In a tribute to Angus Deaton, the 2015 Nobel laureate in Economics, Diane Coffey and Dean Spears – former graduate students of Prof. Deaton at Princeton University – review some of his work on the well-being of the poor in India, and discuss the paradoxes and puzzles that still remain.
About two weeks ago, Angus Deaton became the 2015 Nobel laureate in Economics. Apparently, when the prize committee called, Angus could not manage to stop thanking them. Eventually, the committee kindly had to put a stop to it: they thanked him, and explained that he burnishes their Nobel prize by accepting it.
The field of development economics certainly has much for which to thank Angus Deaton. Indeed, he helped build the micro-empirical platform upon which essentially the entirety of contemporary development economics stands. While constructing such tools and teaching us all to use them by his example, he has contributed clarity and wisdom to some of our field’s most important questions.
He challenges us all to choose paradox, hard work and honesty when it comes to our own empirical questions. The two of us were lucky enough to receive these lessons in person as graduate students at Princeton. And, of course, whether one is in Princeton or India or anywhere else, Professor Deaton challenges us through the clarity and wisdom of his writing. Much of this writing has been about well-being in India. In this article, we highlight some of Angus Deaton’s ideas for India. As a tribute to Professor Deaton, we review some of his work on what measurements reveal about the well-being of the poor, and what paradoxes and blind spots still remain.
Household surveys and aggregate statistics
Some of the work that Angus Deaton is most widely known for throughout the field of economics is careful thought about the relationship between aggregate statistics and what is learned from household-level survey data. This is particularly important for India, where it is widely publicised that GDP (Gross Domestic Product) growth has recently been fast, but fewer people are aware that consumption growth, as measured in household surveys, has been much slower. For example, India Human Development Survey data
were recently released, which allow a comparison between consumption in 2005 and 2012. Over this seven-year period, World Bank data claim that GDP per capita grew at a rate of over 10% per year; yet, the survey data show that household consumption per capita grew by less than 5% per year. This is still a remarkable improvement in average standard of living, but paints a somewhat different picture than what the aggregate data say.
A part of what Deaton has taught economists is how to use surveys to answer economic questions. Before there could be impact evaluations and identification strategies to compute the effects of programmes and policies on households’ outcomes, we had to learn how to measure those outcomes. Deaton never hides from those details, or their consequences: for example, whether consumption is measured with a 30-day recall period or a seven-day recall period makes a difference. If we want to use subjective well-being survey questions, Angus warns, we must understand the large context effects of the prior questions asked in our survey form. If we want to compare the real cost of a rural latrine that costs Rs. 5,000 in India with one that costs Rp. 600,000 in Indonesia, we cannot ignore the fact that consumption surveys tell us that people buy different things in those different places (Deaton et al. 2004
). But Deaton’s work on survey data is far from a mere list of warnings and forbiddances: he has received over 1,000 citations on Google scholar (over 100 in 2014) for his 1985 paper proposing linking of repeated cross-section surveys (such as a panel dataset) with geographic fixed effects (Deaton 1985
). This technique is now so widely used in applied microeconometrics - typically with a difference-in-differences identification strategy layered in between – that many economists implementing it in this fall’s batch of job market papers may not realise that it ever had an origin as a creative idea.
In his more recent work, Deaton has moved beyond consumption to an attempt to understand well-being in all its multidimensionality. A good example of Angus’ ability to say things that are obviously true - yet somehow different from both sides of an existing debate - is his contribution in The Great Escape
to the debate over whether Pareto improvements should really be considered improvements . Political progressives sometimes want to overthrow the conservative idea that only improvements that make nobody worse off are improvements. Angus suggests that Pareto improvements in well-being should indeed be considered social improvements, but that improving wealth or consumption is not the same thing as improving well-being. Making a rich person richer in a way that expands his political power over the poor (but doesn’t actually make him noticeably more satisfied with his life) may not be a Pareto improvement after all – but if not, the fault is not with Pareto. Deaton’s work with psychologist Daniel Kahneman suggests that this is no mere theoretical possibility: reported life satisfaction plateaus as income increases. Moreover, changes in average health can be very different from a country’s changes in average wealth – unfortunately, here too, India is a case in point (Deaton and Kahneman 2010
Puzzles and paradoxes of Indian development
Despite all of Professor Deaton’s contributions, there is still much that is not understood about well-being in India. Some of this is because of open puzzles and unsolved paradoxes. Puzzles about the failure of wealth to translate into health for Indians are unfortunately especially abundant. Deaton has shown that, cohort after cohort, year after year, younger generations of adult men in India are becoming taller than older generation at a faster rate than Indian women are becoming taller. This strongly suggests some form of discrimination against girl children in health or nutrition, and many researchers, including Professor Deaton himself, have documented forms of gender discrimination in India, but much remains to be learned about the growing height gap (Subramanian and Deaton 1991
Another nutritional puzzle is that average calorie consumption in India has been declining over the past two decades, even though households are getting richer, as Deaton has shown in collaboration with Jean Drèze (Dreze and Deaton 2009
). This seems to be at odds with basic household economics; in particular, the fact that the decline in calorie consumption has accompanied expanding incomes suggests that it is a choice, or a revealed preference. Dean, in joint work with Josephine Duh, finds that some of the decline can be accounted for by an improving disease environment which causes a smaller fraction of the calories consumed to be wasted by disease, such as diarrhoea – but even this leaves much to be explained. Similarly, other recent research by Shari Eli and Nicholas Li find that declining requirements of energy for work cannot explain more than a fraction of the calorie decline.
Data and democracy
Unfortunately, some of the open questions about today’s India are not open because the data present paradoxes: on too many matters, our ignorance increasingly reflects the fact that there is almost no health data at all. Deaton has written much about height of people in India and around the world, but this analysis does not include Indian height data collected since 2005 – because this was the last time India bothered to collect a Demographic and Health Survey. In today’s India, data are increasingly unavailable, in what appears to be a policy choice by successive governments. When data are eventually released (such as in the case of the Rapid Survey of Children) it is only in the form of aggregated summary statistics, the inadequacies of which have been demonstrated by Deaton’s work on inequality and the difference between households surveys and aggregates.
In this environment, perhaps one of Deaton’s most important ideas for India is what he said in his Nobel prize press conference: such policy decisions are political, and often reflect “vested interests”. It is no coincidence that the words state and statistics sound so similar. As Deaton wrote in his statement to the Indian press last week: “my work shows how important it is that independent researchers should have access to data, so that government statistics can be checked, and so that the democratic debate within India can be informed by the different interpretations of different scholars. High quality, open, transparent, and uncensored data are needed to support democracy.”
In a country of over a billion people, there are thankfully many ideas for Indian development; data would allow them to be checked against fact.
- Deaton, Angus (1985), “Panel data from time series of cross sections”, Journal of Econometrics, 30: 109-126.
- Deaton, Angus and Jean Dreze (2009), “Food and nutrition in India: Facts and interpretations”, Economic and Political Weekly, 44(7): 42-65.
- Deaton, A, Alatas, V and J Friedman (2004), ‘Purchasing power parity exchange rates from household survey data: India and Indonesia’, World Bank.
- Kahneman, Daniel and Angus Deaton (2010), “High income improves evaluation of life but not emotional well being” Proceedings of the National Academies of Science, 107(38): 16489–93.
- Subramanian, Shankar and Angus Deaton (1991), "Gender Effects in Indian Consumption Patterns", Sarvekshana, 14:1 - 12.