Above: Nobel laureate Abhijit Banerjee with Prime Minister Narendra Modi in New Delhi/Photo: PIB
The pioneering work of the Nobel-winning trio in alleviating poverty using the medical model of Randomised Controlled Trials can help the government deliver services more efficiently and tackle difficult challenges
By Sanjiv Bhatia
The 2019 Nobel Prize in Economics was awarded to three economists, including one of Indian origin, Dr Abhijit Banerjee, for their unique approach to identifying effective ways of providing services like education and healthcare and delivering welfare subsidies like food rations, minimum employment guarantee and universal basic income.
Underlying their research is the development of what is known as experimental economics. Unlike the physical sciences, where one can get exact answers through repeated and controlled laboratory experimentation, economics deals with people and does not lend itself to such controlled testing. As a result, traditional economics has generally been normative, relying exclusively on theory, logic and analysis of past data.
The experimental approach proposed by Banerjee and his cohorts uses randomised controlled trials (RCTs) similar to those used in testing the efficacy of new drugs. Two groups of people are created, one the experimental group to which a new policy initiative is administered and the other a control group which is administered an alternative initiative. Differences between the responses of the two groups are then used to understand the efficacy of the applied policy initiatives. Such trials are being used to evaluate the impact of a wide variety of projects—everything from water purification tablets to microcredit schemes, financial literacy classes to teachers’ performance bonuses.
The answers from these experiments could eventually benefit the poorest of the poor—the 865 million people (roughly 13 percent of the world’s population) who live on less than Rs 16 per day. Some critics of their work have argued that their experiments use subjects that may not be “aware” enough to make rational choices. The science of economics is based almost entirely on the assumption that people make rational choices—economic models do not fit nicely if this assumption is violated. Banerjee defends his work by arguing that because the poor have so little, they have to put more careful thought into their choices. With Rs 16, they have a choice of two pounds of rice or four bananas, and they have to choose wisely. Rationality of choice is, therefore, built into their compulsion.
An example of the kinds of issues Banerjee and his cohorts study is the problem of teacher absenteeism in government schools in India. According to a national survey, on any given day, almost one in four (25 percent) of teachers are absent from their jobs. To better understand how to incentivise teachers in these schools, Banerjee created two groups: schools in which the teachers were offered short-term contracts with incentives for extension tied to performance and in the other, lower pupil-teacher ratio. They found significant improvement in both the attendance of teachers and the performance of their pupils when teachers were hired on short-term contracts with incentives tied to performance.
Other experiments done by Banerjee on improving educational outcomes show that addressing children’s current learning gaps, rather than following an over-ambitious uniform curriculum, can lead to significant learning gains. The “chunauti” programme introduced by the Delhi government in 2016 to check student dropout rates and improve education quality was based on this experimental research. The programme divides students into groups based on their learning abilities, with remedial tutoring by the better teachers to the weakest students. Subsequent research shows that targeted remedial education provides educational outcomes far superior to free books and mid-day meals.
Banerjee’s research on the rural job guarantee programme, MGNREGA, is also quite revealing. Contrary to popular belief, he finds that MGNREGA has been extremely helpful to the rural poor. Almost 2.2 billion person-days of work was generated by the programme, covering around 50 million households and over 200 million people. Nearly 80 percent of the money ended up in the right hands. It also reduced urban migration, by allowing people in rural areas with access to MGNREGA to stay at home rather than go to the city to find work. But as the programme is inflexible—people can’t find work whenever they need because public projects to accommodate their labour cannot be created instantly—Banerjee recommends universal basic income which pays every poor person a minimum guaranteed income so they have something to fall back on without having to deal with the vagaries of MGNREGA. This understanding was the genesis of the NYAY scheme introduced by the Congress party before the 2019 election.
Another large government welfare programme where experimental economics can provide useful answers is the food distribution system in India which provides subsidised rations to the poor. Given the many difficulties in storing and moving food and the “losses” in the distribution process, should the focus be on providing food? Is it possible to think of it as a general income support programme for the poor and leave the design of individual programmes to the states? Let each state decide, based on outcomes from RCT studies if they wish to give out food or food stamps or just skip it all and hand out cash. Or perhaps, combining MGNREGA and the food distribution programme into a universal income scheme might be the most efficient and cost effective way to help the poor.
These kinds of experiments in social economics can greatly help the government deliver services more efficiently. The efficacy of alternative approaches can be studied using sample groups and the data analysed to see which method works most efficiently and consistently. RCT is a tool that must be incorporated into the provision of government services, especially as it relates to the distribution of welfare among the poor.
Governments have the power to do enormous good but also tremendous damage. Policy makers that use a thoughtful and evidence-based approach to policy making can effectively tackle difficult challenges. But those that force ad hoc, top-down politics-driven decisions on the people can cause irreparable harm.
Banerjee’s approach to finding localised solutions also tangentially endorses the Gandhian philosophy of local government (Panchayat Raj). Banerjee’s work provides intellectual support for Gandhi’s vision by arguing that experiments like the ones they conduct involving local people must be an integral part of policy making, especially when it pertains to the delivery of services to the poor. Evidence-based policy making must replace politics-based policies. This will require politicians and bureaucrats to spend significant time in impoverished areas to better understand what works and what doesn’t. Making policies sitting in an air-conditioned office in New Delhi will never solve the problems of poverty.
The RCT approach is not without its detractors. Critics like economist Jagdish Bhagwati argue that while the experimental approach may appear ideal for small-scale projects, it is unsuitable for evaluating broader macro-economic, political and institutional issues that are the root cause of poverty. These larger systemic issues—increased economic freedom, stable capital markets, labour reforms—may have a more profound impact on poverty than focusing on individual behaviour. Others, like Nobel-winning economist Angus Deaton, dismiss the RCT approach entirely stating that “demonstrating that a treatment works in one situation doesn’t necessarily imply that it will work the same way elsewhere”.
There are also issues of possible disconnect between the time and money required to study policy interventions and the short-time horizons of donors and governments. A policy intervention may prove to be highly effective in a small control group but maybe rejected on political grounds. Notwithstanding the criticism, Banerjee’s research shows that many government programmes proclaimed to be magic bullets of yesterday become failed ideas of today.
It is, therefore, vital that countries like India, with millions in poverty and limited public resources, have an independent project/budget evaluation office to provide policy makers with non-partisan facts on the true cost and effectiveness of government welfare and poverty alleviation schemes. Banerjee’s work suggests that small pilot studies can be conducted by specialists to understand the benefits and the costs of proposed government interventions. Once the government starts a programme, it is seldom terminated, and it is, therefore, best to evaluate its effectiveness upfront using experimental studies.
Dr Abhijit Banerjee, his wife Esther Duflo, and Michael Kremer rightly deserve the prestigious Nobel Prize in Economics for their pioneering work in alleviating poverty using the medical model of RCTs to aid in answering the question “what works”. Their research adds another dimension to the expanding field of economics by dividing big issues into outcome-focused experiments at the individual and group level. In acknowledging their contributions, the Nobel Committee wrote: “The research conducted by this year’s Laureates has considerably improved our ability to fight global poverty. In just two decades, their new experiment-based approach has transformed development economics.”
—The writer is a financial economist and founder, contractwithindia.com