The Indian Economy Blog

August 23, 2005

Six Eggs A Year

Filed under: Growth,Labour market — Amit Varma @ 4:13 am

A common canard of the Left:

Rural India is in acute distress, which is bound to turn to turmoil if its crisis is not addressed. It is not too late. There is a strong case for a universal employment guarantee and a universal Public Distribution System.

This is from the strap of a recent op-ed article by Utsa Patnaik in the Hindu. Well, Aadisht Khanna finds fault with some of her assertions, and takes her on. First, he summarises the points she is making:

All right, so what is actually going on? Basically, Professor Patnaik has made the following assertions:

* Rural India is facing an employment crisis
* This is because of the economic policies pursued in the past fifteen years.
* The proof of this is that people are eating much less grain.
* The assertion that people are eating less grain is borne out by data from the National Sample Survey, which measures consumption and expenditure across India.

Okay. Let’s take this up.

Aadisht goes through the data, some of which he lists, and finds that Patnaik’s conclusions are drawn from selective data, and are, thus, simplistic. He writes:

There is a decline in rice and wheat consumption, and also in the consumption of dal… But at the same time, the consumption of other stuff has risen- milk, vegetables of all sorts, meat of all sorts (though fish has shown the most dramatic rise), and most notably eggs- the consumption of those has doubled.

And this suggests something that you would expect a Professor of Economics to know- the consumption pattern looks suspiciously like that of Giffen goods.

Normal goods are the ones which you buy more of when you have more money. Giffen goods, on the other hand, are goods which you buy less of when you have less money- because you now cut down on your consumption of that good, and use the savings to buy more of something else.

What’s the classical example of Giffen goods used in economics textbooks? That when your income rises, you buy less bread and more meat- exactly what we see happening in rural India from 1988 to 2000.

He examines if the averages are skewed by the rich getting markedly richer, but finds, in other NSS reports, that “the consumption of people in the lowest five income percentiles has also increased.” And how much more does rural India consume now? About six eggs a year.

(Link via email from Ravikiran Rao, who also pointed to two relevant articles by Swaminathan Aiyar: “Do the Poor Face Higher Prices?” and “It’s not just calories, stupid“.)


  1. It’s not as if a Professor of Economics did not know about Giffen goods. Let me quote from an article of hers from October 2004 [Link].

    “[A]ny notion of grain becoming an inferior good refers to the expenditure on grains and its products for direct consumption, and not to the total absorption of grains which includes both direct use as well as indirect use (the latter as feed for livestock to produce milk, eggs, meat and so on, plus use of grain as industrial raw material). The absorption of foodgrains per capita, because it is for all uses, is always found to rise, not fall, as the consumer’s average income rises. Our figures of availability (which are a little higher on purpose than the official figures of availability) refer to absorption of grain for all purposes. (These are calculated directly from output data, which is the hardest data we have, which is adjusted only for trade and for change in stocks, so by definition it has to meet all possible final uses).”

    “The fact that the availability, or absorption of foodgrains per head, always rises as a nation’s per capita income rises, because it is for all uses, is supported by an extensive literature on the responsiveness of demand for cereals to rising incomes, and by the FAO ‘food balance sheets’ which give data over time for output, trade and stocks by individual crops, and cover virtually every country. China, with about double India’s per capita income, absorbed 325 kg. per capita of foodgrains (excluding tubers) in the mid-1990s compared to India’s less than 200 kg. at that time. (but China’s grain absorption too has been falling in recent years as unemployment grows and inequalities rise ).”

    “Mexico absorbed 375 kg.per capita, high income Europe absorbed over 650 kg. per capita and USA absorbed the maximum, 850 kg. per capita of which less than a quarter was directly consumed and the rest converted to animal products, processed or put to industrial use (Calculated by the author from FAO, Food Balance Sheets for 1992-94.)”

    “The recent trend in this country of sharply declining foodgrains absorption per head while average per capita income has been rising, is thus highly abnormal, not only in the light of international experience but also in comparison with our own past experience – we have always seen rising grain absorption per capita as average incomes rose in the past in India.”

    In a post with several links, it would have been convenient for the reader if you linked to the
    other NSS reports as well.

    Comment by Anand — August 23, 2005 @ 7:00 am

  2. If she knows about Giffen goods, Anand, and doesn’t take it into account while writing her piece, then it would seem that she is not ignorant but deceitful. I wonder which is worse.

    Comment by Amit Varma — August 23, 2005 @ 7:22 am

  3. 1) The very data she is relying on should tell her that the same people who are consuming less of foodgrains are consuming more of other items. Whether this makes it a case of Giffen goods or not, the fact remains that this is evidence that she should have considered before claiming that there is some sort of crisis in the rural economy.

    2) Anand, when we were arguing over John Dreze’s report, you took great care to argue that just because the reduction in poverty occurred during a period of liberalization, we shouldn’t attribute a cause-and-effect relationship between the two. Will you extend the same courtesy to evidence that apparently (but, as it turns out, not actually) supports your viewpoint? i.e. will you take care to say that just because the reduction in calorie consumption happened during the time of liberalization, we shouldn’t jump to conclusions and attribute a cause and effect relationship between the two?

    3) The second Swaminomics link (about it not being the calories stupid) implies that there has been a secular reduction in calorie consumption for the past thirty years. If this has been the case, then don’t you think it is highly dishonest to attribute the reduction to liberalization? And if calorie consumption has been reducing right through a period when there was no doubt about the reduction in poverty (i.e. during the green revolution) then surely there is some other explanation for the reduction than unemployment and some sort of rural crisis? Isn’t it more likely that Aiyar’s explanation is right, i.e the poor are shifting to superior, but lower calorie foodstuff?

    Comment by Ravikiran — August 23, 2005 @ 7:43 am

  4. Anand, that paper you’ve linked to is bizarre. Nowhere does she even mention the fact that the consumption of other foodstuffs is actually on the rise (which is the crucial fact Aadisht points out.) She has simply ignored the data that does not fit her conclusions.

    Comment by Ravikiran — August 23, 2005 @ 8:01 am

  5. Hi Amit
    I think this is interesting. The issue is how do you know if a person is better off or worse off by just looking at his diet. I don’t know for sure if that can be determined. Obviously, if he starves, he must be worse off. If he gets fatter, he probably is better off. But other than the extremes, it might be difficult. For example, if someone has been eating little other than rice and daal for years and not enough of that to fill his stomach, the first thing he might eat if he gets a little more income might be candy. That’s not a good nutritional choice but I guarantee that the marginal utility of that first piece of candy in years would be enormous.

    This is why it is a good thing to try to measure income and not calories, although I admit that income might be hard to measure accurately if a lot of goods are bartered instead of bought in the market. If income goes down, there is no doubt the person is worse off. If income goes up, then there is no doubt that the person could be better off, although I suppose it is possible that he might make poor choices with that extra income (but that isn’t for me to judge).

    If the income measure is essentially impossible to calculate in the village economy, then I think the way I would try to measure diet in a much more detailed fashion. I would poll maybe a 1000 people and really track their diet in detail and try to understand that diet as well as possible. But I can see that trying to infer income from diet is a complex subject.

    Comment by Michael H. — August 23, 2005 @ 9:47 am

  6. Amit — I think Prof Patnaik was neither ignorant nor deceitful. In an Op-Ed piece, her priority was naturally in outlining her thesis. In theory she could have added the paras that I have quoted, in her piece too. But then I myself thought about that part only when Aadisht raised that issue.

    Ravi — (1) I don’t disagree with you on that right now as I haven’t seen the relevant data myself. I have requested Amit and Aadisht to provide the relevant links so that anyone who’s interested in the details can see it for oneself. I’ll probably agree with you on that count. But Aadisht’s assertion that the consumption of people in the lowest five income percentiles has also increased, does this really follow from the very data she is relying on?

    (2) My point was that there was no drastic reduction in poverty in the post reform period to attribute it to reforms. The rate of change was more or less the same in the pre and post reform period. In this case, by Utsa Patnaik’s article, rate of change of absorption per capita of grains was positive in the pre-reforms era whereas it’s negative in the post-90′s. So there’s a difference between the two scenarios. I also had a small comment here based on the Deaton-Dreze paper which will help to clarify what exactly was my stand — right or wrong — on that issue.

    (3) I find Aiyar’s argument a bit problematic as he uses the reduction from 55% to 26% (people below poverty line) as the basis for his calculations. Those figures are not comparable as methodologies of computation were different. Strictly speaking, Aiyar’s stats says that there’s a reduction in the calorie consumption in the last 30 years. It doesn’t give the pre-91 – post-91 split which is what is important for this discussion.

    Now suppose Aiyar is right (the way you have interpreted him). And the Giffen goods argument holds. Coming back to point #2, that says that people have been shifting to superior foods more or less the same way in the pre-reform era as in the post-reform period. In that case the assertion that the poor have progressed far better under reforms than before does not hold, right? (Amit, I’m quoting you there).

    (4) Again I’ll be happy to get the link to the source which says that the consumption of people in the lowest five income percentiles has also increased. Otherwise I’ll comment on it once I dig up that data myself.

    Comment by Anand — August 23, 2005 @ 9:59 am

  7. Anand, I just did a back of the envelope calculation as to how much the consumption in the lowest five income percentiles has increased. Assuming you consume nothing but rice, you would have consumed 44 kilos more in 2000 than in 1994.

    The calculations are there as a comment in my original blogpost.

    Comment by Aadisht Khanna — August 23, 2005 @ 10:50 am

  8. Umm… no. I was relying on Aiyar’s assertion that there has been a “gradual” reduction in the number of calories consumed between 1972-73 and 2000. If “gradual” means what it means, then the change must have been more or less steady over the past 30 years.

    And if you read the article further, you will realise that the fall in calorie consumption has been highest for the middle and upper categories, while for the bottom 30%, it has been almost the same. This has nothing to do with his figures for poverty.

    Secondly, if you read virtually any report, you will realise that the rural consumption of cereal is invariably higher than the urban consumption. As the urban guys are usually richer than the rural folks, there is some evidence that cereal consumption actually reduces as people move out of poverty.

    And if you think of it, it should be obvious why. My mom, when she used to go to school used to have ganji in the morning, ganji for lunch, and if anything was left, ganji for dinner. Now as they have moved to the middle-class, people in those villages have moved to the idli-dosa in the morning and rice-sambhar for lunch routine we southies are familiar with. That of course, means a shift away from a cereal-based diet. What should logically happen is that as people start moving from “starvation” diet to “adequate” diet, the cereal consumption should increase. As they move beyond that, it should reduce, because they supplement their diet with other choices.

    Comment by Ravikiran — August 23, 2005 @ 11:14 am

  9. Note to North Indians: “Ganji” is a rice gruel. Not to be confused with banians or vests. They don’t consume items of clothing in South India.

    Comment by Ravikiran — August 23, 2005 @ 11:20 am

  10. No doubt everyone here is aware of the NSS controversies which make inter temporal comparisons diffficult. The latest piece by Himanshu and Sen in the EPW suggests that the biases suggest that poverty has not fallen as previously asserted by Deaton-Dreze. Ravallion suggests (I forget where) that the trend of poverty reduction has reverted to its pre 1980 level (not surprising given the massive public spending on Indian agriculture in the 80s). In any case, there is no question about the employment figures (I’m assuming). Nor is there a question about the collapse in prices, and the rise in input prices (is there?).Nor am I seeing any challenging of the question of absolute consumption thesis she raises (which is most relevant for the debate here).


    Comment by Arjun — August 23, 2005 @ 12:03 pm

  11. I am no expert in any of these matters. However, it is not surprising to me if reforms have not dented the deepest Indian poverty. Firstsy, agricultural reform has been non-existent. Secondly, states like Bihar remain trapped in a medieval hell. How do the stats look if we leave out Bihar and UP?

    Comment by Vish Subramanian — August 23, 2005 @ 7:10 pm

  12. Aadisht — Thanks a lot. I’ve also posted my comment there. I must say that your post gave the impression that you had access to data that speaks of the bottom percentiles, whereas what you tried to do was to do computations based on the “total” (and hence skewed) average. As I too do not have the relevant data, my comment relies upon your methodology of computation as well.

    Ravi — Thanks for that explanation. I agree with you that as people start moving from “starvation” diet to “adequate” diet, the cereal consumption should increase. As they move beyond that, it should reduce. And let us rely upon Aiyar’s assertion of “gradual” reduction though it would have been much better if he also backed it up with some hard stats. In particular, there’s no reduction in the bottom 30%. Then, is it okay to conclude that the bottom 30% haven’t moved out of the starvation diet? How can we conclude in that case that the poor have done far better under reforms than before? What are your thoughts on this?

    Arjun — Thanks for those pointers. I too think, as you pointed out, it’s important to note that Prof Patnaik is talking about “total” consumption, as in my earlier long quote.

    Vish — I wasn’t surprised either to see Utsa Patnaik’s conclusions.

    Comment by Anand — August 24, 2005 @ 7:00 am

  13. incidentally, and this is being a bit pedantic, since we are not considering any price vectors in this argument, the question is not about giffen goods, but about inferior goods.


    Comment by Arjun — August 24, 2005 @ 10:10 am

  14. But Anand, the evidence seems to indicate that calorie consumption by the bottom 30% has remained constant for the past 30 years. The controversy is over poverty has reduced over the period from 93-99. There is no doubt that poverty did reduce between 73 and 88, during the green revolution period. Now any theory you come up with has to explain this fact too, as well as the fact that calorie consumption has reduced for the upper 70% of the population. Swami’s explanation is that poverty has been reducing steadily, and the richer among the bottom 30% have crossed the hump and moved into reducing-calorie consumption region, while the calorie consumption for the poorer among them is still increasing, and these trends cancel each other out, thereby keeping the average the same. Now if you think this explanation is wrong, you’ll have to come up with a theory that explains all the observed facts, not just the convenient ones.

    Comment by Ravikiran — August 24, 2005 @ 11:35 am

  15. Ravi — We need to have the relevant stats for a fixed time period for these discussions. We can’t be taking 1973-2000 as the period for one analysis, and 73-88 for another. Agree that Aiyar asserts “gradual”, but to continue along we need to verify it ourselves. Also I have no idea about the percentage of “the richer among the bottom 30%”. Aiyar’s manipulation is based on the reduction from 55% to 26%, which to say the least is very suspicious. (Incidentally, I would like to know whether you yourself buy this “dramatic reduction stats”.)

    If one’s willing to believe Utsa Patnaik, the conclusion is that per capita consumption wasn’t decreasing at any point in the pre-reforms period, whereas drastic reductions were visible in the post-reforms period. Cancellation in the bottom 30% is taking place because Aiyar does the computation from 73 to 2000 together. (And Patnaik was talking about village India.)

    In any case, how do you explain the poor progressing far better under the reforms?

    Comment by Anand — August 24, 2005 @ 9:55 pm

  16. I am not out to claim that the rate of reduction of poverty has been drastically better under reforms. The argument here is to prove or refute the claim that poverty and malnutrition have worsened under reforms. For the moment, I think that Aadisht’s analysis makes it clear that there is no such “crisis” and the data yields to alternative explanations.

    As for why we haven’t seen accelerating reduction of poverty under reforms, there is a simple explanation. For that to happen, the poor have to get jobs. For that, industries have to be set up. Outside of the big cities, it is still a nightmare to set up industries. Land acquisition is a bureaucratic nightmare because in order to “help” the poor villagers, the government doesn’t allow them to sell their agricultural land to set up industries. So naturally it has to be done through government sponsored middle-men. The number of permissions needed are still huge. In other words, what reforms?

    Also, the crisis in the agricultural sector has nothing to do with reforms. The green revolution had started running out of steam in the late 80s. It is one thing to give incentives to increase food production when we were facing shortage of food. But by the 80s, we were facing an excess of food. So the government had to, and still has to buy high and sell low. It was the buying at high prices that got landed farmers out of poverty. There is a limit to how much any government can do that and we had reached a limit to getting people out of poverty through that route.

    In fact, this policy is now counterproductive, because by buying at high prices, the open market price for food goes up, hurting the landless labourers who have to buy their food from there, worsening poverty. There is no easy solution to the agricultural mess. The only way is to manage a smooth transition of people from the agricultural sector to other sectors, something that has been happening (too slowly for my taste) under reforms. If we hadn’t carried out reforms when we did, we would have simply seen a stagnation in or even a worsening of the poverty rate.

    Comment by Ravikiran — August 25, 2005 @ 1:39 am

  17. I don’t think Aadisht’s analysis makes it clear. By alternative explanations, if you mean the ‘giffen goods’ argument, that’s not clear from Aadisht’s post at all. As Aadisht said:

    “Of course, these are average figures and don’t necessarily prove that everyone in rural India is consuming more. There is always a possibility that the rich have become much richer, and their consumption skews average consumption upwards while the poor are still in bad shape. It’s also possible that there are some states in which consumption has increased and they’re skewing the numbers.”

    Now Aadisht’s refutation of the above para is:

    “But if you look at other NSS reports, you find that that hasn’t happened- the consumption of people in the lowest five income percentiles has also increased.”

    Neither Aadisht nor I have been able to check the consumption figures for the bottom percentiles to claim that skewness hasn’t happened. What Aadisht does is a computation — based on the absolute (and hence skewed) averages available — to conclude that purchasing power has increased even in the bottom percentile. That argument doesn’t look right to me and I have detailed this in a comment at the orginal post.

    Utsa Patnaik blames the drastic reductions in the State’s spending on rural development for the agrarian crisis. If I understand right, you too agree with that, and your third para in the above comment suggests so (though you also say that the state doesn’t have another option).

    Comment by Anand — August 25, 2005 @ 5:18 am

  18. I did not say that State spending has reduced. I don’t know if it has. Even if it was kept as a constant proportion of state spending, poverty rate would have stagnated or worsened. That is because those who could have been helped by the previous policies have been helped. Now the same policies are hurting the landless poorest.

    Yes it is frustrating that we don’t have the correct reports. I tried searching using the keywords Aiyar gave and came up with a blank.

    Comment by Ravikiran — August 25, 2005 @ 8:17 am

  19. The more I thought about it, the more I became convinced that the whole idea of trying to infer someone’s income from his diet is nonsense. It completely ignores caloric burn rate. That has to be going way down as India industrializes. Actually Swaminathan Aiyar makes this point in the second paper (Its not just calories, stupid). I make reiterate that argument on my blog.

    Comment by Michael H. — August 25, 2005 @ 9:13 am

  20. Michael,

    Its a seemingly cool idea, but I’m simply not convinced by it. I would like to know the degree of mechanization in agriculture for the rural poor before believing the caloric burn thesis. Given the lack of public investment in agriculture, and the other signs of deprivation- increasing unemployment, more evidence of farmer suicide and so on, I would be loath to assume that you have full scale industrialization proceeding in the hinterland. I would also assume that regardless of the caloric burn rate, as industrialization proceeds, there should be at least a small increase in total consumption, and not a fall. No doubt calorie intake is a flawed method, but its at least a more reliable measure of inter termporal poverty calculations than we have with the NSS reports. By the way, back of the envelope calculations on unadjusted NSS data will tell you precisely nothing. Its not worth it because the statistical problems with NSS are tremendously difficult.

    Ravi, there is of course no counterfactual to your claim that without reforms the poverty rate would have been worse. Its like the justification for the Iraq war. I am also not sure that you can attribute increased agricultural output and sharp reductions in poverty rates in the 1980 to price support mechanisms (it appears that irrigation and public investment had a lot to do with it).

    Comment by Arjun — August 25, 2005 @ 10:20 am

  21. Ravi — Thanks.

    Michael — Probably poverty levels in India will shoot up if you measure, if you can that is, income instead of consumption patterns. I have a comment at your post too taking “nutritional deficiency” into consideration. Some discussion is happening at my blog as well. For the sake of other readers, my comment is based on an article by CP Chandrashekhar and Jayati Ghosh that appeared in The Hindu Business Line in 2003.

    Arjun — I too tend to believe that back of the envelope calculations on unadjusted NSS data will tell you precisely nothing. I’ve tried to show this by adopting Aadisht’s method of computation to the “palak and other leafy vegetables” section. A simple computation shows the enormous skewness that happens. This is in a comment at Aadisht’s original post.

    Comment by Anand — August 25, 2005 @ 11:24 am

RSS feed for comments on this post. TrackBack URL

Leave a comment

Powered by WP Hashcash

Powered by WordPress