The Supply of Advice Is Gendered—and This Is a Problem for Development Organizations

In our piece earlier this year about how and when policymakers engage with research, Janeen Madan Keller, Erin Collinson, and I made a point in passing that attracted some attention: evidence uptake in organizations is relational. What we meant was that, for the most part, policymakers don’t make decisions after simply sitting at their desk with a research paper or data, or at least don’t solely do so. They usually talk to people. But who they talk to matters a lot for what evidence they engage with, and what perspectives they hear.

Sometimes the people they talk to will be specialist researchers, but most decision-makers, most of the time, don’t have access to the world-leading researchers in their field. They may sometimes have a relationship with one or two academic researchers , but this will usually make up a small part of the advice they receive. Instead, they most commonly rely on advisers within their organization  whose job is to analyse the data and evidence and provide advice or counsel to more senior decision-makers.

But the relationship between policy advisers and the senior decision-makers they report to can be complicated. In forthcoming research with Pieter Serneels and Benedetta Musillo, which I presented preliminary results of at the Center for the Study of African Economies Annual Conference on March 19, we use a survey experiment with a large sample of policymakers to show that—in an anonymous organization that makes substantial spending decisions in developing countries—women are systematically less likely to volunteer advice to senior decision-makers than men. We asked them to consider a situation where senior officials are discussing a policy problem; one of them makes a questionable inference from the evidence presented, and advisers are asked if they would respond (and if so, how).

Among the 320 respondents to the survey, women are significantly less likely to say they would speak out. The effect size is large, around 9 percent of the average willingness to offer advice across all of those surveyed. Though the context we put them in is quite specific, with an implied challenge to a senior decision-maker, this is often what policy advisers need to do: challenge preconceptions with evidence or a closer or more specialist reading of the data. And the finding tracks with a great deal of research about the gendered working environment of large organizations. A 2020 survey in the US, for example, found that women were more likely to report ‘being overlooked’ in online meetings; a Harvard Business Review survey looked at 360 feedback on staff in senior roles in US private sector companies and found that performance in meetings was more likely to be a concern for women (findings corroborated by a follow-up survey). Victoria Brescoll’s study using archival data from the US Senate found a positive relationship between power and speaking time—but only for men, not women. Christopher Karpowitz and Tali Mendelberg’s The Silent Sex finds much the same. In academia too, women either speak less or are interrupted more.

Why does this happen? In our study, we try to dig a little deeper, and while we can’t be definitive, our findings are consistent with women having internalized that they will be taken less seriously than their (male) peers; such internalized norms can have large negative effects on behaviour, as Karla Hoff and Priyanka Pandey have found in a different setting. As they wrote, “the mechanisms of discrimination operate, in part, within the individuals who are members of the groups who have been discriminated against.” This idea of ‘stereotype threat’ was first proposed by Claude Steele, and has been extensively studied since then.

And if you think that different life experiences lead to different perspectives, and the propensity to ask different questions or look at different evidence (and the evidence suggests this may be so), then this suggests a problem: we want decision-makers to hear a range of views before making choices. If the evidence or advice they hear is systematically selective, then their decisions may be biased and suboptimal.

Our study didn’t stop there, however. We also investigated how randomly varying the gender composition of the senior decision-makers to whom advice was to be offered affected the supply of advice. The results here are less clear-cut, but they do provide pretty consistent (but weaker) evidence that more gender-diverse leadership boards induce a higher supply of voluntary advice provision. This result is not gendered: it holds across the sample we survey, and there is no specific effect of diverse leadership for women (at least, not from a one-off ‘engagement’ with a more diverse leadership group). In other words, a diverse senior management board induces more advice from across the organization.

Putting both of these results together suggests something really important to investigate further in leading development organizations. As my colleague Eeshani Kandpal recently pointed out, the leadership of IFIs exhibits a striking lack of gender diversity  (the research underlying this point is here). If our results generalize, this suggests that not only are these leaders getting too little input from the women in their organizations, but they aren’t getting the most out of the full set of experts and advisers they employ. Diverse leadership is not just a signal to the outside world about the organization; it’s also a signal to the people who work inside it, and for a field that is rightly preoccupied with gender disparities, that can really matter.

There’s much more to come from our ongoing research, which looks at how decision-making and gender interact. Watch this space.