Gender in the MDB Social Protection Response to the COVID-19 Pandemic

Recent work by the CGD researchers through the COVID-19 Gender and Development Initiative suggests that the COVID-19 pandemic has impacted women particularly severely, disproportionately affecting their paid and unpaid work, income, and food security, while also exacerbating the risk of gender-based violence. In response to the pandemic, multilateral development banks (MDBs) rapidly expanded their financing across a range of areas, including health and social protection, but little is known regarding whether new operations have been designed to address the crisis’ disproportionate impacts on women. This note builds upon previous CGD research to examine how MDB operations in response to COVID-19 have considered gender differences in their design, and how that design has evolved across institutions, with a particular focus on social protection measures in lower-income countries. We focus our analysis on social protection projects, recognizing this as the primary policy lever of choice by governments and supporting donor institutions as they sought to mitigate the risks of rising poverty and food insecurity in the COVID-19 context, as well as critical component of policy packages capable of promoting an inclusive recovery. Similar analyses could also be undertaken for projects related to broader economic development, health, education, or other sectors.

For now, the data at our disposal merely reflects whether MDB operations teams have designed projects with the intention of measuring projects’ reach and benefits by gender, not the ultimate effectiveness of projects in narrowing gender gaps. Our analysis is limited by the data publicly available to us through MDB project documents. Recognizing that it will take some time for MDBs to track the implementation of their COVID response projects and make this information publicly available, we take the initial step of reviewing project appraisal documents (PADs) or the equivalent to gauge the extent to which projects include gender-specific indicators and/or targets. We recognize the limitations of this approach and intend to build upon this initial analysis in future work as more project implementation data becomes available.

Considering loans to selected countries from March through December 2020, we find that, of the 64 projects with adequate documentation and identified social protection elements, 47 (73.4 percent) contained targets focused on reaching and benefiting women and girls. Projects utilizing such gender-specific targets were generally of higher dollar value than their counterparts in the sample. We find also that the presence of gender-specific targets varied over time and institutions, with the Asian Development Bank (ADB) and African Development Bank (AfDB) consistently incorporating them from the outset, and the World Bank (WB) only later developing a gender-specific focus.

The prevalence of gender-specific indicators and targets

CGD researchers, through the COVID-19 Gender and Development Initiative, conducted an analysis of project appraisal documents (PADs) or equivalents from the World Bank, Asian Development Bank, and African Development Bank for countries that are IDA eligible or in which women were identified through a Data2X and Open Data Watch analysis as particularly vulnerable to the direct and indirect impacts of COVID-19. Publicly available PADs for projects related to COVID-19 within this sample of countries were assembled and scrutinized for both gender-related indicators (those that call for data to be disaggregated by gender or those specific to a given gender) and targets (those indicators which include a goal). These provisions were then categorized and tallied. (Several PADs from the Inter-American Development Bank were also reviewed but the number was too small to derive meaningful conclusions. This partly reflects the fact that the IADB has fewer IDA clients than the other three MDBs.)

The review included a total of 142 operations in the sample countries with publicly available PADs dated from March to December, 2020, representing an estimated commitment value of $15.61 billion (Figure 1a). For context, early in the pandemic the stated intention of the World Bank was to issue $104 billion in COVID-19 related financing. Sixty four of the examined 142 projects (45.1 percent) included social protection measures. Of those 64 projects, 52 (81.3 percent) were found to have at least one indicator focused on reaching or benefiting women and girls and 47 projects (73.4 percent) included at least one gender-specific target. Only 29 projects utilized targets aiming for women’s inclusion at or above parity with men (e.g., “Number of persons receiving temporary unemployment benefits; of which 56% {are} women”).

Figure 1b shows the corresponding information but weighted by commitment value. The percentages of MDB operations with gender-specific indicators and targets are higher in this second figure, indicating that larger operations are more likely to include gender-specific elements. Out of the $15.61 billion committed to our sample operations, $12.15 billion (77.8 percent) was allocated to projects with at least one social protection measure. Of those $12.15 billion, $11.25 billion (92.6 percent) was dedicated to a project with at least one social protection-related gender indicator, and $11.03 billion (90.8 percent) was assigned to a project with at least one social protection related gender target. These numbers do not, however, identify the commitment values actually subject to gender targets because a single project can include several components.

Indicators or targets that aim to collect intersectional data (i.e., those that call not only for disaggregation by gender but also by migrant status, disability status, or age, for example) were present in very few of the social protection-related projects that we identified. While most social protection programs include some targeting measures meant to identify populations with multiplicative forms of disadvantage, such as women in poverty, which is inherently intersectional, we have included in this analysis only those projects which collect data on intersectionality in their results frameworks. Intersectional data is needed to determine which populations are being reached and impacted by project activities, particularly where the goal of the project is to mitigate negative impacts that are borne by the most marginalized populations in any given context. Research has shown that in addition to gender (and indeed in combination with gender), the pandemic and ensuing economic crisis has had differential impacts along race, class, migratory status, and age lines. Of the 64 projects with social protection measures in our sample, just 17 (26.6 percent) included intersectional indicators, as opposed to 52 projects (81.3 percent) that collected data on women and girls overall. Only 11 (17.2 percent) of the 64 projects included an intersectional target, far fewer than the 47 projects (73.4 percent) with a target for women and girls generally. Unlike the indicators and targets regarding women and girls overall, the share of dollars dedicated to projects with intersectional indicators and targets is not substantially greater than the share of projects that include them. Of the $12.15 billion committed to projects with social protection measures, $3.46 billion (27.7 percent) went to projects with intersectional indicators, and $3.03 billion (24.2 percent) to projects with intersectional targets.

Figure 1a. Number of MDB Projects with PADs. Figure 1b. Value of MDB Projects with PADs

The nature of gender-specific indicators and targets

Within the 64 operations incorporating social protection measures, a total of 105 gender indicators related to social protection were identified (Figure 2), including 79 specific targets. Nearly half of these indicators related to cash transfers, by far the most frequent type of social protection activity. Such measures are typified by the World Bank’s response program in Guinea, which required that 30 percent of the recipients of emergency cash transfers be women. Food and other in-kind support measures were prominent as well, representing 16 percent of the gender-related indicators identified. An ADB program for the Marshall Islands ordered that “Food baskets (be) delivered to 2,380 households in the neighboring islands and atolls—of which 25 percent are headed by women—every month for at least 6 months.” The remaining indicators encompassed a range of other areas, including health insurance, housing, and utility assistance.

Few of the targets related to programs designed to benefit women exclusively. When such targets were incorporated, they often related to the scaling up of existing programs aimed at women, rather than the creation of new ones. An ADB loan in Bangladesh, for instance, called for an additional 250,000 women to be enrolled in the National Allowance Scheme for Widowed and Husband-deserted Distressed Women. Indeed, the majority of indicators measured women’s inclusion in programs targeting broader populations, as did an ADB loan to Samoa stipulating that “Unemployment benefits (be) paid to at least 3,000 workers, of whom at least 30 percent are women.” Roughly one third of the targets for women’s inclusion in social protection programs singled out women as heads-of-household. The AfDB’s loan to Cape Verde, for example, called for the “Expansion of the Social Inclusion Income Programme for vulnerable households…with at least 40 percent headed by women.”

Figure 2. Gender Indicators by Social Protection Measure

The timing of gender measures

Within our sample of countries, the overall MDB response to COVID-19 proceeded largely in two waves, and reflected somewhat different approaches across the institutions. (Project dates were compiled on a month-year basis. Project approval dates or documentation dates were used, depending on availability.) An initial rush of lending activity occurred in the 2nd quarter of 2020, with the first two COVID-19-related loans coming in March, followed by 50 in April (valued at $3.11 billion), the largest project total of any month (Figures 3a and 3b). Lending remained elevated through the 2nd quarter of 2020, but gradually declined into the 3rd, reaching a nadir of three loans in September. The 4th quarter of 2020 then witnessed a second surge in lending (a total of $3.3 billion in November), which subsided by the end of the year. These cycles were driven primarily by the actions of the World Bank and ADB. During the first phase, the World Bank produced a large number of relatively low dollar value loans with little apparent emphasis on addressing the gendered impacts of the crisis. The ADB, in contrast, produced a smaller number of larger value loans during this period, all of which did incorporate gender-specific targets. The World Bank was largely responsible for the second wave of lending, producing a smaller number of high value loans that demonstrated greater emphasis on reaching and benefitting women than earlier World Bank projects. The ADB maintained its gender focus in this period, directing nearly all of its resources to projects with gender targets. The AfDB, while unable to match the loan value of its two peers, deployed their loans following similar seasonal patterns, while consistently incorporating gender-specific targets into its operations.

Figure 3a. MDB Projects Over Time, by Bank & Gender Target Inclusion

Figure 3b. MDB Project Value Over Time, by Bank & Inclusion of Gender Target

The 64 projects that included social protection measures largely followed the same pattern (Figures 4a and 4b). Between April and June of 2020, 33 projects with social protection measures were approved, with a total commitment value of $6.96 billion. Of these, 24 (roughly 73 percent) included gender targets specific to social protection. The World Bank produced 14 of the projects in this period, totaling $1.4 billion, half of which came in April. Only five of these 14 projects contained gender targets specific to social protection. In contrast, the Asian Development Bank generated 11 projects totaling $4.49 billion between April and June, all of which included gender targets. The AfDB produced the remaining eight projects during this period, valued at $1.05 billion, with a sudden peak in May that tailed off in the following months. Although small in comparison to the other two MDBs, the AfDB incorporated gender targets into all of their operations in which social protection measures were present.

There was then a lull in social protection related lending through the 3rd quarter, followed by a second surge in the 4th. Between July and October of 2020, only 17 projects were generated between the three banks, roughly half of the previous three months. November, however, saw 13 projects totaling $3.17 billion produced, 77 percent of which had gender targets. The World Bank was largely responsible for this second wave, representing eight projects (six with gender targets) and $2.77 billion. The Asian Development Bank contributed five projects, all but one containing gender targets, but totaling a comparatively small $0.39 billion. The AfDB mustered a small increase in their lending during the 3rd quarter as well, but in October, not November. Their 3rd quarter boost consisted of two projects incorporating social protection and gender targets, totaling $0.07 billion.

Figure 4a. Number of SP Projects Over Time, by Bank & Inclusion of Gender Target

Figure 4b. Value of SP Projects Over Time, by Bank & Gender Target Inclusion


The main conclusion to emerge from this initial assessment is that a substantial fraction of social protection operations launched to support countries through the COVID-19 pandemic have included some form of gender-related indicators and/or targets. Such targets generally considered gender in isolation, as few included intersectional elements. In terms of social protection, the most prevalent form of provisions by far involved cash transfers, in large part assigning quotas for support to women and girls (for example, the World Bank stipulating that 30 percent of emergency cash transfers in Guinea be for female recipients). It was also not unusual for gender-related measures to be embedded in in-kind programs, generally those providing food parcels or rations. Measures incorporating gender are more prevalent when weighted by loan value than by loan numbers, indicating a tendency for them to be more common in larger operations. Indeed, weighted by loan value, gender-related indicators and targets were present in the vast bulk of lending. However, we are unable to reliably measure the specific allocation of funds within the projects for social protection versus other areas of operation, nor can we accurately determine what percentage of the funds dedicated to social protection supported gender-sensitive measures.

Breaking down the total by institutions and over time suggests some patterns. The Asian Development Bank was determined to act aggressively and did comparatively well at considering gender concerns from the beginning of the crisis. Urgency did not entirely crowd out consideration of the pandemic’s impact on women and the need to find ways to help cushion it. The World Bank, while lagging initially, developed a greater focus on gender over time, and arguably did a good job of phasing and deploying resources over the course of what has turned out to be an extended crisis. The AfDB has been the most consistent in including gender-related measures in its operations.

The assessment does not address the effectiveness of these MDB activities, nor the quality or appropriateness of the indicators and targets chosen by operations teams. Rather, these results speak only to the presence of an attempt by the MDBs to respond to the gender divide exacerbated by COVID-19.

Nor does this study delve deeply into the intentions that drove lending patterns. In some cases, operations drew on previous programs that had included gender elements. In others, it is likely that they reflected concerns over the impact of the pandemic and subsequent lockdowns on families and children. Fears for the well-being of women and their families are well-grounded. School closures, for instance, disrupted many feeding programs, creating an additional burden on household budgets and food preparation requirements. It is possible that the traditional role of women as care-givers was also considered. Multiple studies (see here and here) have demonstrated that women are performing more unpaid care work, both relative to before the pandemic and relative to men. Such concerns were rarely made explicit in projects’ results frameworks, however, as only one of the 109 gender-specific indicators identified focused on expanding access to childcare.

Other important limitations to our conclusions should be noted as well. First, our findings do not speak to whether gender considerations were included at the behest of MDBs, of recipient governments, or in response to shared concerns. Second, the study does not encompass all MDBs or other financial institutions. Third, not all lending within the MDBs examined is included, only loans that are within our sample and with publicly available PADs as of December, 2020. Fourth, as noted above we are unable to identify how much of the approved lending was directed specifically towards social protection measures or subject to gender-sensitive provisions. Fifth, we have not yet tracked the timing of the disbursements. Sixth, information is not yet consistently available confirming progress towards the gender-related targets that we have identified.

For all these reasons, our findings should not be interpreted as a judgment on the success of MDB lending in cushioning gender-related impacts of the pandemic, but instead as an initial rough sketch of quantifiable metrics. Moving forward, we intended to continue this analysis, using new information to address these questions as opportunities arise.

Read the PDF version of this CGD note for annexes listing MDB operations in response to the COVID-19 pandemic and a list of gender indicators.

Rights & Permissions

You may use and disseminate CGD’s publications under these conditions.