Economic growth is associated with a relative shift in the structure of the workforce away from agriculture, towards industry and services. This happened not only to primary product importers like Great Britain, where agriculture's share has declined to less than 4 per cent, but also to specialized primary producers like Australia, Denmark and New Zealand, where agriculture's share has declined to 10 or 15 per cent. Simon Kuznets, who restricts his comparison of long-term trends in the structure of the labour force to twenty-five countries for which reasonably comparable information is available, finds India to be the only case of a virtually unchanged employment structure. This makes the Indian experience an important one in the history of economic development, worthy of careful, detailed examination.

Reasonably reliable information on the structure of the workforce is available for the twentieth century from the population censuses. Census data also exist for 1871–2, 1881 and 1891, but the 1871–2 data which relate to British India and some native states are clearly unreliable as the adult male workers exceed the adult male population by about 4.6 million; the 1891 data relate to the occupations of the entire population and cannot therefore be compared with the twentieth century estimates which generally relate to the workforce. The 1881 evidence can be used provided its limitations are realized. The 1881 estimates relate to males and to a territory which comprises Ajmer, Bengal, Berar, Bombay, Central Provinces, Coorg, Madras, North-West Provinces, Punjab, Baroda, Central India, Mysore and Travancore. A further deficiency is that the proportion of ‘general labour’ (i.e., unspecified workers) is as high as 8.3 per cent of the workforce. We compare the 1881 estimates with similar estimates for 1911, a fairly normal year.