Professor Sen's explanations examined


Sen's first causal hypothesis is that the famine was caused by factors related to wartime inflation:

The increase in the rice price in Phase I [January 1942 to March 1943] was essentially related to demand factors. . . The price increase in the phase I period, while not confined to Bengal, was much more acute in Bengal than elsewhere (see Singh, 1965, pp 95-99; Palekar, 1962). This was, to a great extent, the result of general inflationary pressure in a war economy. The fall of Burma had brought Bengal to the war front and Bengal saw military and civil construction at a totally unprecedented scale. The war expenditures were financed to a great extent by printing notes.' (Sen, op cit, Ref 2,1977, p 50, quoting A. Singh, Sectional Price Movements in India, Banaras Hindu University, Benares, India, 1965, and S.A. Palekar, Real Wages in India, 1939-1950, International Book House, Bombay, India,1962)

Sen's explanations are generally ones that have been around for centuries. The inflation explanation for famines was very popular after the Central European hyperinflations of the 1930s, where they did appear to cause deaths. It was this very popularity that encouraged the Bengal Government to adopt it. The explanation appeared in economics textbooks in the 1950s. It even appeared in chess books of the time: challenger to Lasker later died of hunger in Austria during hyperinflation. I do not know of anyone who would deny the possibility that this could cause a famine. PB 1997

Sen incorrectly uses the Working Class Cost of Living Index (from Singh) as a measure of general inflation in comparison to food prices. To a large extent the Working Class Cost of Living Index is the food price. Even if it was permissible to use it, it would not support his thesis. Figure 2 shows that it rose at much the same rate in Bengal as elsewhere in India up to the cyclone and the stoppage of interprovincial trade, after which it rose rapidly in line with rice prices.(lndeed, throughout the war, government saw the rise in rice prices due to shortages as being a cause rather than an effect of inflation.)

Sen does not explain why this massive inflation should only have hit Bengal, nor does he explain why it should have stopped after the December 1944 crop was harvested. He does not explain how paying a good wage to a few hundred thousand factory workers should have increased the grain price paid by 60 million people by a factor of four to 20 times. It is particularly strange as he quotes evidence that this extra demand was offset by declining demand and employment in the agricultural sector. (Sen, op cit, Ref 2, 1977, pp 43, 44, 51.) Nor does he explain why the part of Orissa also hit by the cyclone should also have had a famine, though it did not have the same inflationary pressure.

He also fails to explain the enormous scale of the price increase. The population normally spent perhaps 90% of its income on food. Agricultural wages fell, other agricultural incomes were static from one harvest to the next and industrial wages were kept down by giving an issue of food in place of a price increase. Accumulated personal possessions were sold on a massive scale to pay prices that were four to 20 times normal.

I am inclined to accept that it would probably be an underestimate to say that two thirds of the total population were affected by it' (Department of Anthropology, Calcutta University, quoted by Rajan, op cit, Ref 18). An independent estimate was made by P.C. Mahalanobis, R.K. Mukkerjee and A. Ghosh, A sample survey of after effects of Bengal famine of 1943', Sankhya, Vol 7, No 4, 1946, pp 337-400, based on a sample survey of the survivors. They estimate that of the 10.2 million families in the rural population, 1.6 million sold some or all of their land or mortgaged it,1.1 million sold plough cattle, and in 0.7 million the head of the house- hold changed to a lower-status occupation (including 0.26 million becoming destitute). These figures are not mutually exclusive: many families suffered loss of land and cattle, and many became destitute be- cause they had sold all they had. Taking an average family size of 5.4, it seems that perhaps 10-15 million people were affected in these ways. However, many more were affected in ways that would not have been recorded in these statistics. Most went hungry; many were hit by disease; many were impoverished but kept the same occupation; many sold all they had except their land. Village labourers and artisans, at a somewhat higher economic level, sold their domestic utensils, ornaments, parts of their dwellings such as doors, windows and corrugated iron sheets, trade implements, clothes and domestic animals if they had any - sold indeed anything on which money could be raised - to more fortunate neighbours' (Famine Commission, op cit, Ref 2, p 67))

If one accepts that there was a shortage, the price rises are easily explained. Prices rose because of a crop failure. When government tried to buy rice for relief with virtually no limit to price, it pushed the price up to astronomical levels because the rice was not there. This is the inevitable result of applying Sen's prescription when there is a shortage.

It must be concluded that his analysis does not support his hypothesis and his facts tend to refute it.

Uneven expansion of purchasing power Closely linked to the above hypothesis is the one that the famine was associated with an uneven expansion of purchasing power, meaning that the rich could buy more, leaving less for the poor.(Sen, op cit, Ref 2,1977b, p 51;1981, p 77) It is set out most clearly as follows:

In a poor community take the poorest section, say, the bottom 20"/o of the population and double the income of half that group, keeping the money income of the rest unchanged. In the short run prices of food will now rise sharply, since the lucky half of the poorest group will now fill their part-filled bellies. While this might affect the food consumption of other groups as well, the group that will be pushed towards starvation will be the remaining half of the poorest community which will face higher prices with unchanged money income. Something of this nature happened in the economy of Bengal in 1943.' (Sen, op cit, Ref 2, 1980b, p 618.)

This change in income did not in fact take place - as shown in Sen's sources. If 10% of the population had increased their consumption from l4oz per day to l7oz per day (see Table 2), this would have caused a 1.8% change in total demand. This indicates that an increase in income will lead to a modest rise in rice consumption, while a large increase in income may lead to a fall in consumption. Indeed, the figures suggest that if, as Sen argues, the increased income was earned by the industrial working classes, total consumption of rice would have fallen.

Table 2. Estimates of per capita consumption of all cereals.

Ounces per day

General average rate for Bengal

Sectional average rates:

Rural population
Calcutta middle classes
Mofussil urban middle classes
Industrial working classes
Families whose monthly expenditure is Rs 10 or less





Notes: a) FIC, see text, op cit, Ref 2, p 204. The estimates were furnished by Professor Mahala- nobis, Honorary Secretary, Indian Statistical Institute, Calcutta. Professor Mahalanobis analysed the results of five different surveys conducted at different times between 1936 and 1942. Some of these were made at the instance of the Bengal government and others were undertaken by the Indian Statistical Institute or the Viswabharati Institute of Rural Recon- struction. b) The number of families whose monthly expen- diture was Rs 10 or less, was 3212 as against a total of 15 409 families in the sample; and the number of persons included in such families was 11 788, as against a total of 81 554 in the sample. c) Other studies reviewed by the Foodgrains Procurement Committee suggests a lower limit to average per capita consumption of 15 ounces per day and an upper limit of 17 ounces.

Sen does not explain how a 1.8% change in total demand could cause a famine affecting 40 million people. Elsewhere he says:

Those involved in military and civil defence works, in the army, in industry and commerce stimulated by war activities, and almost the entire normal population of Calcutta covered by distribution arrangements at subsidized prices. . . could exercise strong demand pressures on food, while others excluded from this expansion or protection simply had to take the consequences of a rise in food prices.'(Sen, op cit, Ref 2, 1977b, p 51.)

The narrowest interpretation of this is that perhaps one million employees used their high incomes to buy more food. They ate enough to cause three million deaths, and serious hunger for 40 million people. This implies their eating perhaps six times as much as their normal intake. (In his reply to my Development Studies Association paper, Sen said that this six- fold discrepancy was due to the fact that I had ignored dependents. This was another misquotation: I did in fact allow for them - see for example, Famine Commission, op cit, Ref 2, pp 30, 31, 63 and my calculations presented in Table 2. Even if I had neglected them, a substantial discrepancy would remain.) The broadest interpretation implies that some six million people in Greater Calcutta were eating twice as much as usual on average. As many did not, the others would have to have eaten more than this.(Note that with any assumption but zero carryover, this discrepancy would be larger.) Not only were they eating this fantastic amount of food, but they were willing and able to pay from four to 20 times the normal price for it. Sen implies that this odd demand was confined to Bengal, and that people suddenly switched back to normal demand functions when the December 1943 crop was harvested.

The facts given in Sen's sources are different. It is not true that almost the entire population of Calcutta was covered'. The preferential schemes never covered more than a quarter of the population and they were often cut because of shortage of food.(Famine Commission, op cit, Ref 2, pp 31, 32, 63.) Preferential supply schemes, plus the controlled and approved markets, received 32% of the grain available in Calcutta in the first quarter, 43% in the second, 23% in the third and 18% in the fourth (see Table 3). Furthermore, the consumption of Calcutta actually fell by 12% - 45% during the famine, depending on the population estimates. The hypothesis as far as the army is concerned is dismissed by the Famine Commission and elsewhere Sen himself accepts this.(Famine Commission, op cit, Ref 2, p 18; and Sen, op cit, Ref 2,1976, p 1279.)

This hypothesis, which is central to his argument, must be rejected on two grounds. First, it is impossible that such changes in distribution could have taken place. Second, Sen's own sources make it clear that the movement was in the opposite direction. His whole thesis must be rejected on these grounds alone.

Table 3. Imports of grain to Calcutta 1943, compared with normal consumption and consumption under rationing.

Second quarter Third
Net imports of paddy and rice (a,b) tons 31,912 88,568 61,038 91,8242 73,342
Imports of wheat (a) tons 26,000 38,000 99,000 176,000 339,000
Less: wheat sent to country (c) tons -20,000 -100,000


Total grain available tons 57,912 126,568 140,038 167,824 492,342
Grain required at 1944 ration level (Greater Calcutta) (d) tons 118,193 118,193 118,193 118,193 472,772
Net imports as % of ration level % 48 107 118 141 104
Normal consumption (High estimate) (e) tons 181,835 181,835 181,835 181,835 727,342
Net imports as % of normal consumption % 43 94 104 174 91
Normal consumption (low estimate) (e) tons 134,310 134,310 134,310 134,310


Net imports as % of normal consumption % 43 94 104 124 91
Grain required at 1944 ration level (CalcuttaTrade Area) (d) tons 92,191 92,191 92,191 92,191 368,762
Net imports as % of ration level % 62 137 151 182 133
Normal consumption (high estimate) (e) tons 141,832 141,832 141,832 141,832 567,326
Net imports as % of high estimate % 40 89 98 118 86
Normal consumption (low estimate) (e) tons 104,762 104,762 104,762 104,762 419,048
Net imports as % of low estimate % 55 120 133 160 117
Supplied through employers' organizations tons 12,487 36,063 17,902 20,164 86,616
Supplied through controlled shops and approved markets tons 6,988 18,262 14,344 10,868 50,482
Total special distribution tons 19,475 54,345 32,246 31,032 137,098
Controlled supply as % of total supply % 33 42 23 18 27
Controlled supply as % of 1944 ration level (Greater Calcutta) % 16 45 27 26 28

Notes: a) FIC, see text, op cit, Ref 2, pp 219-33. Based on trade statistics and figures supplied by the Civil Supplies Department of the government of Bengal. b) Stocks at the beginning of the year were very low because of reduced imports in 1942 (FIC, see text, op cit, Ref 2, p 219). Paddy has been converted to rice equivalent. Bengal government figures suggest that net imports were lower by 22 000 tons. c) The 120 000 tons sent to country areas are assumed, quite arbitrarily, to have been exported in the last two quarters. d) Initially, it is assumed that grain imports were spread through the Greater Calcutta area. This would not have appeared in the statistics because road traffic, personal baggage and illegal shipments were ignored. Later the assumption that all grain was consumed in the Calcutta Trade Area is considered. This area was assumed to consume 78 % of the Greater Calcutta total (FIC, see text, Ref 2, op cit, p 219). e) The per capita consumption under rationing was between 65% and 88% of the normal level, depending on assumptions about the total population (FIC, see text, Ref 2, op cit, p 219). f) The records for the third and fourth quarters do not include arrivals by countryboat, as no records were maintained for two months, and as the amounts were small for other months.


Inequalities in distribution

Those who believe in the FAD approach which Sen is attacking place great emphasis on the change in purchasing power during a famine, which means that the poor cannot buy food and that the people who had previously been moderately well off are impoverished. They emphasize the need for relief works, soup kitchens, special agricultural loans, loans for artisans and weavers, etc. See, for example, Famine Commission, Bengal Famine Code, op cit, Ref 4. The Bengal Famine Code is the only document I know which says that all famines are FAD. This appears to be a reaction against the disaster caused by diagnosing the 1883 Orissa famine as a Sen-type famine, and applying the measures Sen advises. The people who believed in FAD in 1942 considered that most farmers in 1942 had a reasonable yield with high prices and so were better off. However, rural indebtedness meant that the crop often went to a moneylender or landlord who made all the profits. The indebted farmers had to buy back their food, on credit, at an inflated price. Consumers spent most of their money on food and so could not afford other goods or services - those who supplied these goods and services died.

Sen denies that the FAD approach recognizes these phenomena, though his examples are drawn from sources which adopted the FAD approach. He then makes the unwarranted assumption that because these switches in distribution accompanied the famine, they caused it. His analysis of Mahalanobis, Mukkerjee and Ghosh, on which he spends so much time is, therefore, irrelevant to his thesis. The analysis is, in any case, wrong, as he has used raw, unweighted data derived from a heavily stratified sample (Mahalanobis et al, op cit, Ref 44, Tables 4.2 and 7 A5) rather than the adjacent weighted figures. The question is not just why should one group have starved rather than another, but why should anyone have starved at all. Sen does not answer this question, while the Famine Commission answers both. His hypothesis must be rejected.



Sen brings up the old bogeyman of speculation as one of the most important causes of the famine. He talks of speculative withdrawal', especially between December 1942 and March 1943, but also up to November. There was also vigorous speculation' from March to November.(Sen, op cit, Ref 2, 1977, p 50; 1981, p 76.) There is an enormous literature on speculation, hoarding and storage dating back at least to Adam Smith. One thing is agreed - that the uninformed layman's criticisms of speculation are unfounded. Yet Sen does not provide a model to show why the uninformed layman should be correct in this instance. Instead he quotes the Famine Commission in his support. This is a mistake because the Famine Commission described speculation at a different time, a type of speculation that could not have caused the famine. Speculation would only have caused the famine if it reduced the total supply on the market. Neither the Famine Commission nor anyone else has suggested that speculators reduced total supply either by exporting or holding stocks until the next season (which would have lost individual traders a lot of money). On the contrary, the traders imported (legally or illegally) all the grain they could buy. Furthermore, house-to-house searches for grain in mid-1943 showed that there were no enormous stocks. The failure of speculators to respond to the government's market intervention also suggests that there were no large stocks. There was normal speculation. Speculators did buy grain and standing crops in late 1942, hoping to profit as prices rose. This certainly raised prices, though it may be argued that this was far less important in raising prices than government procurement in 1943. The price rise impoverished millions of people and determined that the poor would die because they had no money to buy food. However, if there was a second or third degree shortage, it was inevitable that millions would die in the absence of large imports. The market mechanism determined that one group of people would die rather than another, but it did not increase the death rate. In fact, given that the government did not import, it reduced deaths. As Adam Smith argued, the speculator puts everyone on thrift and good management' from the beginning of the season, effectively imposing rationing. (Smith, op cit, Ref 1, p 24.) He also ensures that all the crop is not consumed at the beginning of the year and that prices are lower and supplies are higher at the end of the year than would be the case without speculation. It should not be forgotten that if the Bengal government had been successful in their efforts to get speculators to release stocks and bring down prices at the beginning of 1943, or if it had seized and distributed stocks, then Bengal would have run out of food before the next crop and tens of millions would have died. Had there been a second degree shortage and normal storage with perfect knowledge, there would not have been the steady rise in prices that actually occurred. Speculators would have bought in stocks immediately after the cyclone, pushing up prices. They would then have released a more or less constant amount per month over the season with a small price rise over the year to cover storage costs (Figure 3). Had there been no shortage and excessive speculation, as Sen suggests, the price would have risen sharply, then would have fallen to near zero as traders competed to sell surplus stocks - anyone left with surplus stocks when the price went down would lose money (Figure 4). The fact that prices continued to rise throughout the season suggests one or both of the following:


Sen is also at variance with the facts given in his sources on the subject of the timing of the speculation. There was little left on the market for speculators to buy between December and March, and after that there was little indeed available at a very high price - hardly the time for vigorous speculation'. The Bengal government's belief in the existence of enormous speculative stocks was a major reason for their failure to act effectively. The belief in speculation, rather than the speculation itself, was the villain. The popular fear of engrossing and forestalling may be compared to the popular terrors and suspicions of witchcraft'.(Adam Smith, op cit, Ref 1, p 24.) It is concluded that Sen's bald statement that the famine was caused by speculation is contrary to accepted theory, and he has given no reason why in this case the accepted theory is wrong. He is also contradicted by the facts in the sources he cites. The point is firmly refuted.



Sen states that hoarding was a cause of the famine and that there were panic purchases between December 1942 and March 1943. He also talks of panic hoarding from March to November 1943 as being the cause of the famine. For hoarding to be the cause of the famine, it would have had to either change the supply in the season or change the distribution, with the rich consuming or storing more than their needs and leaving less for the rest of the population. The fact that stocks were held by farmers or consumers rather than traders would not have this effect. Nor would the fact that they continued to maintain excessive stocks built up in previous years.

The transient shortages caused by hoarding in a developed country are different. A supermarket tries to have not more than three days' supply in shop and warehouse, while a factory tries to keep stock down to a few week's supply. Any slight hiccup in demand may cause scarcities lasting a couple of weeks. In Bengal, 6-12 months' grain was produced at once, and most of the marketed surplus was stored by traders. The shortage lasted over a year. If one accepts that there was a major shortage, one can argue that demand became extremely inelastic at these levels of supply, so that even a small amount of hoarding could have affected prices. Sen does not accept the shortage though.

In fact, officials complained of hoarding in 1939 when war broke out, in 1941 when Japan entered the war, in 1942 when Burma fell and during the famine of 1943. Presumably, by 1943 it was not the accumulation of hoards but their maintenance that was talked of, which would not have worsened the food situation. It is difficult to understand how, as Sen says, there could have been panic hoarding from March to November 1943 - the famine was already under way, very little was on the market, and very few people could afford to buy even their urgent requirements at the going prices. If anything, the accumulation in previous years meant that an unusually large amount was carried over in private hands from the excellent crop of the previous year, mitigating the effect of the shortage. How many people could hoard? In Bengal, only 20% of the population were well fed even in a normal year; most of the salary and wage earners lived from hand to mouth, and most cultivators borrowed against their next crop to buy food. Some people could have accumulated three or four months' supply at a time over a normal year, but this would have run out in the middle of the famine. Very few could have or would have accumulated enough to last from one harvest to another with some left over. However, Sen states that not only did people do this in the high price year of 1943, but they were able to carry on building up their hoards until November 1943, as prices rose to 20 times the normal level.

To put this into perspective, if l0% of the population had built up their stocks to 6 weeks supply in 1942, this would have increased total demand by just over 1 % in that year. In the unlikely event that they doubled their stocks instead of eating them in the famine year, it would still only add 1.1% to total demand in 1943, if, as Sen says, there was no shortage. If there was a serious shortage though, there would have been a 1.6% increase in total demand, a significant proportion of the marketed surplus from the production sector.

Sen's assertion that there was hoarding is made without any supporting evidence or analysis. What evidence there is makes it unlikely that there was any increase in the amount hoarded in 1943, and virtually impossible that there was in March to November 1943 as he states. This hypothesis is refuted. Hoarding, like speculation, is a bogeyman invoked by politicians and administrators. Because they believe that there are large private hoards and that there is really plenty of food, they take no effective action. The result can be famine.


Failure to import

Sen says that a contributory cause of the famine was the failure of the government to import more grain. This contradicts his statement that there was adequate food available. Logically, he can only say that it was a failure to take one of the many possible, non-essential measures. The Famine Commission, on the other hand, believed that there was a second or third degree shortage, so it was a culpable failure to take the only action which could have ameliorated the famine.


Borderline between two price regimes

Sen presents the following as a causal explanation of the famine. There is no further elaboration of the argument.

Finally, it is perhaps significant that the Bengal famine stood exactly at the borderline of two historical price regimes. Prices had been more or less stationary for decades (the 1941 rice price was comparable to that in 1914), and the price rises (especially of food) that began in 1942 were to become a part of life from then on. Institutional arrangements, including wage systems were slow to adjust to the new reality.'(Sen, op cit, Ref 2,1971 b, p 51.)

Why was it that Bengal alone should have been devastated by famine when the same applied to virtually every country in the world?

Previous | Return to main | Next