HOW AMARTYA SEN'S THEORY CAUSES FAMINES
The problem of famines and food shortages is one of the most acute facing agricultural economists. Today fifteen countries have famines and thirty million people face starvation. Since the mid 1970s Professor Amartya Sen's approach to the economics of famine has become influential. He has argued at some length that a major cause of famines is not a sudden decline in food availability, but a sudden redistribution of what food is available. It will be argued here that there are major weaknesses in his theory which mean that it is more likely to cause famines than to cure them. It will be argued that his theory and analysis are wrong and that there are inconsistencies between the arguments he presents. Furthermore, the implications of his theory are contradicted by the facts given in the sources he uses. There are also ubiquitous and systematic inconsistencies between the facts he gives and the facts given in his sources.
This monograph will concentrate on Sen's analysis of the Bengal famine of 1943, as it is the one he gives most attention to, it is the best-documented one, and it is the one for which his theory is most plausible. To be absolutely fair to him, the analysis will rely entirely on the sources he quotes and no new evidence will be presented.
The analysis is presented purely as a refutation of Sen. It is not a complete analysis of the Bengal famine. It would take a book to do justice to so important and so complex a subject, and the book would not overlap with Sen's analysis to any degree. The analysis will not cover the practical problems of administration, physical distribution or rationing, though they were important in the Bengal famine. Nor will it consider the failures of long-term agrarian and food policy which made the situation so critical and so difficult to deal with. However, the points that are discussed are not trivial: the failure of the authorities to understand them caused three million deaths in 1943.
The language of normal economic theory will be used rather than that of Sen's entitlement theory. There are several reasons for this. First, Sen himself used this language when dealing with the Bengal Famine, with his occasional mentions of entitlement declines etc. being external to his analysis. Second, we are concerned with what actually happened, rather than with the labels put on some of the effects. Third, the use of the value-loaded vocabulary of entitlement would confuse those who are not familiar with it, or who do not agree with it. Finally, it has become clear in discussions that different people interpret his entitlement theory in very different ways.
2 CAUSES OF FAMINE
Both Sen and the economists who worked in the mainstream tradition accept that famine can arise in one of three ways. The first is when there is a fall in the aggregate food supply of a country or region. This may happen because of drought, flood, plant disease(the potato blight), pests (the plague of locusts), blockade, post-harvest losses, or disruption of production by migration, epidemic or war. Problems with storage or transport or the manipulation of stocks by speculators can mean periods of famine even in a year in which there are adequate supplies.
The second way is an increase in aggregate demand. This may arise from immigration, an influx of refugees or an invading army. It has happened that foreign buyers or an occupying army have been able to export the food supply. A change in farmers' willingness to sell may be best treated under this heading, as a change in reservation demand.
Third, a change in the distribution of what food is available between people in different occupations may bring about a famine when there is no food shortage. For example, hyperinflation, unemployment or enclosures have left some people without the means to buy food. Changes in income distribution may mean that some people eat more, leaving less for the others. These changes should not be confused with changes in geographical distribution because of transport failures etc. which would come under a fall in supply. Nor should these changes be confused with the maldistribution that exists to a greater or lesser extent in every society, which may cause chronic malnutrition among the poor, but which does not cause famine. It is a truism that the world could feed itself, so all famines are due to a change in relative distribution, if the famine area is defined sufficiently widely. To do this ignores reality: there are political, social, physical and economic constraints to redistribution which cannot be assumed away.
The classical view, expressed by Adam Smith (1776, ii p 26) for instance, is that, while all these causes are possible, in practice a shortage in supply was the cause of all the famines in the two hundred years before he wrote. The modern mainstream view has been less extreme. It has been thought that shortage of supply is the most usual reason, but that other factors do, on occasion, cause famines. Moreover, it has been noted that even if a famine is caused by a change in supply alone, there will be evidence, during the course of the famine, of a change in demand and a change in distribution. For example, a shortfall brought about by a flood will change demand, both reservation demand and people's expectations. It will also change the relative purchasing power of different socio-economic groups and will provide a strong incentive for speculation. In fact, the sudden shock to the system can be expected to bring to a head all those agrarian and social problems that have been festering below the surface for decades. These are effects of the supply change, not the cause of the famine.
Sen disagrees strongly with both the classical and the modern traditions. Since the mid 1970s he has argued at length that many or most famines are caused by a subsection of the third type, by a change in people's ability to buy an adequate supply of food. He is scathing about those people who argue that a reduced supply is the reason, and particularly about those who believe that it was a major cause of the Bengal famine of 1943 (Sen 1976, 1977, 1979, 1980a, 1981a, 1981b, 1984). He calls their view the Food Availability Decline (FAD) view. It should be noted that Sen is not proposing a new explanation for famine, but is saying that it is normal, rather than very rare, for a famine to be caused by redistribution when adequate supplies are available.
3 WHY DOES IT MATTER?
The disagreement between Sen and mainstream economists is not of mere academic interest. It strongly influences the action a government will take to prevent famine and the action it will take if a famine occurs. Millions of lives depend on it.
This monograph is not concerned with long-term food policy. Suffice it to say that the traditional food security and famine prevention measures, of producing a surplus in normal years and building up an emergency stockpile, are not appropriate if famines are not caused by supply shortages.
It will be argued throughout this monograph that Sen's theory of the causation of famines, and the analysis he considers appropriate to it, will lead to a misdiagnosis of the cause of famine, the degree of famine and the appropriate action to deal with the famine.
It would be easy enough to show that economists in the threatened country are likely to misdiagnose the problem and the appropriate solution using Sen's analysis, so causing famine. They are not highly-trained theoretical economists. They have limited resources and extremely unreliable statistics; they are working under extreme time pressure and political pressure; and they have an awesome responsibility. Instead, I propose to show that Professor Sen himself has consistently reached the wrong conclusion, and moreover that the bias has been one that would lead to famine. His books and papers were not subject to the time constraint that face the economist in the field. He has dealt with some of the best-documented famines in history and he has had the advantages of hindsight and ex-post data. He was far better placed than the people in the field to make a diagnosis, yet he was wrong.
In the next sub-section it will be shown that the effects of misdiagnosing a famine as a Sen-type famine are serious. The result will be that government will take action which may be totally ineffective or which may worsen the situation. Making the opposite mistake and misdiagnosing a Sen-type famine as a Food Availability Decline (FAD)famine does not matter. The action taken will ameliorate the famine.
3.1 Food Availability
Unlike Sen, I consider that one cannot discuss famines without constantly taking into account aggregate food supply. For this reason I should like to distinguish several degrees of food shortage
NO SHORTAGE. If there is plenty of food to go round, a famine can only result from a major change in distribution.
FIRST DEGREE SHORTAGE. There is sufficient food to provide a barely adequate diet for everyone, provided that there is rationing. If there is not, some sections of the population will suffer from serious malnutrition, or starvation.
SECOND DEGREE SHORTAGE. The population has insufficient food for long-term survival. Rationing would leave everyone hungry and suffering from deficiency diseases, but surviving until the next harvest, though there would be some excess mortality among the sick, weak and old. In the absence of rationing, there will be widespread starvation.
THIRD DEGREE SHORTAGE. Insufficient food for survival. Starvation is inevitable without imports. If there were no imports and everyone were given a bare survival ration, food would run out before the next harvest, and the whole population would die. Historically, populations have avoided this fate by leaving the young, the old and the weak to die, so that the able-bodied adults might survive. In Bengal in 1943, the young, the old and the poor were left to die, wives and families were abandoned, children were sold into prostitution, and useless consumers were murdered (see for example Famine Commission, 1945a; Rajan, 1944; K.C. Ghosh, 1944). Usually a population will resort to emigration, war or cannibalism before accepting annihilation. No doubt the human race has survived only because our ancestors made such decisions.
A critical limitation of the application of Sen's theory is that a famine can only be caused by a change in distribution when there is no shortage, or, just possibly, when there is a first degree shortage. With a second or third degree shortage, there is a famine situation even with normal distribution. Indeed, with a third degree shortage, changes in distribution are needed if anybody at all is to survive.
3.2 Changes in the Degree of Shortage
The degree of shortage, the food availability per head per day, does not remain constant from one harvest to another. Imports can improve the situation. So can migration, though at an enormous social cost. Mismanagement can easily worsen things. For example, if a government does not impose any rationing when there is a first-degree shortage, so much food may be consumed at the beginning of the year that a second-degree shortage is created. If it issues the rations appropriate to a first-degree shortage when there is in fact a second degree shortage, then later in the year there will be a third degree shortage, with insufficient food for survival. Mass starvation is a very real threat.
Again, if government decides that the famine is the result of hoarding by merchants (and such complaints are made during every shortage), it might force them to release enough grain onto the market to keep supplies and prices at the normal level. If supplies were in fact 25 % below average, the result would be that the food supplies would be exhausted three months before the next crop.
3.3 The Effect of Misdiagnosis
The traditional view, and the view of the most hardened believers in FAD, is that all famines should be dealt with by issuing food to the poor, by relief works with payment in food, and by loans to prevent the impoverishment of farmers and artisans (Appendix I gives details of the provisions of the Bengal Famine Code). Market intervention (i.e. price control, control of stocks, seizure of stocks, prohibition of exports etc.) is also normal.
In addition to this, rationing is taken to be a possible complete solution for first and second degree shortages, though the practical problems are such that it would be unsafe to rely on it. There are major problems deriving from poor administration, corruption, political interference or difficulties with physical distribution. There may not be sufficient maldistribution for fair shares' to produce much improvement; in Bengal in 1933, for instance, only 22% of the population were well nourished (Famine Commission, 1945 p6) and since the middle classes ate considerably less rice (and more protective foods)than the average, rationing of rice, the staple food, would not have had much effect on the proportion of people who were well-nourished. Rationing is more likely to be effective in reducing aggregate consumption than in producing optimum allocation.
Imports are essential for a third degree shortage and, in practice, for most second-degree shortages.
Sen disagrees. As will be shown below, he states that it is wrong to concentrate attention on the degree of shortage when determining how to tackle a famine, and indeed that it is wrong to consider degree of shortage at all. A Sen-type famine, caused by changes in distribution alone, can be dealt with by the basic measures alone - issuing food, relief works, loans and market intervention. Seizure of stocks and a formal rationing system may be needed. Misdiagnosis of a FAD famine with second or third degree shortage as a Sen-type famine means (a) no rationing at all or excessive rations, and (b) no imports. The result will be that no effective measures are taken and the situation is worsened.
Unfortunately, the uninformed laymen, whether politicians or administrators, are easily convinced that the high prices are due to something they think they understand, like speculation, inflation and hoarding, and something they can deal with by administrative action. They feel that they have acted decisively and usefully if they shoot a few speculators and seizes their stocks for distribution. There is a reluctance to accept the horrifying responsibility of dealing with a third-degree shortage, and to accept that one is totally powerless in the face of forces beyond one's control.
4 THE BENGAL FAMINE
In 1943 Bengal suffered from a famine that resulted in perhaps 1.5 million deaths from hunger and the same number of deaths in the epidemics that hit a population weakened by hunger. The Commission of Inquiry that was appointed on the famine produced two remarkably detailed reports on the famine, the efforts that were made to control it, and the measures needed to prevent further famines (Famine Inquiry Commission - referred in future as FIC - 1945a, 1945b). The reports have been the main source of information for subsequent studies. They were highly critical of the Bengal Government, the Government of India, the Viceroy, the Imperial Government and the grain traders. Indeed only Wavell, who took over as Viceroy in October 1943, escaped with his reputation untarnished. An indication of the quality of the report is that it accepted and documented nearly all the criticisms of the authorities made in the highly critical, highly political books on the famine written by Hindu and Muslim nationalists (e.g. Dutt, 1944; T.C.Ghosh, 1944; K.C.Ghosh, 1944; Rajan, 1944), and it added a good number of criticisms of its own. The Viceroy was only too well aware of the political embarrassment that would be caused by exposing the inefficiency of the administration (Wavell in Moon(1973) pp 36-7), and it was widely believed that the Indian Government tried to limit the damage by printing only a few copies on inferior paper (Aykroyd, 1974). The English editor of the Calcutta Statesman, who had done most to bring the famine to the notice of officialdom in Calcutta, Delhi, London and Washington (defying censorship and the wrath of the local officials to do so) commented "The Famine Commission's report is as complete, painstaking and balanced an account of what happened and why, as will ever be achievable." (Stevens, 1966).
Very few people have had the opportunity to check Sen's work against its sources, and only a handful have done so - myself, my referees and a few researchers.
When Pergamon proposed to publish The Famine Commission Report as a classic and relevant account some years ago, Sen made a strong recommendation that the report was so flawed as not really to be worth putting out as a book. As a result it is still virtually impossible for most people to check his work.
Not only does this mean that misstatements have gone undetected, it means that Sen has been credited with presenting new insights, new analysis and new data when they appear fully developed in the Famine Commission Report.
Other sources must be collected from specialist libraries in London, Oxford and Cambridge for instance. It is time consuming and expensive to get it. Again, it is impossible for most of his readers to get hold of.
I would be very pleased to assist readers in obtaining source documents.
Rice is the staple food of Bengal, accounting for 80-90% of the calories consumed. There are three rice crops. The aman' crop, harvested in November or December, provides 74% of the rice supply, with the upland aus' crop harvested in August and September providing 24 % and the boro' crop harvested in February or March providing 3 %.
In December 1940 there was a much reduced aman' crop, and, as a result, there were shortages and isolated outbreaks of famine in 1941, which were easily handled with relief measures and market intervention.
In December 1941 the rice crop was well above average, but prices started high and rose throughout the following year.
In July 1941 Japanese assets in the Empire were frozen. On 28th July Japan invaded Indo-China. On 8th December war broke out. There were air raids on Rangoon on 23rd January 1942 and on 7th March Rangoon fell. On l5th February Singapore fell. On 5th and 6th April Ceylon and Eastern India were bombed.
This meant a serious reduction in supply to India, as Bengal and other areas such as Madras relied on imports from Burma to make up their normal deficit, as did Ceylon. Indeed, it meant that India had at least a first-degree shortage which required strict control and rationing throughout the war to prevent further famines (Aykroyd, 1974; Moon, 1973 pp 32, 174). The threat of invasion meant that many producers held onto some of their stocks as an emergency supply. It also caused demand to rise. There was an influx of refugees from Burma, and at the same time consumers tried to build up their emergency stocks. These supply and demand changes caused prices to rise throughout India. As Bengal had had a good December 1941 crop, some surplus rice was exported to the other areas of India which had been hit by the loss of the Burma rice.
On October l6th 1942 a cyclone accompanied by torrential rains hit West Bengal, causing wind and rain damage and flooding. Three tidal waves laid waste a strip of land seven miles wide along the coast and three miles wide along river banks. The resulting high tides increased the flooding caused by the rain. Some 4000 square miles were affected.14,500 people and 190,000 cattle were killed; crops and grain stores were damaged. Fungus disease and root-rot then hit the sodden crops, causing even more severe damage than the flooding itself.
When this poor winter crop was harvested, prices rose rapidly, doubling within a month in the country areas. By March 1943 there was hunger throughout Bengal, and from July to November, the famine was in full spate. Relief measures of the kind mentioned in Appendix I were introduced immediately in the area affected by the famine (FIC pp 32,65, 66, 236). They were extended throughout the country as the famine hit. The measures were totally inadequate in most areas and were only really adequate in Calcutta.
In December 1943 the Bengal Government thought that the shortage was mainly psychological. (See Appendix II for a detailed description of the actions of the Bengal Government through the famine.) They believed that there was no real shortage and that the price rises were caused by speculators. As a result, their main strategy throughout the famine was to "break the Calcutta market", releasing imported, purchased or seized grain on the market in an attempt to frighten speculators and hoarders into selling off their surplus stocks. They had great difficulty in buying any rice in the country areas in the first quarter of 1943, even when officials were told that there was no limit on the price they could pay. They were able to buy only 23,000 tons out of a normal annual consumption of 9.6 million tons, and releasing this had no impact on market prices. A further 28,000 tons was imported in March, which, again, had no effect. In June and July a further 90,000 tons were imported, again with no effect on prices.
In April and May there was a propaganda drive to persuade speculators and hoarders to release their stocks, with no apparent effect.
In June 1943 there was a house-to-house search for secret stockpiles, but it was found that there was very little in stock, not enough to last to the November harvest. It was around this time that the Bengal Government came to believe that there was in fact a serious shortage.
Throughout the famine imports into Bengal were low. Initially, this was because the Bengal Government thought that the shortage was no more serious than that of 1941, and thought that it would be possible to manage with the food available. In the second quarter of 1943 it was thought that imports of 300,000 tons would be sufficient to cure the famine (This was the quantity eventually imported). From the middle of the year government began to believe that much bigger imports were needed. It was found though that other provinces refused to permit exports to Bengal. India as a whole had a first degree shortage and prices were rising, so any exports from a province would cause further price rises and would cause civil unrest. For a brief period in June and July, Bihar and Orissa permitted exports to Bengal (after much pressure from the Indian Government), but they stopped these when their own prices rose sharply. The Imperial Government would not provide grain or shipping.
In November 1943 the new Viceroy, Wavell, increased exports to Bengal and sent in the army to improve the physical and organisational distribution of the grain. The September crop helped, but the famine did not end until an exceptionally good, though rather late, crop was harvested in December. However, epidemics now took over as the main killer, and as many people died between November 1943 and July 1944 as in the previous period.
It is estimated that between one and two million people starved and the same number died from disease, giving a total of between two and four million deaths. Sen (1977 p 33)calls this famine "possibly the biggest famine in the last hundred years". However, Masefield (1963 pp l2-14), who he quotes on the history of famine, mentions half a dozen where the death toll was higher, nearly ten times greater than the official death toll in Bengal (which includes death by disease) in one case:
The Sahel Famine of the 1970s could be said to have had a smaller death toll, but only if redefined into many small famines.
Again, Sen's statement clearly conflicts with the facts in his source.
How many people were affected by the famine, apart from those who died? 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 (1944)). An independent estimate was made by Mahalanobis, Mukkerjee and Ghosh (1946), 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 household 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 because they had sold all they had. Taking an average family size of 5.4, it seems that perhaps 10 to 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.
5 CAUSES OF THE FAMINE
There are several different versions of the causes of the famine. I will set them out here and examine them at length later.
5.1 The Famine Commission Version
The Famine Commission argues that the basic cause of the famine was a reduction in the food supply. This was due to the poor December 1942 crop, and to the fact that there was a reduced carry-over of supplies from previous years, resulting in at least a second-degree shortage, with insufficient rice available to keep the population healthy, no matter how it was distributed.
Gross mismanagement of the crisis, particularly by the Bengal and Indian Governments, meant that nothing effective was done to alleviate it. They took only the action appropriate to a first-degree shortage. The relief measures were totally inadequate for a problem of this scale. There should have been massive imports. A rationing system should have been introduced in Calcutta at least. The government should have seized all grain stocks and should have taken over the whole grain trade.
This brief summary does the Famine Commission an injustice. Their report is rich and closely argued and by no means as simplistic as I have suggested.
5.2 Professor Husain's Version
Professor Husain, a member of the Famine Commission, argued that the shortage was even worse than the rest of the Famine Commission believed. He argued that the carry-over of old stocks from 1942 to 1943 was very low indeed, so there was a third degree shortage, and serious famine was inevitable in the absence of major imports.
The grain trade generally made the same assessment, both because of their knowledge of stock levels, and because they thought that the Department of Agriculture had over-estimated the December 1942 crop. They invested accordingly, and made a lot of money.
5.3 Professor Sen's Version
Professor Sen presents his version in support of his claim that this famine and many others were not caused by a decline in food availability, and that the Food Availability Decline approach to famine analysis is inappropriate. He states that there was plenty of food available in Bengal in 1943, at least 9% more per capita than there had been in 1941, when there had been no famine, so the famine did not arise because of a decline in the availability of food. In particular he doubts that there was a reduced carry-over at the beginning of 1943, as the Famine Commission says. Instead, the famine occurred because of a change in the distribution of existing food supplies, arising from wartime conditions, particularly inflation. This meant that some groups of the population got higher incomes and ate more, leaving little for the rest of the population. At the same time others did not have sufficient money to buy food. He provides the following causal hypotheses (Sen 1977 pp 50, 51; 1981a pp 75-78), which will be considered in a later section:
The first six of these explanations are all seen as having the same result, concentrating purchasing power and food consumption on a small section of the population, so that there was insufficient food available for everyone else.
Sen's whole argument tends towards his conclusion that "the failure to anticipate the Bengal famine . . . and indeed the inability even to recognise it when it came, can be traced largely to the government's overriding concern with aggregate food availability statistics." (1984 p477) He suggests that if the problem had been analysed using his approach, the famine would not have occurred.
It will be shown in some detail in this monograph that all Sen's explanations were discussed in the source documents, notably the FIC. They were part of the popular explanation at the time. Furthermore much of the policy of the Bengal Government at the time was based on them, with disastrous consequences. It will also be pointed out that they were popular during the 1888 Orissa famine and as far back as Adam Smith at least. The inflation explanation for famines was very popular after the Central European hyper inflations 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: a challenger to Lasker for the world championship later died of hunger in Austria during hyper inflation. I do not know of anyone who would deny the possibility that this could cause a famine.
6 PROFESSOR SEN'S EXPLANATIONS EXAMINED
In this section Professor Sen's explanations for the Bengal Famine will be compared with those of the Famine Commission. In assessing the explanations it will be borne in mind that it is always possible to support a hypothesis or "explain" a single phenomenon with two or three pieces of selected evidence and a minimum of theory. Accordingly, the competing explanations will be examined to see if they are in accord with all the facts in the sources, not just a select few. The different explanations given will be examined to see if they are compatible with each other, with the facts, and with economic theory.
Sen's first causal hypothesis is that the famine was set off by factors related to wartime inflation, though the fullest and clearest statement I have been able to find in his works on the subject is
However the facts as presented in his sources give very little support for this. First, from the beginning of the war until the cyclone hit Bengal, the Working Class Cost of Living Index in Calcutta (which was very heavily weighted by the price of grain) rose slightly less than that in Madras and Bombay. (Singh, 1965 pp 95-99). From October 1942, when the cyclone struck, it rose much faster (See Figure 3). Sen does not explain why this inflationary pattern should suddenly change and affect Bengal so much more seriously. Nor does he explain why the change should happen to coincide with the cyclone.
Again, up to the cyclone, the rice price rise was not more acute in Bengal than elsewhere in India. Rice prices were higher elsewhere in India and there was a thriving export trade from Bengal (FIC pp 17, 18, 21, 22, 23, 28, 29). The United Provinces and the Central Provinces had crops that were poorer than even the crops of the previous year (FIC p17).This, together with the loss of Burma imports caused prices to rise, and would go some way to explain the price rises in Kanpur, which started before the cyclone (Figure 3).
To some extent inflation caused rice prices to rise. However, it must be questioned how far the price rises even before the famine were caused by inflation, rather than by the shortages. The Government of India was, of course, obsessed by wartime inflation, but was particularly concerned with grain prices as being a cause of inflation rather than a result. The Famine Commission gave a detailed and convincing account of the factors causing the rise in the rice price during 1942. They included panic hoarding by farmers and consumers following the outbreak of war with Japan; a shortfall of rice in India as a whole (but not in Bengal) as a result of the loss of the Burma rice; the loss of rice stocks in the cyclone; changed expectations after the cyclone; and the reduction in wheat imports by Calcutta. Their explanation of how grain prices started to rise after the cyclone had destroyed 30 % of the crop and after inter provincial trade had been banned is completely convincing. It also explains why the famine was confined to Bengal and to those parts of Orissa hit by the cyclone.
Sen himself quotes figures showing that the price of rice rose much faster than the price of fish, umbrellas, milk, haircuts etc. in spite of the fact that the supply of fish was much lower than usual. (1977 pp45, 46). Yet these goods were presumably equally hit by inflation.
Sen makes no attempt to explain why the inflation should have affected Bengal alone. Nor does he provide any market mechanism to explain how the inflation was transmitted to the price of grain - one would have expected the impact to have been concentrated on scarce consumer goods. He does not explain how paying a good wage to a few hundred thousand factory workers would result in an inflationary price rise in the grain bought by 50 million people. He himself quotes evidence that this extra demand in the industrial sector was offset by a falling demand in the agricultural sector.
On page 51, again, he quotes the Famine Commission as saying that generally the agricultural labourers were not affected. He then talks of a decline in employment in the sector (p44). It is not clear from this how the inflationary pressure would be put onto food prices, with so few people affected. Finally, he does not explain why the inflationary effect should cease to apply on the harvesting of the December 1943 crop.
It must be concluded that the Famine Commission provides by far the most satisfactory explanation of price rises both before and after the cyclone. No doubt inflation had some effect, but it was minuscule by comparison with the rises when the shortages began to bite (see Figures l, 2 and 3).
A major reason for the rise in prices in 1943, which is not mentioned by Sen, is the purchasing of grain by the Bengal Government. They bought grain for relief, for public works and for "breaking the Calcutta market". For part of the period they were bidding against the trade without much limit on price. The danger of this practice was recognized explicitly by Malthus (1800) and by the Bengal Famine Code. It would not matter if there was no real shortage or a first degree shortage - everyone would get food, though prices would rise. However, if there is a more serious shortage, it would lead to a never-ending price spiral, because there is just not enough food to go round, however much money is paid. This happens both if the money is given to the poor to buy grain, or if the government buys grain on their behalf. If the Famine Commission's belief that there was a shortage of grain is accepted, then this must certainly have happened. This another major disaster brought about by misdiagnosis of the cause of the famine.
Again and again Sen's explanation is contradicted by the facts in his sources.
Sen places speculation after inflation 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 1943 (1977 p50; 1981 p76). There is an enormous literature on speculation, hoarding and storage dating back to Adam Smith at least. It is agreed on one thing: that the uninformed layman's criticisms of speculation are unfounded. Yet Sen does not provide a model to show why the uninformed layman's criticisms should be correct in this instance. Instead, he quotes the Famine Commission in his support, though in fact the Famine Commission describes a speculation which happened at another time, and which, moreover, was a form of speculation which could not have caused the famine.
The type of speculation that could have caused a famine is where traders buy up grain and withhold it from the market to push up the price. They reduce the total supply, and the surplus is exported, destroyed or carried over to the next season. There has been no suggestion in the sources that this happened, and nor has Sen argued that this happened. On the contrary, they imported all they could as soon as they saw that the December 1942 crop had failed. They imported all they could for the three months when there was free trade between Bengal, Bihar and Orissa. They smuggled in whatever they could throughout the period. In addition, from September 1942 to October 1943 the trade was trying to get the Bengal Government to import more. Since there was no suggestion that speculators reduced the amount of grain on the market during the famine year, one cannot argue that speculation caused the famine.
It is agreed by the sources that there was the other type of speculation based on normal storage, where merchants buy and store grain and then release it through the season. If the selling price is greater than the purchase price plus the storage cost, they make a profit (and my own observations in Africa and Asia confirms the theoretical prediction that they often make a loss). The Famine Commission was concerned with this kind of speculation. As civil servants in wartime, they were concerned about speculation and profiteering and its effects on price levels generally. As human beings, they were concerned that Government's failure to introduce famine relief on a sufficiently large scale had condemned to death those people who could not afford to pay the high prices. They pay a great deal of attention to the fact that traders and large farmers bought up the crop at the beginning of the season and made enormous profits out of selling it, bit by bit, at famine prices (and, of course, the fact that they liquidated their stocks and prices still continued to rise confirms that there was a shortage). It meant that the people who starved were the poorest classes, such as landless labourers and very small farmers, as well as those whose ability to purchase grain was altered by the famine, like small traders, suppliers of services and employees. Those who had goods to sell impoverished themselves to survive, and as a result there was a further concentration in land ownership (and Bengalis still suffering the after effects of this transfer of resources).
However, the fact that there was this enormous transfer of resources does not mean that speculation increased the death rate. If we accept the Famine Commission's view that there was at least a second-degree shortage, we accept that some people were certain to die, in the absence of large-scale imports. The high prices determined that it was the poor who died rather than the rich. Had the same quantity of food been sold at a price only50% above the normal price, the same people would have died. There would not have been the same enormous transfer of property, though.
In fact, given that the Bengal Government was doing nothing which could relieve a second or third degree shortage, this speculation saved a lot of lives. If traders had not held on to stocks for a higher price, there would have been a worse catastrophe. If, at the beginning of the year, they had released the amount of grain the Bengal Government wanted released, the country would have completely run out of grain well before the harvest. In the absence of any effective government intervention, high prices are beneficial. They get everybody to reduce consumption, and to ration supplies through the season - and during the Bengal famine prices rose so high that virtually everyone had to restrict consumption.
I shall take hoarding to mean increased retentions by farmers and increased stockholding by consumers. It is possible for both to occur without altering the supply of grain throughout the year: farmers rather than merchants hold the stocks and consumers try to buy early in the year rather than stagger their purchases over the year. Some market disruption will, of course, be observed. This hoarding would not cause famine.
It is possible, though, that farmers and consumers would hoard by increasing their stocks until they had a substantial surplus at the end of the year. This change in distribution would mean that there was less available for other people, so it could be taken as a cause of famine. This would only be so if the hoarding happened during the famine year; if it happened the year before it would reduce the problem. Sen states that it did in fact happen during 1943.
Sen's chronology conflicts with that of the Famine Commission. They point out that the entry of Japan into the war, and the fall of Singapore and Burma took place in 1941 and at the beginning of 1942. As early as May 1942, it was reported that "cultivators on the one hand were becoming very cautious and unwilling sellers, and speculators on the other hand were operating on a larger scale than in normal times and circumstances, with only one consequence, a steady rise in prices" (FIC p28). The bulk of the hoarding took place then.
It is difficult to accept Sen's statement that there were panic purchases between December 1942 and March 1943 and that there was panic hoarding from March to November 1943. Prices were then too high for any but the very wealthy to buy in excess of their own needs. Supplies on the market were low, well below normal levels (FICp11). For example, imports by Calcutta in the first quarter of 1943 were half the normal level, and in the second quarter were still below normal (Table 1) - the figures are readily available in the Famine Commission report - so the purchases by consumers for hoarding were not higher than normal in Calcutta at any rate. There may have been increased retentions by farmers out of the December 1942 crop. It is unlikely though that people who had built up an emergency reserve in 1942 would have added to it in 1943.
All in all, it is likely that increased hoarding, in the sense of purchasing or retaining more than one's needs until the next crop, was largely confined to 1942. It was then that there was a threat of invasion, it was then that the grain was available at a price that made hoarding possible. If this were so, it would largely explain the price rise during 1942,when there was a good crop. It cannot, however, explain the famine; on the contrary, it means that Bengal would have gone into 1943 with a higher than usual reserve supply in private hands.
How important was hoarding? Who was in a position to hoard for personal consumption? Wage and salary earners, small traders and craftsmen lived from hand to mouth and would not be in a position to buy more than perhaps a week or two's extra grain. If they had accumulated it bit by bit over 1942, buying a little extra each month, they might have accumulated a month's supply - which would have been consumed early in the famine year. In 1943 high prices made any such accumulation impossible. Wealthy traders, professionals and large farmers could well have hoarded enough grain to last them to the end of the year, with a couple of bags over. Some medium-sized farmers could have done the same, but most farmers did not produce enough grain to live on, and relied on paid work to fill the gap. Many were so seriously indebted that the bulk of their crop went to moneylenders, traders or landlords immediately it was harvested, so they had to borrow money through the season to buy food. They were in no position to accumulate reserve stocks.
From the fact that only 20 % of the population was well-nourished in normal times, and that only perhaps 1 million out of 7.5 million farming families had enough land for subsistence (FIC p10), we can conclude that not many people were in a position to hold stocks of the order of a year's supplies, or to hold substantial emergency reserves. Very few can have been in a position to buy eleven months' supplies in December, enough to see them through to the autumn crop, with a bit in hand. If they had bought three or four months' supply at the beginning of the year - a major investment - it would have been exhausted well before the next crop was harvested. This is confirmed by the fact that two-thirds of the population was hit by the famine.
Table 1. Imports of grain to Calcutta 1943, compared with normal consumption and consumption under rationing.
The order of magnitude we are considering is that perhaps 10% of the population may have stored 10 % extra in 1942 - an increase in total demand of 1 %. In 1943 perhaps 5% did, an increase in demand of ´ %. A case could be made for a much lower figure.
This gives no support at all for Sen's hypothesis that hoarding substantially reallocated food supplies in any of the periods, January to March 1943, March to June 1943 or June to December 1943. If anything, it suggests that grain hoarded in previous periods would have been consumed in these periods, reducing demand.
Hoarding, like speculation, is a bogeyman invoked by politicians in time of scarcity. Because they believe that there are large private hoards, and that there is plenty of food really, they take no effective action. The result is famine.
6.4 Uneven Expansion of Purchasing Power
Closely linked to the above hypotheses is Sen's causal hypothesis that the famine was caused by an uneven expansion of incomes and purchasing powers (Sen 1977 p 51; 1981 p77). It is set out most clearly as follows
This change in incomes did not in fact take place. However, it would be instructive to look at the implications of the model. If 10 % of the population increased their consumption from 14 oz. per day to 17 oz. per day, there would be 1.8 % more consumed in total (See FIC p204 for the consumption figures in five surveys in the years preceding the famine). Sen states that this 1.8 % would all come from one group, presumably those who died. This supply and demand response is quite unlike that normally assumed in economics, where the increased demand would affect the prices paid and the amount consumed by everybody. The effect described could arise from armed robbery, but not from the workings of the market. Sen's market mechanism is not explained, nor is it explained why two-thirds of the population should have been hit by famine as a result of a 1.8% increase in demand.
Elsewhere, he states a somewhat more credible hypothesis, that the army, industries and commercial firms got preferential supplies of food. As a result, their employees ate more, leaving insufficient food for the rest of the population. As far as the army is concerned, he is clearly wrong: the army, mainly wheat-eaters, consumed very little extra in relation to India's supplies, and the army in Bengal was supplied externally (FIC p18).Furthermore, the soldiers would have eaten even if they were not in the army. Indeed, elsewhere Sen himself accepts this (1976 p1279).
Sen says that "almost the entire normal population of Calcutta [was] covered by distribution arrangements at subsidised prices" (Sen 1981 b p77). This is untrue. There were indeed preferential supply schemes for the employees of industrial and commercial firms, but these never covered more than a million people out of six million in Greater Calcutta (FIC p30). These schemes, plus the controlled shops and approved markets, got 32% of the grain available in Calcutta in the first quarter of 1943, 43 % in the second quarter, 23 % in the third quarter and 18 % in the fourth (See Table 1. The figures are set out at length in the FIC pp219-233). In no sense can it be said that almost the whole population was covered.
Nor can it be said that those who were covered were fully insulated against the famine:
It will be noted too that the issue given (3.58 seers per employee per week) was totally inadequate for an employee with a family.
It is difficult to square these facts with Sen's statement that one million employees ate so much extra because of the special issues of food that a famine hit the remaining 59 million people of Bengal. This would imply that each and every one of them ate six times as much rice as in normal years.
Sen can be interpreted as presenting a third possibility, that the people of Greater Calcutta, 6 million of them, ate so much more that there was not enough food to go round for the rest of the population. This is an all-embracing hypothesis. It includes the hypotheses above, that half of the very poor got a higher income and ate more, and that industrial workers ate more. It is the end result of all the other hypotheses, that inflation and high prices meant that one class ate more, leaving less for the rest. It includes the hypotheses that the preferential distribution of grain plus the disposal of public purchases and seizures on the Calcutta market meant that an unduly large proportion of total supplies ended up in Calcutta.
Again, Sen appears to be asserting that six million people ate so much more in excess of their normal consumption that most of the remaining 54 million suffered from acute food shortages, and three million people died. This implies that each and every one of them ate twice their normal amount of food. Indeed, since many people in Calcutta had little or no increase in income, and had no access to the special schemes, it implies that the others would have increased their consumption far more than this.
However he is not just asserting that people were eating twice their usual ration, he is asserting that they were willing to pay from four to twenty times the normal price in order to get this extra food - only a fraction of normal consumption was available at special rates: the rest would have had to be bought at famine prices. The prices reported rose to Rs.60 to 80 per maund of rice containing 20% stones and 20 % dirt, according to the Statesman and a price as high as Rs.120 a maund was reached at the height of the famine. It will be noted too that industrial wages did not rise with inflation - the subsidised food was a substitute for wage rises.
He implies further that people suddenly switched back to their normal demand functions as soon as the December 1943 crop was harvested, with wartime inflation no longer having the same effect.
These hypotheses are ridiculous, and it might seem to be unnecessary to disprove them with figures. However, the figures are given in some detail by the Famine Commission, Sen's primary source. They are presented and analysed in Table l. They give no support whatsoever for Sen's doubling of the grain consumption by Calcutta. They suggest a fall in consumption of 12% to 45%, depending on population estimates, to approximately the level of the 1944 ration (rationing was not introduced until 1944). Even if one makes the unlikely assumption that there was no transfer of grain from the Calcutta Trade Area to Greater Calcutta, if one ignores the unrecorded road traffic, and if one ignores the influx of soldiers, refugees and starving people from the countryside, one gets a picture of sharply reduced consumption, though at a level above that of rationing.
Table 2, again from his sources, gives further reason to doubt that a change in distribution of incomes of the type he describes would cause the swing in consumption he states happens. It suggests that the increase in incomes of the industrial working classes, as a result of war industries, would actually have led to a decrease in their rice consumption.
Table 2. Estimates of per capita consumption of all cereals.
6.5 Inequalities in Distribution
Sen argues that the famine was caused solely by changes in distribution. In order to demonstrate that this is so, he quotes evidence to show that
This evidence does not prove his point. All famines, however caused, will have these effects, so the phenomena may just as well have been the result of a change in demand. I cannot accept Sen's statement that a famine can occur without such changes in distribution (1977 p35). A substantial fall in supply will inevitably change the relative amounts received by the different classes, even if it is only the change brought about by rationing.
Some of the changes in distribution that may arise from a shortage and that were cited by the Famine Commission, with its FAD view, are as follows. Most farmers in Bengal had a high yield and a high price because of the famine and so were better off. Even those with a reduced yield may have got a higher total revenue than usual. However, rural indebtedness meant that often the crop went to a moneylender or landlord who made all the profits. The indebted farmers had to buy back the food they had produced at the inflated price, and had to borrow money to do so. Those families who could not or did not buy their rice at the beginning of the season were unable to buy rice at the new price with the money they had. Consumers spent most of their income on food and could not afford other goods or services, so traders and suppliers of goods and services were impoverished. Many died.
The introduction of famine relief schemes in rural areas and of rationing for some employees in urban areas meant a change in distribution not linked to market forces. At the same time, the collapse of traditional village charity during the famine because of the high price of foods altered other non-market distribution (as in the Irish famine where landlords reneged on their social and legal responsibilities for poor law relief once the need became abnormal. Rights vanished under the pressure of scarcity).
All these examples are compatible with the Food Availability Decline (FAD) hypothesis which Sen so dislikes, so his argument falls down. Since the theoretical basis of his argument is wrong, the accuracy of his data on changes in distribution is irrelevant. However, his data are wrong. They are calculated from Table 4.2 and 7(A5) of Mahalanobis, Mukerjee and Ghosh (1946). Unfortunately Sen used tables of raw, unweighted data derived from a heavily stratified sample, rather than from the adjacent tables giving weighted results.
6.6 Selective Impact on Certain Groups
Sen gives as a cause of the famine the fact that from March to November, the demand for crafts, services and "superior" foods fell, so that the people supplying them were plunged into destitution. As shown above, the Famine Commission showed how this effect would arise as a result of shortages. Their FAD analysis shows quite clearly why one group of people rather than another should have starved. There is however another question that should be answered - Why should anybody at all have starved? The Famine Commission explains this quite simply with the fall in aggregate food supply. As Sen does not accept that there was a shortfall in supply, he is stating that the redistribution caused shortages, not just that it aggravated the effects of shortages. The logic is not explained.
6.7 Failure to Import
Sen states that a contributory cause of the famine was the failure of the Bengal Government, the Indian Government and the Imperial Government to increase Bengal's imports. As this was a failure to increase supply rather than a reduction in supply, it cannot logically be called a cause of the famine.
The Famine Commission was highly critical of government for failing to take the only action that could have relieved the serious shortage. As Sen does not accept that there was a serious shortage, he cannot logically take this view either. The most he can say is that, in failing to import, government failed to take one of the many courses of action which would have produced the desired effect.
One argument was put to the Famine Commission which suggested that the famine was caused by provincial governments preventing free trade in grain. Before the restrictions on exports the economic unit was the whole of India, afterwards each province was separate. Instead of the whole of India having a first-degree shortage, Bengal had a third-degree shortage. This argument is, of course, at variance with Sen's on most points. The Famine Commission is doubtful of such a sweeping assertion, and considers that it would have been politically unacceptable not to intervene in the grain trade.
6.8 Borderline between Two Price Regimes
Sen presents the following as a causal explanation of the famine. There is no further elaboration of his argument.
Why was it that Bengal alone should have been devastated by famine when the same applied to virtually every country in the world during the Second World War?
6.9 Boat Denial Policy
Sen considers the boat denial policy to have been a cause of the famine (1984 p4611;1980b p619). (See, however Sen (1976 p1279) where he expresses the opposite view.)In May 1942 orders were issued for the removal of boats capable of carrying more than 10 passengers from the coastal areas of Bengal in order to deny them to the Japanese if they invaded. The Famine Commission was very critical of the Bengal Government for their operation of the scheme (FIC pp26, 27), as it reduced fish catches and made transport difficult, hampering relief measures. Both the Indian Government and the Bengal Government considered that physical distribution was a serious constraint on relief measures. Indeed, the main effect of Wavell's intervention with the army was that four times as much per week was distributed (See FIC, Aykroyd, Wavell (in Moon 1973),and Mansergh 1973 p361) It slightly reduced the quantity of food available, and to this extent it was a cause of famine. The appalling mortality among fishermen must be putdown to the fact that the compensation paid to them, while possibly adequate when given to them, was grossly inadequate once the prices rose.
Sen accepts the general view that the boat denial policy was of little importance in reducing total supply. However, since Sen believes that transport problems were overstated, and since he believes that anything hampering transport from the starving country areas to the overfed towns was a good thing, it is difficult to see why he considers it to have been harmful.
6.10 Rice Denial Policy
Among the "factors working negatively on the supply of rice" Sen talks of
Curiously enough, elsewhere (e.g. Sen 1977 p45; 1981 p 67) he quotes some of the facts directly from the Famine Commission, and concludes merely that "it did contribute to local scarcities". Elsewhere, too, (1976 pl279) he states that it did not contribute to the famine (though stating that it did result in a loss of food).
However, if one refers to the original source, one gets a very different story, that a very small quantity of rice was moved from a surplus area to a deficit area, in a year with a record crop, and that this was done in May and June 1942, well before the cyclone. The rice denial policy implied
Clearly, there are major factual discrepancies here between Sen and the source he cites. Furthermore, it will be noted that the policy carried out was exactly that which Sen recommends - public authorities buying surplus rice and distributing it to deficit areas.
The denial rice was taken from the area which was later hit by the cyclone, which caused extensive damage to food stores. If it had not been removed, some at least would have been destroyed. To this, very small, extent the denial policy actually increased food supplies in Bengal.
6.11 Evaluation of Sen's Causal Hypotheses
Nowhere does Sen present a cohesive model designed to explain all the phenomena. There is no economic model in the normal sense. He does not have any model of the market or marketing system or the institutions involved. He presents instead isolated points, each intended to explain only a limited range of phenomena.
These points are not developed in any way. There is no attempt to show how inflation could have caused a rise in prices to famine level or how speculation could have caused starvation and high prices through the year. The moment any attempt is made to develop them, they run foul of demand theory - for example supply remains constant, demand plummets and prices rocket.
In addition, the moment any attempt is made to check these hypotheses against the facts, they are shown to be incorrect. It is necessary to show that the defence workers ate six times as much as usual or that the citizens of Calcutta ate twice as much as usual - and paid up to twenty times as much as the usual price for their food. The statistics in Sen's sources show a substantial fall in consumption instead. Time after time, it is seen that the facts in Sen's sources do not support the implications of his hypotheses. Time after time there is no mention of those facts in the source documents which do not support his hypotheses. Time after time the facts in the source documents conflict with those that Sen gives. On occasion there is a conflict between the facts as cited by Sen, and the facts as given in the document he gives as a source for those facts.
For this reason, in the next section I shall examine the factual accuracy of the two statements which are basic to his whole argument
7 WHAT THE BENGAL GOVERNMENT BELIEVED AND DID
The main thrust of Sen's argument is that the overriding cause of the famine was the approach of the Bengal Government. He claims that they adopted the FAD approach, acting on the belief that the famine was caused by a sudden sharp decline in food availability. As a result, they failed to anticipate the famine; they failed to recognize it when it came; and there were disastrous policy failures in dealing with it.
This is untrue in all respects, as is shown in Appendix II which gives a detailed description of the Bengal Governments actions. The sources are agreed that the Bengal Government made much the same assessment as Sen of food availability (until, in July or August, when the famine was reaching its peak, they started to realize that there was a major shortage). They believed that there was only a first-degree shortage. Surprisingly, in support of his claim that the Bengal Government was obsessed by the FAD approach, Sen gives two pages of evidence showing just the opposite: that the Bengal Government was firmly convinced that there was adequate food available, and that the hunger was due to changes in distribution (1977 00 53,54; 1981 pp 80-82). They also had the same theory of famine causation, and of the appropriate way of dealing with the famine. They believed that lack of purchasing power rather than lack of food caused starvation. They believed that price control was necessary under wartime inflation to prevent certain groups getting more than their fair share. They believed in public relief schemes. They believed that a large supply of food had to be distributed through the public distribution system. They believed that some degree of rationing was desirable. They believed that speculation and hoarding were major causes of the famine. They attempted therefore to provide the population, and particularly the population of Calcutta, with the purchasing power necessary to obtain the food. They instituted public relief measures. They intervened on the market. Had they been right in their assessment of food supplies, and on the cause of the famine they might have been successful. However, they held the same views as Sen, and they acted accordingly. Three million people died.
If they had held the FAD view, as Sen states, their logic would have been as follows: "There is widespread hunger and even starvation. Under the FAD approach, the only possible reason is a shortage of food. Ergo, we must import one and a half million tons immediately." Whether their analysis was right or wrong, their response would have saved three million lives.
7.1 Monitoring the Shortage
Sen is indignant that the government should have spent any time at all on monitoring available supply, once it had been decided that the famine was due to maldistribution:
I must disagree in the strongest possible terms. The effects of treating a serious food shortage as merely a change in distribution are horrendous. It would be criminal negligence bordering on the genocidal to treat any famine as merely a first-degree shortage without constantly reconsidering the possibility that either initial estimates were wrong or the degree of shortage had changed. This at least the Bengal Government was not guilty of.
8 WAS THERE A SHORTAGE?
Sen is emphatic that the shortage in food supply was not a cause of the famine. He says that there was at least 11 % more food available than in 1941 when there was no famine, and he ends by saying "It seems safe to conclude that the disastrous Bengal famine was not the reflection of a remarkable overall shortage of food grains in Bengal. (1977 p42)The view of the Famine Commission, on the other hand, was that the food shortage was the major precipitating cause of the, famine, and other people, such as Professor Husain, himself a member of the Commission, thought that the Commission had underestimated the shortage.
In this section it will be shown that the data available were so bad that the Bengal Government were criminally negligent to base their decisions on it. Evidence will be presented which suggests that official estimates understated the degree of the shortage. Sen's refusal to believe in reduced carry-over stocks will be examined. It will be concluded that the figures available cannot be used as meaningful evidence in favour of Sen's hypothesis. It will be shown that Sen can be read as overstating the reliability of his data, and so, in effect, overstating the probable amount of grain available in 1943. It will be shown that Sen misrepresents the facts in his sources.
8.1 The Size of the Harvest
In view of the catastrophic effects of underestimating the degree of shortage, decisions should not be based on the assumption that best guesses at production, stocks, consumption, etc. are correct. They should take into account the effects if these figures happen to be optimistic. It would be appalling irresponsibility rely on best guesses when deciding how much to import or what rations to issue. Accordingly I shall be examining here the reliability of the data on which the Bengal Government acted, and the data on which Sen bases his calculations.
The key information for dealing with the famine was the estimate of the December 1942 crop. This was based on the crop forecast, not on a post-harvest estimate. It was based on subjective estimate of how a) yield, and b) area deviated from the norm. As the estimates were based on area planted it is not clear to what extent the forecast would have allowed for damage from disease, flooding and cyclone. There would have been a much bigger discrepancy than usual between planted acreage and harvested acreage, and both reporting error and failure to modify the estimates right up to harvest are probable.
Sen got his figures from the Famine Commission, which is at pains to state how unreliable they are:
No attempt ever appears to have been made to check these results against anything else, or to amend them in the light of experience, so nobody knew if they even indicated the direction of change. It would in any case have been difficult, as only a proportion of the crop was marketed, and as the marketing system was not monitored. The cyclones, tidal waves and disease outbreaks which caused so much damage in 1942 had never occurred before, previous famines being due to drought in other areas of Bengal. As a result there was no experience on which to base estimates of their effect. It is just conceivable that the system might have given a rough indication of the percentage change in total production - but only if all areas of Bengal were equally affected by low rainfall etc. In 1942, however, areas were differentially affected, so there would have been major aggregation errors, if, for example the worst affected areas were high-yielding areas, densely populated areas or areas where the acreage was accurately known.
The collectors of statistics did not know the normal acreage or yield, only people's estimates of the deviation from the norm, so there was a substantial error in the estimate of total production, as well as an unknown aggregation bias. At the time parts of Bengal, notably those hit by the cyclone, were simmering on the verge of an insurrection, and the army was active, burning villages etc. Under these circumstances, it is doubtful whether the collection procedure was reliable.
There was a sample survey of the commercially valuable Bengal jute crop, but not of the rice crop. "Although by that time, owing to Japan's entry into the war, the food situation in Bengal had already become difficult, I failed completely to persuade the Government to extend the sample survey to cover the paddy crop in Bengal. The Bengal famine occurred in 1943. Since that year we have had the opportunity of carrying out a sample survey of both jute and rice crops throughout the province." (Mahalanobis, 1946 p333)
The Governor of Bengal, in a document cited by Sen (no. 158 in Mansergh, 1973), talks of "the dubiety of all available statistics and therefore lack of accurate knowledge of what the real shortage is".
The Famine Commission changed the official Bengal Government acreage figures by 20% and adjusted the yield figures, because Mahalanobis's sample survey, conducted after the famine, produced very different figures to the official figures, figures which may well have been more accurate. Of course this did nothing to correct the aggregation bias.
Desai (1953 p8) gives a useful review of the agricultural and other statistics of this period, and his rigorous use of them is an exemplary. He compares the official estimates and those obtained by scientific sample surveys carried out by Mahalanobis. He shows that the discrepancies are large, with survey estimates being between 47% and 153% of the official estimate. The discrepancies also vary from year to year, with the sample estimate of the jute crop being 2.6 % above the official estimate in 1941 and 52.6% above it in 1946. Since there was no sample survey of rice carried out in Bengal until after the famine, we have no real idea of the accuracy of the 1942 Department of Agriculture estimates, except the belief that the error was greater than 20%. There is no way of correcting it with hindsight. Desai's assessment may be compared with Sen's statement that "the Raj was, in fact, more or less right in its estimate of overall food availability."(Sen 1977 p53; 1981b p80)
As well as this error in area and aggregation procedure, an upward bias in subjective eye estimates of prospective yield of a growing crop may be expected. Mahalanobis (1946)talks of an upward bias even with controlled sampling methods on a mature crop, because enumerators tend to select the best fields and the best areas of damaged fields. Again, the bias would have been very different in a year of cyclone, tidal wave and disease damage like 1942.
We must also be concerned with the quality of the data entering the system. This was bad even in the better organised studies:-
There is also a more serious form of bias: the scale of incompetence and corruption was so vast that virtually every politician and administrator had cause to want evidence suppressed or altered. There is some indication that pressure was brought on statisticians to do this:-
It was also strongly rumoured at the time that the Indian Government had deliberately printed only a few copies of the Famine Commission report, to limit the circulation of its criticisms (Aykroyd, 1974).
It must be concluded that the statistical basis of estimates of production is so bad that the Bengal Government was not in a position to say with any confidence whether the December 1942 crop was 11 o/ above the December 1940 crop or 20 % below it. By treating these best guesses as though they were perfectly accurate, they made famine inevitable.
8.1.1 Non-statistical evidence
Even if we had reason to trust the statistics, we should not ignore the non-statistical evidence. A lot of people gave warnings of the famine, warnings which conflicted with the official production estimates. As early as December 1942, after the cyclone and before the aman' crop had been harvested, the trade was talking of the worst crop in twenty years (FIC p33). Traders bought up any stocks they could in Bengal and they went into the neighbouring provinces of Orissa and Bihar to buy grain and standing crops. They were prepared to smuggle the rice into Bengal if trade restrictions prevented them from doing it legally. The trade had its own way of estimating supplies (including stocks)and did not rely on official estimates. In this case, they were so certain of their estimates that they invested all they could borrow, and of course they made a lot of money as a result. The Bengal Government ignored their warnings.
Bhatia (1967 p35), quoting from the unpublished evidence to the Famine Commission, states that public men and organisations had warned the government. Rajan (1944 p15)quotes a European member of the provincial assembly as saying
Sen himself quotes pressure from "a businessman much involved in rice trading" to increase imports by one million tons, as late as October 1943 (Sen 1977 p54, quoting from Document 174 in Mansergh, 1973 p390). This would of course have been against the businessman's interests if he had large speculative stocks.
During the famine it became increasingly apparent that the shortage was much more serious than the production statistics suggested. The total failure of all the government's intervention measures to bring down the price is particularly significant. The government's attempts to "break the Calcutta market" by dumping large quantities on the market failed, both because it proved impossible to buy or seize any large quantities and because what was put on the market vanished without a ripple. During the period of free trade with Bihar and Orissa, 91,000 tons was imported and dumped on the market (and some was bought illegally and smuggled in). The effect was to push up prices sharply in Bihar and Orissa, but there was no noticeable effect on prices in Calcutta (FIC p52). This suggests that 91,000 tons was a very large proportion of any surplus in Bihar and Orissa, but was a very small amount in relation to the Bengal deficit. Calcutta would consume only 21-24,000 tons of rice and 15,000 tons of wheat a month (FIC p203) The house-to-house search for stocks (see page 14) showed only that the stocks were much lower than expected.
8.1.2 Why the information was ignored
Why, one may ask, did the Bengal Government ignore the evidence that there was a serious shortage? At this stage one can but guess.
First, there is a type of official mind that will believe statistics in preference to any other evidence. They believed that the crop was adequate because the production statistics said so. They ignored the warnings of the trade and the failures of their intervention policy because there were no statistics on them. (This has happened elsewhere. For example van der Laan (1975), mentions the Sierra Leone Government in 1919 ignoring the traders' warnings of a shortage for six months until food riots broke out, apparently believing that price control was a substitute for imports.) They refused to consider carry-over because there were no statistics on it. When the famine persisted they did not question their statistics, but tried to find other explanations compatible with their statistics, producing the old bogeymen of speculators and hoarders.
Second, the officials were very pleased with their success in preventing famines in 1928,1936, 1941 and 1942, (though, as will be shown in the section on carry-over, the achievement was not as big as they thought - carry-over stocks were consumed to makeup for the fall in production, so consumption did not fall dramatically). They were, perhaps, psychologically predisposed to see this as another task they could handle easily.
Third, they were very keen to blame the famine on a conspiracy by hoarders and speculators, rather than on their own inadequacies, or on forces they did not understand, as politicians and officials always are - I have seen the same reaction in Europe, Africa and Asia.
Fourth, the officials were certainly understaffed and overworked before the famine, because of the war. The famine stretched the administrative system past breaking point(Aykroyd, 1974).
Finally, in 1985 we can perhaps discuss two points that the Famine Commission considered to be too hot to handle in 1945. First, corruption was rife. Everyone writing at the time commented that nearly every public servant who handled any famine relief managed to make a personal profit out of it, and the trading firms who were appointed government agents defrauded the government and cheated the people. The Indians writing at the time did not hesitate to accuse the politicians, both Hindu and Muslim, of being involved. It would follow that most of the people who handled information and made decisions had a strong financial interest in the continuation of the famine. Secondly, Rajan (1944 p44) gives quotations which suggest that both right-wing British officials in the Government of India, and right-wing politicians in Britain were predisposed to accept any evidence that suggested that an elected Indian government, be it only a provincial government, with Indian administrators, was so corrupt and incompetent, and that the Indian businessmen were so greedy, that they could create a famine in the midst of plenty. As a result, they observed the constitutional niceties, and did not check the facts on the ground or overrule the Bengal Government. (Even when Wavell did intervene, he first had to persuade the local politicians to let him give them the extra resources). It is forty years too late to confirm these hypotheses, but experience elsewhere shows that they are only too likely to be true.
Professor Sen rests his argument that there was sufficient food to go round in 1943 on his belief that there was at least 9 % more food per head available than there had been in 1941. To reach this conclusion he rejects all the arguments of the Famine Commission on carry-over, though they were put at considerable length. Indeed, he makes the very serious allegation that
He provides no evidence to support his contention.
The Famine Commission argued that, both because of the need for normal stocks and because the rice is not palatable for some months after harvest, it was not normal to start eating the December crop until February or March. Some varieties of rice were kept for a year or two before consumption. (See Professor Husain's minority report (FIC pp 179-199) and the Report on Rice Marketing (Government of India 1942)). In addition there would have been the normal commercial stocks. They argue that normally there was a three month carry-over of grain at the end of the year.
I cannot accept Sen's suggestion that we should ignore the carryover on the grounds that we have no statistical data on the stock position (1977 pp42, 55). Still less can I accept his assertion that variations in carry-over were mythical. I would find it difficult to believe that there is any country that does not aim at least a two-month carry-over of grain. I find it very difficult indeed to believe that exactly the same amount is carried over after a famine year as after a series of bumper harvests.
It might be mentioned that all the explanations Sen gives for the famine are equally unquantified (mythical?) variations in equally unobserved (non-existent?) items. The difference is that where the Famine Commission gives ten pages of argument and facts in support of their carry-over explanation, Sen gives only a sentence or two in support of his explanations.
The very poor crop of December 1940 meant that the people ran out of rice earlier than usual and started eating the crop of December 1941 as soon as it was harvested. It is difficult not to accept that there was a reduced carry-over from such a short crop. This means that the consumption of rice in 1941 was well above the "adjusted current supply of rice" quoted by Sen, while the consumption in the next year was well below it. (The "adjusted current supply" is the addition to stocks, not the supply available to the consumer.) This means that while "adjusted current supply" might have been 13 % higher in 1943 than in 1942, it does not follow that supply to the consumer or consumption were.
Sen says that even if this were so, which he denies, the December 1941 crop was so big that the trade would have restored their stocks to the normal carry-over levels by the beginning of 1943.
I shall not recapitulate the arguments put forward by the Famine Commission and Professor Husain to show that there was a much reduced carry-over into 1943. Instead, I shall show that Sen's own figures destroy his case. Table 3 is based on exactly the same figures that Sen uses (FIC p215), except that I have arbitrarily chosen a carry-over of two million tons at the beginning of 1941. The table shows that this large surplus changes into a deficit of nearly a third of a million tons by November 1943 - enough to explain the famine. All that has been necessary is to assume a level of consumption per head equal to the mean for the years 1928 to 1942. This figure is well below the mean for the first half of the century. A low population growth rate of 0.46% has been assumed: a higher deficit would have been noted with 1 %, the figure Sen used.
The deficit shown in Table 3 would have been much greater if allowance had been made for the unrecorded exports of rice by road and country-boat, during 1942, when rice prices were much higher elsewhere in India.
9 SEN'S USE OF THE STATISTICS
The Famine Commission recognised many of the weaknesses of the statistics, and strongly criticised the Bengal Government for basing their estimate of import requirements on them. The Famine Commission itself, writing after the event, could place rather more emphasis on the best guesses, and give somewhat less to what would happen if these were wrong. One would expect that Sen would be equally circumspect in using such unreliable statistics when he is advising people on tackling future famines. However, he bases his arguments on these guesses, which he uses as if they were perfectly accurate, drawing major conclusions from a few percentage points difference. In particular, he bases most of his argument not on an unreliable production forecast, but on the difference between two unreliable forecasts. This is quite unacceptable: the data do not support the conclusions.
Table 3. Stocks of rice in Bengal, 1939-1943
Notes: Opening stock is an arbitrary assumption. Consumption per head is the mean current supply less seed for the years 1928-42. This is below the mean for the first half of the century. A population growth rate of 1 % has been assumed(Sen,1977, p 40). The remaining 64 000 tons of 1943 imports are assumed to have come in November and December. No allowance is made for the unrecorded exports by road and country boat in 1942. These would have increased the deficit substantially.
Sen also appears to be claiming much greater reliability for his statistics than is justified, to be giving conservative figures rather than best guesses - and with such inaccurate statistics this amounts to a misstatement of 30% or more. He bases his calculations, he says, on "a careful tally of food availability in Bengal". He talks of presenting "the results of a food supply calculation, taking into account local production and trade, choosing -wherever the data permit - an assumption as unfavourable to 1943 as possible". He concludes that "Current availability of food was at least 11 per cent higher than in 1941,when there was nothing remotely like a famine" (Sen 1984 p461). Elsewhere (1977 p40),he says "This is most certainly an over-estimate for 1941 vis a vis 1943, but this is an acceptable bias as it favours the thesis we are rejecting", "To bias the figures as much as possible against 1943 . . ." He may also be interpreted as claiming a much greater accuracy for them than is justified, because he frequently quotes different secondary sources as giving much the same estimate of total production or import needs (see for example Sen 1977 pp53-4). Since these secondary sources are all based on the same primary source, official production estimates, no added confidence is given. His scathing comments on those who consider that the famine was caused by shortages emphasise the impression that he is totally confident of his figures.
Sen (1977) quotes Document no 265 p357 in Mansergh (1971) as stating that "the rice crop in Bengal was recognised to be indifferent rather than exceptionally bad". In fact, the document stated, as early as 9th December 1942, that there was both cyclone damage in certain areas and an indifferent crop in Bengal generally. The combined effect was seen as being exceptionally serious. This is a particularly clear example of Sen's misrepresentation of the facts in his sources.
In fact, the figures he gives are not in any sense conservative. The output figures are, as shown above, wildly unreliable. The import figures are no more reliable than such figures usually are, and in addition they fail to cover trade by road and country-boat. For these figures he uses the Famine Commission guesses, and not a conservative figure. (Note that the Famine Commission assumes, and Sen accepts, an identical unrecorded net import in 1941, a year of shortages and recorded net imports, and 1942, a year when Bengal had a surplus and the rest of India a shortage and when Bengal had substantial recorded net exports.) His conservative adjustments consist of making a slight adjustment to allow for unrecorded wheat imports, an alteration of a fraction of one per cent of the total. Again, he makes much of choosing a 1 % population growth rate instead of 0.46 %, which makes a difference of 1 % when he uses it for comparing 1941 with 1943. These conservative adjustments do not make any noticeable improvement to the accuracy of the aggregate figures he uses.
10 THE FOOD AVAILABILITY DECLINE APPROACH
Sen denounces the Food Availability Decline (FAD) approach in no uncertain terms. At times he is denouncing an unbelievably narrow approach which I am quite certain no economist ever held. At other times he is denouncing a more balanced view, like that of the Famine Commission. At others, he is denouncing the view of the Bengal Government, which was virtually the same as his own. However, I do not think it would be unfair to say that through most of his writings he considers the approach of the Famine Commission to be typical of the approach he condemns.
In my opinion the Famine Commission wrote an excellent report. They sought the truth rather than evidence in favour of their hypotheses. They entered into their study with no preconceived ideas as to whether it was a FAD or a distribution famine and they reached a conclusion that was not in accordance with the official view. The report was written forty years ago by non-economists, so it is to be expected that we should think that their economic analysis was naive or even wrong in parts. In view of this it is surprising that they should have made few major errors and that they should have been broadly correct in their conclusions. Certainly their analysis had more depth than Sen's. In spite of the deficiencies of their market analysis, I would not be ashamed to have written such a report.
It is very disturbing, therefore, that Sen should have attacked the report in terms that would persuade inexperienced economists that it is foolish or even immoral to examine the problem in the way the Famine Commission did. It is even more disturbing that they should be led to think that it is foolish or immoral to diagnose a famine as being due to a decline in food availability. This can only lead to failure to take the appropriate action, as a result of which millions will die. The following quotations from Sen's works confirm that he has condemned the FAD approach in terms which could not be described as dispassionate.
It will also be noted that the above quotations say that the FAD approach ignores phenomena and explanations which are in fact covered in depth by the Famine Commission report and others which Sen describes as FAD. Sen states, for instance, that the FAD approach avoids dealing with the change in relative purchasing power and the impoverishment of people of some classes and occupations, and ignores the fact that, for example, "a sharp decline in the relative price of a commodity vis a vis food can jeopardise the ability to survive of the people who live by selling that commodity". In fact, a look at his sources will show that the FAD approach recognises these phenomena and pays a great deal of attention to them, and, indeed, he gets all his examples from sources that adopt the FAD approach.
The Famine Commission, for instance, while taking an essentially FAD approach to the famine, gave a lot of attention to the hardship caused by redistribution. They present all the examples given by Sen and they analyse the distribution shifts in far more detail than he does. In most cases the shifts can only be explained by a large change in supply or demand, and elsewhere this is a possible explanation. In no case can the phenomenon be explained only by Sen's hypothesis, and Sen makes no attempt to show that it can.
The Bengal Famine Code is the only book I know of which states unequivocally that all famines are caused by shortages (and, even so, this appears to be not so much a matter of belief as a reaction against those officials who took no effective action in the famine of 1886, believing it to be a Sen-type famine). Its uncompromising FAD approach involved making food available, issuing ration books and giving people food or the means to buy food. It laid down special measures to protect artisans, weavers and manufacturers, recognising that the swing in purchasing power could destroy their markets. It laid down measures to prevent farmers from impoverishing themselves. It even laid down special treatment for "respectable" men who became destitute. Changes in distribution were seen as among the biggest problems arising out of a Food Availability Decline famine. (See Appendix I for details of the Bengal Famine Code.)
It must be concluded therefore that Sen is incorrect in his claim that the FAD approach avoids these matters
It most certainly does not, as Sen's sources make very clear indeed.
The only way to be sure of curing a famine, however caused, is to import more food. Any analysis of a famine that underestimates the degree of a shortage is very dangerous. It means that market intervention and relief measures will be used when imports are urgently needed. It means that no rationing or mild rationing will be imposed when severe rationing is urgently needed. The result will be a worsening of the degree of shortage, making famine inevitable.
It has been shown here that Sen's analysis of the Bengal famine led to the wrong conclusions. It led, in fact to the conclusions reached by the Bengal Government. The Bengal Government acted on these conclusions, and three million people died.
Sen's hypotheses collapse as soon as the implications are examined. His main hypothesis is that the famine was caused by inflation and by increased earnings to industrial workers, resulting in their buying more food so that there was less food available for others. This collapses when it is realised that it implies their eating two to six times as much rice as normally, and paying four to twenty times the normal price to do so - when the statistics show a decline in consumption. He uses the old bogeymen of speculators and hoarders, but does not explain how they could have caused a reduction in supply. Again, the evidence did not support his hypotheses. Some of his hypotheses, like the failure of the Imperial, Indian and Bengal Governments to increase exports to Bengal, conflict with his overriding hypothesis that there was no shortage. Elsewhere, he confuses the results of the famine with its causes.
Throughout the examination of the hypotheses it was found that the facts given in Sen's sources conflicted with those implied by his hypotheses. Sometimes the facts he cited conflicted with those given in his source. Sometimes, as with the denial' rice and with subsidised distribution in Calcutta, there was a clear contradiction between what he said the Famine Commission said, and what it did in fact say.
His two overriding hypotheses were, first, that the Bengal Government adopted a FAD approach and, second, that there was at least 11 % more food available in 1943 than in 1941. An examination of his sources shows the clearest possible evidence that the Bengal Government, like Sen, thought that there was adequate food and blamed the famine on inflation, speculation and hoarding. The statistics on food availability are shown to be so bad that it is not possible to say with any confidence that there was any more food produced in 1941 than in 1943, much less that there was more food (including stocks)available for consumption. Certainly the statistics cannot be used as Sen uses them. However, even if one were to assume that Sen's figures were correct, they would suggest a major shortfall in supplies in late 1943 - just when the famine was at its peak. Sen seriously overstates the accuracy of his figures, in effect overstating available supply.
It is concluded that Sen's theoretical explanation of the Bengal Famine is wholly without foundation. What is more, the facts are almost all against him.
The appropriate method of examining a famine has nothing to do with the opposing dogmas of the FAD approach (if such an approach ever existed) and Sen's entitlement theory. The approach normally used in examining price policy and marketing is rigorous and has an enormous explanatory power. A complex model is built up to take into account all institutional factors and other factors relevant to the market. Such a model has the advantage that factual inaccuracies are immediately revealed as inconsistencies. It also has the advantage that it takes into account the agrarian problems, the price policies and the marketing systems that are, all too often, the underlying cause of the famine, and that strongly influence the course of the famine.
However, this approach takes time and not a little skill. If the models are not established before the famine starts, it is unlikely that sound models can be built up in time to influence decisions. If snap decisions are to be made, the FAD approach is the safest: "There is hunger. This must be due to a food shortage. Therefore we must import." It may give the wrong diagnosis, but nevertheless it will save lives.
Aykroyd, W.R. The conquest of famine, London, Chatto and Windus, 1974.
Allen, George, "Famines: the Bowbrick-Sen dispute and some related issues," Food Policy, 11(3) 259-263, 1986.
Bengal Administration, Bengal Famine Code, (Revised edition of December 1895) Calcutta. 1897.
Bedi, Freda, Bengal Lamenting, Lion, Lahore, 1944?
Bhatia, B.M. Famines in India, Asia Publishing House, Bombay 1967.
Blyn, G. Agricultural trends in India, 1891-1947: output, availability and production, Philadelphia, University of Pennsylvania Press. 1966
Bowbrick, P., "Rejoinder: an untenable hypothesis on the causes of famine", Food Policy.12 (1) 5-9, February. 1987.
Bowbrick, P. "A refutation of Sen's theory of famine", Food Policy. 11(2) 105-124. 1986.
Bowbrick, P. A refutation of Professor Sen's theory of famines. Institute of Agricultural Economics, Oxford. 1986.
Bowbrick, P. "How Professor Sen's theory can cause famines", Agricultural Economics Society Conference. March. 1985.
Desai, R.C., Standard of living in India and Pakistan, 1931-2 to 1940-41, Popular Book Depot, Bombay, 1953.
Dutt, T.K., Hungry Bengal, Indian Printing Works, Lahore, 1944.
Famine Inquiry Commission, Report on Bengal, New Delhi, Government of India, 1945a
Famine Inquiry Commission Final Report, Madras, Government of India. 1945b
Ghost, K.C. Famines in Bengal, 1170-1943 Calcutta: Indian Associated Publishing, 1944.
Ghost, T.K.O., The Bengal Tragedy, Hero Publications, Lahore. 1944.
Government of India, Report on the marketing of rice in India and Burma, Government of India Press, Calcutta. 1942.
Harden, B., "`Why Hickenlooper is Ethiopia's Catch-22", The Guardian 4th January 1985.
Mahalanobis, P.C. "Recent experiments in statistical sampling in the Indian Statistical Institute." Philosophical Transactions of the Royal Society, Part iv, pp326-378. 1946.
Mahalanobis, P.C., Mukkerjee, R.K., and Ghosh, A. "A sample survey of after effects of Bengal famine of 1943." Sankhya 7(4),337-400. 1946.
Masefield, G.B., Famine: its prevention and relief, Oxford, OUP. 1963.
Mansergh, N. (ed.) The transfer of power 1942-7 vol. III, London, HMSO. 1971.
Mansergh, N. (ed.) The transfer of power 1942-7 vol. IV, London, HMSO. 1973.
Moon, P. (ed.), Wavell: the Viceroy's journal, OUP, Oxford 1973.
Palekar, S.A., Real wages in India 1939-1950 International Book House, Bombay. 1962.
Rajan, N.S.R., Famine in retrospect, Pamda Publications, Bombay. 1944.
Sen, Amartya, "Famines as failures of exchange entitlements", Economic and Political Weekly, Special Number, August 1976.
Sen, Amartya, "Starvation and exchange entitlements: a general approach and its application to the Great Bengal Famine", Cambridge Journal of Economics, I 33-59 (1977)
Sen, Amartya, "Famine mortality: a study of the Bengal Famine of 1943" in Hobsbawm et al. Peasants in history: essays in memory of Daniel Thorner, Calcutta. Oxford University Press, 1980
Sen, Amartya, "Famines", World Development 8 (9) 613-21. Sept. 1980b.
Sen, Amartya, Poverty and Famines Oxford, Clarendon Press, 1981.
Sen, Amartya, "Ingredients of famine analysis: availability and entitlements" Quarterly Journal of Economics, August 1981b. 433-464.
Sen, Amartya, Resources, values and development, Blackwell, 1984.
Sen, Amartya, "Reply: famine and Mr Bowbrick", Food Policy 12(1) 10-14.
Sen, Amartya "The causes of famine: a reply", Food Policy 11(2) 125-132, 1986.
Singh, A. Sectional price movements in India, Benares, Banaras Hindu University. 1965.
Smith, Adam, The Wealth of Nations, Everyman, 1910, 1977.
Stevens, I. Monsoon morning, London, Ernest Benn. 1966.
van der Laan, H.V. The Lebanese Traders in Sierra Leone, Moulon, the Hague, Paris, 1975.
APPENDIX ONE: FAMINE RELIEF MEASURES
It is not enough to issue ration books if there is no food. It is not enough to import food or to stock distribution depots if people cannot afford to buy the food. In all famines prices rise, so the poor cannot buy food and they starve, while other people sell all their possessions to buy food impoverishing themselves. For this reason famine relief measures typically include provisions like the following, which were laid down in the Bengal Famine Code and which were applied in Bengal in 1941, 1942 and 1943. (See Bengal Famine Code (1895), Famine Commission (1945a p69). See also K.C. Ghost(1944) for similar relief measures being used in Bombay in 1629-30.)
It will be noted that money was given rather than grain in order to reduce the costs of administering distribution to isolated recipients. The same is being done in Ethiopia today(Harden, 1985). This might be appropriate in a purely redistributional, Sen-type, famine. However the possible effects on prices even in a first-degree shortage are worrying, and in a second-degree shortage one might expect a dramatic rise in prices with no improvement in distribution.
APPENDIX TWO: ACTION BY THE GOVERNMENT
The following detailed description of the Bengal Government's actions confirms that it adopted an approach very similar to Sen's. They believed that there was only a first-degree shortage. They believed that lack of purchasing power rather than lack of food caused starvation. They believed that price control was necessary under wartime inflation to prevent certain groups from getting more than their fair share. They believed in public relief schemes. They believed that a large supply of food had to be distributed through the public distribution system. They believed that some degree of rationing was desirable. They believed that speculation and hoarding were major causes of the famine. They attempted therefore to provide the population, and particularly the population of Calcutta, with the purchasing power necessary to obtain the food. They instituted public relief measures. They intervened on the market. Had they been right in their assessment of food supplies, and on the cause of the famine, they might have been successful. Instead, three million people died.
The Bengal Government was convinced that famines were caused mainly by changes in distribution, and that they could control such famines by market intervention. They had experience to support this belief.
When the Government saw that the price of rice was rising in 1942, it felt it should act. In June it fixed a maximum price for rice on the Calcutta market, but prices had already exceeded this level by the time it was implemented on 1st July. As a result, merchants diverted supplies to the high priced markets elsewhere in India (FIC p29). The Government reacted by stopping exports, except under permit, from l6th July, and by seizing and distributing stocks. A week later they raised maximum prices by one rupee. It still had little effect. Then the "denial" rice was put on the market, and the district officers were ordered not to enforce price control. Commercial firms and state bodies set up organisations for distributing subsidised rice to employees. At the same time, good rain in September and October improved expectations for the December crop. The combined effect was to stabilise prices, though at a rather high level. (FIC p30). The Government were reinforced in their belief that market intervention could prevent a shortage from getting out of hand.
All this time, as throughout the war, the Bengal Government and the Indian Government were concentrating on anti-inflationary price control, both for the normal reasons, and because they feared that price rises and shortages would lead to an outbreak of political violence. The price control was instituted in the firm belief that there were adequate stocks.
On October l6th there was a cyclone. The trade saw the implications and started buying. Prices doubled in the country areas between l8th November and 7th December. In December 1942 the Civil Supplies Department of Bengal saw the shortage as mainly psychological (FIC p33). In December too, at the India Food Conference it was considered uncertain whether Bengal needed supplies - it had experienced poor crops before and yet had imported relatively small quantities.
After the cyclone there was an immediate shortage in the cyclone-hit areas, and Government introduced relief measures on the lines set out in Appendix I. As destitution started in other areas, relief started there too. It started in January 1943 in Chittagong, and soon after in Tipperah, Faridipur and Dacca. It is not suggested that the relief efforts were adequate, but they would have been, had there been only a first-degree shortage.
The government saw its basic role in relief as providing purchasing power to those who had lost it
It will be noted that if there is any substantial shortage of supply, this policy leads merely to an uncontrollable price spiral, with government and individuals bidding against each other for non-existent grain.
When there was a shortage in Calcutta, on 27th December, Government requisitioned stocks and sold them through controlled shops. They considered the shortage to be purely a symptom of panic by distributors.
The Government firmly believed that there was no shortage, but prices continued to rise at the beginning of 1943. It diagnosed a shortage due to speculation and acted accordingly.
It is unnecessary to point out the weaknesses in their logic: had they analysed the situation correctly there would have been no famine.
In order to "break the market", Government aimed to buy 7400 tons of rice. In fact, it only managed to buy 2800 tons - even though "District Officers were informed that, if necessary, requisitioning was to be resorted to until the quota fixed for the district was procured". The scheme was replaced by another one aiming at buying 22,000 tons a month at government prices. In spite of embargoes preventing anyone but the Government agents exporting from the district, only 2,200 tons were purchased between 10th and l7th January, and the scheme was abandoned (FIC p36). A Food Grains Purchasing Officer was then allowed to buy direct from the trade, but he managed to get only 3000 tons between l8th February and 11th March (FIC pp 36-8). The fact that such tiny quantities, out of 9.6 million tons annual consumption, were considered adequate to "break the market" confirms that the Government did not believe that there was any real shortage. All that was necessary was to panic speculators into disposing of their stocks. The fact that it was not possible for the government to buy even the small quantities they wanted suggests that there was a major shortage of supply.
The supply situation for Calcutta was desperate: the city received only half its normal supply in the first quarter of the year. The Bengal Government considered it impractical to seize surplus stocks by force, as they would have wished, and decontrolled the market instead. They did it with reluctance because they believed that
They were fully in agreement with Sen on this.
In order to mitigate the effects on Calcutta, the area most at risk because it was at the end of the distribution chain, the Government increased distribution of subsidised grain through employers' organisations, approved markets etc.
District Officers were told to purchase without limit of price any rice and paddy offered to them in the first three days up to a limit of 20,000 maunds. They bought 17,000 tons between l2th and 3lst March, but only 18,000 tons up to the end of August (FIC p39).Again, one is struck by the small quantities considered and the difficulty in purchasing it.
On 10th March it was proposed that 60,000 tons should be obtained from neighbouring states "to break the Calcutta market". Only 28,000 tons were obtained and "the supplies were not sufficient to achieve the primary object of "breaking the Calcutta market" (FIC p43)
During June and July there was free trade between Bengal, Bihar and Orissa. 90,000 tons were purchased and moved into Bengal (38,000 tons by Government agents), but "the effect on prices in Bengal was negligible." (FIC p52). Again, this suggests that the quantity obtained was tiny in relation to the true shortage.
In April and May there was "a propaganda drive for the purpose of convincing the people that the supply position did not justify the high prices prevailing". It was hoped that this propaganda, coinciding with the arrival of imports, would induce a freer flow of stocks into the market and bring down prices. It was intended that this would induce speculators and hoarders to release their stocks. These objectives were not achieved, possibly because there were no excess stocks to release.
In June 1943 there was the Food Drive which aimed at locating surplus stocks and "to organise distribution of local surpluses as loans or by sales to those who were in need of food grains" (FIC p55). In fact it was found that there was very little in stock (see page 14).
In August the Government re-controlled the market.
As late as l9th October 1943, when the famine was at its peak, Wavell noted in his journal "On the food situation Linlithgow [The Viceroy] says chief factor morale."(Moon, 1973 p34).
It should never be forgotten that if the Bengal Government had been successful in its market intervention, the death toll would have been much higher. If they had forced traders to sell off their stocks until market prices fell to the 1942 controlled price, if they had seized and distributed hoarded grain, if they had persuaded farmers and hoarders to sell their stocks, then Bengal would have run out of food completely.
APPENDIX THREE: THE IMPACT OF SPECULATION ON PRICES
Figure 4 shows how rice prices moved during the period of the famine. They started to rise as soon as the cyclone hit. They rose during the year with the speed of the rise being related to government action on price controls, imports, grain purchases etc. as detailed in Appendix 2. They did not fall until the new crop was harvested in December 1943.Even then the prices remained firm, well above the 1942 level, and supplies remained tight in spite of the excellent harvest.
Figure 5 shows how prices would have moved had there been normal storage, assuming that there was perfect information, and assuming, contrary to Sen, that there was indeed a major crop failure. Speculators would have bought up the available crop the moment the news of the cyclone became known. They would then have released stocks throughout the season, reducing the supply slightly towards the end of the season so that the increased price covered the storage cost. The price would have fallen when the new crop arrived. The 1944 price would have been no higher than the 1943 price because supplies were good. In drawing these curves it seemed reasonable to ignore the fact that a monopolist would have had a somewhat different optimum supply when demand was not of constant elasticity.
Sen states that there was no shortage in supply and this implies prices like those in Figure 6, had there been normal storage. He states, though, that there was not normal storage, but something called speculation which put up prices without reducing supply (He explicitly rejects change in carryover; neither he nor anyone else has suggested that speculators destroyed their stocks or exported them; he ridicules the idea that there was a reduced crop.) This is not the speculation discussed in economic theory. If Sen's speculation is to be in any way compatible with economic theory, it implies that speculators released only small quantities at one period of the year, and released large quantities at others. This would, however, have brought about a price curve like that of Figure 7. Prices would have been high when releases were low and would have been rock bottom when the remaining stocks were released at the end of the year. The price curve actually observed was completely different to this. First, it rose throughout the year, rather than falling, as Sen's theory implies. Second, it was above the normal price throughout the year.
In addition it may be noted that this strategy would not have increased total revenue, given that there were constant elasticity demand curves. In fact the speculators made vast fortunes out of their speculation. Some increased revenue could have been obtained had there been a monopolist facing some linear demand curves, but there has been no suggestion that this was the case.
If enormous stocks had been carried forward into 1944, which Sen explicitly denies, prices would have collapsed then, especially as there had been a record crop. In fact, supplies remained tight and prices firm in 1944.
These models have assumed a constant market demand. This is not in accordance with normal economic theory. Nor, in my opinion, is it compatible with Sen's emphasis on lack of purchasing power as a cause of the famine, though he makes no mention, explicit or implicit, of such a fall in aggregate demand in the Bengal famine. It seems to me that demand must have fallen as people came to the end of their limited resources, and that it is impossible to talk of starvation being caused by limited purchasing power without taking this into account. It might be expected that aggregate demand would fall as the very poor finished their resources and died. It would fall as the less poor came to the end of their resources, having sold off their clothes, their doors, the roofs over their heads. It would fall as the landed peasants exhausted the money from mortgaging their land. This shift in aggregate demand would have pushed down prices in the second half of 1943.This would have happened even if constant supplies had been released over the year, but under the Sen hypothesis which implies larger supplies then, it would have meant that prices fell to zero and stayed there. Figure 8 shows the price changes implied.
Sen's hypothesis of limited purchasing power is also contradicted by the evidence of the enormous amount that was in fact spent on rice in 1943 and the enormous profits made.
Explanations for the observed price shifts
As has been emphasised previously, it is not the purpose of this monograph to do anything but disprove Sen's explanation of the Bengal famine. It is not necessary, or indeed desirable, that I should present an alternative explanation. However, in this section I present some hypotheses that would explain the price shifts actually observed. None of them are compatible with Sen's hypotheses. I make no attempt to assess the relative importance of the hypotheses; this would be impossible, as they all work in the same direction, and there is no way of knowing which were important and which are merely theoretical possibilities. Many of the hypotheses refer to clearly-documented policies and facts, on Government propaganda and purchasing policies for example. Others, on shifts in demand and expectations, are largely speculative - and one could not imagine shifts in demand curves being closely monitored in any famine.
The hypotheses are not mutually dependent, except only that they all rely on the assumption that there was a shortage in supply, which Sen denies.
Factors Affecting Supply
The most obvious explanation for the rise in price throughout the season is that most of the crop was put on the market in the earlier part of the year, and that supplies became shorter throughout the year, pushing up prices. This was taken for granted by most observers at the time.
One reason for this uneven release of rice onto the market was government action of the following kinds
These aspects of Government policy would have had favourable effects if they had increased imports to make up the shortfall later in the year.
Action by Speculators
Speculators could be expected to plan for a slight fall in supplies over the year, so that prices would rise enough to cover their storage costs (including interest and risk).However, they assumed, not unreasonably, that Government would be importing large quantities of food as soon as possible. It would pay them, therefore, to release the bulk of their stocks early in the year. Since these imports did not materialise, total supplies fell through the year. Had speculators realised that very little was to be imported, they would have released less at the beginning of the year, and would have made higher profits).
Shifting expectations would cause a shift in supply functions. In normal years, supply is largely a one-off decision made after harvest and modified by assessments of changes in import and export possibilities, which, in normal years again, are marginal. In 1943,though, speculators reacted to changes in information throughout the season. The most obvious was the realisation that there were not to be major imports. Others, like price control, imports from neighbouring provinces etc., have been detailed in Appendix 2. Not only did these changes have to be taken into account, but they had to be taken into account throughout the season, with an ever-decreasing time before the next harvest. Even the next harvest was not a fixed point in time: a late harvest caused the famine to last longer in some years, so people did not know in November whether they needed one month's supply or two months' to survive until the harvest.
Supply by Farmers
Changes in the supply by farmers appears to have had only a limited effect in causing the rise in prices. The supply by farmers (and the reservation demand) was, no doubt, affected by the famine. As has been explained earlier, changes in the retentions by peasants are not likely to have had any effect in causing the famine. It was said at the time though, that the failure of the September 1943 crop to break the famine was due to the fact that peasants were so frightened by the effect of the famine that they retained enough to keep themselves and their families alive even if the next crop were to fail. I have seen no evidence put forward to support this assertion, and it must be looked at with caution, since, as we have seen, officials everywhere are inclined to blame every shortage on hoarding. In this case, it is indeed possible that the terror of famine would have overcome the attractions of a very high price (especially among those who observed the fate of those who sold off their stocks for a high price in early 1943). It is possible, too, that this terror would have overcome to some extent the institutional constraints, that much of the crop belonged to the moneylenders and landlords rather than the producers, though, again, I have seen no evidence to confirm this.
It was suggested at the time that a switch in the supply function of this sort meant that producers (whether peasants or landlords) kept back much of the December 1943 crop, leaving supplies tight in spite of a large supply. It is difficult to believe the suggestion that a substantial switch in the supply function occurred in all of the following
If all these changes had occurred the farmers would have been net buyers by 1944.
Changes in Demand Functions
The price changes recorded could have been caused by changes in the demand function even if there had been a constant supply throughout the season - though obviously a combination of supply and demand changes is more likely.
It has been pointed out above that in the limited-purchasing-power model, which I think must be accepted if one accepts Sen, the aggregate purchasing power would fall off as the poor died and as other people impoverished themselves. This would have caused a fall in the price, rather than the rise actually observed. In the previous section some supply factors which might have outweighed this effect have been mentioned. Below are set out some demand factors which would also have tended to outweigh it.
The most powerful explanation is the effect of the subsistence sector on the market, an effect commonly seen in subsistence economies. If the subsistence sector normally produces a surplus of 10%, the marketed surplus will fall to zero if there is a 10% fall in yields and will double if there is a 10% rise in yields. Government believed at the time that some six million tons were marketed out of a total crop of 90 million in India. Such figures can be little more than guesses. They also depend enormously on definitions, where, for example one draws the line between trade within the subsistence sector and trade from the subsistence sector, or agricultural sector, to consumers, and whether sales to village consumers and deficit producers are included.
However, this effect is modified by changes in levels of consumption. Between surplus and deficit there is a range of subsistence where no food is bought or sold by subsistence producers. At the highest, people eat well, and, in some cultures, use part of the crop for ritual purposes or for alcohol. At the lowest, the family eats a bare survival ration once a day and eats up its accumulated reserves. Below this level, people have to buy in food if they are to survive. The very poor December 1942 crop meant that people switched early to survival rations, but had to come onto the market for supplies at varying stages through the year. The effect was an enormous increase in market demand towards the end of the year, with possibly two to four times the normal number of people trying to buy food on the market. Even if there had been a constant supply on the market, even if there had been a rising supply on the market, supplies would have been inadequate, and prices would have risen sharply.
Changes in Individuals' Demand
It may also be argued that the famine brought about a switch in the demand of individuals. The demand functions before the famine were based on what would be a normal, reasonable, and therefore acceptable price. (The concepts set out in this section are commonplace in marketing.) People would refuse to be cheated. Decisions would be made on the assumption that alternative consumption goods and alternative foods were available. People would be influenced by the fact that they had some food in their larders, and their expectation that they would be able to buy food at a reasonable price if they waited a day or two or went to a different market. Most buyers, no doubt, never questioned that they would always be able to buy all they wanted at a reasonable price if they had the money. The decision would be similar in kind to that made by an Englishman who is asked to pay twice the normal price for a loaf because it is the last one in the shop.
Once there was famine, all this changed. People no longer had the same conception of what was a normal price, nor did they have the same objection to being cheated. One notices in a period of acute shortages a sort of mass hysteria, with people buying goods they do not want just because they are available, with people vying with each other to buy black market goods, and then boasting of the price they paid, taking a perverse pleasure in being cheated. This reaction may be confined to the richer section of the community, but it will affect total demand.
At the same time, there were no longer alternative foods on the market: if anything, wheat was scarcer than rice. People could no longer rely on the fact that they had food at home, nor could they rely on there being plenty of food available at a reasonable price if only they waited a few days or went to other markets.
Their whole time horizon changed. They were no longer planning this week's purchases; they were planning a strategy which would enable them to survive until the next harvest(a problem which was normally confined to the subsistence farmers). Food became not a consumer good in the market, but the only means of survival. Their strategy would have to be based on medium-term expectations and production.
Some of the expectations are:-
All these expectations must be assessed according to the degree of uncertainty and to the cost (usually death) if an error is made. Time preference is important: little weight will be given to the possibility of a famine next year if there is a risk of starvation next month. As the harvest approaches, there will be ever more rapid changes in what one is willing to spend in order to survive another week: at one time only a twentieth of one's worldly goods, but in the last week before the harvest all one's worldly goods.
In the famine, those who survived were the ones whose income, capital and savings bought them enough food to survive until the next harvest (death by disease cannot be planned against). Some people whose family income was below the margin were able to survive by selling, murdering or, most common, abandoning the useless mouths in the family. The apathy caused by hunger meant that robbery to survive was very rare indeed.
No long-term demand curve can involve people spending more than they earn on food, so the people of Bengal could not have spent as much on food if there had been another famine the following year. Clearly this means that there was a major shift in demand curves over the famine period.
The September 1943 Crop
The September 1943 crop produced enough to feed the country for three months , and there were only three months to go before the next crop. It may be asked why the famine did not stop when this crop was harvested, why the starvation lasted until December. At the time it was thought by officials that it would break the famine as soon as it was harvested (Rutherford in Document no. 158 Mansergh (1973) pp361-3). When it did not do so, officials put it down to the fact that peasants would not release any of the crop. Later commentators either ignored the question, or skimmed over it. In fact it can be comfortably explained by the supply and demand hypotheses put forward above.
The September crop was an upland rice crop. It was not produced in all areas of Bengal, and the people who grew it did not necessarily produce a crop in December. In normal years most of it was retained for consumption over the whole of the following year, and only the surplus was retained. That is to say, most of the producers relied on it for most of their food. No doubt some landlords, and even some farmers, calculated that they should retain only three month's supply, and market the rest at famine prices, hoping that they would be able to buy all they needed at a low price when the December 1943 crop was harvested. It would be asking too much to expect all subsistence farmers to make such a radical change in their strategy for survival in a single year. It would mean their taking a risk that they had never taken before. The risk would seem greater because they were surrounded by the famine victims. It would seem particularly grave because they had seen the fate of the farmers who had sold their stocks to cash in on the high prices of March, and had not been able to buy back the food they needed to live on at the higher prices of May and June.
Possibly, too, there was some realisation that if the December 1943 crop was only moderately bad, the famine would continue into 1944. There were already many subsistence farmers who had consumed their accumulated reserves and were selling their possessions to buy food. If the crop was low, these people would be buying food again in 1944 and prices would be high. This might mean that the farmers who harvested in September could not buy back the food they had sold.
Even had the farmers wanted to sell their rice, there were difficulties. The marketing structure did not exist to buy all this rice from these areas, nor did the transport system -indeed, the boat denial policy had made the distribution impossible. Local charity obligations and the political and moral difficulties of exporting from a deficit area would have made it difficult to export. Certainly the local community would have been well-fed before any surplus was allowed out.
Let us suppose that, in spite of these problems, these areas marketed 20% of their crop instead of 10%. The increased supply would have come on the market as more and more of the subsistence producers were being forced onto the market to survive. The supply may have doubled, but the number of people on the market increased more
One may question whether the marketed December crop was expected to last until September, and whether the September crop was expected to last only until the next December crop was harvested (with the delays described elsewhere). It is more likely that in some areas the December crop was marketed all the year, while in the areas where a September crop was grown this crop was marketed for much or most of the year. If this was so, it further limits the possibility that the September crop could have been expected to break the famine.