POVERTY AND FAMINES 67 available". Sen brings in empirical evidence to argue that this was not the case. I am not familiar whh the Famine Inquiry Commission's report and other related material regarding the Bengal famine. Hence I am not in a position to say whether Sen's reconstruction of the empirical evidence is valid or not. I shall proceed on the assumption that it is and raise questions about the nature of Sen's arguments. A brief account of the famine itself is necessary and I give below information gathered from Sen's book, but rearranged to some extent. The famine took place in 1943 during the course of, indeed during the crucial and difficult period of, the Second World War. The output of rice in Bengal in 1940 was 8.2 million tonnes, 6.8 million tonnes in 1941, 9.3 million tonnes in 1942 and 7.6 million tonnes in 1943. Taking into account the net imports of rice and the outputs and imports of wheat, total foodgrains available in Bengal from 1940 to 1943 were 10.2 million tonnes, 8.3 million tonnes, 10.9 million tonnes and 9.2 million tonnes. The indices of total foodgrains (taking 1941=100) were 122, 100, 131 and 111 for these years and the corresponding per capita availability indices were 123, 100, 130 and 109. From these Sen argues that there was no food availability decline in Bengal in 1943 as the total availability and per capita availability of foodgrains in 1943 were higher than in 1941 (when there was no famine). Hence the FAD view is shown to be untenable. Sen now turns to his alternative explanation. If the famine was not the result of the decline in the availability of foodgrains, starvation must have been caused by the inability of some to purchase foodgrains— this is the essence of the entitlement approach which is expressed in the opening sentences of the book: "Starvation is the characteristic of some people not having enough food to eat. It is not the characteristic of there being not enough food to eat" (p 1; emphases in the original). So Sen directs his enquiry to see whether in Bengal in 1943 there were people who were not in a position to purchase foodgrains, or those "who experienced a decline in their exchange entitlements". And he finds that the ratio of the price of rice to wage rates (or the ratio of the price of rice to the price of the goods they were producing) turned sharply against "fishermen, transport workers, paddy huskers, agricultural labourers, those in 'other productive occupations', craftsmen and non-agricultural labourers, in that order" (pp 71-72). And so he arrives at the conclusion that famines are caused not by FAD, but by shifting exchange entitlements. Sen rests this part of his argument by demonstrating that in a famine situation not everybody is equally affected, that the most affected are those with limited resource entitlements and have to purchase foodgrain from the market and that "there is a fairly close relation between the interoccupational orderings of pauperization in the 'immediate pre-famine' and the 'famine' periods" (p 71). Jn sum, a famine has a differential impact on different sections in society and the worst affected are those who are usually on the verge of