diversity within management and staff ranks. In the IMF, the first diversity program was launched in 1996. Since 2000 the IMF has published annual diversity reports discussing national representation in detail. The IMF has increased its focus on benchmarks and diversity goals to overcome uneven implementation. For example, the 2003 Enhanced Diversity Plan introduced greater transparency and accountability via five-year benchmarks. Over time, the IMF has introduced several evaluation tools to enforce these reforms, including the Staff Assessment for Manager Diversity (2009), the department-level Diversity Scorecard (2010, revised in 2011), the institution-wide Diversity Working Group (2014), and the reform of the Diversity and Inclusion Council (2015). Despite these efforts, by the IMF’s own account, people from underrepresented regions (URRs), consisting of sub-Saharan Africa, East Asia, Middle East, North Africa, and European Transition Countries, remain scarce.

The World Bank has similarly sought to enhance diversity and inclusion among its staff and management. For example, the World Bank website argues that “[w]e embrace diversity and look for ways to become more inclusive because diversity has the potential to yield greater work productive and competitive advantages” (World Bank 2021). The commitment to national diversity is even inscribed into the International Bank for Reconstruction and Development (IBRD) Articles of Agreement: “in appointment of the officers and staff the President shall, subject to the paramount importance of securing the highest standards of efficiency and technical competence, pay due regard to the importance of recruiting personnel on as wide a geographic basis as possible” (World Bank 2012a). Beginning in the 1970s, the World Bank expanded its staff’s geographic diversity deliberately, and since 1998 nationality has been a central focus of the World Bank’s diversity initiatives. Nationality is measured by Part I and Part II contributing member status.³ In addition, the World Bank emphasizes hiring from sub-Saharan Africa and the Caribbean as central to overcoming racial discrimination within its hiring practices (World Bank 2016). In 2007 it adopted a five-year Diversity and Inclusion Strategy that included the goal of increasing the number of managers from sub-Saharan Africa and the Caribbean (SSA/CR) as well as Part II countries more broadly. Diversity and Inclusion Compacts were established within this strategy, including precise targets and specific actions to achieve them (World Bank 2016). The Diversity and Inclusion Compacts are still utilized today to measure and report on progress. Since the beginning of the compacts, there has been a steady increase in the number of SSA/CR managers (World Bank 2012b). In financial year 2017, the World Bank reached its target for staff representation from sub-Saharan Africa and the Caribbean for the first time since 1998, for a total of 10 percent of headquarters staff (World Bank 2017).

A second critical dimension of bureaucratic representation in IOs is educational background. Organizational learning, adaptation, and change are tied indelibly to intellectual diversity that encourages debate over policy ideas. Yet in the IMF and the World Bank, economic thinking dominates. This dominance is partly a result of how the institutions’ mandates were written and historically interpreted (including defining development as “economic growth”) as well as the tendency of the institutions to define development and finance as highly technical work. As a discipline, economics provides theories, methods, and empirical tool sets that help large bureaucracies “render legible” complex issues in ways that disciplines like anthropology or sociology may not (Markoff and Montecinos 1993; Mosse 2011; Scott 1999).⁴ It is thus not surprising that a majority of the staff of the IMF, and many staff in key positions at the World Bank, are economists trained in a select number of UK or US universities widely viewed as producing the “best” economists in the world.

Nonetheless, as Scott Page (2019) has argued, the more complex the organizational tasks, the more important it is to have intellectual diversity. The lack of educational or professional diversity among management and staff can lead to disciplinary monopolies (Rao and Woolcock 2007), resulting in insular organizational cultures, groupthink, and organizational inertia (Barnett and Finnemore 2004; Clift 2018; Seabrooke 2007).⁵ As Rao and Woolcock (2007, 479) describe in their account of disciplinary monopolies in the World Bank Development Research Group, “by promoting economics as the sole lens through which to understand and respond to the development process, it restricts what is studied, delimits how those issues are analyzed and thereby offers clients an unnecessarily narrow menu of policy options and strategies.”

Intellectual monopolies create high barriers to entry for alternative viewpoints. “Policy entrepreneurs” within the IFIs often must resort to strategic framing by using the language, models, and methods of the dominant disciplines to gain intellectual traction and influence over policy and practice (Finnemore and Sikkink 1998, 910; Susan Park and Vetterlein 2010). As a result, disciplinary monopolies often lead to path-dependent, incremental change where emerging policy norms fit themselves to the prevailing organizational culture rather than the other way around (Lewis et al. 2003; Weaver 2008; Weaver and Nelson 2016). Such dynamics have been documented in numerous case studies on both IFIs, for instance, concerning debates over women’s empowerment (Weaver 2010) or human rights (Sarfaty 2009) at the World Bank and macroprudential regulation (Baker 2013) or capital controls (Abdelal 2007) at the IMF.

Disciplinary monopolies can also contribute to instances of paradigm maintenance. Paradigm maintenance refers to cases where ideas that conflict with the organizations’ prevailing orthodoxy or “official line” will be explicitly censored or distorted in ways that recast arguments and evidence to favor existing beliefs and policies (Broad 2006). For instance, Wade (2002) discusses paradigm maintenance in the case of the World Bank’s East Asian Miracle Report. He details how the state-led development message of this critical report on the region’s economic growth was eventually recrafted to conclude that this growth was due to laissez-faire economic policies consistent with the World Bank’s set of policy prescriptions at the time.

Furthermore, for the IMF and the World Bank’s borrowing countries, disciplinary monopolies may even lead to unequal treatment of member states. Existing research shows that IMF staff treat countries more favorably if recipient country officials share training in US economics departments (Chwieroth 2015; Nelson 2017). For example, Nelson (2017) argues that these departments socialize students into similar policy beliefs as the IMF itself, and the IMF recruits recurrently from these departments. Consequently, the IMF staff is more comfortable relying on neoclassically trained policymakers to design loan programs. This reliance translates into bigger loans, fewer conditions, and more waivers. Similar research shows that when national policymakers do not share the outlook of the World Bank, staff need to spend substantially more effort in preparation and supervision to get the recipient country on board (Smets, Knack, and Molenaers 2013). Therefore, disciplinary monopolies can undermine productive working relationships with recipient counterparts.

However, breaking up disciplinary monopolies does not necessarily lead to a more diverse intellectual culture within IOs because of self-selection, socialization, and the interests of powerful member states. First, staff members may self-select into IOs that share their views on key policy issues (Hooghe 2005). This might lead to hiring staff who have different educational backgrounds but who still believe in similar policy norms. Second, exposure to the intellectual cultures inside the IMF and the World Bank may socialize staff into prevailing norms (Murdoch et al. 2018). Third, powerful member states have had a substantial interest in establishing and preserving dominant economic paradigms if they serve their interests through the selection and screening of staff (Kentikelenis and Babb 2019; Wade 2002). Despite these caveats, overcoming the disciplinary monopolies at the IFIs is often seen as central to creating space for more diverse perspectives.

Nevertheless, efforts to overcome disciplinary monopolies at the IFIs have been much less the focus of reforms than initiatives to diversify staff nationality or gender. As a result, the diversification of educational backgrounds at the IMF has progressed slowly, despite the fact that the IMF publicly espouses a commitment to educational diversity, stating that “[it] introduces diversity of thought, different approaches to problem-solving, and enables us to provide more innovative solutions to our members” (IMF 2020). Data on staff educational background first appeared in the 2010 Diversity Annual Report—a decade after the first diversity strategy (IMF 2011). The IMF’s reform efforts around education are clouded by its differentiation of “Special Career Streams.” These positions, which include administrative work such as communications, human resources, accounting, and library administration, are siloed from the traditional economics work of the IMF and have specific sub targets within reform initiatives. Without data distinguishing diversity between these positions and others, it is difficult to tell from official IMF reports alone how much educational diversity reported in annual reports is attributable to changes in these non-economics roles. The World Bank has similarly fallen short to overcome disciplinary monopolies. World Bank staff have discussed the need to measure diversity and inclusion more broadly by focusing on professional and educational backgrounds. However, the World Bank does not seem to have started tracking educational background systematically or formulated explicit goals to diversify it (Das, Joubert, and Tordoir 2017; World Bank 2012d, 2018).

A third essential dimension of IO representativeness focuses on gender, specifically the number and positions of women within IO staff.⁶ Incorporating more women into IO staff is paramount for three main reasons: representation, expertise, and inclusion.

First, gender representation is necessary to maintain legitimacy. Gender mainstreaming movements and gender rights advocates have long emphasized that women deserve fair and equal representation in IOs. However, studies have shown that women are still systematically underrepresented in IO staff and leadership (Haack, Karns, and Murray 2020; Woods 2015). Some prominent IOs, like the IMF, the European Commission, or the World Trade Organization, have appointed their first women leaders in the last ten years. Others, like the United Nations or the World Bank, have still not selected a woman as their executive head. While the share of women among IO staff has made more progress toward greater representation, gender equity remains elusive (Haack, Karns, and Murray 2020; MacRae 2012).

Second, representative bureaucracy theory posits that a bureaucrat’s social and demographic characteristics (particularly race and gender) will affect the likelihood of pursuing policies and acting on behalf of others with whom they share identities (Bishu and Kennedy 2020). Under certain conditions, increased descriptive representation of women in organizations can increase the probability of substantive representation, manifested in prioritizing interests and issues that distinctly affect women (Keiser et al. 2002). Indeed, some studies have provided evidence that women’s representation in leadership positions translates into an improved focus on gender issues in IOs (Blackmon 2020). Moreover, the expertise of women also leads to different decision-making in various policy areas beyond gender issues—a “positive spillover” or “co-benefit” effect—as well as improved decision-making and performance in public-sector institutions (Barraza Vargas 2019; Sanghee Park 2013; Voeten 2020). Increases in women’s presence and power within IOs are thus seen as essential to achieving gender mainstreaming goals and avoiding “check the box” exercises.

Third, IOs need to ensure inclusion of women’s interests in policy-making. Descriptive representation does not automatically translate into substantive representation of women’s interests. Women are a heterogeneous group with diverse interests that may or may not align with other women. This heterogeneity begs the question of who is authorized to speak on behalf of women or represent the interests of a collective of women inside IO staff. And even if they were permitted to do so, other women would also need to be able to hold them accountable (Diamond and Hartsock 1981; O’Rourke 2012; Sapiro 1981). Such questions are particularly relevant when considering the underrepresentation of women in combination with other types of underrepresentation at the IFIs. Importantly, women from recipient countries are even more underrepresented. It is not immediately apparent that women from donor countries will represent the issues that women from recipient countries care about. Furthermore, IFI staff from recipient countries are often from national elites, and their interests might not necessarily align with those of women beneficiaries. Therefore, it is central for IOs to empower women to participate in stakeholder consultations and compensate for inequalities in access to inspection panels rather than simply increasing women’s descriptive representation.

Despite the importance of women’s representation, progress in hiring and promoting women into key decision-making positions remains uneven. For example, at the IMF in 2001, women were the heads of only three departments and one out of twenty senior personnel manager positions (IMF 2001). In 2002, 10.8 percent of women held the economist stream’s B-level (senior level, often division chief or higher) positions. The IMF set out to double this share in its 2003 benchmarks with a goal of 20 percent for women’s share of all B-level positions. It also set a 15–20 percent goal for economist roles and 25–30 percent for its specialized career streams. It benchmarked 30 percent for women economists at the lower employment levels and 50 percent in specialized career streams. By 2007, however, a year from the benchmark date, women’s share of B-level positions in the economist stream was only 11.5 percent. The goal was reached only in 2010. For 2014 the IMF aimed at 25–30 percent for all B-level positions and 20–25 percent for economists (IMF 2017). This goal was updated in 2015 to 30 percent share of all B-level positions and 25 percent of economist positions. The IMF achieved its second (2014) benchmark for B-level women economists and appointed its first women chief economist and general counsel in 2017. Since then, it has kept pace with its gender benchmark for 2020, achieving its goals in 2019 (IMF 2020).

The World Bank first articulated the need to increase the number of women on staff in the early 1970s. In 1973 a Status of Women working group was created, and by 1982 the World Bank had appointed its first woman vice president. In 1989 the bank adopted an action program to increase the number of women on its staff and their representation in management and other senior programs. It tracked progress through three goals, which included filling one-quarter of all vacancies at Grade 25 and above with women. It also set a goal of 25–35 percent of Young Professionals being women. In 1991 an Advisory Group on Higher-level Women’s Issues assessed progress and recommended overcoming constraints (World Bank 1992). In 1992 the Stern Report on Gender Equality was published, and a gender equality initiative was launched (World Bank 2010b). The first woman managing director was hired in 1995, and in 1997 the first woman regional vice president attained the position. By 1997 the bank reached the targets set in the 1992 Stern Report, but gender imbalances persisted strongly for women from Part II countries. They noted that “…the closer we come to operations and decision-making, the smaller the share of women” (World Bank 1997).

Finally, in 2000 the first African woman managing director, Ngozi Okonjo-Iweala, was appointed—just one year before the launch of the bank’s first Gender Mainstreaming Strategy, Women’s Equality as Smart Economics. In 2009 President Robert Zoellick set a goal of gender parity in management by 2012. The goal was not met, but during his tenure from 2007 to 2012, the ratio of women in management across the Bank Group rose from 29 percent to 35 percent (World Bank 2012c). In 2016, the same year the bank launched its second Gender Mainstreaming Strategy, Gender Equality, Poverty Reduction and Inclusive Growth, the bank became the first IFI to receive the “Economic Dividends for Gender Equality” (EDGE) Gender Equality Certification. EDGE, an initiative launched at the 2011 World Economic Forum, is a leading certificate for gender equality of private- and public-sector organizations. In March 2017 they joined the UN Women’s HeForShe movement, committing to close the gender gap in senior management by 2020, achieving gender parity in technical roles by 2022, and attaining the second level of EDGE certification by 2020.

Despite these espoused commitments and explicit reforms to enhance representation within the ranks of management and staff at the IMF and the World Bank, the perception that the IFIs are predominantly run by white, male economists from high-income countries remains. Testing this presumption, however, is difficult. Publicly available information on staff demographics is hard to find. Moreover, if data are available, they tend to be provided in the aggregate and often unlinked to specific organizational positions. This lack of transparency on their workforce composition impedes a careful assessment of the different dimensions of representation and their interplay.

In this article, we seek to redress this data deficit. We are particularly interested in senior management and operational leadership roles. We do so in recognition that it is insufficient to merely reach a critical number of different identity groups for descriptive representation to become substantive representation. Instead, it is important where these individuals are in the organizational hierarchy and what authority and influence they command. We thus gather data at the individual level, using a sampling strategy designed to capture staff members who hold substantial influence over decisions regarding program selection, management, and results.

Bureaucratic representation theory emphasizes the importance of representation for interactions with the clients that public administrations are meant to serve. The IFIs’ principal clients are governments that received IMF and World Bank loans, grants, technical assistance, and expert advice. The focus on this particular group, whom we call “client-facing” staff, allows us to emphasize staff groups whose decisions have central importance for the policies of member states. In the case of the IMF and the World Bank, this includes, most importantly, IMF resident representatives and regional directors and World Bank country directors and regional vice presidents (Chwieroth 2015; Weller and Yi-chong 2010). By focusing on this selected sample, we can narrow the data collected from all members of each IFI (including support staff and consultants, a number that surpasses well over one hundred thousand over our twenty-year time frame) to a more focused and critical subset of permanent staff.

Data was collected in a two-step process. We first extracted the names of individuals working for the IMF and the World Bank from publicly available sources provided by each institution. For the World Bank, we coded the names of all regional vice presidents named in annual reports and collected the names of all country directors overseeing country operations between 2000 and 2020. For the IMF, we extracted the names of all regional directors from annual reports and collected data on all resident representatives that manage the relationships with member states. In a second step, we coded each individual’s nationality, education, and gender. We coded demographic data by searching for information manually via numerous publicly available sources, including World Bank and IMF websites, bibliometric information (such as dissertation or thesis databases), or professional websites. Through these means, we obtained data on staff members who shaped key decisions between 2000 and 2020. For the descriptive patterns in this article, each staff member was counted for each year listed in the respective documents. For each organization, this group of client-facing staff made up between 100 and 150 individuals each year between 2000 and 2020. We interpret our findings in light of the data released by the IMF and the World Bank in various reports on the overall composition of their workforces.

How has representation changed (if at all) over the past twenty years, and what does representation look like in the World Bank and the IMF today? Concerning our first dimension, nationality, figure 1 displays patterns of representation of nationals from high-income, middle-income, and low-income countries in the IMF and the World Bank. The data shows that client-facing positions have historically been primarily occupied by staff from high-income countries, especially in the World Bank.

Nonetheless, both IFIs have made progress in meeting goals for national representation over time. The number of high-income staff members in client-facing roles has decreased, and the share of staff from middle-income countries (MICs) has increased. There have been more minor changes in the representation of lower-income country (LIC) nationals, especially in the World Bank. These figures align with the information released by the World Bank and the IMF on the workforce composition of their larger staff force. When we combine MICs and LICs, managers from middle- or low-income countries have become more numerous (in relative terms) over time. In the World Bank, they made up around 36.3 percent in 2000 and 41.4 percent in 2005, and the count has been stable at around 42–43 percent of World Bank managers since then (IMF 2019; World Bank 2015b). For the IMF, approximately 40 percent of management staff was from middle- or low-income countries in 2019 (IMF 2019). This number has also steadily increased from about 25 percent in 2000 (IMF 2001). However, our data on client-facing management shows that the share of people from LICs has declined somewhat, while the percentage of people from MICs has increased. The data show that the efforts of both IFIs to incorporate more staff from middle- and low-income countries into their workforce seem to have led to increased representativeness of their workforce (albeit uneven across country income brackets). However, US Americans make up an outsize share of administrative and management staff in both cases.

Figure 2 illustrates changes in the educational background of the major client-facing staff between 2000 and 2020. Here, we simply report whether staff have an economics degree from the United States or the United Kingdom. That way, we can see whether there has been any shift in the share of the most dominant educational group. The figure shows that most client-facing staff in both IFIs had a degree from US- or UK-based economics departments in most years. However, that share has decreased markedly (by around 15–20 percent for both IFIs) since the year 2000, suggesting that perhaps the “hegemony of Anglo-American economists” may be somewhat waning, at least in terms of staffing numbers. This diversification has different sources. The World Bank has expanded the disciplinary background of its client-facing staff more substantially than the IMF. It hires from a somewhat more diverse portfolio of degrees, including, for example, engineering, political science, or international development. The few IMF staff who are not economists tend to have closely related degrees in economic history or management, marking more minor changes in disciplinary balances. However, the IMF has hired more client-facing staff who studied in countries other than the United States and the United Kingdom than the World Bank, which leads to only around 10 percent differences in US- and UK-educated economists between the two organizations. However, most people not educated in the United States or the United Kingdom have at least one degree from other European countries, Canada, or Australia.

Overall, the educational makeup of both institutions’ staff has seen minor changes. Nevertheless, they remain dominated by economists and staff educated at universities in the United States and the United Kingdom. Similar patterns are observable when comparing the data with overall patterns of representation in official reports. The numbers align with the general educational workforce composition of the IFIs. According to the IMF’s diversity report, 58 percent of graduate degrees of IMF staff were awarded from US American universities in 2019 and 11 percent from British universities—with continental European and Canadian universities accounting for nearly all other graduate degrees (IMF 2020). The number of US- or UK-trained economists has slightly decreased (5 percent) compared to earlier years (IMF 2016). However, the IMF’s published education data is rather patchy over the years and does not differentiate by degrees. Therefore, disaggregating different economics subdisciplines to assess disciplinary monopolies further is difficult.

As expected, given differences in organizational mandates and tasks, the World Bank is much more diverse than the IMF in terms of disciplinary backgrounds. Economists are still in the majority; roughly 48 percent of staff in our sample had some kind of economics, business, or finance degree in 2010. However, there have also been around 10 percent social scientists and approximately 10 percent natural scientists among World Bank staff (World Bank 2010a). As in the IMF, most staff members have degrees from the United States or the United Kingdom. According to World Bank official reports, in 2012 roughly 45 percent of graduate degrees were awarded by US-based universities and around 12 percent by UK-based universities (World Bank 2010a). The trend we observe in our data on client-facing staff implies that the share of UK- or US-educated economists in the World Bank has been somewhat stable since then. However, we see a decrease in the number of US- or UK-educated staff in client-facing positions in the IMF. These changes are primarily due to an increase in staff educated in other European countries, potentially because staff with a focus on Europe who joined the IMF during the European debt crises rose through the organizational hierarchy.

The challenge in interpreting this data is that there is no clear consensus on the optimal balance of disciplinary expertise in the IFIs. For example, for many, it is not surprising that the IMF rarely hires from disciplines outside of economics, given the nature of its work on global monetary and financial stability. Therefore, to see a continued dominance of economists, broadly defined (rather than disaggregated by type of economics expertise, such as labor economics or finance), may not cause concern for some audiences. The numbers also do not reveal the extent to which the institutions may be pushing for greater heterogeneity within disciplines. For example, in major staffing reforms in 2007 at the IMF, there was a concerted effort to hire more staff with training in the banking sector rather than more general macroeconomic theory and policy. These staff would still appear as “economists” in our dataset but would likely bring different interests and ideas to the table. Nonetheless, in light of historical criticism of the IMF’s intellectual homogeneity and management’s call for a greater diversity of thought through hiring more midcareer staff with substantial practitioner experience rather than PhDs or staff from cognate fields such as political economy (a subfield of political science), these numbers signal at a minimum the need for further investigation into persistent hiring norms that give preference to theoretical economic training over other skill sets and professional experiences.

Concerning gender representation goals, both IFIs have increased the number of women in client-facing positions of authority, but the share is still notably low. Figure 3 visualizes the gender distribution of the main client-facing IFI staff since 2000. The IMF shows fewer women working as resident representatives or regional directors (less than 15 percent by 2020). On the other hand, the World Bank has doubled the share of women country directors and regional vice presidents, from around 20 percent in 2000 to 40 percent in 2020. The data we show for gender representation of client-facing staff is aligned with the overall patterns of representation at both IFIs.

The lack of women’s representation in the IMF has been a long-standing issue. In 2000, 13.9 percent of managers (B-level positions) were women. The share of women managers increased to 15.6 percent in 2005, 21.5 percent in 2010, 23.6 percent in 2015, and 30 percent in 2019 (IMF 2006, 2011, 2016, 2020). The lack of women is particularly apparent in the IMF’s economist stream, where there are even fewer women than in the organization at large (a phenomenon that may reflect the relative scarcity of women with graduate degrees in economics). Together with the lower representation of women in director positions (23.8 percent in 2019), these patterns can explain the disproportionately low number of women IMF resident representatives and regional directors. On the other hand, the World Bank’s full-time staff who identify as women has been around 50 percent since 2000. However, this includes a substantial share of administrative and support staff who have little impact on the decisions taken in the interactions with World Bank clients. When looking at staff with a manager title, the numbers look more like the patterns of representation among client-facing staff shown below. Around 25 percent of managers at the World Bank were women in 2000 (World Bank 2015b). This number has increased steadily to 26.7 percent in 2005, 34.3 percent in 2010, 38.3 percent in 2015, and 44.1 percent in 2019 (World Bank 2015b, 2019). Therefore, the World Bank has made substantial progress in hiring women as managers, although women are less represented in client-facing roles of authority than in overall management.

So far, we have treated the different dimensions of representation separately. However, people simultaneously differ on multiple dimensions. Such intersectionality has been an important concept used to illuminate how the experiences of people sharing a certain demographic characteristic are not homogeneous (Davis 2008).⁷ Staff who combine multiple underrepresented dimensions tend to be even more scarce within organizations. Therefore, we look at the combinations of staff characteristics that are not dominant in the IFIs’ organizational culture.

Table 1 displays the interplay of representation across the three dimensions for all people in the database for the IMF, and table 2 for the World Bank. For example, in both IFIs, there are as many or more staff members from lower- or middle-income countries among male economists trained in the United States or the United Kingdom. However, differences start to appear when changing any of the other dimensions. There are more women economists trained in the United States and the United Kingdom from high-income countries than from low- or middle-income countries. While this pattern emerges irrespective of gender, the differences are more pronounced for women than for men. Some of these differences could be explained by a lack of supply in candidates who share multiple dimensions, yet organizational differences are difficult to explain. The IMF and the World Bank have similar mandates and hire from similar talent pools. Many staff members work for both IFIs during their careers. However, the World Bank employs considerably more women economists than the IMF (especially from high-income countries). At the same time, the differences between income groups for women economists are more pronounced in the World Bank than in the IMF. These examples underscore the importance of going beyond simple aggregated patterns and actively facilitating hires from underrepresented groups in multiple dimensions.

Intersectionality also opens up avenues for exploration beyond the focus of this article. For example, substantive representation implies that staff members who share certain biographic backgrounds with beneficiaries may be more receptive to their interests. However, the few IMF and World Bank staff from LICs are typically elites whose life experiences differ considerably from the life experiences of beneficiaries in these countries. Therefore, it remains essential to ensure access for beneficiaries through accountability mechanisms like inspection panels (Zvobgo and Graham 2020), as well as participatory engagement like beneficiary consultations, even if representation goals have been reached.

This data analysis indicates that the strategies the IMF and the World Bank have enacted to increase staff diversity across the dimensions of nationality, disciplines, and gender are working—albeit likely more slowly and unevenly than desired when viewed in the context of espoused goals. Viewing the data over time in a disaggregated fashion helps illuminate areas of achievement and continuing challenges and opportunities.

However, this does not help us resolve the fundamental question of what representation should look like in the IMF and the World Bank, given their distinct political and task environments. Such a question must grapple with three countervailing pressures. First, there is the pressure of principles and the desire to attain broader inclusion and equality among nationalities, ideas, and people to counter Western dominance. This is often seen as critical for sustaining member states’ support, cooperation with the institutions, and deterring contested multilateralism or exit threats. Yet, second, there is the reality of the IMF’s and the World Bank’s principals and their interests, deeply embedded in the institutions’ financial, governance, and operational mandates. Any representational reforms that threaten these status quo powers and their entrenched influence over the IMF and the World Bank are likely to meet resistance and may undermine much-needed financial and political support for the institutions from their wealthiest members. Finally, achieving a more diverse IMF or World Bank—with respect to ideology, disciplines, genders, and nationalities—runs into the very practical matters of inherent limits in the market for staff. These limits are often rooted in the culture and practices of academic pipelines or barriers to educational access, as well as conflicting opinions over what “expertise” is really needed to tackle the finance and development mandates of these organizations. Until some compromise is reached on these issues, it may be impossible to establish a clear benchmark for representation in all of its dimensions.

As such, this early study is limited to description and exploration. Nonetheless, it provides necessary data to generate further analysis that will, hopefully, inform deliberations on the above issues. With this disaggregated data, we are able to identify critical questions regarding representational patterns that we can unpack and explain through qualitative research, key informant interviews, and more careful process tracing and analysis of primary documents. Furthermore, it allows for quantitative analysis examining the consequences of different representational trends. For example, Heinzel, Weaver, and Jorgensen (2021) show that women task team leaders and country directors incorporate more comprehensive gender mainstreaming into their project design.

In conclusion, we would like to offer a few prescriptions for ongoing work on representation at the IMF and the World Bank based on our study thus far. We believe three actions could be taken to further progress toward both institutions’ current stated goals of greater representation, diversity, and inclusion. First, both institutions should continue defining explicit representation goals subject to regular monitoring, public reporting (through, for example, annual progress reports), and enforced feedback and learning mechanisms (for example, managerial responses and updated implementation plans). Such steps will enhance transparency, external accountability, and organizational learning. Explicit comparisons of institutional progress, not only between the IMF and the World Bank but also other major international organizations, may also incite productive peer competition.

Second, the IMF and the World Bank should be more proactive in reporting on the workforce composition of their staff. The IMF’s annual diversity and inclusion reports are an essential first step. These efforts are substantially more developed than the World Bank’s reporting of its workforce composition. However, both lag behind other IOs in the UN system regarding the detail of the data. Disaggregated demographic information on workforce composition is necessary to understand where underrepresentation manifests and to hold institutions accountable for their representation goals.

Third, the IMF and the World Bank should explicitly incorporate intersectional underrepresentation into their diversity strategies to ensure that progress toward one goal (for example, more women in senior management positions) does not somehow cancel out progress in other goals (for example, more nationals from Part II countries in these ranks). This data, in turn, would also reveal risks of tokenism: hiring a few women, non-Western non-economists who can simultaneously “check the right boxes” while not making substantial advances in genuine representation. In addition, the IMF and the World Bank can help create pipelines for diverse staff through direct support for training in universities in low- and middle-income countries.

Ultimately, we hope that external accountability and advocacy, in addition to scholarly research, can help the IFIs recognize and address persistent challenges in representation. In this way, the IMF and the World Bank may join the ranks of other IOs making more rapid progress in similar reforms, such as the United Nations and the European Union. Moreover, if such representation gaps can be closed and effectively demonstrated to external stakeholders, legitimacy and effectiveness gains may be realized (McDowell et al. 2020). And in an era of proliferating multilateralism and challenges to the postwar institutional order, maintaining the image of being fair and inclusive of all member states may be essential to these institutions’ futures.

The authors thank the participants of the Global Perspectives Workshop on the Future of Multilateralism for helpful comments and suggestions. We thank Niina Nibling, Rohin Balkundi and the members of the Innovations for Peace and Development Data4Development lab at The University of Texas at Austin for excellent research assistance. MH gratefully acknowledges funding by the A.​SK foundation. The study was approved by the Institutional Review Board of The University of Texas at Austin.

The authors have no competing interests to declare.

Dr. Catherine Weaver is associate professor at the LBJ School of Public Affairs and co-director of Innovations for Peace and Development at the University of Texas at Austin. In fall 2022, she held the Fulbright Canada Chair in Global Governance at the Balsillie School of International Affairs. Dr. Mirko Heinzel is a postdoctoral researcher at the University of Glasgow and A.SK Fellow at the WZB Berlin Social Science Centre. Samantha Jorgensen received her MA in global policy studies with a specialization in international development, monitoring, evaluation, and learning from the LBJ School of Public Affairs at the University of Texas at Austin, where she also served as a project team lead at Innovations for Peace and Development. Joseph Flores received his MA in global policy studies, with a specialization in immigration policy and global governance, from the LBJ School of Public Affairs at the University of Texas at Austin, where he also served as a project team lead at Innovations for Peace and Development. He previously worked at the Council on Foreign Relations and the Migration Policy Institute.