The Politics of Gender Mainstreaming in Foreign Aid

Abstract

Gender mainstreaming—the incorporation of a gender equality perspective into the design, implementation, and evaluation of all aid projects—has become a signature policy tool among Western donors. However, advancing gender equality can be politically contentious and lead to backlash, particularly in autocratic regimes where women’s socioeconomic status is low. We argue that donors’ desire for recipient government buy-in creates incentives for them to pay attention to domestic policy cues, whose salience varies across regime types. Employing detailed data from the OECD’s Gender Equality Policy Marker, we show that donors engage differently with democracies and autocracies. Among democratic recipients, those with higher legal status for women have less gender mainstreaming aid, suggesting a “needs-based” logic. Conversely, in autocracies, donors respond positively to policy cues indicating the (domestic) political acceptability of gender equality. Our findings underscore the importance of treating gender mainstreaming as a distinct category of assistance whose application is attuned to domestic implementation problems. Beyond the study of foreign aid, we offer insights into how international audiences may interpret policy cues differently depending on regime type.

La incorporación de la perspectiva de género (es decir, la incorporación de una perspectiva de igualdad de género en el diseño, la ejecución y la evaluación de todos los proyectos de ayuda) se ha convertido en una herramienta política distintiva entre los donantes occidentales. Sin embargo, los avances en cuanto a la igualdad de género pueden resultar políticamente polémicos y provocar reacciones negativas, especialmente en regímenes autocráticos donde el estatus socioeconómico de las mujeres es bajo. Argumentamos que el deseo por parte de los donantes de que el Gobierno receptor acepte crear incentivos para que presten atención a las señales en materia de política interna, cuya importancia varía según el tipo de régimen. Utilizamos datos detallados del Marcador de Políticas de Igualdad de Género (GEPM, por sus siglas en inglés) de la OCDE, lo que nos permite demostrar que los donantes se relacionan de manera diferente con las democracias y con las autocracias. Entre los receptores democráticos, aquellos con un estatus legal más alto para las mujeres reciben menos ayudas para la incorporación de la perspectiva de género, lo que sugiere una lógica ”basada en las necesidades”. Por el contrario, en las autocracias, los donantes responden positivamente a las señales en materia de política que indican la aceptabilidad política (interna) de la igualdad de género. Nuestras conclusiones remarcan la importancia de tratar la incorporación de la perspectiva de género como una categoría distinta de la asistencia, y cuya aplicación se encuentra en sintonía con los problemas de implementación nacional. Además del estudio de la ayuda exterior, ofrecemos información con respecto a cómo las audiencias internacionales pueden interpretar las señales políticas de manera diferente según el tipo de régimen.

La popularisation du genre - l’incorporation de la perspective d’égalité des genres dans la conception, la mise enœuvre et l’évaluation de tous les projets d’aide - est devenue un outil politique signature des donateurs occidentaux. Cependant, faire progresser l’égalité des genres peut s’avérer polémique sur le plan politique et générer un retour de bâton, notamment dans les régimes autocratiques où le statut socioéconomique des femmes est faible. Nous affirmons que si le donateur souhaite l’adhésion du gouvernement bénéficiaire, il est d’autant plus incité à s’intéresser aux indices politiques à l’échelle nationale, dont l’importance varie d’un type de régime à l’autre. En employant des données détaillées de la base Gender Equality Policy Marker (GEPM) de l’OCDE, nous montrons que les donateurs interagissent différemment avec les démocraties et les autocraties. Les bénéficiaires démocratiques où le statut légal de la femme est élevé reçoivent moins d’aide à la popularisation du genre, ce qui indiquerait une logique ” fondée sur les besoins ”. Á l’inverse, dans les autocraties, les donateurs réagissent de façon positive aux signes politiques qui indiquent une acceptabilité politique (nationale) de l’égalité des genres. Nos conclusions soulignent l’importance de traiter la popularisation des questions relatives au genre en tant que catégorie distincte d’assistance, dont l’application s’adapte aux problèmes nationaux de mise enœuvre. Au-delà de l’étude de l’aide étrangère, nous offrons des renseignements sur les différentes interprétations possibles des indices politiques par des publics internationaux en fonction du type de régime.

Introduction

Women’s economic empowerment has become a major policy goal of Western development assistance. Around the world, women’s economic potential is under-utilized due to the ill effects of low educational attainment, low labor force participation, and worse health outcomes, relative to men. Legal systems in many countries do not grant women equality in inheritance, property, and employment rights, further limiting their economic participation. In 1995, the Beijing World Conference on Women brought gender equality to the forefront of the international development agenda. Soon thereafter, the Development Assistance Committee (DAC) of the OECD initiated the Gender Equality Policy Marker (GEPM), setting an expectation for DAC members to increase their efforts toward gender mainstreaming, that is, assessing the implications for women (and men) of any planned action throughout the design, implementation, and evaluation stages of all projects.¹ By the 2010s, nearly all Paris Club donors—including European Union (EU) members, Australia, Canada, Japan, and the United States—had adopted gender mainstreaming policies.² This marked a transition from “first generation” efforts to advance women’s empowerment—which focused on descriptive representation and capacity building for female politicians—to a “second generation” approach that sought to tackle underlying societal gender inequalities (Brechenmacher and Mann 2024). Though the United States moved away from a gender-focused approach in 2025 at the start of Trump’s second term, the commitment to gender mainstreaming among other DAC donors continues.

The mandate of gender mainstreaming policies is strikingly broad, calling for the consideration of impacts on women in all recipient countries and in all types of projects, irrespective of sector or channel of delivery. The EU’s third Gender Action Plan (2023) illustrates the universal nature of gender mainstreaming policies among DAC donors: “All external assistance across all sectors, including infrastructure, digital, energy, agriculture, and blended funds, etc., should integrate a gender perspective and support gender equality.” The integration of gender equality goals into such a wide range of projects represents a consequential departure from the older practice of dedicating particular projects to women’s rights under the banner of democracy assistance. We refer to such projects as primary-targeted gender aid, which the OECD describes as those for which women’s empowerment “is the main objective... and is fundamental in its design and expected results” (OECD 2016). Such efforts have met with backlash due to their more overtly political implications (Henderson 2000; Carapico 2002; Sundstrom 2005). Mainstreaming holds out promise for a less direct, yet ultimately more comprehensive, approach toward engaging recipient governments on the issue of gender equality.

But how consistently are these ambitious goals applied in practice? Some studies of gender mainstreaming effectiveness contribute, by omission, to an impression that implementation is an apolitical, technocratic affair,³ or that variation in effort is shaped by the preferences of donor agency staff (Jones and Swiss 2014; Heinzel, Weaver, and Jorgensen 2025). Yet, there is voluminous evidence that foreign aid is politicized in multiple ways (Kilby and Dreher 2010; Bermeo 2017), including that donors seek to design their programs to enhance buy-in by recipient governments (Bush 2015; Snider 2018). The desire to ensure the cooperation of state authorities is likely to be strong for gender mainstreaming, which—unlike democracy assistance to civil society—covers a large volume of projects channeled directly through the government. We argue that considerations of recipient buy-in are more prominent when engaging with autocratic regimes, where the risk of backlash is higher from some governments than others.

We assess these expectations with data from the OECD’s GEPM, which provides project-level information, for all DAC donors, about whether gender equality is either a primary or secondary (“gender-mainstreaming”) goal. The GEPM is a rich yet still under-utilized data source. Most prior studies rely on purpose codes from the Creditor Reporting System (CRS) to identify projects that are dedicated to women’s empowerment organizations and institutions (WEOI). We show that these data flag only a small proportion of aid related to women and fail to capture gender mainstreaming altogether.

Analyzing patterns of gender-mainstreamed aid over a 22-year time frame (1998–2020), we explore whether DAC donors engage differently with democracies and autocracies. We argue that domestic policy cues—operationalized as laws related to women’s rights and representation—exhibit varying salience across regime types. Because the risk of backlash to women’s empowerment is higher in authoritarian regimes, donors are more attentive to positive policy cues in these contexts. Accordingly, we find that among democratic recipients, gender mainstreaming accords with a needs-based logic, whereby countries with worse performance on women’s de jure rights see higher levels of mainstreamed aid. In contrast, among autocratic recipients, donors engage in more gender mainstreaming in countries where the de jure protection of women’s rights indicates political buy-in for action in this area. The upshot of our analysis is that, for recipients with strong performance on de jure women’s rights, there is no difference across democracies and autocracies in the proportion of foreign aid that is gender mainstreamed. But among recipients with weak legal protection of women’s rights, donors engage in substantially more gender mainstreaming in democracies.⁴

This may imply, concerningly, that donors avoid engaging on gender issues in a large swathe of cases where reform is most needed: autocratic recipients where women have lower legal status. We offer a more nuanced interpretation, namely that donors are attentive to domestic implementation problems and choose to focus their efforts in places where they can gain traction via greater cooperation from state authorities. Moreover, levels of gender-mainstreamed aid remain rather substantial even among the more regressive autocracies that see less of it. In autocratic recipients at the lowest-performing quartile on measures of women’s well-being such as maternal mortality rates and labor market participation, an average of about 20 percent of Official Development Assistance (ODA) commitments are gender-mainstreamed—far higher than the 2–3 percent of commitments that are primarily dedicated to gender issues.

Our analyses therefore underscore the importance of gender-mainstreamed aid as a sizeable (yet under-studied) category of development assistance whose application differs across democracies and autocratic regimes. Unlike primary-targeted women’s rights projects, gender mainstreaming is applied based on the intersection of need and recipient buy-in. This insight carries implications for future research. First, studies on the effects of gender-targeted aid should separate mainstreamed aid from primary-targeted projects, given the different approaches, sectors, and channels of delivery that they employ. Use of the GEPM data, which include information on gender mainstreaming, should become standard practice in the empirical study of foreign aid and gender equality, which has typically relied on CRS purpose codes that capture only a small portion of gender-targeted aid. In the second section, we describe the multiple layers of quality control applied by the DAC in generating the GEPM data, which should instill confidence among researchers.

Second, progress toward a more comprehensive theory of gender-targeted aid should build from our finding that mainstreamed aid is allocated according to logics that differ across recipient regime types. Understanding the conditions under which mainstreaming is applied is essential for a full accounting of its impact—not only whether it is effective, but why. For example, research could further probe how donors assess recipient country buy-in in autocracies, beyond the de jure indicators that we explore here.

Third, our findings suggest synergies with research on authoritarian politics, which has explored how autocracies can reap international reputational benefits from women’s rights reforms, including gender quotas (Bjarnegård and Zetterberg 2022; Bush, Donno, and Zetterberg 2024). Our analysis indicates that the consequences of quota adoption are not merely symbolic: introducing such measures may pull countries into deeper engagement with donors on issues of gender equality.

Tracking Gender Targeting in Foreign Aid Programs

Throughout the 1990s and early 2000s, economists and policymakers increasingly adopted the perspective that empowering women is important not only in its own right but also for unlocking economic development (Mehra 1997; Coleman 2004; Dahlu, Knutsen, and Mechkova 2022). DeWaal (2006) refers to the “efficiency approach” toward gender and development, which focuses on how harnessing women’s labor and entrepreneurship can jump-start economic growth. Although this approach is criticized by some as instrumentalizing women’s rights (Cadesky 2020; Parisi 2020), it is now firmly embedded in the strategies of OECD donors (Holvoet 2010), and features prominently among those, such as Canada and Germany, that have adopted a feminist foreign policy.

This has taken two tracks. The traditional approach is to dedicate entire projects toward gender equality and/or women’s empowerment.⁵ Categorized by the DAC CRS under the purpose code for “women’s equality organizations and institutions” (WEOI), these are primarily democracy assistance projects, many of which are directed to civil society groups and therefore bypass the government. In the GEPM system, inaugurated in 1999, projects designated as primarily targeting gender are those whose main purpose is advancing women’s equality and empowerment. (This includes, but is not limited to, projects falling under the WEOI purpose code.) To provide a few examples: a grant from Canada to Kenya in 2018 to support business opportunities for rural women in east and southern Africa; a grant from the United States to South Sudan to provide reproductive healthcare needs of girls, women, and underserved communities; a grant from Spain to Colombia to implement the restitution of land rights to women that had been displaced by conflict.

Gender mainstreaming entails a different, more holistic approach toward advancing women’s equality and empowerment. It involves incorporating a gender perspective into project design, implementation, and evaluation as an important, but not primary, focus. Crucially, mainstreaming is envisioned as applying across all sectors of assistance, not only democracy assistance but also health, sanitation, education, peace and conflict, and even seemingly distant sectors such as energy and transport. For example, a 2017 grant from the United Kingdom to Myanmar to improve civil society participation in the peace process, including women’s groups; a 2013 grant from the EU to Nicaragua to improve higher education competitiveness; and a 2008 grant from Belgium to Cambodia for the construction of a water navigation center to improve transport on the Mekong River. For these latter examples, we see that the projects are based on general development goals but with a gender equality dimension (UN Economic and Social Council 1997; DeWaal 2006; World Bank Group 2015). Thus, gender mainstreaming policies cover a vast swathe of foreign aid projects that are channeled directly to recipient governments. This is an important point, which we argue has implications for how mainstreaming is applied in practice.

Most research on foreign aid and women’s empowerment has relied on the CRS sector data, examining projects designated under the WEOI purpose code.⁶ But this represents only a small subset of projects with a gender focus. Perhaps for this reason, prior studies have drawn somewhat disparate conclusions on the effects of such aid. Women’s rights aid has been shown to increase female political representation (Baliamoune-Lutz 2016), quota adoption (Edgell 2017), and economic rights (Su and Yang 2022; Minasyan and Montinola 2023), but not women’s health outcomes and fertility rates (Swain, Garikipati, and Wallentin 2020).

In 1999, the DAC inaugurated the GEPM, which asks all donors to screen their projects and mark whether gender equality is either a primary or secondary (“gender-mainstreaming”) goal. To understand the data-generating process, we conducted in-depth research into the organization’s data reporting structure. The CRS provides clear guidelines and standardized spreadsheets through which members report project-level information. Importantly, the organization also supports member reporting on the GEPM via regular meetings of the DAC’s working group on gender equality (GenderNet). An examination of GenderNet meeting minutes reveals that donors share notes on their GEPM use and receive expert reporting tutorials to ensure a common understanding of what it means for a project to “principally” or “significantly” target this policy goal. Moreover, discussions with members of the CRS highlighted the DAC’s quality assurance protocol whereby its staff perform regular checks on GEPM reporting.⁷ While it is in principle possible that donors might inflate these data in an attempt to virtue signal, our research does not provide any evidence to suggest that this is a pervasive issue.⁸

An initial look at the GEPM data underscores that gender-mainstreamed projects are far more numerous than those coded with a gender focus by the CRS WEOI purpose code. Figure 1 tracks all development assistance with a gender-equality component over time: in the top panel, note the prevalence of gender-mainstreamed aid in light gray (i.e., aid commitments marked as pertaining to a project with gender equality as a secondary goal)—nearing 40 percent of all development aid commitments by 2018. In darker gray, both panels depict the percentage of commitments primarily targeting gender equality according to GEPM—consistently hovering around 3–4 percent of total commitments since the 2010s. Lastly, the black dots indicate the percentage of aid commitments identified with the WEOI purpose code—an even smaller subset of development assistance.

The descriptive data in Figure 1 highlight how any analysis of foreign aid action on gender equality that limits its focus to either projects identified with the WEOI purpose code or to projects marked as primarily targeting gender via GEPM misses the very large component of gender mainstreaming in development assistance—an important phenomenon that has not received sufficient attention in quantitative research. Two recent studies that have considered the effectiveness of gender mainstreaming find, encouragingly, that it is associated with women’s legal, social, and political empowerment (Su and Yang 2022; Minasyan and Montinola 2023). Questions remain about whether mainstreaming is applied differently, or works differently, depending on the domestic political context.

Indeed, there is plenty of variation across recipients of donor action on gender equality. Examining the countries that received the highest and lowest proportions of gender-targeted aid (see the Online Appendix) reveals that some of the largest recipients of gender-targeted aid are autocratic regimes, such as Algeria, Angola, and Rwanda; but these are not particularly regressive regimes when it comes to women’s rights. On the other hand, among the smallest recipients are some democracies, such as Botswana, Peru, and South Africa, as well as autocracies where women’s rights are suppressed, such as Iran and Syria. This initial look at the data indicates that gender mainstreaming in foreign aid is not applied based simply on need, nor on a straightforward logic whereby democracies are prioritized over autocracies.

Perspectives on Aid Program Design

How do recipient-country politics shape patterns of gender mainstreaming by OECD donors? Research on aid allocation provides some insights to guide our inquiry. The essential message is that aid is politicized in numerous ways. One element is the interests of donors, namely how their political, strategic, and geopolitical preferences shape where aid is sent (Kilby and Dreher 2010; Bermeo 2017). These studies show that although guided by recipient-country need, aid provision is also meaningfully shaped by donors’ political considerations—a fact that may undermine aid effectiveness because it implies that assistance is regularly sent without serious policy conditionality attached.

Aid effectiveness can also be undermined, at the recipient level, by corruption and poor bureaucratic capacity—an insight that forms the heart of the quality of governance research agenda in the study of foreign aid (Brautigam and Knack 2004; Dietrich and Winters 2021). An expansive view of governance may incorporate a gender perspective: One recent study finds that development interventions work better in countries with greater gender equality (Montinola and Prince 2018). Some have examined the relationship between governance and aid from the opposite angle, that is, whether donors reward improvement in governance (Alesina and Weder 2002). On this point, there is specific evidence that advances in women’s rights—including gender quota laws and reforms to women’s economic rights—are rewarded with increased aid from Western donors (Bush and Zetterberg 2021; Okundaye and Breuning 2021; Bush, Donno, and Zetterberg 2024).

Research on how recipient-level factors shape the design of aid programs—that is, which sectors are targeted and which policy goals are attached—is more limited. Winters and Martinez (2015) and Bermeo (2017) show that donors consider recipient government capacity when determining the sectoral focus of their support, avoiding sectors that require high government involvement (such as direct budget support and infrastructure projects) in recipients with poor governance. Dietrich (2013, 2021) shows that donors alter their delivery tactics in poorly governed states, focusing on supporting non-state actors rather than the government. In these studies, state capacity is treated as largely an apolitical problem.

Coming closer to a political theory of aid composition, Dietrich and Wright (2015) argue that Western donors favor economic development assistance in closed dictatorships, and shift more toward democracy and governance packages once regimes transition to multiparty systems. The underlying logic is that the domestic regime context shapes the anticipated effectiveness of different (sectoral) aid tactics. Bush’s analysis of the “taming” of democracy assistance similarly demonstrates that Western donors think about results when determining which policy goals to prioritize: They seek to identify projects that will win the acceptance and cooperation of the recipient government (Bush 2015). We develop these insights, offering an explanation for the implementation of gender mainstreaming that accounts for the anticipated level of “buy-in” from recipient countries.

Theoretical Expectations

In recent decades, the international development community has moved toward an accepted view on the key markers of women’s empowerment. The Millennium Development Goals, for example, stressed women’s access to health care, equality in education, increased participation in the workforce, and increased legislative representation. The World Bank’s project on Women, Business, and the Law (WBL) identifies a range of legal rights that are essential for women’s full participation in the economy and publishes an index that captures performance on these dimensions. A needs-based logic would predict that gender targeting should increase as countries’ performance on such metrics of women’s rights declines. However, we argue that this is not the only consideration shaping the incorporation of gender mainstreaming in donors’ aid programs.

While governments in some recipient countries are committed to women’s rights, this is not true everywhere—particularly for autocracies where advancing gender equality may disrupt political hierarchies that support the regime. Research on backlash against international norms shows how powerful this phenomenon can be, with real consequences for outside actors’ ability to engage on gender-related issues (Sundstrom 2005; Graff 2014; Bush and Jamal 2015; Krizsán and Roggeband 2021; Zaremberg, Tabbush, and Friedman 2021; Terman 2022). For aid that is channeled directly to state bodies, a lack of cooperation from domestic authorities can lead to delayed or ineffective implementation—a problem for Western aid agencies that face often intense pressure to demonstrate results to their political principals.

Development agencies therefore have reason to consider how their policy goals will be received by recipient governments and adjust their efforts accordingly. We refer to this domestic reception, and the attendant likelihood of backlash, as the domestic buy-in for gender targeting. The OECD’s strategy for advancing gender equality acknowledges the role of favorable conditions in recipient countries (OECD 2020), noting that at the domestic level “[i]mplementing commitments to gender equality and women’s empowerment requires a range of efforts including political will and policy dialogue...” (OECD 2020, 8). Indeed, evaluations of gender mainstreaming implementation offer some criticism that donors may use a lack of domestic ownership as an excuse to avoid integrating gender goals into aid program design (Holvoet and Inberg 2014). In sum, donors seek to design their aid packages in a way that advances their policy goals while also securing the cooperation (or, at minimum, the acceptance) of domestic stakeholders.

For gender-targeted aid, this highlights an important distinction: gender mainstreaming policies cover all aid projects, the majority of which are channeled directly to the state (public aid). In contrast, primary-targeted gender equality projects are channeled to a greater extent toward non-state civil society actors (bypassed aid), for which the question of state cooperation is less relevant.⁹ These discrepancies are illustrated in Figure 2. Due to these differences in channel of delivery, we expect the provision of gender-mainstreamed aid to be sensitive to recipient government buy-in. In what follows, we develop expectations about patterns of gender mainstreaming across democracies and autocracies. In the Online Appendix, we present models for primary-targeted gender aid as a comparison.

Regime Type

The backlash against gender mainstreaming, and the broader equality-enhancing agenda to which it belongs, is more of a risk in autocracies than in democracies. Though authoritarian regimes are not universally opposed to advancing women’s rights, there is wide variation among them, and they tend on average to exhibit lower levels of women’s empowerment than democracies (Beer 2009). This is supported at the micro-level by longstanding findings in survey research of a correlation between authoritarian and patriarchal attitudes (Duncan, Peterson, and Winter 1997; Inglehart, Norris, and Welzel 2003). Women’s rights may therefore pose a political threat to non-democratic systems. Chenoweth and Marks (2022) note the prevalence of women at the forefront of many contemporary pro-democracy movements. In these contexts, “...fully free, politically active women are a threat to autocratic leaders” (Chenoweth and Marks 2022, 4). The extent of this threat can vary substantially across autocracies, as we discuss below.

In contrast, democracies are characterized by larger winning coalitions and more meaningful elections, which tend to dilute the concentrated power of patriarchal elites. Democratic systems feature greater protections for individual rights and equality under the law, making state actors more attuned to the types of goals advanced by a gender equality agenda. The upshot is that democratic regimes—even those performing poorly on women’s empowerment—are less likely to feel politically threatened by gender mainstreaming and its equality-enhancing agenda. The risk of backlash is lower. We therefore expect, ceteris paribus, that mainstreaming will be higher in democracies.

 

Policy Cues

Among authoritarian regimes, there is greater variation in support for women’s rights, making informational shortcuts more valuable as a guide to action. Some regimes, such as Rwanda, Morocco, and Ethiopia, have adopted a modernizing stance toward women (Bjarnegård and Zetterberg 2022; Donno, Fox, and Kaasik 2022), embracing the view that female empowerment is associated with economic growth (Coleman 2004) and international prestige (Towns 2012). In such cases, policy cues via de jure gender equality reforms serve to indicate government buy-in for gender mainstreaming in foreign aid. Yet, for other countries, such as Iran or present-day Turkey, suppressing women’s rights is at the heart of the regime’s ideology and (conservative) support coalition. In such contexts, external efforts to advance women’s empowerment may be perceived as threatening to the regime’s survival. The risk that state actors would resist or undermine aid project implementation is accordingly greater, and donors respond preemptively to this possibility. In a study of US aid to Egypt and Morocco, Snider (2018) shows that assistance programs are “negotiated deals” whereby recipient governments help determine both the goals and form of the projects. She argues that this dynamic is especially prominent when engaging with autocracies, where the risk of backlash against a rights-based and pro-civil society agenda is higher.

How do donors learn about recipient government buy-in? One way is through their direct interactions with domestic officials (cf., Snider 2018, 800). But they also look for policy cues that send a more public signal as to the regime’s openness to pursuing gender equality. We highlight two de jure policy areas to which donors are known to be attentive. First, legislative gender quotas are a prominent priority for Western donors. As Edgell (2017), Bush (2011), and Towns (2010) describe, the post-Cold War drive for gender quotas emanated primarily from international pressure. Moreover, quotas and women’s representation are easily measurable outcomes that are well suited for donors to use as an information shortcut (Okundaye and Breuning 2021).

Second, donors are attuned to legal reforms related to women’s economic rights (Bush, Donno, and Zetterberg 2024), in part due to the advent of the World Bank’s WBL index. These data provide comprehensive cross-national information on women’s legal status along eight dimensions related to labor rights, property rights, and family law. The United States’ Millennium Challenge Corporation, for example, explicitly uses the WBL ratings as criteria for determining eligibility for funding.¹⁰

 

Data and Research Design

To empirically investigate gender equality targeting in different recipient country contexts, we construct two panel data sets utilizing information on ODA from the OECD’s CRS. Starting from the CRS’s project-level data, we collapse observations first into a data set at the recipient-year level. This allows us to explain variation in the percentage of aid commitments targeting gender equality to a given recipient in a given year—the primary focus of our theory. Second, we build a data set at the donor–recipient-year level, allowing us to also account for characteristics of the donor and donor–recipient relationship. We use all available CRS data from 1998 to 2022, encompassing over 2.5K observations at the recipient-year level and over 46K observations at the donor–recipient-year level.

The dependent variable of our main analyses is gender mainstreaming, measured as the percentage of aid commitments (in deflated USD) where the GEPM data indicate a secondary targeting of gender equality. Our analysis concerns ODA, defined by the OECD as government aid that promotes the economic development and welfare of developing countries. The GEPM classification of projects as gender-mainstreamed is donor-reported—as is all other OECD data on foreign aid. As discussed further in the Online Appendix, the DAC provides a set of tools and review opportunities to ensure that its members are using the GEPM in a similar way, and DAC staff perform extensive quality checks on these data. In the Online Appendix, we further present analyses predicting the percentage of primary-targeted gender aid as well as of aid identified by the WEOI purpose code.¹¹ Note that, while there has been a general increase in gender mainstreaming in foreign aid over time, there is considerable variation among recipient countries. Our analysis seeks to explain this.

Recall that our hypotheses anticipate higher average levels of mainstreaming in democracies (Hypothesis 1). In autocracies, we expect policy cues to be positively associated with gender-mainstreamed aid (Hypothesis 2). To operationalize regime type, we use a dichotomous variable equal to zero if the V-Dem “Regimes of the World” dataset classifies the recipient as either a closed or electoral autocracy and one for electoral or liberal democracy (Luhrmann, Tannenberg, and Lindberg 2018). The Online Appendix models present results with alternate measures of regime type.

We operationalize de jure gender equality policy cues in two ways: (i) the presence of a legislative quota (Hughes et al. 2017)¹² and (ii) the World Bank’s WBL index. The former captures provisions that encourage more equal political representation across men and women. As discussed, the adoption of a quota is a single and easily observed policy that is tracked by Western donors (Bush 2011). It is not, however, a measure of women’s de facto empowerment, given the indirect link between descriptive and substantive representation. The WBL Index captures women’s legal equality in the areas of mobility, workplace, pay, marriage, parenthood, entrepreneurship, assets, and pensions. These data are collected via expert surveys with observations ranging from zero, indicating no legal protection for women, to 100, indicating very high legal protection. This measure provides a summary of women’s de jure economic rights.

We estimate a set of linear regressions to assess our hypotheses. All models feature the percentage of gender mainstreaming aid as the dependent variable and two key explanatory variables: recipient regime type and equality-promoting policy cue, which appear both as direct predictors and in interaction with regime type. To make fairer comparisons across recipients, all models include measures of recipient economic development (measured as GDP per capita, logged) and aid dependence (ODA commitments over GDP). All models include year fixed effects, and standard errors clustered at the recipient level. In robustness checks reported in the Online Appendix, we estimate models with country fixed effects.

Our first four specifications (Models 1–4) are estimated on over 3K recipient-year units, 1998–2022. Models 5–6 are then estimated on over 54K donor–recipient-year units over the same period, and include additional covariates for the importance of a given recipient to a given donor (measured as aid commitments from that donor to that recipient, over that donor’s total aid) and for a measure of gender equality in the donor country (percentage of legislature seats held by women), as well as fixed effects at the donor-year level.¹³

Results

Results in Table 1 provide support for our hypotheses. On average, democracies receive a significantly higher proportion of gender-mainstreamed aid. As discussed, we do not have expectations about a direct effect of domestic policy cues (quotas and the WBL index) on the extent of gender mainstreaming. Rather, we are interested in the differing effects of those policy cues in democratic versus autocratic recipients. Models 3–4 include the relevant interaction terms, and Figure 3 visualizes the effects. Here, the positive and significant coefficient for democracy indicates that, among recipients without gender quotas or with a low WBL index, democracies receive a greater proportion of gender-mainstreamed aid than autocracies. Among autocracies, gender-promoting policy cues exhibit a positive, significant effect on gender-mainstreamed aid. We note that the effect of the WBL index is less significant—though still in the expected direction—in the donor–recipient-year model (Model 6), a point that we consider below.

Reflecting on these findings, there is clear evidence that DAC donors apply gender mainstreaming policies differently across democracies and autocracies. Figure 3 shows that among countries where women have low de jure status, authoritarian regimes see much lower levels of gender-mainstreamed aid. One conclusion, then, is that gender mainstreaming is under-provided in autocracies performing poorly on women’s legal rights, relative to similar democracies. However, among regimes with quotas or with higher levels of economic rights for women, there is no difference between democracy and autocracy in terms of the proportion of projects that are gender mainstreamed.

Two aspects of our findings were unanticipated. First, in democracies, higher de jure women’s rights are associated with a lower proportion of gender-mainstreamed aid. We surmise that this reflects a “needs-based” logic, whereby donors reduce effort toward gender equality in (democratic) countries that already exhibit solid performance. But such an explanation warrants further investigation, given that mainstreaming policies are envisioned as applying universally and that there is surely room for improvement even in democracies rated highly on de jure indexes of women’s rights.

Second, in autocracies, the effect of laws related to women’s socioeconomic rights is less significant and robust than quotas. As Figure 3 illustrates, the proportion of mainstreamed projects is essentially flat across levels of the WBL index in autocracies, while the interaction term Democracy × WBL is not significant in the recipient–donor-year sample (Model 6). We surmise that this may reflect the unique status of quotas as a policy that has been a top priority of donors for decades (Bush 2011; Edgell 2017; Bush and Zetterberg 2021). Moreover, the introduction of quotas is a milestone whose effects in terms of female representation are easily measured. The WBL index, in contrast, captures laws in a range of areas—e.g., freedom of movement or citizenship—whose implementation and effects are more difficult to measure (Bjarnegård and Donno 2024). Further research could explore how donors monitor and respond to women’s socioeconomic rights, and whether their engagement on these issues may differ depending on contextual factors, such as whether the recipient country’s legal system grants authority to religious or traditional courts that undermine the implementation of reforms.

Conclusion

Gender mainstreaming is a prominent element in many governments’ development agendas, yet there has been little effort to study its application cross-nationally. The language of mainstreaming policies is universal, requiring donors to implement a gendered perspective in all projects across all recipients. Yet, this vision is at odds with the well-known politicization of foreign aid. In this article, we have focused on one political factor—the degree of recipient government buy-in—that shapes how vigorously donors employ gender mainstreaming across democracies and autocracies.

Our evidence indicates that, when designing their projects, donors anticipate and give weight to the need for gender mainstreaming, but they also anticipate implementation problems. Patterns of selectivity in gender mainstreaming are therefore shaped by considerations of where a gender focus is likely to make a difference. We find that, when recipients do not indicate buy-in to act on gender issues through equality-promoting policies, regime type is an important determinant of the allocation of gender aid: donors engage in gender mainstreaming to a greater extent in democracies than in autocracies. However, when recipients have introduced equality-promoting laws and policies, the amount of gender-mainstreamed aid is comparable across regime types. We hope that scholars will proceed to theorize more fully the various domestic factors—beyond de jure reforms—that indicate buy-in for gender equality programming in recipient states.

Our findings do imply that some of the “hardest” cases—autocratic recipients with poor performance on de jure women’s rights—see lower levels of gender mainstreaming; a potentially concerning trend given that autocracies are increasing in number around the world (Luhrmann and Lindberg 2019). Critics might interpret this, cynically, as donors avoiding engagement in hard cases in order to minimize the appearance of ineffectiveness. We suggest instead that donors adjust to the reality that these cases are those where gender projects have a greater chance of creating backlash and even putting women at risk.

Nevertheless, we emphasize that even among gender-regressive autocracies, mainstreaming remains quite common: Our analysis predicts that among authoritarian recipients at the lowest quartiles on measures of maternal mortality and labor market participation, an average of about 20 percent of ODA commitments are gender-mainstreamed—far higher than the 2–3 percent of commitments that primarily target gender in these cases. We conclude that although gender mainstreaming is applied selectively, it is the most prevalent way in which DAC donors seek to advance the empowerment of women; and that it is indeed employed in many cases where progress for women is sorely needed. Going forward, this may be changing, as the United States abruptly altered its foreign aid priorities under the second Trump administration in 2025. It remains to be seen how other DAC donors respond to the change in US policy and whether, in the longer term, the United States will renew its commitment to advancing women’s empowerment around the world.

Our findings have implications for the study of aid effectiveness. Methodologically, future analyses of gender mainstreaming should account for the selective nature in which it is applied. On a conceptual level, our findings suggest the need for further theorizing about when, and how, donors focus their policy-related efforts on cases where need intersects with domestic receptivity. How do practitioners define and identify such cases? Do practitioner understandings map onto researchers’ estimates of aid effectiveness? We also suggest that efforts to assess the effectiveness of gender mainstreaming may be hampered by a poor fit with existing indexes of women’s empowerment, which tend to capture high-level national trends. Future research may benefit from the development of outcome measures that vary across aid sectors and emerge more directly from the mainstreaming approach.

Another potential avenue for future research would be to explore donor behavior in other policy areas, such as climate issues. Similar to the GEPM, the OECD DAC has tracked aid projects targeting the objectives of the Rio Conventions on biodiversity, climate change, and desertification through the so-called “Rio markers.” The allocation of “climate mainstreaming” aid may follow a similar logic to the one we uncover for gender mainstreaming.

Beyond the study of foreign aid, our analysis offers insights into the importance of domestic policy cues, which may extend to other issue domains. It is known that international actors seek information shortcuts as a way to gauge domestic political will and intentions in a variety of areas, from sovereign creditworthiness (Gray 2009; Brooks, Cunha, and Mosley 2015), to investor protection (Jensen 2012), to commitments to clean elections (Hyde 2011). We suggest that domestic policy cues in democracies and autocracies can be interpreted differently by international audiences.

Funder Information

This project was supported by the Swiss National Science Foundation.

Author Biography

Simone Dietrich is an Associate Professor in the Department of Political Science and International Relations at the University of Geneva, Switzerland.

Daniela Donno is an Associate Professor in the Department of Political Science at the University of Oklahoma, USA.

Katharina Fleiner is a PhD candidate in the Department of Political Science and International Relations at the University of Geneva, Switzerland.

Alice Iannantuoni is an Assistant Professor of International Political Economy in the European School of Political and Social Sciences (ESPOL) at the Université Catholique de Lille, France.

Notes

Authors’ note: The authors thank Sarah Bush, Allison Carnegie, Mirko Heinzel, Helen Milner, Cleo OBrien-Udry, Matthew Winters, Alexandra Zeitz, and participants to the 15th Political Economy of International Organization annual meeting and the 74th Virtual Workshop on Authoritarian Regimes for helpful comments and suggestions.

Footnotes

References