Does development always eliminate gender discrimincation?

Naila Kabeer scrutinises the thinking behind the recent Nobel prize-winning methodology in development economics.

The 2019 Nobel Prize in Economics was awarded to three leading proponents of randomised control trials (RCTs) for pioneering a rigorous methodology for evaluating development interventions and contributing to a new development economics – so-called randomista economics.

RCT practitioners describe themselves as “humble plumbers,” eschewing theoretical frameworks and ideological positions, interested only in evaluating the question of what works in development as rigorously and objectively as possible. But this treatment of theory and method and its implications for how they weave evidence into stories of change creates limitations.

Their faith in the rigour of their methodology leads randomistas to claim that evidence of impact stemming from their trials can be attributed confidently to the intervention. Their view is, since they know what works, they do not need to investigate how it might work, for whom and why. So they dispense with theories that would help them to answer questions about causal processes and how context-specific they might be.

Randomista economics is the antithesis of the interdisciplinary approach and pluralist methodology of the economics that feminists have been working with.”

Again out of faith in their approach, randomistas dispense with alternative ways of researching impact. They dismiss the use of methods that could help them to tease out causal mechanisms. They also dismiss findings of studies that have not used experimental methods, however relevant these might be to locating their own findings in a broader analysis of change.

So randomista economics is the antithesis of the interdisciplinary approach and pluralist methodology of the economics that feminists have been working with. In addition, feminist economics is centrally concerned with the structures of patriarchal constraints and the invisible workings of power in people’s lives – aspects that are difficult to capture within an experimental format.

The Random Way
Randomised control trials consist of controlled experiments that randomly decide who, of a group of people eligible for the intervention, will participate in it and who will be excluded. The impact of the intervention is then determined by comparing selected measures of change between the two groups.  Random assignment is believed to minimise the likelihood that the two groups will differ from each other in systematic ways that might bias the changes they report. These biases were a major limitation of previous evaluation methods.

A critical reading of a recent paper from one of the Nobel Prize winners, Esther Duflo, illustrates the differences between current feminist economics and the randomista’s way of working.

Duflo’s essay deals with what she described as a “strident” debate between two positions. The randomista position argues that broad-based development is sufficient for the promotion of women’s empowerment and there is no need for policies explicitly intended to promote gender equality. The feminst line argues that women’s empowerment makes an important contribution to development. The corollary is that development will benefit from gender-specific equality measures.

In reality, most feminist economists agree that broad-based policies can promote gender equality on certain fronts but argue that additional measures are needed to dismantle discrimination across the board.

Duflo positions herself in this debate as an “honest broker,’”rigorously sifting through the evidence to determine the validity of these competing claims. However, it becomes clear from her selective use and interpretation of evidence that she favours the former position and is sceptical about the need for affirmative policies. Let me summarise her arguments and offer a feminist refutation.

Do economic policies promote gender equality?

In support of the claim that broad-based policies can increase gender equality, Duflo cites Claudia Goldin’s analysis of historical evidence from the US that shows how rising female education and aspirations combined with the rise in “respectable” service sector jobs brought about a “quiet revolution” that transformed family relations and women’s work. Duflo suggests that similar processes at work in the Global South. She cites evidence from China that shows that women’s access to new wage opportunities increased girls’ survival rates and women’s education levels while in India, the availability of IT-enabled service sector jobs was found to increase female education and employment.

She is aware that girls die in larger numbers than boys in India but attributes it to poverty rather than gender discrimination. Broad-based anti-poverty measures would address this problem.  However, she does acknowledge development does not always eliminate gender discrimination. She notes the resort to female-selective abortion in India and China which have given rise to sex ratios at birth that are highly skewed towards males. She suggests that this is a rational response on the part of parents to new technologies that reduce the costs of female-selective abortion in the face of rising costs of dowry needed to marry off daughters.

Duflo is aware that girls die in larger numbers than boys in India but attributes it to poverty rather than gender discrimination.”

As a feminist economist, I have major problems with these arguments. First, I object to her use of Goldin’s study to bolster her own claims. We can have considerable confidence in Goldin’s claim of a quiet revolution in the US economy because it is based on rich longitudinal data going back to the 1880s and combines it with case studies of different phases of this period. By contrast, Duflo’s claim of a quiet revolution in India is based on very thin evidence: a household survey in an urban neighbourhood in 2001 and a three-year RCT of a number of villages near Delhi in 2003. This is not an adequate basis to claim a revolution.

More importantly, there is no quiet revolution happening in India: on the contrary, economic growth in India has been accompanied by a steady decline in female participation in the labour force. It has also been accompanied by worsening sex ratios at birth, as has been the case in China.

Nor does it make sense to attribute this to rational responses by parents weighing up the declining costs of female-selective abortion versus the rising costs of marrying -off daughters. Gender discrimination in these countries is rooted in historical patriarchal practices and explains why girls are still more likely to die than boys in India as well as why parents are resorting to female selective abortion.

Does gender equality promote economic development?

According to Duflo, the evidence for arguments in favour of targeting women in the distribution of development resources rests on studies that found that women’s preferences meant that they were more likely than men to invest resources at their disposal in their children’s health and education and hence in the productivity of the future workforce. While she dismisses older studies using non-experimental methods that reported these findings, the more recent experimental studies she cites come to the same conclusions. Nevertheless, she remains sceptical, pointing out that such behaviour could reflect the fact that women with access to resources were already likely to be somewhat exceptional or that they were more likely to be married to altruistic men: we might be observing the effects of male altruism.

The question of preferences comes up again in her argument questioning political quotas for women. She notes that studies have found that women elected to office allocate public budgets differently, but argues that this does not mean that they allocate it better. Women leaders in West Bengal voted for better roads and improved access to drinking water, while men voted for education, but how do we decide which is better for development?

She concludes that while there might be intrinsic grounds for policies that directly address gender inequality, the instrumental grounds are weak. Policy makers need to be aware that there are likely to be development costs: efficiency might be better served by either targeting men or favouring broad-based interventions.

Women leaders in West Bengal voted for better roads and improved access to drinking water, while men voted for education, but how do we decide which is better for development?”

What might a feminist response to these arguments be? First of all, feminist economists do not share the belief of mainstream economists that individual preferences are idiosyncratic and randomly distributed so that evidence of female altruism can be countered by evidence of male altruism. The generalisation that women are more likely to use resources at their disposal for the benefit of their children is not just a matter of quantitative findings. It has also been suggested by anthropological research which points to strong cultural norms in a wide range of contexts which assign responsibility for family well-being and care of children to mothers. Preferences are thus, at least partly, a product of social contexts and will reflect the gender relations prevailing in that context. This may not always lead to altruistic behaviour by women. Contexts of strong son preference are also contexts where women rely on sons for status and security in old age. In such contexts, it is very possible that mothers will discriminate against their daughters in favour of their sons.

Secondly, it is not clear why elected women leaders are being judged on the basis of their budgetary priorities, when elected male leaders have rarely been held accountable for their use of public money. In any case, the argument for having quotas for women is not an argument about economic development, but about the quality of democracy, an effort to correct for women’s historic disadvantage in the political domain.

In her conclusion, Duflo cites studies that argue that women gained economic rights in the US, not through affirmative policies, but because men as fathers decided at a certain stage of the country’s development that it would benefit their daughters. She appears to believe that the same processes are likely to happen in developing countries. We could take her advice and wait for paternal concern to materialise in these countries– or we could continue to mobilise women and men to support feminist demands for a fairer world.

  • This is a shorter, more accessible, version of Naila Kabeer’s fully- referenced acadmic article, which can found here.

Naila Kabeer

Naila is joint professor in the departments of International Development and Gender Studies at the London School of Economics. She is a feminist economist and works on various issues relating …

Read More »

One Comment on “Plumbers and pedagogues”

  1. The statement that in India more girls die than boys lacks logic. After they both reach a mature age they will eventually pass away. Do you mean to say that the average age that young women die in India is lower than the average age when young men die in India? If so then we need to know at what age limit this no longer applies, after the majority of the un-dead girls and boys have grown to adulthood. The trouble with many humanists doing economics is that they don’t care about how inexact their use of language is and that is one reason why they give this subject a bad name and they themselves are ignored.

Leave a Reply

Your email address will not be published. Required fields are marked *