In a tribute to Angus Deaton, recipient of the 2015 Nobel Prize in Economics, Reetika Khera, who did her post-doctorate research at Princeton University under Deaton, outlines his India-specific contributions both as a rigorous economist and a public intellectual. She discusses Deaton’s great concern with measurement issues, and the over-reliance on randomised controlled trials as evidence for policymaking; and his support of government action for social policy.
Much of the work by Angus Deaton, the winner of this year’s Nobel Prize in Economics, has been focussed on measurement issues. He has questioned the quality of data collected in large surveys and suggested ways of improving the surveys. He has also thought very hard about how these data could or could not be used, how to reduce measurement errors, and what inferences one can, or cannot, draw from data that might suffer from measurement errors.
Boring as that may sound, it has ensured that economists pay attention to detail, and do the hard work that empirical analysis demands. Good data is fundamental to good economics.
One example of this is from India. His contribution to the understanding of price indices and relatedly, poverty estimation, has been very important. He has highlighted the problems with the computation of price indices in India and how these affect poverty estimation. His proposal to use prices implicit in the data collected by the National Sample Survey Office (NSSO) was implemented by the Suresh Tendulkar committee some years ago.
His book on these issues, The Analysis of Household Surveys, published in 1997, remains the best to learn about data issues. Along with poverty estimation, he has applied his deep understanding of several disciplines (ranging from biology to philosophy) to work on mortality, health, nutrition and well-being. In his latest book, The Great Escape: Health, Wealth and the Origins of Wellbeing Inequality, he generously acknowledges Amartya Sen’s influence on this aspect of work.
India-focussed work on poverty
Much of this body of work on nutrition in India (and elsewhere) has come since the 2000s. A large part of it is India-focussed, much of it co-authored with Jean Drèze. In 1999-2000, there was a lot of interest in the poverty estimates as these were the first post-liberalisation estimates. Supporters of liberalisation were keen to show that poverty had declined, and that its rate of decline had accelerated since liberalisation. Those against it were unwilling to accept this.
A change in the methodology for data collection between 1993-94 and 1999-2000 made a straightforward comparison between the two point estimates impossible. Deaton’s work (with Drèze) on comparable estimates disappointed both camps. They found that while the official claim that poverty had declined in the post-liberalisation period was true, the claim that there had been an acceleration in the rate of decline was not.
In 2009, Deaton and Drèze published a paper in the Economic and Political Weekly (EPW) on the nutrition situation in India which once again provoked economists on both sides of the ideological spectrum. On the one hand it led to a debate, in the same journal, with Utsa Patnaik — widely considered to be on the Left — who felt that the decline in calorie consumption was a symptom of rising poverty. They, however, pointed out that calorie intake was declining even at given levels of real per-capita expenditure, especially among the better-off households, and discussed other possible reasons for this pattern.
On the other hand, in 2013, Arvind Panagariya challenged a long-held understanding among economists and nutritionists about anthropometric outcomes. The argument was not so much about whether height is a good indicator of nutrition, human development and well-being. Panagariya’s thesis was that Indians are short, not because they are undernourished but because they are “genetically programmed to be so”. Deaton and his co-authors pointed out in their response, again in the EPW, that “all of his arguments about the role of genetics is residual: if we cannot think of anything else [we assume that] it must be genetics.”
These two debates are illustrative of Deaton’s careful analysis and unwavering honesty, his ability to separate his social commitments from what his meticulous data work suggests. Indeed, at a press conference at Princeton University just after the prize was announced, he said that sometimes his work leads him to “very uncomfortable” places.
Questioning the dominant trend
Another important part of his contribution has been to question dominant fashions in development economics. When ‘instrumental variables1 ’ were a popular tool to establish causality, he wrote “students no longer look for a thesis topic, but for an instrument”.
More recently, a new technique ‘randomised control trials2’ (RCTs) has taken development economics by storm. For some, RCTs are seen as the only form of evidence, disregarding not only other forms of quantitative evidence but also other forms of qualitative evidence. Deaton has been among the few voices to question our over-reliance on RCT experiments, while acknowledging it as a valid body of evidence.
“I argue that experiments have no special ability to produce more credible knowledge than other methods, and that actual experiments are frequently subject to practical problems that undermine any claims to statistical or epistemic superiority,” he said.
The “randomistas”, a term used by The Economist, have been influential in several states in India. The Government of Tamil Nadu has signed a Memorandum of Understanding (MoU) “to institutionalise an evidence-based approach to policymaking”. Some experiments have led to greater hardship for the poor — for example, delays in wage payments to the Mahatma Gandhi National Rural Employment Guarantee Scheme (MNREGS) workers — without any accountability for these adverse outcomes.
For Deaton, the problem is not with RCTs per se — some of his current work is on how to use RCTs — but rather with the view that it is the only form of evidence that matters or that it should be the only driver of policy decisions. He has his differences with arguments like that advocated by Abhijit Banerjee when he said that “the World Bank should cease to fund any activity, including presumably macro policy advice, that has not been previously subject to evaluation by an appropriate RCT.”
Advocate for a greater State role
Apart from claims of statistical or epistemic superiority, Deaton has questioned the divorce between policymaking and public discussion. His contribution has not only been as a rigorous economist, but equally as a public intellectual. He is a firm supporter of government action for social policy. Without being blind to the problems of governments in poorer countries, he forcefully argues for their greater accountability.
“The absence of state capacity — that is, of the services and protections that people in rich countries take for granted — is one of the major causes of poverty and deprivation around the world. Without effective states working with active and involved citizens, there is little chance for the growth that is needed to abolish global poverty,” he argued.
On the consequences of inequality, while some inequality can be a good thing, Deaton has argued that too much could potentially have negative consequences for us all.
He noted that:
“The very wealthy have little need for state-provided education or health care… They have even less reason to support health insurance for everyone, or to worry about the low quality of public schools that plagues much of the country…To worry about these consequences of extreme inequality has nothing to do with being envious of the rich and everything to do with the fear that rapidly growing top incomes are a threat to the well-being of everyone else.”
These messages are important because public debate in the past few years in India have tended to belittle various forms of public support through terms like ‘doles’, ‘freebies’, and ‘handouts’. Repetition succeeded in creating an impression — unsubstantiated by facts — that India has gone overboard in its social spending. We have witnessed strong rhetoric on evidence-based policymaking, combined with a resolute disregard for inconvenient facts. We need to engage with Deaton’s contributions very carefully.
This article appeared in The Hindu on 17 October 2015.
Instrumental variables are used to estimate causal relationships when there are measurement or confounding errors in observational data.
A Randomised Controlled Trial (RCT) is a type of impact evaluation process to determine whether a programme has an impact, and, to quantify that impact. RCTs measure programme effectiveness typically by comparing outcomes of those (individuals, communities, schools, etc.) who received the programme (treatment group) against those who did not (control group).