Foundational Investments for Gender Equality: A Call for the Investments in Structures and Policies that Make Other Interventions Work Better

Financial support to gender equality has had a rough 18 months. Aid budgets have shrunk and skeptics once at the margins of the development conversation have been emboldened. Funding for population and reproductive health was projected to drop 35-53 percent in 2025, and UN Women estimates low- and middle-income countries (LMICs) face a $420 billion annual shortfall in what is needed to achieve gender equality. But gender equality remains Goal 5 of the Sustainable Development Goals and a stated priority for most multilateral institutions and many governments. This makes economic sense: closing the gender gap in labor force participation and management roles could meaningfully raise global economic activity.

In this resource-constrained environment, the pressure to demonstrate returns on every dollar has heightened. That pressure is reasonable, but the class of investments that determine whether those dollars work as intended may not be as readily able to demonstrate returns. This class of investments, which we call “foundational investments,” aims to redress the systems that produce inequality in the first place. This essay is about such foundational investments, which are effectively “public goods” for other investments because they enhance the effectiveness of everything that follows. In the sections that follow, we discuss what such investments consist of, why they sometimes go un- or under-funded, and what we propose funding now.

What we mean by "foundational for gender equality"

Foundational investments in gender equality build or strengthen the institutions, norms, infrastructure, data, and public systems that determine who has access to rights, opportunities, and power. The return on these investments is often diffuse, shared by society, and long-term. Their returns also often accrue to actors other than the one who pays. Whereas a targeted intervention delivers something discrete that has value to recipients (cash, bed nets, vaccines, teacher training), these investments change the conditions—by leveling the playing field—that drive the success of future interventions. The defining feature is complementarity: foundational investments raise the returns on subsequent investments by generating the right enabling conditions for the interventions that follow.

Funders are often willing to pay for targeted interventions because they can point to the specific goods and services delivered (and the associated costs). Improvements in the enabling environment do not always work this way, which is why they may go unfunded. The benefits accrue over time and across institutions. They may also enhance the effectiveness of later interventions funded by institutions that did not contribute to the foundational investment itself. As a result, these types of investments lack clear ownership and accountability. And attribution is also harder, making it tricky for funders to report to stakeholders on the impact of a foundational investment. This results in a market failure of sorts and means that these investments can be overlooked even when the aggregate returns can be substantial.

Reforming discriminatory laws around property, inheritance, and labor market participation—and changing the norms that leave laws unenforced even after they change on paper—are all examples of foundational investments. Others include the rules, norms, and accountability that shape who reaches leadership positions—which ultimately influence the broader effectiveness of development institutions that implement interventions. Some governments have also tried embedding gender analysis directly into budget decisions, known as gender-responsive budgeting, though more evidence is needed on its effectiveness. Collecting the data—such as the Demographic and Health Surveys (DHS)—is also essential to understand and ameliorate women’s and children’s well-being, gendered effects of policy. An interesting additional case is that of the childcare, eldercare, and health systems that determine women’s unpaid work burden. Even though child- and elder-care slots are consumed by individual families—and thus have impact that can be attributed to individual interventions—care systems also deliver a significant portion of their returns to outcomes that whoever funded them cannot claim.

Foundational investments enable cost-effective interventions to function. Health and family planning interventions are a clear example of smart, cost-effective interventions: injectable contraceptives, menstrual cups, and emerging products including lenacapavir (a new long-acting antiretroviral medication used to prevent and treat HIV) are all very cost-effective on paper. But stigma and social norms shape whether women can access the care and technology in question, which in turn determines whether the expected returns for the intervention will be realized. And the returns to reducing stigma are captured by every intervention targeted at women that follows. In each of these instances—reforming laws, shifting norms, investing in care systems—decentralized, intervention-specific funding thus delivers less of all three than the aggregate social return would justify.

What cost-effectiveness can miss—and how to fix it

Prioritizing investments based on cost-effectiveness makes sense when the benefits can be easily attributed, but many efforts to advance gender equality resist such clean measurement. Interventions often measure impacts at the household level rather than at the individual level, obscuring the within-household dynamics that gender programming is trying to shift. And interventions designed to address another objective—such as improving early childhood development or safe water access—frequently have gendered effects, especially on women’s time, which can be overlooked by conventional outcome measurement. For instance, such effects may only show up in time-use surveys, and not every country publishes these regularly. And many of the benefits of gender equality—greater agency, freedom from violence, or more time to study or work—are multidimensional and have historically been difficult to measure.

The encouraging news is that we have, in fact, made a lot of progress in understanding which data are needed to capture such gendered effects. The Women’s Empowerment in Agriculture Index and its successors have standardized multidimensional agency measurement, and more recent work has focused specifically on embedding empowerment modules into nationally representative, multi-topic surveys rather than bespoke survey instruments. Standardized time-use modules capture more than we have ever known about women’s lives in LMICs. Measurement of intimate partner violence has improved substantially. The binding constraint is no longer the “how to measure”—it is that the instruments are not deployed at scale and not funded.

One actionable and high-impact action the development community could take right now is to fund the measurement infrastructure. The DHS is the clearest illustration of the complementarity we have been describing. It was designed to provide high-quality, nationally representative data on population, health, and nutrition to help countries make evidence-based policies. And it has indeed been used to detect the effects of policies and programs across health, education, and social protection that its designers never anticipated. Yet DHS funding has fallen dramatically since January 2025, and the program risks lapsing when a temporary Gates Foundation grant ends, despite the fact that 30 SDG indicators depend on its data and countries with limited statistical capacity have no substitute for it. The same can be said about the funding of time-use surveys and dedicated violence-against-women surveys, whose funding is also on shaky ground. Investing in these data systems is cost-effective and immediately actionable.

A more fundamental problem is that cost-effectiveness analysis takes the status quo as its starting point, and that status quo already reflects existing gender inequality. Take the example of the disproportionate amount of unpaid care work performed by women. If we treat women’s unpaid work as a fixed condition of the status quo, not something that can be changed via policy, the most "cost-effective" intervention could be one that directs resources toward men simply because they have fewer constraints on their time. The analysis may be sound in terms of what generates the highest return on investment, but it does nothing to question or change the unequal underlying conditions. Thus, investments to shift the baseline, such as reducing and redistributing unpaid care work through care policies, remain critical to ensure that we are not further entrenching gender inequality into the system.

An illustrative example of why we need foundational investments: Cash transfers and intimate partner violence

Cash transfers are among the most evaluated instruments in development economics, and their effects on intimate partner violence (IPV) have been studied rigorously. If a gender-related outcome were likely to yield a broadly applicable cost-effectiveness estimate, we should find it here. What the evidence offers is arguably more useful: a clear picture of what cash does, and a reasonably good idea of how to make it do more.

A meta-analysis finds meaningful reductions in physical, sexual, and emotional IPV and in controlling behaviors, and a mixed-method review found most studies showed cash decreasing IPV. This is good news and it is missing from many cost-effectiveness calculations for cash, which means we are systematically understating the returns from transfers to women. So, the first thing to do is straightforward: include IPV impacts in the cost-effectiveness calculations for cash transfers.

The second is to use what we already know about design. The Transfer Modality Research Initiative in Bangladesh randomized cash and food transfers, with and without an intensive nutrition behavior change communication (BCC) component. Transfers alone produced no sustained effect on IPV. Transfers with BCC reduced physical violence, even four years after the program ended. These reductions operated through durable increases in women's threat points and in the social costs to men of perpetrating violence. The BCC arm did not include any gender or violence messaging at all. Such design choices are inexpensive and can be the difference between a transfer that reduces violence and one that does not.

How cash affects IPV thus depends on transfer design, the spouse’s response, household structure, and prevailing norms and laws. It is important to keep in mind that "the effect of cash transfers on IPV" is, thus, a function of local context and can evolve over time, rather than a parameter that we can estimate in one setting and apply in another.

The heterogeneity itself is informative. The more useful question is which features of a setting determine what cash does—because those features are themselves things to invest in, particularly in places that lack those features. In other words, what determines whether cash reduces violence is whether a woman has somewhere to go, some legal recourse, and a social environment that doesn't treat her earning as a threat. Those conditions are worth funding, and they are worth funding precisely where they are the weakest: the impacts of cash may well look the weakest where norms work against women, but these are the women that most need it.

The international financial institutions still have a lot of leverage

Governments and international financial institutions (IFIs) have the mandates to internalize such spillovers. Whereas a bilateral donor is accountable for its own portfolio, a government is accountable for outcomes across its whole population, and IFI lending is typically in support of a country's policy framework rather than merely towards the results of a single project. That difference in mandate is why responsibility falls where it does. Whatever the "end of multilateralism" conversation concludes, the IFIs remain among the largest actors in the gender financing space, and continued investment in them must be part of any “future of development” strategy.

The broadest form of financial support that an IFI like the World Bank provides to a government is general budget support—Development Policy Financing (DPF). DPF disburses against prior policy and institutional reforms: specific reforms a government completes before funds are released. This is what makes DPF distinctive for our purposes. Investment lending can and does build foundations, but what it cannot do is change laws, budget processes, or institutional mandates. DPF, in contrast, is one of the few instruments that can do these things, including strengthening public financial management, embedding gender-responsive budgeting into a country's budget process, and building institutional capacity.

The World Bank's own retrospective on DPF names gender equality as one of the areas where the instrument has been put to use—and goes on to suggest that DPF should be used to fund childcare policies, systems, and institutions to address gender-based violence and women in leadership and decision-making. Now, because DPF funds a government's whole budget, tracking exactly what happened to the "gender" dollar within it is close to impossible. This is not to say that tracking gender at the IFIs is not important—it should still be a priority. But this is in itself an argument for building gender into a government's budgeting systems and prior-action frameworks, rather than only as donor-dependent project funding.

Concretely, we would like to see gender reforms appear more often as prior actions in DPF operations, and we would like the World Bank’s outcome-oriented gender tagging applied to policy lending, not only to investment projects.

What evidence do we need for foundational investments?

Make no mistake, these foundational investments in gender equality cannot be made blindly. No "vibe funding": we recognize that we need a lot more evidence here. But evidence on foundational investments will require more than the randomized controlled trial, which is the field's dominant evaluation tool for good reasons—but a poor fit for the diffuse benefits provided by foundational investments.

Take inheritance law reform. Such reform cannot be randomized, so it resists clean identification. Besley and Burgess (2000) used the staggered adoption of land reform across Indian states to estimate effects on poverty, and their strategy would be viewed askance today—largely because we now understand how unreliable two-way fixed effects can be when treatment timing is staggered and effects are heterogeneous. Arguably, the rising methodological bar has also made such questions less likely to be asked.

But India's inheritance reforms also show why questions around laws and norms still must be asked, even if through a broader range of empirical methods. Despite a national amendment and gains to girls’ education, labor supply, and physical autonomy, evidence suggests that it raised the cost of having daughters and worsened female feticide. Implementation research also finds that eight years after the law change, women still do not inherit land at equal rates. The point is that legal reform needs to go in tandem with norm change—and that we can only see this gap in the Indian case because someone funded three decades of non-experimental work on a staggered rollout. Assessing foundational investments therefore requires a different research toolkit—process tracing, comparative case studies, difference-in-differences and quasi-experimental designs, long-horizon monitoring of institutions and norms. Such research can offer policymakers a broader toolkit to achieve the structural change necessary for more equal societies. This research is valuable in itself and requires deliberate investment.

Conclusion

Rigorous evidence on cost-effective interventions must guide the development community, and this is certainly not an argument against such investments. It is, however, an argument that these interventions will underperform without certain enabling conditions—conditions that can only be made feasible by foundational investments in leveling the playing field. Governments and multilateral organizations have the mandate to fund these—and the evidence that assesses them—and must do so deliberately and in a coordinated manner.

We are encouraged by new entrants in this space alongside the institutions that have funded gender equality work for decades. But focusing solely on the most immediately impactful intervention risks missing the opportunity to achieve longer-lasting and more systemic change. There is room for both types of investments, and we encourage donors, including emerging tech philanthropists, to consider making foundational investments for gender equality. The measurement infrastructure is a good place to start: it is inexpensive, the tools already exist, and it allows us to understand the state of investments in gender equality.

With thanks to Kathleen Beegle, Diva Dhar, Dave Evans, Rachel Glennerster, and Clem Landers for feedback.