Although employment in general has been a major failure of the macroeconomic policies of the past decade, the decline in employment generation in agriculture is probably more startling. The Census of India 2001 and the 55^th Round of the NSS, both show a dramatic slowdown (and in some states, actual decline) of employment in agriculture.

The NSS data suggest that the employment elasticity of agricultural output (the rate of change of employment per unit change of GDP in agriculture) has fallen from 0.7 in the period 1987–88 to 1993–94, to only 0.1 in the period 1993–94 to 1999–2000. This sharp fall means that the employment elasticity is among the lowest yet observed in Indian agriculture since such data began to be collected.

Chart 1 gives some indication of the evidence from the NSS. Employment growth rates are provided for the periods 1987–88 to 1993–94 (hereafter the first period) and 1993–94 to 1999–2000 (hereafter the second period), using the 56^th Round survey results according to various definitions and superimposing them on Census population figures. Note that these reflect people employed rather than actual work, since there is no real estimate of days/hours worked.

Chart 1 >> Click to Enlarge

The chart shows a very substantial decline of employment by all indicators, to rates that are far below those of growth of population over these periods. The usual status definition, which refers to what a person usually does over the year in question, can be interpreted as a stock measure of employment. Both in terms of principal activity and principal plus subsidiary activities, there is evidence of very significant falls in employment growth.

But there are even sharper falls, by the 'flows' definitions, of weekly and daily status employment growth, suggesting that even for those who saw themselves as usually employed, there were difficulties in getting jobs on a weekly or daily basis.

Furthermore, the data indicate that the sharpest fall has been in the number of those classified as 'self-employed' in agriculture, which has actually exhibited an absolute overall decline. This means that the number of persons working on household-operated holdings has gone down.

What explains this sharp fall in agricultural employment growth? A wide range of explanations and associated interpretations have been offered. A common explanation hinges on the expansion of the non-agricultural sector in rural areas. According to this argument, labour has been moving out of agriculture in the standard way predicted by the Lewis model, without affecting productivity in agriculture because of the prevalence of disguised unemployment in agriculture. This means that the tendency is a wholly positive one, reflecting acceleration of growth and development.

The problem with this argument is that while non-agricultural employment has indeed increased as a share of population, the increase is nowhere near enough to explain the dramatic decline in agricultural employment generation. In fact, both the increasing role of education for the younger age cohort and the apparent increase in non-agricultural employment growth, are inadequate to counterbalance the sharply decelerating growth in agricultural employment.

So much so that not all the increase in such employment may be the result of positive pull factors. Rather, several observers have described at least some of the increased non-agricultural employment as possibly reflecting a distress phenomenon (or 'push' process), with rural residents desperately searching for any jobs that are available or whatever can be described as work.

Another explanation has to do with changes in technological and cropping patterns that have reduced labour demand in agriculture. These factors may certainly have played an important role. Much of the more recent technological change in Indian agriculture has taken the form of mechanization that is labour-saving. Indeed, the rate of labour-saving mechanization appears to have increased over the past decade, and this has both reduced labour demand and made smaller farmers relatively worse off.

Similarly, cropping-pattern changes are also likely to have played a role because even while they can have divergent effects, the general thrust of such shifts (especially towards horticulture and floriculture at the margin in some areas) may be said to have reduced demand for labour.

An additional, and important, reason for the slowdown in employment generation and the large fall in employment elasticity of agricultural output growth may have to do with the pattern of land relations in rural India. It is now clear that this period witnessed a significant degree of concentration in terms of operated holdings, which reflected changes in both ownership and tenancy patterns. Many small and very marginal peasants lost their land during this period, and have therefore been forced to look for work as landless labourers; micro-level surveys have reported increasing leasing-in by large farmers from small landowners.

This is indicated, on an all-India level, in Chart 2. There has been a steep increase in landless households as a percentage of total rural households, from around 35 per cent in 1987–88 to as much as 41 per cent in 1999–2000.

Chart 2 >> Click to Enlarge

It is well known that, for various reasons, persons occupying small holdings tend to use land and labour more intensively than those with larger farms, to achieve higher per land unit productivity. Typically, this means that they employ more household members at least in some agricultural work, whether it be principal or subsidiary activity. If those who previously occupied land are now effectively dispossessed and are forced to hire themselves out as wage labour, the chances are low of a similar level of household employment being achieved.