Agrarian stratification in India as it emerges from NSO 2021 Report

C. N. Subramaniam

The National Statistical Office (NSO) published a report in 2021 called, Situation Assessment of Agricultural Households and Land and Holdings of Rural India, 2019. This was based on the 77th National Sample Survey done in 2018-19. The following is a partial analysis of the data presented in the report.

Profile of rural society in India

Based on the survey it is estimated that there are in all 172.4 million rural households in India. Of these 93 million are estimated to be ‘agricultural’ households and another 79 million households are engaged in non-agricultural work in the rural areas. In other words 54% of rural households are engaged in agriculture while as much as 46% households are engaged in non-agricultural work. It is important to note that there has been a significant shift of households from agriculture to non-agricultural work without shifting to urban centres. There has been a pronounced proliferation of rural non-agricultural employment in recent years.

40% of rural households are ‘self-employed’ in agriculture, implying that they are peasant farmers. 42% of rural households are entirely dependent upon wage labour. (16% of rural households are engaged in agricultural labour and more than 26% in casual or regular non-agricultural labour.)  9% are said to be self-employed in non-agricultural enterprises implying small shopkeepers or artisans. Quite clearly non-agricultural employment is becoming significant for wage workers in the countryside. In other words proletarianization is a marked phenomenon in the countryside today. But this is not all. As we shall see a little later, even a large section of the so called ‘self-employed’ in agriculture have to resort to selling their labour power to make their ends meet.

The survey report gives information about educational qualifications of rural inhabitants. Total literacy is still a distant dream in the rural areas with the literacy rate still pegged at 74% of all rural persons above the age of 7. Female literacy is as low as 65%.  Only 32% of all rural persons above 15 years have completed 12-year schooling. This of course has implications for the kind of employment they can access and even the kind of rational agriculture they can practice.

The rural population also has a caste profile. 12% of these rural households are of Scheduled Tribes (ST); 22% are Scheduled Castes (SC) while the Other Backward Castes (OBC) constitute about 44% and upper castes (Others) about 22%. We should remember that by and large the STs do not inhabit the same territories as the SC, OBC and Others. The latter inhabit mainly plain agricultural villages while the STs usually inhabit forested and hilly terrains with less intensive agriculture and also fewer multicaste villages. That is why NSO’s clubbing of STs with the rest is a bit misleading.

The rural population is thus roughly equally divided between farmers and farm and non-farm workers. While the issues faced by the farmers and the internal stratification among them have been much talked about and represented in movements, the substantially large workers and especially those engaged in non-farm sectors who actually represent the rural proletarians and semi proletarians, has not attracted the attention of either scholars or labour organisers. Of course we do have a few powerful and militant organisations representing health and childcare women workers, most of them working in rural areas. This only gives us an indication of the precarity of their working conditions and the radical potential of organising them and fighting for their rights. We should also note that the caste dimension of the rural population. A substantial number of the rural wage workers belong to the SC or ST category, the latter largely working for either the forest department or public work contractors. Age old caste prejudices create wedges between the working people as the depressed castes or Dalits are employed in low status jobs and the upper caste workers would like to stay aloof from them. Any rural working-class movement that seeks to unite all casual wage workers has to address this issue.

Landownership

The NSO report gives details about different grades of landownership which is of great interest. The following table summarises some of the findings.

Table 1. Percentage distribution of rural households and percentage distribution of area owned by size category

ST SC SC Other All
Land holding ha land area
000 ha
Households
00
% of
STs
land area
000 ha
House % of
SC
land area
000 ha
House % of
OBC
land area
000 ha
House
% of
other
land area
000 ha
%
Households
00
%
0 0 18928 9 0 33425 9 0 65101 8.5 0 23531
6
0
 -
140984
8
<.04 53 47324 22 186 175771 47 265 236305 31 113 91487 25
616 1 550887 32
<1 4750 104205 49 4511 139663 38 13560 341763 45 6862 176529
48
29683 34 762159 44
<2 3783 28447 13 2139 15543 4 9951 72134 9 6045 43831
12
21919 25 159954 9
<4 2308 9584 5 1544 6389 2 8880 34646 5 6561 25638
7
19294 22 76257 4
<10 1107 2169 1 618 1162 0.3 6656 11717 2 4503 8201
2
12882 15 23248 1
>10 355 180 0.1 3 3 0.001 2129 1511 0.2 974 730
0.2
3460 4 2424 0.1
Total
12356 210837 100 9001 371956 100 41441 763177 100 25058 369947
100
87854 100 1715913 100
% of all
land and
households
14 12 - 10 22 - 47 44 -
29 22
-
- -
-
-

Based on Statement 2.2 & Table 31 of Appendix A, Situation Assessment

Of all rural households (both ‘agricultural’ and non-agricultural), 8% are entirely landless (without even homesteads), presumably engaged in wage labour. This is the national average. Some states like Andhra Pradesh diverge significantly from this. Andhra – 18%; Telangana, West Bengal, Odisha, Gujarat, Haryana etc have between 11-14% landless households. Uttarakhand has as much as 21% landless households. We also need to see how much the figures will change once we remove the tribal districts from this ‘national average’. The caste dimension operates here as well. While landlessness among the OBC households is around the average, it is around 9% for the SCs and ST households. Only 6% of upper castes are landless in comparison. (NSO – Statement 2.2: Percentage distribution of rural households and percentage distribution of area owned by size category of ownership holdings for each household social group)

It is generally acknowledged that on an average 2 to 8 hectares of land is needed to sustain a family. (1 ha = 2.4 acre). Naturally, this will vary a lot depending upon the fertility, irrigation and location of the land. With this ballpark figure in mind let us look at the situation of the farming households.

The report (Statement 2.2) also tells us that of all rural households 73% own less than one hectare of land. These 73% households together own about 35% of lands owned by all rural households. This effectively means that their land will not be sufficient to sustain the household and they have to work as wage labourers on other farms etc. The NSO categorises them as ‘marginal’ households. 76% of SC and OBC households are in the marginal category while the corresponding figure for the upper castes it is 60%. In other words the overwhelming proportion of SC and OBC agricultural households are marginal.

Let’s go to the next category of landownership, or small landowners owning one hectare to less than 2 ha. Over 9% of rural (and 16% of all agricultural) households are in this category and together they own 25% of all agricultural land held by rural households. We should remember that when it comes to making ends meet, the small farmers fare no better than the marginal farmers and have to resort to wage labour to sustain themselves. If we club these two categories we get the following results:

Percentage of Rural households in the marginal and small category: 85%

Percentage of land held by households in the marginal and small category: 60%

In short nearly 85% of rural households depend upon wage labour and the 60% of land that they hold is not likely to produce substantial marketable surplus and these households enter markets as wage workers and as consumers of agricultural produce rather than as sellers of agricultural produce. This figure will be significantly higher if we only take into account the demographically and economically larger states. Not that they do not sell their produce in the market, which they have to do due to the overall commercialisation of agricultural inputs and specialised farming that they do. What it implies is that due to the small size of their holding they will not be able to take advantage of economies of scale and the quantum of produce they will be marketing therefore will be small and also produced at a relatively higher cost.

Farmers with 2 to 4 hectares (with somewhat viable farm size) account for a mere 4% of rural (8% of agricultural) households and own about 22% of the lands.

Farmers whom we can place in the middle category, who own about 4 to ten hectares account for a mere one percent of the rural households and own about 15% of lands. These then really constitute the ‘rural elite’ as the large landholders who own more than ten hectares are too few numerically, being only 0.1 percent of rural households and more importantly, own only about 4% of lands. (However, this proportion is significantly different in AP – 9%; Punjab – 11%; Maharashtra – 7%)

This 1.1 percent of rural households own about 20% of lands and it is they who produce for the market and call the shots.

This is somewhat confirmed by the data on operational holdings. As per Table 36 (p.A 3308) average size of operated area per holding is well under 0.83 ha. The average size of 43% of operational holdings is 0.22 ha. and 24% operational holdings are .68 ha and 16% are 1.2 ha. These three categories together control more than half of all land in use. 8.2% of all land is held by operational holdings above 7 ha. In other words while about 8% of farms have a competitive edge due to economy of scale, more than 50% of the farms are below 1.2 ha and hence highly uneconomical in a competitive market.

Leasing of land and landlordship

Several scholars have announced the demise of rentier landlordism and renting in of land in India. At least scholarly literature ignores the role of rent and tenancy in Indian agriculture in recent decades. However the NSO reveals a different picture. According to Table 32 (p.A 2238) 12% of agricultural land is leased in by the farmers. And a very large proportion of marginal and small farmers are reported to be leasing in land. Understandably 83% of the land used by the landless is leased in. This includes the house sites / homesteads they live in. Likewise, 76% of the land operated by marginal farmers owning less than .04 ha is leased in from others. 20% of the land operated by the farmers owning more than .04 and less than .5 ha is also leased in. On the other hand it can be seen that large landowners (ie those owning between 10 to 20 ha) rent out a substantial part of their land, upto 10%. Even those owning 4 to 7.5 ha lease out as much as 6% of their lands. While these figures may appear to be low, we should remember that much of land renting is invisible and what has been reported may be just the tip of the iceberg.

 However, it should be noted that several states show a larger percentage than the national average: thus Andhra Pradesh and Odisha have more than 30% households renting in land while Assam, Bihar, Kerala, Telangana, Tripura, UP and West Bengal are all have more than 14% households leasing in land. These are the more populous states where semi-feudal relations persist significantly. The national average of 14% would also include significant amount of land leased out by marginal households who find it uneconomical to till the tiny plots of land. Thus the category of ‘leasing in or out’ contains both large landlords and also miniscule landowners. It is therefore difficult to distinguish between capitalist renting in of land by large farmers employing wage labour and producing for market and distress renting in of land by marginal and small farmers just to be able to survive.

On the face of it such data indicate the waning of ‘semi-feudal’ landlordism; marginal, small and middle farmers who own less than four hectares of land appear to constitute about 95% of agricultural households own not less than 80% lands. However, we need to realise that formal landownership titles and statistical surveys tend to make rentier landlordism invisible. It requires a lot of in depth anthropological investigation at grass root level to unearth patterns of land control exercised by upper caste large landlords. FAS-PARI survey of two villages in Kaveri Delta, Tamil Nadu, showed that the top landlord families, which constituted 2 per cent and 1 per cent of the households owned 38 per cent and 18 per cent of the agricultural land held by all households. Another survey in Bihar showed that though their land has been subdivided over the years, landlords sit atop a hierarchy where acute wealth inequality persists. “Landlords and capitalist farmers”, accounting for a small proportion of the total households, owned and controlled bulk of assets (including land) in the villages surveyed. Despite the fact that there has been indeed a significant shift in the pattern of landownership from upper caste large landlords to small and medium OBC households, it is too early to declare rentier landlordism of the upper castes as a thing of the past.

Even as rentier landlordism is becoming invisible, the enmeshing of the peasant in multiple networks of financial dependence and bondage including caste networks, has serious implications for old style ‘land to the tiller’ kind of anti-feudal struggles. Feudal rent may not be as burdensome to the peasants as interest payments to banks etc and dependence upon input suppliers and produce purchasers.

Marketing of agricultural produce

Since the survey does not tell us about how much produce each category sells in the market we are left to speculate on how much the households having less than one or two ha of land would be selling in the market and how much foodstuff they would be purchasing from the market. If they qualify as ‘subsistence’ farmers then most of them would have little to spare to the market after fulfilling the domestic need. However, since most of them would be forced in different ways to use agricultural inputs provided by the market, and would be taking loans for this purpose, they would be also selling a portion of their produce in the market. However as mentioned above, these issues are not addressed by the data presented by the NSO. We do not have household category wise break up of marketing and only have overall averages for crops and states. Access to institutional markets is highly uneven regionally, it being concentrated in a few green revolution states like the Punjab or Haryana. From the aggregated data it is evident that most farmers sell their produce in the local market and not in formal marketing institutions like the APMC or cooperatives etc. Significant exceptions are sugarcane and cotton and soybean producers who seem to be selling substantial portion of their produce to institutions and also to some extent paddy producers. The rest including wheat producers appear to be selling in local markets more. Another finding of significance is that most of the farmers who sell in the local markets are aware that they are getting a low and unsatisfactory price.

If we were to interpret the information we have so far, it would appear that while the agricultural market is likely to be dominated by the 1.1% of all agricultural households who own 20% of all lands, a substantial percentage of marketed produce would also be from the large sea of marginal and small farmers who are work under double disadvantage. Firstly they operate highly marginal and unsustainable holdings which pushes up the amount of labour time put into producing the crops. In other words they produce at a higher cost per unit of produce. As such they will tend to sell at a higher price in the market. However, since the large producers who produce the crops more economically and at a lower cost control bulk sales in the market and can afford to push down the prices a little. This will force the mass of small producers to sell at a price lower than sustainable price thus transferring the value produced by them to the larger operators. The long and short is that the laws of the market operate in such a fashion that the small producers end up being exploited by a network of financial and trading agencies (banks, agribusiness and traders) and also the larger crop producers.

We had speculated that 87% of agricultural households who are marginal and small farmers need to engage in wage labour to sustain themselves. This is confirmed by the data on average monthly earnings of average farmers. A disaggregation of the income sources for an average rural household thus shows a very high proportion of income from wage labour as compared to income from crop production: 40% from wages and 38% from agricultural produce and 15% from livestock. While the income from crop production exceeds income from wage labour in a few states like Haryana, Panjab, Karnataka, MP etc, most other states the share of wage labour is significantly higher than the income from crops or more or less equal. In other words, most of the farmers of the country are both purchasers of foodstuffs and sellers, perhaps they are in the market more as buyers than as sellers of foodstuffs. Understandably for those owning minimal lands (like less than 0.4 ha) income from wage labour (and livestock) is of prime importance compared to income from the crop produced. As one goes up the ladder especially holdings above 2 ha, income from wage labour is less than half that of income from crops.

Income and indebtedness of agricultural households

The NSO also gives information about the sources of income and total monthly income of ‘agricultural households’ (excluding those who are termed non-agricultural). These are clubbed as hereunder.

Table 2. Average Estimated Monthly Income of Agricultural Households
from different sources 2018-19

Average estimated monthly income of agricultural households
Land wages Crops others total % of
wages
< .01 6435 1435 2113 9983 64
< 0.4 4491 657 1241 6389 70
< 1 3906 2042 1002 6950 56
< 2 3647 4313 1228 9188 40
< 4 3548 7945 1504 12997 27
< 10 4273 16914 1265 22452 19
> 10 3943 37639 8821 50403 8

NSO Statement 5.1 B

It is evident that all categories of agricultural households under two ha, depend heavily upon wage labour (40 to 70% of household income). These constitute about 87% of agricultural households who typically earn less than Rs 10000 a month. It is little wonder that a substantially large proportion of the farmers are indebted. The pattern of indebtedness is revealing.

Indebtedness is an issue flagged by the NSSO survey. About 50% of the agricultural households are indebted; average household outstanding loan being to the order of Rs. 75000/; over 57% pf the loans were taken for agricultural purposes; 70% of the loans were from formal institutions like banks. Some states like Telangana and Andhra deviate substantially from this national average. More than 90% of agricultural households of these two states and indebted, the average outstanding loan per household being to the tune of Rs 1,50,000/ to 2,50,000/.  

The small and marginal farmers owe six to ten times their monthly earnings.  Interestingly, the amount of outstanding loans increases manifold as one goes up the landownership ladder. The percentage of households indebted in each category too increases. While 38% of all household holding less than .01 ha of land is indebted to the tune of Rs. 26800/, the figure jumps to 81% and Rs. 791000/ in the case of those with above 10 ha of land.

Table 3 Percentage of indebted agricultural households and average loan amount

           


loan Rs
% of
households
% of all
indebted
< .01 26800 38 0.5
< 0.4 33200 41 28
< 1 52000 48 34
< 2 94500 57 20
< 4 175000 70 12
< 10 326700 79 4
> 10 791000 81 0.6
All 74121 50 100

NSO, Statement 5.7

While almost all categories of farmers raise loans for both capital investment and also circulation capital, the marginal and smaller farmers raise a significant proportion of loans for consumption purposes. (Statement 5.9). Similarly it is seen that while almost all categories of farmers rely to a large extent on formal financial institutions like Banks it is the marginal and small farmers who resort to loans from private money lenders, presumably for consumption and emergency loans. (NSO Statement 5.8) Banks and financial institutions have emerged as important creditors to the farming community and represent the penetration of Banking and Finance capital into the agrarian sector.

Given such a profile, the conclusion that Indian peasant agriculture is increasingly becoming unsustainable appears unavoidable. Miniscule farms worked by domestic labour forced to compete in a volatile ‘open to the world’ market and hence condemned to use machines (hired tractors), commercial seeds and chemical farm inputs. Indebtedness is the consequence and a further entanglement with finance capital, forcing the peasant into an increased self-exploitation.

Households in the comfortable zone of more than two ha of land merely constitute about 5% of rural households. However, together they control about 41% of land. These then are the rural elite who also control the supply in the agricultural market with their marketable surplus. Understandably they would also be employing the landless and the marginal farmers as wage workers.

Caste Dimension

There is a caste dimension to the rural agrarian society which this survey brings out to some extent but also obscures tantalising details. We will go over some of the important aspects that come to light from the data.

To begin with we note that ST households which constitute about 12% rural households hold about 14% of the lands; SC households constitute about 22% households hold only 10% of the lands; the OBCs who constitute 44% of the households hold about 47% lands and finally the other castes who constitute about 22% of the households hold 29% lands. Quite simply, the SCs have far smaller share of the lands compared to other social groups. Secondly we note that when it comes to landlessness, every community has an almost equal share. Only in the case of the ‘Other’ castes, it is likely that the landlessness is mitigated by capital availability for business etc. However, as we proceed up the scale of landholding we find a marked divergence. Among marginal households (those owning one ha or less we see that about 85% of SC households are in the marginal category, as against others (ST – 71%; OBC – 76; Others – 72%). At the same time, it should be noted that the average land possessed by these households is broadly of the same order, less than 0.3 ha per household. These two issues form the basis for a possible class solidarity cutting across caste lines in the categories of both the landless workers and marginal farmers.

As we go up the landholding scale we see that the proportion of SCs holding land diminishes significantly in comparison with other communities, especially the OBC and Others. While only 2% of SC households have land between 2-4 ha, the figure for OBCs is 5% and 7%. In the next category of 4-10 ha we see a further decline in the share of SCs – it is now 0.3%, while for the OBCs and Others it is as much as 2% of their respective populations. A difference of six times. The presence of SCs in the large landholdings (above 10 ha) is negligible, about 0.001% compared to 0.2% of the OBCs and others. A difference of 245 times. The presence of a disproportionately large number of SCs in the labour market, with minimal cultural and social capital makes them a bank of labour available at a less than market price of skilled and dignified labour. This is the harsh reality despite all the narratives fostered by the upper caste employer farmers.

The very high prevalence of wage labour among the STs also indicates that despite apparently owning comfortable amounts of land, they depend upon wage labour to make ends meet. In other words the lands yield too little for the families to subsist upon, given the very low investment done by the state in developing tribal agriculture. These are marginal lands with very low productivity. The following table gives some idea of the community composition of labouring households. We can see that 44% of ST and 58% of SC households engage in wage labour as against 38% OBC and 33% Other households. (There is a slight statistical anomaly in the case of others owning more than 10 Ha, where a large number of estimated persons are shown as engaged in casual wage labour, which distorts the overall picture somewhat.)

Table 4. Percentage of rural households engaged in other than ‘Self-employment’

Wage labour etc


ST SC OBC Oth All
<  0.01 66 77 64 61 68
< 0.4 61 66 52 47 55
< 1 33 24 17 16 21
< 2 20 15 10 11 12
< 4 8 15 8 8 9
< 10 7 14 7 6 7
> 10 0 0 3 20 11
% of
households
44% 58% 38% 33% 42%

Based on NSO, Situation Assessment ..2021 Table 5, A2 224-230

Two points emerge strongly from this table. Firstly more than half of all marginal farmers (those owning less than one ha of land) engage in wage labour. This corelates well with the table on the share of wage earnings in the overall earnings of rural households given above. Secondly, we see a far larger proportion of STs and SCs in each category (except the above ten ha which has no ST or SC household) doing wage labour. Thus even while owning a similar amount of land, it is the SCs and the STs who are forced to engage in wage labour to supplement family income. The Others do it the least. Two factors can explain this situation. Firstly the fact that the quality of land that the SCs and STs possess, may be of much lower quality than the land possessed by the OBCs and the Others. Secondly it may be due to much lower social and cultural resources available with the SCs and STs, which mean that they can only survive by selling their labour power. Either way this points to an institutionalised creation of subordinate and low paid labour supply for upper castes.

Some conclusions

i. Given the miniscule farm sizes, it is obvious that the farmers cannot take advantage of economies of scale and even though forced to use farm machines, do so in an uneconomical manner.  ‘Scale neutrality’ which was the promise of the green revolution technological package, is rapidly turning out to be an illusion.

ii. Indian farmers who could have been self-sufficient peasant households having been forced to participate in world capitalist market, are increasingly faced with the reality of their lower productivity and ‘irrational’ use of resources. This is not to say that peasant farms are in themselves low in productivity or irrational. They were highly productive and rational in a non-capitalist market context. But having to compete with the same products produced through economies of scale and heavy state subsidies, Indian farmers are increasingly facing a crisis.

iii. The reality is that vast segments of Indian farming population are simultaneously engaged in wage work to supplement their income and the expanding rural non-farm sector is providing much of the additional employment. At the same time the technologies employed and the wage levels are too low to offer a stable alternative to farming. We thus find small farmers resorting to wage work and on the other hand completely landless households desperately renting in small parcels of land to sustain themselves.

iv. Social movements have so far only focussed on rural population as a farming population and little attention has been paid to wage work. Of course there are important exceptions in some states like Rajasthan where rural workers have been organised.

v. Vast segments of farmers thus are in the market simultaneously as both sellers and buyers of agricultural produce. If one divides their income into three parts, agricultural produce retained for domestic use, agricultural produce sold in the market and wage income, it is possible that the wage income is much more than the income from sale of produce for most marginal and small farmers who constitute the vast majority of rural households. As such their identity as workers may be stronger than their identity as petty producers despite the tave of being ‘farming households’.

vi. This has serious implications for the market protection for the farming sector with mechanisms like ‘minimum support price’ and state procurement of agricultural produce when it goes beneath the minimum support price. Artificially propping up agricultural prices may not actually be in the interests of most of the wage dependent households. At the same time depressing of agricultural prices or even their volatility may deprive the wage earners who are also farmers, of the viability of their miniscule farms. In other words these marginal and small farmers are in a kind of no win situation when it comes to agricultural marketing.

vii. While larger farmers are more heavily indebted to financial institutions, the land less, marginal and small farmers are indebted to private money lenders and financial institutions. This indicates that the larger farmers are using the financial institutions for capital investments and working capital, while the smaller ones are forced to use institutional loans for working capital and private loans for consumption purposes. This enmeshing of the entire farming sector in heavy indebtedness for both consumption and production requirements indicates the spreading of the tentacles of finance capital in the agrarian sector. We need to see in this light the various ‘insurance’ programmes being promoted by the government, like crop insurance and also health insurance.

viii. We had pointed out the caste dimension of agrarian distress, and the particularly vulnerable position of SCs and STs who have the most marginal lands and least cultural and educational resources putting them at a disadvantage both in agriculture and the labour market.

ix, Finally it should be remembered that there is a gender dimension to Indian agriculture which the NSO report does not engage with. The state of feminisation of the agrarian sector and the changing role of women in agriculture and wage labour. This needs to be looked into more carefully.

x. The above discussion is inconclusive and we need to delve more deeply into regional variations, caste dimensions and other aspects to get a clearer picture of the agrarian situation.

Click here to return to the April 2025 index.