Swaminathan S. Anklesaria Aiyar's article in the Economic Times makes a good point, but not, I argue, in the manner which he intended.
The article critiques the work by Utsa Patnaik (see this work for the upgraded version of her work), as methodologically wrong. I note here a number of weak points in his argument.
a) The article only takes into account the issue ratio of people below the "poverty line". However, Utsa Patnaik's article makes the case for the poverty figures by giving many interrelated and important statistics, including debt figures, decreasing rural credit, decreasing rural development spending, falling output and many others. Those are not addressed, a typical case of dismissing other people's critique as unidimensional.
b) The example which he gave (of Burger King etc.) is totally misleading. Those don't correspond to anywhere near the average dietary figures for Americans, or even the recommended ones. Also, again, he uni-dimensionalizes obesity. Obesity is not just caused by over-eating. I suggest Aiyar try out that diet sometime.
c) The main point of Patnaik's work is not that 1800 calories or 2400 calories is the "correct" poverty line, but that the line has been "clandestinely" decreased over the years. With that, the poverty figures don't make any sense. That hasn't been addressed.
d) There are hypotheses for explaining the fall in calorie requirements. How accurate are they? No discussion.
e) The article actually uses circular reasoning. There's rising income and falling poverty, so calories have fallen, therefore there's falling poverty.
f) Relying one just work which seems contrary to someone else's doesn't make for good argument. A variety of different issues/measures/theories have to be considered.
This, to me is not acceptable. We should have more debate on these issues. Some may not apportion blame to neoliberal reforms, but honest scholars should confront the issues nonetheless.
8 comments:
I don't think Aiyar's article is about estimating poverty at all (where other measures such as indebtedness might come in); it's about estimating malnutrition. As he puts it: "I do not think we need to revise the poverty line. What we do need to revise is the notion that our poverty line indicates insufficient food." This doesn't mean that people are growing wealthier (i.e. more asset-rich); it does however mean that fewer people are underfed.
If he means that, he does a poor job of that too. India does lots of malnutrition studies and Amartya Sen himself says that malnutrition in India is worse than sub-Saharan Africa. In India, 40-60 percent of children are malnutritioned. Anaemia among Indians is extremely high. These figures don't square with Aiyar's "happy trends". There's no systematic examination of malnutrion done by Aiyar. To take a perfectly obvious case: why aren't India's public health surveys cited? Just one statement doesn't do justice to this issue.
Now that's a different issue altogether -- if you want to argue that the levels of malnutrition in India are higher than Aiyar claims, fine. I'm aware that the World Bank restimated in 2005 that 60 million children were underweight in India. However, they also noted that food insecurity was not the primary problem; the primary problem was high levels of exposure to infection and inappropriate infant and young child feeding and caring practices, especially in the first two or three years of life. They suggested that more emphasis should be placed on changing family-based feeding and caring behavior than on just expanding coverage by the food distribution system. As I recall, they also reported that 80% of cases were in just 5 states, suggesting that the problem is mostly concentrated in North India, although no doubt virtually every Indian state has some high-poverty districts. Personally, I don't think sufficient reliable data exists to make a case for or against poverty reduction -- the problem is one of data collection, not ideology. As far as I can tell, most of this debate is based on the National Sample Survey and a few other datasets. But the sample selection and sample sizes leave us unable to work with enough hard data -- the NSS samples about 40,000 households, in a country of over a billion people; once ever five years, they sample 115,000 households. As I recall, district-level data is not even available from the NSS, only "region" data and "all-India" data. So if we really want to settle this, doesn't it make sense to actually collect more comprehensive data first? The census strikes me as the best mechanism for this.
Also, I'd be curious about your views on this piece by Pronab Sen, Principal Advisor to the Planning Commission: http://nutritionfoundationofindia.res.in/archives.asp?archiveid=210&back=bydate.asp
1. First about Aiyar and malnutrition. My principal point was that Aiyar is not talking about malnutrition, but poverty. I gave some issues with his critique.
If he's talking about malnutrition (whatever the truth may be), he's still not doing a good job. He should've been talking much more about health data and so forth. I don't want to go into malnutrition figures now. I just want to point out that if he's analysing malnutrition, he does a very poor job.
To my mind, he's talking about poverty and the relevance of calorie levels, not malnutrition.
2. I read the article by Pranab Sen with interest. I think it's a good critique of using calories as determining poverty levels. Table 3 quite convincingly makes the point.
Ah -- so then you believe that his data has reduced the "scope for interpretation and judgement"? ;) ;)
Actually, I'm not really arguing that economists don't make mistakes in practice (of course they do, as, for that matter, do physicists; in both cases, the errors are detectable over time through extended peer review), or, more fundamentally, that all problems have clear-cut solutions and that mathematical economists are competent to find them; that was actually more P.C. Mahalanobis' position (or that of the central planners more generally) that that of any Friedmanite or Hayekian. :) The Lucas Critique, for example, which was particularly damaging to Keynesian macro theory, was built on Adam Smith's simple insight that economic actors, unlike chess pieces, move on their own; i.e. economies are complex, dynamic adaptive systems. In other words, if you meddle and institute a policy experiment based on high-level aggregated historical data, (say, deliberate inflation to minimize unemployment, based on the historical correlation between inflation and unemployment), eventually your experiment will fail because economic actors will adjust their expectations accordingly (such as by adjusting their employment levels in response to upward-adjusted inflation forecasts). This is especially apt in financial economics, where the efficient markets hypothesis has been particularly influential because of the recognition that the competitive behavior of investors ensures that any particular anomalies (which represent opportunities to make windfall profits) are quickly arbitraged away by bandwagoning. My favorite article on this is Burton Malkiel's "Reflections on the Efficient Market Hypothesis: Thirty Years Later," available here:
http://www.e-m-h.org/Malkiel2005.pdf
It's pretty short; you might find it interesting to read. I've often thought that insights from financial economics could be fruitfully applied to macro.
I would read and give my comments on the article you gave, after my comments on Shleifer's article.
I'm certainly not discounting the process of peer review. That's why I tried to look in available scholarship for critiques of Patnaik's paper.
I should say that
a) Sen's critique of a very specific issue is one of the good ones I've seen yet. Many of the scholars have not commented upon it (perhaps they feel it's been already settled).
b) Table 3 in the paper is very convincing, but I'd like more supplementary evidence. There are still issues, for example, they use the 55th round data, in which the NSS changed the methodology, showing a much sharper poverty reduction. It would be good if subsequent round data are analysed. Himanshu and Sen have tried to analyse 61st round data and give their opinion that poverty reduction is much less than claimed. I haven't seen the paper yet.
c) The article rightly also examines and dismisses (rightly or not, I don't know), the critiques put up by Aiyar and others. In other words, even if the theorem is correct (maybe it is, maybe not), the proof by Aiyar (and many others) is not.
d) The article only addressed a specific question. As I tried to point out, the issues are much larger, including rural employment, debt and many others. I don't think a lot has been discussed about it in the media.
So yes, it's helpful, but far from settles the issue. The paper which had the most impact one me was by Angus Deaton and Jean Dreze (one of the foremost experts on hunger) in 2002 in which they point out that inequalities have been sharply rising and many social indicators like infant mortality decline are actually slowing down, even if official povety figures are mostly correct.
Sure; another paper of Malkiel's I'd highly recommend, which addresses various objections to the efficient market hypothesis in depth, is "The Efficient Market Hypothesis and Its Critics" (2003), at:
http://www.e-m-h.org/Malkiel2003.pdf
It's an interesting counterpoint to work on information asymmetries.
Post a Comment