How can famine strike a country with food in its granaries?

In March 2020 the shelves went bare. Whole aisles of them: flour, pasta, tinned tomatoes, and the thing that became the joke of that spring, toilet paper. People photographed the empty racks and sent them to each other, and the photographs carried a question nobody quite finished asking out loud. If it can go this fast, what stops it going all the way? That time, nothing had been destroyed. Households had concentrated a few weeks of ordinary buying into a few days, and a supply chain tuned to steady demand needed a few turns of the wheel to catch up. It did, and the scare left behind a suspicion it had not earned: that hunger, when it comes, comes because the food is gone.

That is worth testing against the case where it went all the way. In 1943 the province of Bengal went through a famine that killed on a scale nobody has ever settled. What stands out in the record is how much of the argument about it, then and since, has been conducted over a difference in the supply of a few percent, while the dying was sorted by something else entirely.

What the record says was there

Start with the accusation. The Famine Inquiry Commission, the official body that reported on Bengal in 1945, did not find that the province had plenty. Its stated first conclusion was that “During 1943, there was a serious shortage in the total supply of rice available for consumption in Bengal as compared with the total supply normally available,” which it traced to a poor winter rice crop at the end of 1942 and to a thin stock of old rice carried forward. Anyone who says the granaries of Bengal were full in 1943 is contradicted by the commission that investigated it.

The Commission also counted, and its own counting is where the difficulty starts. It measured supply in weeks: how many weeks of the province’s requirements the year’s crops and imports would cover. For 1943 the answer was about 43 weeks. It then gave the comparisons, and the comparison year that matters is 1941, when Bengal had about 39 weeks and no famine at all. On the measure the Commission itself put in its report, the famine year stood four weeks above the year nobody remembers. The Commission knew how that read. It argued that the missing quantity was the carry-over of old rice, larger in 1941 and smaller in 1943, which would put the famine year below after all. Then it conceded the problem with its own argument: “We have no information about the actual size of the carry-over in any year,” it wrote, apart from what could be inferred from the gap between each year’s current supply and its requirements. One of its members, M. Afzal Husain, put the point harder in a minute appended to the report. Absolutely no data are available, he wrote, regarding the stock position of rice or any other foodgrain from month to month or year to year, in Bengal or any other part of India.

So the granaries of the chapter’s title are not an inventory. Nobody knows what was in them, and the report that is supposed to tell us says so twice. What can be compared is the flow: what was grown, what came in, and what a family could get hold of. Amartya Sen did the first of those in 1981; the second is what this chapter is about.

An argument about a few percent

Sen’s calculation ran the Commission’s own figures forward, adjusting them at each step in the direction that made his case harder. He added wheat to rice, since Bengal ate both. He scaled for population growth using a rate he chose as an overestimate; the index of food available per head in 1943 then stood about 9 percent above 1941. Finally he deducted the rice and wheat that arrived late in 1943, after the worst of the starvation, on the grounds that food which came too late should not be counted as available; that brought his index to 109 for supply and 107 for food per head, both still above 1941. His conclusion is the sentence the whole literature has been arguing with since: “It seems safe to conclude that the disastrous Bengal famine was not the reflection of a remarkable over-all shortage of foodgrains in Bengal.”

Read the words. Sen is not claiming there was no shortfall. He is claiming the shortfall was not remarkable, and that a famine of this size is not what a shortfall of this size produces. The opposing case attacks both halves.

The first attack is on the arithmetic. M. Mufakharul Islam found that the acreage figures underlying the adjusted series were handled one way for 1938 to 1941 and another way for 1942 and 1943, so that the two halves of the series are not comparable. Correct the discontinuity and the index moves: food available per head in 1943 comes out 4 percent below 1941 rather than above it. Figure 7.1 puts both corrected series on one axis, because the argument is not about magnitude. It is about which way the line points.

The attack does not stop at the arithmetic. Mark Tauger argues that the figures everyone is correcting were never worth correcting: the official crop estimates, in his words, are “one set of unreliable statistical data that Sen and others have incorrectly described as ‘production data.’” In their place he puts measurements from rice research stations, published by the mycologist S. Y. Padmanabhan, which he says “show that Bengal had a major harvest failure in 1942 and a significant shortage of rice.” And Cormac Ó Gráda states the conclusion without hedging: “in the case of Bengal, I have argued that there was indeed a ‘FAD’ and that denying this has led to an undue historiographical focus on hoarding.” FAD is the trade’s abbreviation for food availability decline. He is saying the food did fall.

That is the opposing case, and this chapter does not adjudicate it, because of what the same scholars say next. Ó Gráda’s summary of the shortfall he insists on runs: “In normal times, Bengal might have been resilient enough to cope with such a shortfall; in 1943, given military requirements and war-related disruption to trade and communications, this was a disastrous deficit.” His conclusion, four pages later, is that the failure was “of the imperial power to make good a harvest shortfall that would have been manageable in peacetime. The famine was the product of wartime priorities.” The Famine Inquiry Commission had structured its own findings the same way forty years earlier. Its first conclusion was the shortage; its second was that out of the supply that did exist, the requirements of large sections of the population “were not distributed to them at a price which they could afford to pay.” In the paragraph between them it wrote this: “The supply position was not, however, such as to make starvation on a wide scale inevitable, provided the trade was capable of maintaining the distribution of available supplies subject only to a moderate rise in the price level.”

Every party to this argument, including the ones who think Sen was wrong, agrees that the shortfall was of a size a functioning system absorbs. What they are arguing about is a few percent. What killed on the scale of a million people and more is in the second half of the Commission’s sentence, and it has a name.

The claim, not the food

Sen’s term for it is entitlement, which in his usage has nothing to do with deserving. It means the set of things you can legally get hold of, given what you own and what the market will give you for it. Food does not arrive because it exists. It arrives because you have a claim on it, and there are only a few ways to hold one. You can grow it. You can trade something for it. You can sell your labour and buy it. Or someone can give it to you. Figure 7.2 lays the four out.

Now watch the routes break, one at a time, in the record. The price of rice in Calcutta was between 13 and 14 rupees a maund, a weight of about 82 pounds, on 11 December 1942. By 12 March it was 21. By 21 May it was above 30. By 20 August it was 37, and in October, when official quotations had stopped, there are reports of rice selling at 80 rupees a maund in Chittagong and 105 in Dacca. A fisherman’s catch did not treble in price. A barber’s haircut did not treble. A labourer’s daily wage was, in most cases, fixed in money terms. Every one of those people held a claim on food denominated in something that was not rice, and the claim shrank by nearly two thirds while the rice sat in the same markets it had always sat in.

The cultivator was in a different position, and not because anyone protected him. He held his claim in the form of the crop itself, which is the one denomination that cannot be repriced against rice, because it is rice. That is the whole of the gradient in Figure 7.1. It is not that fishermen were poorer than cultivators to begin with, though many were. It is that a fisherman’s claim on food ran through a market and a cultivator’s did not, and 1943 was a year in which the market moved and nothing else did.

The transfer route is where the machinery is plainest, because in Bengal somebody operated it. The government held that keeping Calcutta’s industrial area fed was among its highest wartime obligations, and from August 1942 the Bengal Chamber of Commerce ran a scheme, with the government’s help, guaranteeing food to the grain shops of firms connected to it. By December it covered 620,000 employees; other chambers covered another 170,000, and arrangements for government, railway, Port Trust and Corporation staff covered 300,000 more. Sen’s summary of what those schemes achieved is this: they “guaranteed freedom from starvation to more than a million employees and their dependants.” A million people in the same province in the same year were handed a claim on food that did not depend on a price, and they lived. The districts, where nobody had been handed one, filled up with people walking toward the city. By July the streets of Calcutta were full of them, and in October alone relief organizations disposed of 3,363 unattended bodies, a count Sen reproduces from Ghosh’s 1944 account.

What a state can and cannot do when a shortage arrives is the subject of the countries volume’s chapter on why some states can tax and enforce while others cannot, and Bengal is the case that shows the capacity being used, precisely and successfully, on behalf of a million people who were not the ones dying.

How many, and the trouble with knowing

The death toll has never been settled and every figure in circulation belongs to somebody. The Famine Inquiry Commission concluded “that about 1·5 million deaths occurred as a direct result of the famine and the epidemics which followed in its train,” built from an excess of “the order of one million” in 1943 plus the deaths of 1944. It was candid about its foundations: “All public health statistics in India are inaccurate,” it wrote, before recording that 1,304,323 deaths were registered between July and December 1943 against a five-year average of 626,048. It rejected Professor K. P. Chattopadhyaya’s estimate of 3.5 million on the ground that its method “cannot be accepted as statistically sound.” W. R. Aykroyd, the member chiefly responsible for the commission’s figure, later said he had come to think it an under-estimate. Sen, recalculating in 1981 with a longer window and a correction for under-registration, wrote that “we may be inclined to pick a figure around 3 million as the death toll of the Bengal famine.” Ó Gráda, surveying the field in 2007, reported that “Subsequent estimates of mortality in Bengal range from 0.8 million to 3.8 million; today the scholarly consensus is about 2.1 million.” The span between the low and the high estimate is three million people, which is what it means to count the dead in a place where deaths were recorded by the village watchman, who in 1943 was sometimes unpaid, sometimes gone to work on military projects, and, as the Commission itself records, “not immune from starvation and disease”; some died, and weeks could pass before anyone replaced them.

One more fact from the record belongs here, because it is the chapter’s thesis stated by the calendar rather than by an economist. In December 1943, Sen notes, Bengal reaped a harvest larger than any in its past. It was also the month in which the death rate in Bengal reached its peak in that century. By then the starvation phase had given way to the epidemic phase, and people who had spent a year being destituted were dying of cholera, malaria and smallpox. The food had come. It was not that kind of problem.

What the empty shelf was, and was not

The 2020 shelves were the opposite case. Nothing had happened to anybody’s claim on food: wages were still arriving, prices barely moved, and the transfer machinery was intact and in several countries about to be enlarged. A two-week hiccup in a flow is visible, photographable and self-correcting. What happened in Bengal was invisible for months. There was no photograph to take in March 1943, when the price of rice had gone from 14 to 21 and a fisherman’s income had not, because a claim shrinking is not a thing you can see on a shelf. By the time it was photographable, in the form of people lying in the streets of Calcutta, the mechanism had been running for a year.

Which is why the granary is the wrong place to look. Famine is not the absence of food from a territory. It is the absence of a claim in a household, and the two can coexist for as long as it takes, with the rice in the warehouse and the market open and the price posted on the wall. The next question follows the same logic upward. If a claim can fail while the goods are still there, what happens when enough people’s claims fail at once, and the state that could have answered them cannot? That is the subject of the next chapter, on when inequality breaks a regime.