Markets & Famine: Amartya Sen is not the last word !

Whether markets help cause or exacerbate famines is one of the great questions of political economy. Cormac Ó Gráda’s recent book Eating People is Wrong, and Other Essays on Famine, its Past, and its Future, along with his earlier volume, Famine: A Short Historyquietly, calmly, and unostentatiously undermines many of the key empirical observations about markets and famines made by Amartya Sen. Yet few seem to have noticed his disagreements with the Nobel laureate who transformed the thinking on the subject. This post includes remarks on the Bengal famine of 1943, the Great Irish Potato Famine, and some of the ‘Victorian’ famines of British India in the late 19th century.

j10449In Diane Coyle’s review of Eating People is Wrongshe remarks Ó Gráda “broadly agrees with Amartya Sen’s famous conclusion [about the Bengal Famine of 1943] that it was a famine of policy rather than nature”.

I find that statement puzzling. Yes, Sen and Ó Gráda do agree that British war-time policies helped bring about the famine through intentional military choice, callous indifference, and unintentional blunders. But in Sen’s analysis, the prime mover behind the initial decline in food availability was not a major shortfall in grain output within Bengal, but a spike in food prices induced by speculation and hoarding in response to the perception of shortfall. There was enough food, but prices rose too high.

Ó Gráda, by contrast, observes that a combination of cyclone and blight (Helminthosporium oryzae) did actually cause a major drop in Bengal’s normal rice harvest. He finds the pattern of prices in Bengal over time inconsistent with hoarding and speculation. The man-made part is the impediments placed by British officials in the way of normal, peace-time functioning of markets which would have brought more food imports into Bengal from other Indian provinces.

To put it as starkly as possible, Sen cites ‘natural’ market forces and Ó Gráda cites man-made market failure. I must stress this point: both agree on the British response/nonresponse. But they disagree fundamentally on the ur-cause of the initial decline in food availability.

In his review Charles Mann at least says “Sen got it partly right and partly wrong” on the Bengal famine of 1943. But he then fails to trace O’Grada’s continuous divergence from Sen on the role of markets in famine.

§  §  §

41q2yFeFyJL._SY344_BO1,204,203,200_Sen famously argued that famines have never occurred in self-governing democracies, because the free flow of information and popular pressure induce governments to take steps to prevent harvest failure or other disasters from turning into actual famine. Today, very few disagree that many famines have had obviously political causes. Or that politics and ideology have in the past inhibited action to relieve the starving. Even fewer disagree that markets do not avert famines all by their magical selves. Thanks largely to Sen, famines are now routinely regarded as a result either of policy blunders, or inaction and indifference, or willful engineering.

But there’s more to Sen’s work on famine than that. Why, in the first place, do populations which in one period manage to feed themselves without relief, suddenly become in the next period unable to do so ? Before Sen turned his gaze on the subject, the culprit had usually been identified as “food availability decline” due to climate or war. But then he shifted the emphasis away from famine as natural disaster to “economic disaster”.

In Sen’s approach to famine, ‘entitlements’ are defined as the “set of alternative commodity bundles that a person can command in a society using the totality of rights and opportunities that he or she faces”. Famine occurs when entitlements decline; and in Poverty and Famines, Sen focused on “exchange entitlement declines”, e.g., when market incomes fall in relation to the price of subsistence foods. Food becomes not physically unavailable, but unaffordable.

In EPW Cormac Ó Gráda produces a convenient punnet square of the different possible attitudes toward famine:


Amartya Sen adopts 1,2 — market failure (due to poor information or uncertainty) is often the original cause of food insufficiency; and 1,1 — when markets are working, they often make famines worse.

But I want to be clear. I am NOT saying Sen=> markets=> evil nor Ó Gráda=> markets=> wonderful. Both treat as an empirical, case-by-case matter, how markets actually behave in specific famines. In Hunger and Public Action, for example, Drèze and Sen devote chapter 6 to both the positive and negative impacts of markets in famine situations, citing both theoretical reasons and historical evidence. But in practise, Sen’s interpretations of famines in history have generally had an anti-market bias. [Edit: Maybe the phrase ‘anti-market’ is too strong. But clearly for Sen markets have often worsened famine risk through hoarding, speculation, and food exports.] 

By contrast, Ó Gráda’s empirical investigations generally find that unless a government or a war has disrupted them, markets usually function well in famines. Nor have well-developed markets, in practise, worsened famines. That doesn’t mean markets alleviate them all that much. They just don’t seem in practise to make things worse. (The full exposition of these views is found in the best and most technical part of EPW, chapter 3, whose analyses range from famines in the European past to contemporary Africa.)

Moreover, in Ó Gráda’s analyses, famines have not primarily ensued from “exchange entitlement declines” divorced from supply shocks. The pendulum had clearly swung too far in the direction of Sen’s approach, and there’s been too much emphasis on the market distribution of food during famines and too little on physical food availability. Even in famines which everyone agrees are man-made, such as Ukraine in the 1930s or Mao’s Great Famine, Ó Gráda finds that supply shocks have been underestimated. “The paucity of evidence for ‘pure’ entitlement famines–famines with no FAD dimension–suggests that modern scholarship may underestimate the role of food supply” (Ó Gráda 2007).

§  §  §

Amartya Sen attributed the “exchange entitlements decline” that purportedly caused the Bengal Famine of 1943 to a combination of war-induced inflation, speculative profiteering, and market overreaction to perceived harvest failure:

…demand forces [leading to inflation] were reinforced by the ‘indifferent’ winter crop and by vigorous speculation and panic hoardings. The hoarding was financially profitable on the basis of even ‘static expectations’: rice prices had more than doubled in the preceding year, while the ‘bazar bill rate’ in Calcutta still stood around 7 per cent per year (the bank deposit rate was below 2 per cent per annum).108 There was a abnormally higher withholding of rice stock by farmers and traders from the winter harvest of 1942–3; the normal release following the harvest did not take place.109 A moderate short-fall in production had by then been translated into an exceptional short-fall in market release. The ‘current supply’ figures of the Famine Inquiry Commission no longer reflected supply to the market.

But against this view Ó Gráda is simply devastating, like a Mongol horde riding roughshod over Eurasia. Both quantitative and qualitative evidence strongly suggest hoarding was not a major factor in Bengal’s price fluctuations

First, a major campaign to search and root out illegal stockpiles failed to discover substantial grain hoards. Yet the Muslim League government of Bengal had every incentive to find them, having publically blamed the food shortage on merchants and moneylenders in the cities (who happened to be mostly Hindus).

Second, if there was a lot of speculative hoarding of rice in 1943, then the market release of hoarded supplies in early 1944 (when there was a bumper crop) should have caused a much bigger price drop than was normal. Although the market price of rice converged with the official controlled price at the end of 1943, prices remained higher than before 1943. From Ó Gráda’s FSH:

bengal prices

Finally, a key implication of Sen’s “exchange entitlement decline” is the great inequality of suffering in times of food crisis, both in terms of mortality and financial distress. The landless labourers suffer. Consumers of food, not producers, suffer. In the Bengal famine of 1943, those classes did suffer the most, but the number of land-owning food producers who were forced to sell their land, or carried abnormal debts, or turned into wage labourers, or migrated to the cities as paupers, was simply too high to be consistent with the Sen interpretation of the famine. It was a general decline of food availability.

§  §  §

A classic trope of well-functioning markets making famine worse is the export of food during famines. The most notorious instance comes from the Great Irish Famine, about which Drèze and Sen noted:

…as it happens, quite a few famines have taken place without much violation of law and order. Even in the disastrous Irish famines of the 1840s (in which about an eighth of the population died, and which led to the emigration of a comparable number to North America), the law and order situation was, in many respects, apparently ‘excellent’. In fact, even as the higher purchasing power of the English consumers attracted food away, through the market mechanism, from famine-stricken Ireland to rich England, with ship after ship sailing down the river Shannon laden with various types of food, there were few violent attempts to interfere with that contrary—and grisly—process. [Drèze & Sen pp 22-23]

However, this ignores the fact that Irish export of relatively expensive wheat and oats financed the import of relatively cheap maize. From EPW:

ograda figure 3.1And from FSH:

irish famine

Of course all of that was too late and took too long to happen. And it would have been much better if all the food produced in Ireland could have been requisitioned and provided to the starving. But it’s also an open question how much the British state in the 1840s was capable of doing that in time, if it had been as activist in its inclination toward famine relief as any modern government today is.

§  §  §

In the second half of the 19th century, somewhere between 15 million and 35 million Indians (out of a population of 150 million or so) died in climate-related famines. A staple of Indian historiography has been to blame the introduction of railroads for making the famines worse — again, by facilitating the export of food grains from a famine-stricken region, not only overseas, but also to other regions within the country. Sen has frequently mentioned this in his works, e.g., Drèze & Sen [pg 90]. (Ó Gráda also addresses this issue, but below I rely on other sources.)

The food-exports-by-rail idea is plausible in the short run and in principle. But compare the extent of the railway network circa 1880, with the map of “famine intensity” in the 1870s when the Great Famine of 1876-78 took place. (Source; I don’t know why one map includes all of British India while the other only shows modern India.)


By the end of the 1870s there were still enormous swaths of India — the size of European countries several times over — which were completely unserviced by rail. Before rail access, Indian commerce naturally relied on roads and rivers, though most of the country is not as dense with waterways as Bengal, let alone China’s Yangtze valley; and food grains typically travelled to market by bullock carts. In the pre-rail era, therefore, prohibitive transport costs fragmented regional markets, so that prices for the same product could vary by double-digit factors even in adjacent regions.

So even if the British Raj had banned grain exports out of the country as soon as famine was noticed (usually when starving migrants started showing up in the cities), that would not have meant the grain shipments could have been transported to the most famine-stricken regions of India in time. In 1876-8, most districts in the northwest and the south would have been hundreds of miles from the nearest rail depôt.

However, a modest-sized literature does find that market integration did reduce India’s famine vulnerability in the long run. Over time, what had been a group of fragmented regions became a national market, as can be gleaned in grain prices from Studer 2008:

studer rice prices

(You can also see wheat prices, as well as the coefficient of variation of grain prices over time.)

The above could imply that — over the long run — grains moved from surplus to deficit regions in order to exploit price arbitrage. But whether India’s famine vulnerability was ultimately reduced by railways development is more directly addressed by Burgess and Donaldson, “Can Openness mitigate the effects of weather shocks? Evidence from India’s famine era“. B&G found that, in the period 1860-1920,

“[O]n average over our sample period, a local rainfall shortage in a district would increase the probability of observing a severe famine in that district… In particular, prior to obtaining railroad access, rainfall shortages had a large ( i.e., statistically and economically significant) effect on famine intensity. But after a district obtained railroad access, famine intensity fell significantly… For a given reduction in rainfall, if a district were connected to the railroad network it faced a significantly lower probability that a food scarcity or famine would occur in that district. This is the key result of the paper. Indeed the ability of rainfall shocks to cause famines almost disappears after a district has been penetrated by the railroad.4”

The above is even though rainfall was not noticeably less volatile after 1900 than before. You can see map visualisations of changing famine intensity and railways mileage extracted from B&G’s slide presentation:

Postscript: Neither this post nor Ó Gráda addresses the claim argued in a well-received book called Late Victorian Holocausts: El Niño Famines and the Making of the Third World by Mike Davis. Davis argues that in 1850-1914 the coerced integration of developing countries, including British India, into global capitalism made their populations more vulnerable to catastrophic weather-related famines than they had been in previous eras. The claim draws on an older tradition in Indian historiography and has recently been reprised in Sven Beckert’s Empire of Cotton: A Global History. This thesis, though a favourite amongst western historians studying colonialism, has been mostly neglected by economists. But there is a small amount of economic history literature offering a different perspective on this question, which both Davis and Beckert have suppressed in their books. That is the subject of a blogpost coming in the near future.

PS #2: James Fenske is apparently now working on cotton, famine, & land tenure institutions in British India after 1850.

This entry was posted in Economic History, Famines and tagged , , , , . Bookmark the permalink.

14 Responses to Markets & Famine: Amartya Sen is not the last word !

  1. Francisco says:

    Thank you. I like reading your posts.


  2. It took me, a marketing economist with some knowledge of how Third World and other markets work, 15 minutes to realize that Amartya Sen could not possibly be right: his model just did not work. It took me two weeks to realize that he had systematically faked his evidence – remember this was when the evidence was hard to get, in a handful of libraries rather than on the internet. I was slow to realize this because I thought nobody would fake research any more than they would cheat at cricket – which shows how long ago it was. My work and Tauger’s showed all Sen said about 1943 was faked. As more contemporary evidence becomes available, more and more discrepencies appear. There are excellent contemporary (e.g. Pinell, Braund, Knight) and near-contemporary (Greenough) analyses. And other people point out misstatements in his other work. Sen’s hypothesis was not just common in 1943 – the belief in it contributed to the failure to take action – it was commonplace in the 1870s – Hunter, Bartle Frere, the Indian Famine Code etc Strachey and Litton mention it in infinitely better analyses than Sen’s. Which raises the question why so many economists with no knowledge of market economics or the Third World should have spent the next forty years pontificating on Sen, in the sure and certain knowledge that he had been accused of faking his evidence. And why they refused to spend the three or four hours needed to double check the discrepencies.


  3. unknown128 says:

    Wernt there several massive famines in Mughal India or in post Mughal but pre British India? I dont know how well documented they are but propably far worst then the famines during british times…. The question thous is if the British can be held responsible for all or even most of the famine dead
    There were several massive famines in independant China at the same time after all and it could be argued that a divided and wartorn india unqonquered by the british would have also had famines with millions of dead.

    Sorry for my bad English


    • Yes, there were famines before the British but there is very little concrete information about it. The British held famine commissions after each major famine, so we know a lot about those.


    • Yes. Famines were normal everywhere before 1850. They do not appear in history books because they were seen as an act of God which nobody could do anything about. And because the class of people who could write did not go hungry. Recent scholars who have looked through local records say that a region of Britain could expect serious hunger once every 14 years, and famine every 55 years or so. And this in a small country with a stable climate. So India must have had dozens of famines most years. Tauger quotes the Russian Empire of the 1890s having to deal with three famines a year.
      What changed things was the changed shipping following the Napoleonic war, which meant that something could have been done about the Irish famine, though not much about most Indian famines.
      The railways suddenly changed everything: grain could be transported from the mid west of the USA to the ports. In India the Raj subsidized the enormously expensive rail system as a famine relief exercise, so that grain never had to go by ox cart more than 25 miles. (Though much of the money was stolen by London bankers – no change there).
      The massive construction of government financed irrigation works in Punjab etc meant that there were large surplusses available. So for the first time in history transport of grain long distances to deal with famine was possible. For the first time it was considered a failure of government to let a famine happen.
      The politics were that an Empire, like Russia or Britain, could shift food within the empire. Small countries had to buy it on the world market, which was impossible with a widespread shortage covering most of Europe for instance.
      The grain supply and famine stocks of the middle east from 3 or 4000 years ago, based on local supply and storage are a different system. See ‘The Long Hot Summer’ for a fascinating analysis in the light of climate change.


  4. To put it into perspective: in Tanzania in the 1980s the early warning system identified a famine situation in the Lake provinces, expected to start a year later. We had to allow 6-8 months to get the food aid from the coast to the affected area. The transport system of the pre-industrial economy just could not take any more. It had not been designed to take large exports to the coast – any were high value to bulk – and there was no demand for imports. Yes, the transport system had been run down after independence, but similar things happen in wartime, after worldwide depressions and after national economic collapse. See also Griffiths, ‘The Economist’s Tale’ for a similar situation in Sierra Leone. That is to say many countries had this major transport constraint up to this century, and perhaps twenty or thirty did not at the beginning of the last century


  5. “The uncertainties surrounding such famine chronologies have been summarized elsewhere (Ó Gráda 2009: 26). Still, they have their uses and Clarkson (1989) has tabulated the data presented by Wilde, relying on his own judgment to distinguish between outright famine and less severe crises. By this reckoning the first half of the fourteenth, seventeenth, and nineteenth centuries were the worst affected by subsistence crises. Wilde’s chronology suggests a crisis or famine on average every thirteen years, but famines in the strict sense were less frequent: by this reckoning there were only 29 during this 550–‐year period.” (O Grada Famine inIreland, 1300–‐1900


  6. unknown128 says:

    Well while unfamiliar with all the details my logic goes this way: By 1913 there were several political entities that were never colonised by europeans, Ethiopia, Liberia, The Ottoman Empire, several tribal entities on the arabian peninsula, Persia, Afghanistan, China, Siam and ofcourse Japan.

    Of this about 10 countries only 1 (Japan) actualy carried out a large scale westernisation that alowed it to surpass all other non western entities (bouth colonised and uncolonised) in wealth and power while all the other 9 basicly remained poor and backward and basicly only became uncolonized by chance or because of circumstances outside their control…All of this countries were in no way better developed then their neighboring colonies (Siam was in no way better of then french Indochina or the Malay confederation, Afghanistan in no way better of then Russian central asia or british Punjab, Ottoman Arab territories in no way better of then British egypt and yemen in no way better of then British Oman). Now had the many states on the indian subcontinent remained independant there would have been only a maximal 10% chance that they would have gone Japans way and a 90% chance they would have gone the Ottomans or even Chinas way…..Now in case of a traditionalist India frozen in ancient forms of social organisation and unwilling to westernize more then absolutly necesary (the 90% chance) it would have been hit by the very same famines to the same or even a greater extent (because of the wars between the different states) as British india was….Only an India going japans way could have handeled the malthusian preassure better then the British Raj did and in my opinion such a chance would have been very low for an independant India.

    Also its quiet possible that it would have endured even greater humanitarian catastrophies like independant China did with the Taiping rebellion.


  7. Joel says:

    Did you ever end up doing anything on Late Victorian Holoucausts? Would be very interested, it’s a fascinating book; obviously polemic, but very well written. I came away from it with a righteous anger.


    • no not yet but it’s coming at some point. the problem is that pre-19th century Indian famines, which undeniably existed, are poorly documented compared with the “late Victorian famines” — which the British documented with many famine commission inquiries. but a post saying “there’s no evidence famine mortality increased under the British” wouldn’t have too much meat


    • but Davis’s claim that famines were exacerbated by rising British taxes is utter nonsense easily belied with data. it’s an old claiming going back to the early 1900s but it’s not taken seriously anymore


    • Joel says:

      Cool. Felt parts of the book based on Indian historiography were kind of weak anyway, but would love a longer rebuttal. Anything you’d recommend in the meantime as an antidote? Still think a lot of the anecdotal stuff is horrifying and damning!


      • Cormac o’Grada quotes analyses of the famine and serious hunger episodes in Britain and Ireland. If I remember a serious famine every seventy years or so and major hunger at least once every twenty years. So common that they are not mentioned in British or Irish history books. So imagine what it was like in areas with more fluctuating climates. The mid Victorian Irish Famine was the first time there was a public perception that something could and should have been done about it: globalization and the merchant marine made it possible. (Transport had been, and remains, a key constraint to relief). This idea rapidly spread throughout the empire. For example Indian railways were heavily subsidized with the objective that they could get grain within 25 miles of any famine (though the city of London bankers ripped them off – I have a reference on this). The raj was good at handling these famines by the 20th Century – it was largely a matter of transport from surplus areas elsewhere in India – but thought they could handle a famine in wartime India in the same way. Tauger points out that well into the 20th century they expected three famine situations a year in the Russian Empire. Yes they managed them well by then, but the revolution changed everything and we got the Ukraine famine of the 1930s (see Tauger on this as well as the Indian famines).
        Do not be misled by the literature on markets and famine. None of the people quoted could get, let alone hold down, a job as an agricultural market economist. People like Sen or O’Grada would be chucked out immediately. Nor can I imagine an academic being able to manage such a job – the real world is infinitely more complex. The economic analysis is infinitely more complex. The data assumed by academics does not exist.The reality of doing real world economics is infinitely more complex. See a book I wrote under another nom de guerre, Peter Griffiths, The Economist’s Tale, which covers the international politics, the national politics, the internal office politics, sex, violence, corruption, plague etc, which had to be taken into account in dealing with a real famine situation. And, yes, I was able to avert the famine.


  8. Pingback: Heavy Lifting: The Potato, the Poor, and Pleasure in Ireland – Rachel Laudan

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s