Featured image: A view of Accra, Ghana. 2016. Photo taken by Sara Stevano.
Abhijit Banerjee, Esther Duflo, and Michael Kremer won the 2019 Nobel Prize for Economics. They can be credited for a revolution in development economics driven by the use of experiments to design social policy for development. They have been praised for their contribution to poverty alleviation. Their work has been central to the elevation of randomised controlled trials (RCTs) as the gold standard method to design development interventions.
When I showed Duflo’s TED talk to my students, most of them were persuaded by the narrative put forward. An empirical approach yielding straight-forward, concrete, and effective answers to well-defined questions – a clear pathway to fight poverty. The success of this approach, corroborated by the decision taken by the Royal Swedish Academy of Sciences yesterday, is evident much beyond my classroom, with the proliferation of experiments across the development industry, among policy makers and in academia. The power of the narrative is something that we, critics, need to take more seriously.
The influence of the randomistas’ work is undoubtable. If there is anything to be celebrated, it is probably that this Nobel Prize is directing the attention of the wider public to the persisting global challenge of poverty at a time when nationalistic and inward-looking sentiments are on the rise across much of the West and in some fast-growing economies in the South.
However, this is also a momentous time in which demands for a radically new way of thinking and doing economics are expanding and getting heard. Many economists now agree that we desperately need a new economic agenda (see this on the UK, for example). The shift is palpable in the mainstream too, and, only last month, the Financial Times launched its ‘Capitalism, time for a reset’ agenda. In this light, the celebration of Banerjee, Duflo, and Kremer appears to be unaware, or inconsiderate, of the voices asking for change.
It is true that many consider behavioural approaches to economics, on which the randomistas’ methods rely, to break away from standard economic theory, mainly through a departure from standard rationality assumptions. However, heterodox economists and political economists have written on how behavioural approaches to economics and, for that matter, development, continue to draw on the same framework, and, in fact, crown the rational agent as the benchmark all others need to aspire to (see Mehta, 2013).
In development economics, this approach merits further scrutiny for the moral connotations it casts on the poor. Poverty is seen as creating a condition of scarcity of psychological resources, as put even more simplistically by the World Bank, ‘the constant, day-to-day hard choices associated with poverty in effect tax an individual’s […] mental resources. This cognitive tax, in turn, can lead to economic decisions that perpetuate poverty.’ Thus, poverty – or specific aspects of it – can be addressed by correcting the poor’s cognitive biases, as argued in this paper on Kenyan farmers by Duflo and colleagues.
This approach draws attention away from the structural conditions underpinning poverty – such as the production of primary commodities, trade imbalances, lack of productive employment opportunities, scarce public provision of social services, to name just a few – and instead focuses on individual behaviours. In this sense, the randomistas’ approach, far from being theory-free and exclusively empirically-driven, ignores much work done in the heterodox streams of development economics focusing on the structural causes of poverty and inequality, and hence falls in line with the neoclassical focus on individuals (see Fine et al., 2016). Alongside the structures, the role of politics in development processes is also ignored.
Several ethical issues entailed in the use of RCTs have been also exposed (see Ravallion, 2018). For example, the arbitrary exclusion of some people – the control group – from receiving a ‘treatment’ they may need or want. Here, I would like to stress an overlooked point: the double standards in the North and in the South (see this blog by Fletchner). For much discussion on safeguarding the ‘libertarian’ in libertarian paternalism – i.e. ensuring freedom of choice while nudging people in the right direction – in the North, the same concern does not apply for people in the Global South. In this context, it is simply assumed people should do as the intervention says, for their own good – of course, we know better! The racialised and colonial dimensions of assuming what is best for others cannot go unnoticed.
Finally, the limitations of RCTs as a method, its domain par excellence, have been highlighted. The critics have debunked the myth of RCTs as a gold standard, showing that other methods allow for the same type of inference and exposing the pitfalls of measuring the average treatment effect. It has also been argued that other methods are needed to understand ‘why something works or not’, in addition to ‘what works’ (see Cartwright, 2007; Deaton and Cartwright, 2018; Kabeer, 2019). This highlights the urgent need for researchers and practitioners to recognise the relevance of case studies that provide a thorough understanding of context, process and relations.
The 2019 Nobel Prize is a celebration and validation of the randomistas’ work. But it is also an opportunity to draw the attention of the development community and the wider public to the plurality of voices that constitute the field of development economics. The debate between the randomistas and their critics is important and it is necessary to engage with it. But, this should compel us to strengthen our ability to understand and analyse the complexities of economic development, rather than jump uncritically on the randomistas’ bandwagon.
Other resources providing a critical perspective on RCTs and behavioural approaches in development economics include Barrett and Carter (2010), Reddy (2011), Maxwell (2012), Ravallion (2012), Akram-Lodhi (2013), Sial and Alves (2018), Bédécarrats et al. (2019), de Souza Leão (2019), and an upcoming paper by Stevano, Johnston, and Codjoe – watch this space!
Sara Stevano is currently a Research Associate at King’s College London and will join the SOAS Economics Department in January 2020 as a Lecturer in the Economics of Africa.