MacroScan (Printable Version) - Slowdown in employment growth: Disturbing trends of the 1990s

One of the great failures of the Indian development strategy over the decades, along with the persistence of poverty and the slow rate of increase in human development indicators, has been inadequate employment generation. This is not just a problem of welfare, since it represents a huge waste of human resources that are crucial to building the economy, and suggests that Indian growth could have been both faster and more equitable if only the enormous labour reserves had been productively utilised.

In the 1990s, it became fashionable among critics of the mixed-economy planning-based strategy to argue that it was this very strategy that was responsible for the slow rate of employment growth. It was suggested that export pessimism and an inward looking import substitution policy had discouraged employment-intensive export production and imposed high-cost capital-intensive production which had low linkage effects with the rest of the economy and did not lead to more use of labour.

These critics argues that opening up the economy to more liberal external trade and foreign investment would not only generate a higher rate of output growth but also automatically create a restructuring of production which would mean a significant increase in labour-intensive production and therefore also substantial increases in employment.

The initial evidence from the early 1990s suggested that this promised increase had not really materialised, especially according to the small samples carried out by the NSS. But this was dismissed by the votaries of neoliberal reforms, who maintained that not only was it likely that the effects of the reforms would take some years to work through, and that therefore it was then too early to come to any judgement, but that in any case the small samples were did not provide sufficient statistical basis on which to come to any conclusion.

That is why the results of the first large sample NSS round since 1993-94, that is the 55th Round which was conducted over 1999-2000, are so significant. These provide the first comprehensive estimate of changes in employment growth and patterns over the 1990s, since the other available data on organised sector employment covers such a small proportion of the total labour force. And since the survey deals with a large sample, it has relevance not only for the all-India pattern but also the experience of different states.

The key results of the 55th Round regarding employment and unemployment have just been released. These reveal a sharp, and even startling, decrease in the rate of employment generation across both rural and urban areas. Indeed, so dramatic are the fall of work force participation and the slowdown in the rate of employment growth that they call into serious question the pattern of growth over this decade.

Thus, as Chart 1 shows, the per cent of rural population, both male and female, showing some usual status employment, has not only declined over the 1990s, but in 1999-2000 was at the lowest level in thirty years. Even in urban areas, as shown in Chart 2, there is evidence of decline since 1993-94, although the levels are not the lowest over the period as in rural areas.
Chart 1 >> Chart 2 >>

This is indeed remarkable, especially given that the CSO's national accounts data indicates that the decade of the 1990s shows a marginally higher rate of growth of output than the previous decade. Thus, the fall in work force participation has come in a period during which economic activity in the aggregate is supposed to have accelerated.

What these employment rates imply in terms of the growth of employment is shown in Chart 3. This chart shows growth rates of employment which have been calculated using extrapolated annual population data based on Census figures up to 1991, and thereafter using the now standard projections of annual rates of growth of 1.73 per cent for the rural population and 1.84 per cent for urban population for the period after 1991.

Based on these estimates of population and the employment data of the NSS, estimates of the annual compound rate of employment growth in rural and urban India are provided in Chart 3. This shows a very significant deceleration for both rural and urban areas, with the annual rate of growth of rural employment falling to as low as 0.67 per cent over the period 1993-94 to 1999-2000. This is not only less than one-third the rate of the previous period 1987-88 to 1993-94, it is also less than half the projected rate of growth of the labour force in the same period. In fact, it turns out that this is the lowest rate of growth of rural employment in post-Independence history.
Chart 3 >>

We have also attempted to calculate the employment elasticity of rural output growth based on these growth rates and on estimates of the rural share of GDP. (The methodology of these estimates of rural GDP growth is the same as that used for a previous Macroscan, in Businessline February 22, 2000, and the details are provided there.) This yields an employment elasticity of rural output growth of only 0.13 for 1993-94 to 1999-00, compared to 0.38 for the previous period.

In other words, the employment elasticity of rural output growth has declined to less than one-third of what it was in the earlier period, which itself represented a decline compared to past trends. It also means that it can no longer be assumed that the process of economic growth itself will necessarily generate much more employment even in the rural areas, as has commonly been supposed by advocates of the economic liberalisation process.

If this is correct, this clearly points to some major problems with the pattern of growth that has emerged in the rural areas in particular. Note that these rates of employment growth refer to all forms of employment in the rural areas, and what is emerging is that there is no sector in which employment is growing fast enough to take up the slack that is brought about by falling labour use patterns in agriculture. This point is discussed in more detail below.

Even in the urban areas, the rate of employment generation has slowed down very dramatically over the various periods considered here, and in the latest period it appears to have been also well below the estimated rate of growth of the urban population. Urban elasticity of employment calculations have not been attempted here, but since it is widely felt that economic growth in the 1990s has been disproportionately higher in the urban areas, the significantly lower rates of employment growth here also suggest very low and falling elasticity of employment generation.

It is worth examining in some more detail the data on types of employment. The past several decades have been marked by a fairly steady process of casualisation of labour, and the past decade turns out to have been no exception. As Charts 4a and 4b indicate, for both rural men and women, the share of casual employment has gone up, in fact quite substantially for men.
Chart 4a >> Chart 4b>>

This is probably no surprise, given the lack of more regular employment generation in the countryside following from the more general decline in per capita public spending on the rural areas. This has naturally led to a drying up of regular employment opportunities, and the reduced multiplier effects of this decline also would have had an effect. But it is also not surprising given the lower rate of aggregate employment generation in the countryside, since periods of excess supply in labour markets are typically associated with worse conditions for the workers, in terms of less secure or regular contracts at the very least.

Chart 4c and 4d give the related patterns for urban men and women workers. Here the pattern is much more mixed. For men, there is only a very slight increase in casual contracts. For women workers, in fact there is a substantial increase in regular work, although this increase is greater when all workers (principal and subsidiary status) are taken together, rather than for principal status alone. This suggests that at least some of the increase may be a reflection of regular but secondary work undertaken by women to supplement household income, for example through manufacturing activities on a putting out basis at home, or part time service activities.
Chart 4c>> Chart 4d>>

However, the chart on urban women's employment pattern should not be taken as definitive simply because the data (which are reported as given in the NSSO report) for the year 1993-94 appear to be inaccurate since they do not total to 100. Assuming that the other years' figures are correct, there appears to be an increase in the share of regular employment and a decline in casual labour, thus reversing the trend of the earlier decades.

One of the much discussed features of rural employment generation over the 1980s was the diversification of employment away from agriculture and primary activities generally, towards secondary and service sector employment. In general this was seen to be a very positive feature, especially as it was accompanied by overall growth of rural employment in most regions of the country and was also associated with a trend decline in the incidence of rural poverty.

In this context, the latest evidence suggests that the move away from primary sector work has become even more marked in the late 1990s, after a period in which such diversification appeared to have reversed between 1987-88 and 1993-94. Thus, Charts 5a and 5b show that primary sector employment has declined for both men and women in the rural areas by 1999-2000, and that the decline has been especially sharp in the case of men, for whom tertiary sector employment has increased its share significantly.
Chart 5a >> Chart 5b >> Chart 5c >> Chart 5d >>

On the face of it, this should be a welcome tendency, suggesting that dependence on agriculture is decreasing and that the rural economy is experiencing much-needed diversification of employment opportunities. However, this latest pattern needs to be viewed in conjunction with the evidence on falling work participation rates and decelerating employment generation, as presented in Charts 1,2 and 3.

With this background, it becomes less obvious that the decline in primary sector share of employment is unambiguously positive. It may reflect the fact that the employment intensity of agricultural production is falling quite sharply especially with the new crops that are being introduced, thus leading to reduced growth of demand for agricultural labour. In such a case the service sector may effectively become a residual sector, in which there is underemployment occasioned by the fact that rural workers are forced to take up some occupation, however low the returns.

If the service sector were indeed a source of substantial rural dynamism, pulling labour in from other sectors, then the likelihood is that there would not have been such a decline in the rates of growth of overall employment. But instead, there is actually a decline in overall work participation rates in rural areas. This suggests that the observed shift in shares is simply a result of the fact that the deceleration of employment growth has been even sharper for agriculture than it has been for non-agricultural rural activities.

For the urban sector the data do not throw up any surprises, since the earlier and now predictable pattern of increase in the share of tertiary employment is evident for both men and women workers.

However, the data that do come as a surprise, given the earlier results, relates to the information on unemployment. It would have been expected, given the fairly sharp decline in usual status employment, to find that unemployment rates had gone up commensurately. In fact, however, as Chart 6a shows, usual status unemployment has increased only marginally for rural men and women, and has actually decreased for urban men and women workers. This points to a substantial decline in labour participation rates overall.
Chart 6a >>

The reason for this is hard to comprehend, since it is usually typical of poor countries with low to non-existent social security benefits, that the working population cannot afford the luxury of open unemployment. Rather, they have no option but to take up some gainful activity, however low paying, simply in order to survive, and thus this usually expresses itself as underemployment. Instead of this, we find that according to the latest NSS Round, there is actually a withdrawal from the labour force, a feature which is more commonly associated with developed post-industrial economies.

Chart 6b, which shows unemployment rates according to the daily status definition, finds a different tendency. According to this, unemployment rates have gone up for all workers except female urban workers. While this may be closer to expectation, it does reflect a different indicator, a flow measure rather than a stock measure of the labour market conditions.
Chart 6b >>

What does all this add up to ? First of all, it suggests that there is a major difficulty with an economic growth process that not only does not increase the rate of employment generation, but actually brings it to a historic low. Clearly, the production restructuring that has occurred in the Indian economy has not been of a type which has created more labour-intensive productive activities, nor has there been sufficient dynamism to make the overall employment grow faster because of sheer volume increases in output.

This is quite contrary to the expectations that were generated at the start of the neoliberal reform process, as mentioned at the outset. But they are also at variance with the apparent increases in output growth rates. This is possible because of declining employment elasticities in production, which would in fact suggest quite the opposite type of production restructuring from that which was anticipated.

These data also have to viewed in conjunction with the evidence on household consumption which was also collected during the 5th Round. As is now well known, this round found substantial increases in per capita household consumption (and presumably, therefore, there would be associated declines in the incidence of absolute poverty) on the basis of using together two recall reference periods of one week and one month.

The implications of this have already been discussed at length in an earlier edition of Macroscan and elsewhere (see Macroscan, Businessline September 19,2000 and Abhijit Sen, "Consumer spending and its distribution : Statistical priorities after NSS 55th Round", Economic and Political Weekly December 16-22, 2000) and we do not propose to go into this matter again.

Suffice it to say that the evidence on employment generation - or rather, the lack of it - that emerges from these results, suggests that the basic economic problems of unemployment and low productivity unemployment not only remain unresolved, but have actually grown more acute, after a decade of marketist reforms.



Themes > Features	23.01.2001


Slowdown in Employment Growth: Disturbing Trends of the 1990s

One of the great failures of the Indian development strategy over the decades, along with the persistence of poverty and the slow rate of increase in human development indicators, has been inadequate employment generation. This is not just a problem of welfare, since it represents a huge waste of human resources that are crucial to building the economy, and suggests that Indian growth could have been both faster and more equitable if only the enormous labour reserves had been productively utilised. In the 1990s, it became fashionable among critics of the mixed-economy planning-based strategy to argue that it was this very strategy that was responsible for the slow rate of employment growth. It was suggested that export pessimism and an inward looking import substitution policy had discouraged employment-intensive export production and imposed high-cost capital-intensive production which had low linkage effects with the rest of the economy and did not lead to more use of labour. These critics argues that opening up the economy to more liberal external trade and foreign investment would not only generate a higher rate of output growth but also automatically create a restructuring of production which would mean a significant increase in labour-intensive production and therefore also substantial increases in employment. The initial evidence from the early 1990s suggested that this promised increase had not really materialised, especially according to the small samples carried out by the NSS. But this was dismissed by the votaries of neoliberal reforms, who maintained that not only was it likely that the effects of the reforms would take some years to work through, and that therefore it was then too early to come to any judgement, but that in any case the small samples were did not provide sufficient statistical basis on which to come to any conclusion. That is why the results of the first large sample NSS round since 1993-94, that is the 55th Round which was conducted over 1999-2000, are so significant. These provide the first comprehensive estimate of changes in employment growth and patterns over the 1990s, since the other available data on organised sector employment covers such a small proportion of the total labour force. And since the survey deals with a large sample, it has relevance not only for the all-India pattern but also the experience of different states. The key results of the 55th Round regarding employment and unemployment have just been released. These reveal a sharp, and even startling, decrease in the rate of employment generation across both rural and urban areas. Indeed, so dramatic are the fall of work force participation and the slowdown in the rate of employment growth that they call into serious question the pattern of growth over this decade. Thus, as Chart 1 shows, the per cent of rural population, both male and female, showing some usual status employment, has not only declined over the 1990s, but in 1999-2000 was at the lowest level in thirty years. Even in urban areas, as shown in Chart 2, there is evidence of decline since 1993-94, although the levels are not the lowest over the period as in rural areas. Chart 1 >> Chart 2 >> This is indeed remarkable, especially given that the CSO's national accounts data indicates that the decade of the 1990s shows a marginally higher rate of growth of output than the previous decade. Thus, the fall in work force participation has come in a period during which economic activity in the aggregate is supposed to have accelerated. What these employment rates imply in terms of the growth of employment is shown in Chart 3. This chart shows growth rates of employment which have been calculated using extrapolated annual population data based on Census figures up to 1991, and thereafter using the now standard projections of annual rates of growth of 1.73 per cent for the rural population and 1.84 per cent for urban population for the period after 1991. Based on these estimates of population and the employment data of the NSS, estimates of the annual compound rate of employment growth in rural and urban India are provided in Chart 3. This shows a very significant deceleration for both rural and urban areas, with the annual rate of growth of rural employment falling to as low as 0.67 per cent over the period 1993-94 to 1999-2000. This is not only less than one-third the rate of the previous period 1987-88 to 1993-94, it is also less than half the projected rate of growth of the labour force in the same period. In fact, it turns out that this is the lowest rate of growth of rural employment in post-Independence history. Chart 3 >> We have also attempted to calculate the employment elasticity of rural output growth based on these growth rates and on estimates of the rural share of GDP. (The methodology of these estimates of rural GDP growth is the same as that used for a previous Macroscan, in Businessline February 22, 2000, and the details are provided there.) This yields an employment elasticity of rural output growth of only 0.13 for 1993-94 to 1999-00, compared to 0.38 for the previous period. In other words, the employment elasticity of rural output growth has declined to less than one-third of what it was in the earlier period, which itself represented a decline compared to past trends. It also means that it can no longer be assumed that the process of economic growth itself will necessarily generate much more employment even in the rural areas, as has commonly been supposed by advocates of the economic liberalisation process. If this is correct, this clearly points to some major problems with the pattern of growth that has emerged in the rural areas in particular. Note that these rates of employment growth refer to all forms of employment in the rural areas, and what is emerging is that there is no sector in which employment is growing fast enough to take up the slack that is brought about by falling labour use patterns in agriculture. This point is discussed in more detail below. Even in the urban areas, the rate of employment generation has slowed down very dramatically over the various periods considered here, and in the latest period it appears to have been also well below the estimated rate of growth of the urban population. Urban elasticity of employment calculations have not been attempted here, but since it is widely felt that economic growth in the 1990s has been disproportionately higher in the urban areas, the significantly lower rates of employment growth here also suggest very low and falling elasticity of employment generation. It is worth examining in some more detail the data on types of employment. The past several decades have been marked by a fairly steady process of casualisation of labour, and the past decade turns out to have been no exception. As Charts 4a and 4b indicate, for both rural men and women, the share of casual employment has gone up, in fact quite substantially for men. Chart 4a >> Chart 4b>> This is probably no surprise, given the lack of more regular employment generation in the countryside following from the more general decline in per capita public spending on the rural areas. This has naturally led to a drying up of regular employment opportunities, and the reduced multiplier effects of this decline also would have had an effect. But it is also not surprising given the lower rate of aggregate employment generation in the countryside, since periods of excess supply in labour markets are typically associated with worse conditions for the workers, in terms of less secure or regular contracts at the very least. Chart 4c and 4d give the related patterns for urban men and women workers. Here the pattern is much more mixed. For men, there is only a very slight increase in casual contracts. For women workers, in fact there is a substantial increase in regular work, although this increase is greater when all workers (principal and subsidiary status) are taken together, rather than for principal status alone. This suggests that at least some of the increase may be a reflection of regular but secondary work undertaken by women to supplement household income, for example through manufacturing activities on a putting out basis at home, or part time service activities. Chart 4c>> Chart 4d>> However, the chart on urban women's employment pattern should not be taken as definitive simply because the data (which are reported as given in the NSSO report) for the year 1993-94 appear to be inaccurate since they do not total to 100. Assuming that the other years' figures are correct, there appears to be an increase in the share of regular employment and a decline in casual labour, thus reversing the trend of the earlier decades. One of the much discussed features of rural employment generation over the 1980s was the diversification of employment away from agriculture and primary activities generally, towards secondary and service sector employment. In general this was seen to be a very positive feature, especially as it was accompanied by overall growth of rural employment in most regions of the country and was also associated with a trend decline in the incidence of rural poverty. In this context, the latest evidence suggests that the move away from primary sector work has become even more marked in the late 1990s, after a period in which such diversification appeared to have reversed between 1987-88 and 1993-94. Thus, Charts 5a and 5b show that primary sector employment has declined for both men and women in the rural areas by 1999-2000, and that the decline has been especially sharp in the case of men, for whom tertiary sector employment has increased its share significantly. Chart 5a >> Chart 5b >> Chart 5c >> Chart 5d >> On the face of it, this should be a welcome tendency, suggesting that dependence on agriculture is decreasing and that the rural economy is experiencing much-needed diversification of employment opportunities. However, this latest pattern needs to be viewed in conjunction with the evidence on falling work participation rates and decelerating employment generation, as presented in Charts 1,2 and 3. With this background, it becomes less obvious that the decline in primary sector share of employment is unambiguously positive. It may reflect the fact that the employment intensity of agricultural production is falling quite sharply especially with the new crops that are being introduced, thus leading to reduced growth of demand for agricultural labour. In such a case the service sector may effectively become a residual sector, in which there is underemployment occasioned by the fact that rural workers are forced to take up some occupation, however low the returns. If the service sector were indeed a source of substantial rural dynamism, pulling labour in from other sectors, then the likelihood is that there would not have been such a decline in the rates of growth of overall employment. But instead, there is actually a decline in overall work participation rates in rural areas. This suggests that the observed shift in shares is simply a result of the fact that the deceleration of employment growth has been even sharper for agriculture than it has been for non-agricultural rural activities. For the urban sector the data do not throw up any surprises, since the earlier and now predictable pattern of increase in the share of tertiary employment is evident for both men and women workers. However, the data that do come as a surprise, given the earlier results, relates to the information on unemployment. It would have been expected, given the fairly sharp decline in usual status employment, to find that unemployment rates had gone up commensurately. In fact, however, as Chart 6a shows, usual status unemployment has increased only marginally for rural men and women, and has actually decreased for urban men and women workers. This points to a substantial decline in labour participation rates overall. Chart 6a >> The reason for this is hard to comprehend, since it is usually typical of poor countries with low to non-existent social security benefits, that the working population cannot afford the luxury of open unemployment. Rather, they have no option but to take up some gainful activity, however low paying, simply in order to survive, and thus this usually expresses itself as underemployment. Instead of this, we find that according to the latest NSS Round, there is actually a withdrawal from the labour force, a feature which is more commonly associated with developed post-industrial economies. Chart 6b, which shows unemployment rates according to the daily status definition, finds a different tendency. According to this, unemployment rates have gone up for all workers except female urban workers. While this may be closer to expectation, it does reflect a different indicator, a flow measure rather than a stock measure of the labour market conditions. Chart 6b >> What does all this add up to ? First of all, it suggests that there is a major difficulty with an economic growth process that not only does not increase the rate of employment generation, but actually brings it to a historic low. Clearly, the production restructuring that has occurred in the Indian economy has not been of a type which has created more labour-intensive productive activities, nor has there been sufficient dynamism to make the overall employment grow faster because of sheer volume increases in output. This is quite contrary to the expectations that were generated at the start of the neoliberal reform process, as mentioned at the outset. But they are also at variance with the apparent increases in output growth rates. This is possible because of declining employment elasticities in production, which would in fact suggest quite the opposite type of production restructuring from that which was anticipated. These data also have to viewed in conjunction with the evidence on household consumption which was also collected during the 5th Round. As is now well known, this round found substantial increases in per capita household consumption (and presumably, therefore, there would be associated declines in the incidence of absolute poverty) on the basis of using together two recall reference periods of one week and one month. The implications of this have already been discussed at length in an earlier edition of Macroscan and elsewhere (see Macroscan, Businessline September 19,2000 and Abhijit Sen, "Consumer spending and its distribution : Statistical priorities after NSS 55th Round", Economic and Political Weekly December 16-22, 2000) and we do not propose to go into this matter again. Suffice it to say that the evidence on employment generation - or rather, the lack of it - that emerges from these results, suggests that the basic economic problems of unemployment and low productivity unemployment not only remain unresolved, but have actually grown more acute, after a decade of marketist reforms.



© MACROSCAN 2001