Aspects of Structural Change since the 1980s 

An often-noted feature of the Indian economy is a lack of correspondence between trends in shares in GDP of the principal sectors and their shares in employment. In particular, while the share of the agricultural sector in GDP has fallen over time, the proportion of the population earning a livelihood in the rural areas has more or less stagnated. One obvious reason for this is the failure of urban economic activity, in particular organised manufacturing, to absorb labour to any substantial degree. In fact, rural-urban migration even to the extent that it has occurred has been accompanied by a burgeoning of the urban non-formal sector, as we note below.
 
However, what is even more disconcerting is that agriculture too has been losing its capacity to absorb labour in adequate measure. In fact, an analysis of the National Sample Survey data on agriculture not only points to a deceleration in the rate of growth of agricultural workers during the 1990s, but for some time now analysts have reported a decline in the elasticity of employment in agriculture with respect to output.
 
Based on such evidence, it has been argued that rural non-agricultural employment must be promoted both as the principal means of absorbing surplus rural labour and as a vent for such surplus labour within the rural sector itself. Not surprisingly, when the NSSO’s surveys of employment relating to 1972-73, 1977-78, 1983 and 1987-88, pointed to a sharp rise in the share of rural non-agricultural employment in total rural employment (especially in the case of males), many analysts welcomed the trend as a positive development that needed to be sustained. The reversal of that trend between 1987-88 and 1993 and its revival subsequently, have also been a matter of much debate.
 
The positive assessment of the growth of rural non-farm activity stems from the perception that under certain conditions there can exist rural linkages of a kind that permit a ‘holistic’ development of the rural areas so that the concomitant of agricultural growth, which generates additional incomes and new demands, some of which can be met with local resources and skills, is a process of occupational diversification in the rural sector itself. This reduces the dependence on the development of urban centres for generating a significant quantum of employment outside agriculture and offers an alternative to the ‘dualism’ characteristic of development in the present-day developing world.
 
In practice, however, there are a number of constraints to the realisation of such a process of rural development. Principally, increases in rural incomes lead to an increase in demand for manufactures of a kind produced in factories located at urban centres. This would be particularly true either when per capita incomes exceeds some critical minimum level or when income inequalities ensure that increases in income tend to be concentrated more among the well to do sections of the population. This implies that non-farm activities in the form of rural manufacturing would be more the exception than the rule.
 
The evidence suggests that rural industries and services are of many kinds. The three main forms of rural industry remain: (i) the production of low quality and cheap varieties of goods meeting certain kinds of needs using locally available raw materials (for example, beedi-making, bamboo work and earthenware production); (ii) agro-processing such as rice milling and production of puffed rice; and (iii) the transitional location of modern industry in rural areas which leads over time to these areas being absorbed as urban centres. (The only exceptions to this are crafts like handlooms, which historically have tended to be concentrated in certain centres).
 
As far as the first of the above-noted versions of rural industry is concerned, it reflects the low level of per capita income among the lower income deciles of the population in a region, which has prevented the diversification of their demand in favour of modern manufactures. The number of such deciles would of course depend on the average level of per capita incomes in the region and the degree of inequality. However, given the fact that modern factory-based production does not tend to be located too close to sources of demand, but rather compensates for higher transportation costs through the higher productivity and better quality associated with mass production, we should expect that as the share of the population whose real per capita income exceeds a critical minimum increases, there would be a tendency for rural-based manufacture of basic consumption goods to actually decline, with an associated decline in the employment generated by such activities.
 
Unlike in the case of manufactures, any increase in the productivity and degree of commercialisation of agriculture and/or increases in incomes from other sources in rural areas is bound to be accompanied by an increase in demand for construction services, repair services, consumer services of various kinds, transportations services and services in the communication and storage areas, which are best met with locally-based labour. Hence an increase in the share of employment and output of these sectors would be an inevitable corollary of improved rural incomes.
 
The difficulty is that the reverse need not be true. That is, an increase in employment in services need not be a reflection of an increase in rural incomes. This is because non-farm employment in services can be the result of a distress-driven spill-over into non-farm activities of a labour force that is faced with inadequate opportunities in crop production. Thus a complex of factors rather than the mere effect of a slow or fast growth in agriculture appear to explain trends in non-farm rural economic activity. Assessing any trend in rural non-agricultural employment requires therefore an examination of the structure of non-farm activity and its evolution.
 
The availability of the results of the Fourth Economic Census relating to 1998, and the comparability of figures yielded by the three censuses relating to the years 1980, 1990 and 1998 provides such an opportunity.  The Censuses provide information on the number of and employment in enterprises engaged in all activities outside crop production and plantations in both urban and rural areas. They however include figures on a number of activities allied to agriculture such as livestock production and agricultural services including hunting, trapping and game propagation, forestry, logging and fishing. These are identified as agricultural activities in the Economic Censuses. These activities have been registering an increase in importance in the rural areas. As Charts 3 and 4 indicate the share of these activities as captured by both their share in the number of rural enterprises and their share in rural employment outside of crop production and plantations has risen consistently leading to a decline in the share of non-agricultural activities between 1980 and 1998. This points to a shrinking of the relative role of “typical” non-farm activities in rural economic activity.

This shrinkage of the relative role of what are considered typical non-agricultural activities could have been interpreted as an indication of rural dynamism in the form of the diversification of agricultural activity but for other disconcerting trends revealed by a comparison of the Economic Censuses. The first striking trend revealed by the figures reported in Charts 1 and 2 is a deceleration in activity in the non-crop producing and non-plantation segments of both urban and rural economy. To start with the urban areas, the evidence points to a decline in the growth of the number of non-farm enterprises from an annual average rate of 3.6 per cent to 2.5 per cent between 1980-90 and 1990-98, and a sharp decline in the number of workers in the non-farm sector from 3.2 to 1.4 per cent. This points to a deceleration in the expansion of employment in the urban sector. While a part of this may be because of improvements in productivity and the movement of child workers into education, the sharpness of the fall does point to inadequate employment growth in urban non-farm activity.

When combined with the evidence on the falling elasticity of employment growth with respect to output growth in the agricultural sector referred to earlier, these urban trends increase the significance of rural non-farm activity. However, even here there are clear signs of deceleration. The average annual rate of growth of the number of rural enterprises only fell marginally from 2.7 per cent during 1980-90 to 2.3 per cent during 1990-98. However, the rate of growth of workers engaged in non-farm activity in the rural areas fell from 3.3 per cent to 2.3 per cent during these two periods. This suggests that the expansion of non-farm activity has occurred through the growth of a number of smaller enterprises, suggesting that at least some of this expansion could be the result of a distress-driven shift to non-agricultural activity.
 
In fact, as Chart 5 makes clear, the growth of non-farm activity has not been accompanied by the growth of small capitalist enterprises based on hired labour, which would also be indicative of dynamic diversification of economic activity. Rather, the share of own-account enterprises (OAEs) or enterprises that do not use hired labour has remained constant at around 77 per cent of all non-farm enterprises in rural areas right through the 1980s and 1990s. Interestingly, even in the urban areas the share of OAEs has declined by just 4 percentage points between 1980 and 1998 and still remains at the relatively high of 62 per cent of all urban non-farm enterprises even in 1998.

Another indicator of the nature of evolution of the structure of economic activity, is the industry-wise distribution of non-farm activity. As Chart 6 shows, three kinds of ‘industries’ dominated rural non-farm activities during the period under study: manufacturing, the wholesale and retail trade, and community, social and personal services, which together accounted for around 90 per cent of non-farm enterprises. However, there have been significant changes in the shares of these three sectors over the 18 years that the three censuses under discussion spanned. The share of manufacturing enterprises in all non-agricultural enterprises fell from 39 per cent in 1980 to 25 per cent in 1998, while that of trading enterprises increased from 31 to 38 per cent and of enterprises engaged in community, social or personal services from 21 to 26 per cent. Thus, it was not areas like storage, transportation, and repair services, where growth would be triggered by agricultural dynamism, that non-farm activities expanded, but in areas like trade and social and personal services.

It could be argued that it is not the expansion in the number of enterprises but in employment that matters, but as chart 7 relating to 1998 shows, the domination of the three sectors mentioned above was even greater in employment than in the number of enterprises and manufacturing’s share in employment was not way out of line from its share in non-farm enterprises.

The domination of trade and service activities in rural non-farm employment is accompanied by a domination of small enterprises in non-farm activity. Own account enterprises that do not employ hired workers accounted for 47 per cent of all non-agricultural activity in 1998 and non-directory establishments or those employing less than six workers (hired and non-hired) for another 21 per cent (Chart 8). Medium sized capitalist enterprises with more than six workers accounted for just 32 per cent of total employment. What is more, even in urban areas, OAEs and NDEs accounted for 50 per cent of non-agricultural employment in 1998 (Chart 9).

Finally, if we examine the distribution of non-agricultural workers by enterprises classified according to size class of employment, we find that even in 1998 small enterprises overwhelmingly dominated the total number of enterprises and substantially dominated in terms of share in workers.  Enterprises employing less than 6 workers accounted for 84 and 79 per cent respectively of rural and urban non-agricultural enterprises and 41 and 33 per cent of rural and urban non-agricultural workers (Chart 10 and 11).

How is all this information to be interpreted? First, the evidence seems to suggest that the kind of diversification conventionally associated with capitalist development has not occurred in India even after five decades of post-Independence development. That process of diversification in the developed countries was captured by Kuznets as involving “the shift away from agricultural to non-agricultural pursuits” and only recently “away from industry to services”, besides “a change in scale of productive units, and a related shift from personal enterprise to impersonal organisation of economic firms, with a corresponding change in the occupational status of labour.” Here, the shift to services has occurred well before manufacturing matured, and the shift to larger scale and impersonal forms of organisation has been extremely gradual.
 
Second, there are signs that even this gradual process of diversification of economic activity has lost momentum during the 1990s.
 
Third, the structure of non-agricultural activity in the rural areas and its evolution suggests that the pace and pattern of growth of agriculture has not adequately contributed to a “dynamic” diversification of rural non-agricultural activity.
 
Finally, semi-capitalist forms of organisation and smaller scale units still dominate the non-agricultural sector in both urban and rural areas to an extent where they appear indicative of structural stagnation in India’s development.