SPECIAL ARTICLEEconomic & Political Weekly EPW april 11, 2009 vol xliv no 1549Peasant Classes under Neoliberalism: A Class Analysis of Two StatesArindam BanerjeeWhile declining real product prices faced by primary commodity producers was one of the central causes of rising farm indebtedness, the gradual shrinkage of formal credit institutions in rural areas has simultaneously caused increasing dominance of private players in the credit market, rendering producers all the more vulnerable. A class analysis of household-level farm production data from two states reveals the pattern of income depression and rising indebtedness in rural areas. The deleterious implications these processes have for future agrarian development call for effective price stabilisation operations and a comprehensive debt-relief policy in the short run.I thank Utsa Patnaik for her insightful comments. I also acknowledgethe Foundation for Agrarian Studies for their financial supportduring this work. The usual disclaimers apply.Arindam Banerjee (arindam@cds.ac.in) is with the Centre for Development Studies, Thiruvananthapuram.The trade-liberalised era since 1991 has witnessed the emer-gence of indebtedness as a grave problem for the Indian peasantry. In less than a decade since the introduction of a neoliberal economic regime, incidents of farmers’ suicide were reported from different corners of the country. This tragic and unprecedented phenomenon caused by increasing debt-driven vulnerability of peasant households started with the cotton farmers of Andhra Pradesh and gradually afflicted farmers – primarily growers of various commercial crops in other parts of the country. The initial confusion that prevailed within the circles of the ruling establishment regarding the causality of the farmer sui-cides was more due to a “denial” syndrome than any serious investigation of the issue. There was an outright non-acceptance of the fact that the trade policies of the government have been instrumental behind the catastrophe witnessed in rural India. The overriding tendency was to attribute the suicides to social problems such as family disputes or alcoholism. The attitude towards the issue has gradually changed not only due to the sheer magnitude of the disaster and the political and social outrage that it generated but also due to sustained enquiries and report-ing of the same. The announcement of the Agricultural Debt Waiver and Debt Relief Scheme, 2008 (henceforthADWDRS) by the government assumes significance in this regard. The much-awaited debt-waiver scheme has arrived much later than required, and regrettably after more than 1,60,000 farmers have ended their lives over the last decade; it is nevertheless a welcome measure.1While much has been written on the issue of farmer indebted-ness, it is worthwhile to revisit the phenomenon from the per-spective of peasant classes. A deeper look at the structure of ris-ing indebtedness in rural areas is necessary for the purpose of assessing the impact it has for the agrarian question in the his-torical time frame. The tightening constraint of debt burdens on the surplus generated in agricultural production and its conse-quent impact for agrarian development needs to be assessed. This exercise is all the more indispensable once we recognise the incidents of farmer suicides as only the extreme manifestation of debt-driven vulnerability. The agricultural production in regions, where farmer suicides are not occurring in multitudes, is not nec-essarily unconstrained by debt burdens. The variance in the eco-nomic and socio-political structures across regions is more the reason why the symptoms of spiralling indebtedness have been divergent in their extent across the country. In this context, this paper attempts to appraise the impact of transforming credit relations on agricultural production. Addi-tionally, an analysis of the structure of rural indebtedness across peasant classes and size-class groups reveals the significance of a 
1991-92 1993-94 1995-96 1997-98 1999-00 2001-02 2003-04 2005-06 1991-92 1993-94 1995-96 1997-98 1999-00 2001-02 2003-04 2005-06 
1991-92 1993-94 1995-96 1997-98 1999-00 2001-02 2003-04 2005-06
1993-94 1995-96 1997-98 1999-00 2001-02 2003-04 2005-06
1991-92 1993-94 1995-96 1997-98 1999-00 2001-02 2003-04 2005-06
SPECIAL ARTICLEEconomic & Political Weekly EPW april 11, 2009 vol xliv no 1551pulses whose real prices rapidly declined post 2001-02 to the early 1990s levels. The trend for maize has been more volatile, but importantly, the real value of maize prices has consistently remained below the 1991-92 level throughout the period. The trends for the real prices of the more commercially culti-vated crops are significantly different. The real producer prices for oilseeds secularly declined in the 1990s to low levels. The sub-sequent rise in oilseeds prices after 2000-01 was more due to inadequate supply and poor quality production, owing to the drought conditions in 2001-02 and 2002-03; this implies that no real benefits were accrued by producers due to this increase. Raw cotton, which has been at the centre of the agrarian crisis in the country, exhibits an unambiguous declining trend in real prices ever since the markets were liberalised in the mid-1990s. From a high in 1994-95, the real prices have fast dwindled and even slipped under the low value that existed in 1992-93.These price trends for the major crops explicitly reveal a more systemic income deflationary process under trade liberalisation for commercial crops rather than mere intermittent shocks. For major food crops, there is a clear erosion of real value of prices from around the year 2000. In the event of rising cultivation costs in agriculture unlike what we have assumed, the decline in the pur-chasing power of the producers is even greater than what we observe from the graphs. A basic implication of these trends is the necessity to stabilise both food and non-food prices with increased interventions in both agricultural commodities’ markets in the form of enhanced procurement and distribution operations so that it improves the net purchasing power of the primary sector producers.Financial liberalisation is the other facet of the current eco-nomic regime responsible for mounting rural household debts. The financial reforms adopted under the aegis of the Narasimham Committee (1991) ensured a gradual shrinkage of the institu-tional banking sector in the rural areas. The committee’s efforts to redefine and gradually phase out priority sector lending norms led to an immediate decline in priority sector lending by banks in the early 1990s. A redefinition of the norms to include infrastruc-ture funds, special bonds of state financial corporations and food processing companies including multinational corporations (MNCs) under the priority sector meant that lending to agriculture con-tinued to decline in the late 1990s and later even when priority sector lending increased. Enough literature discussing these issues and their adverse impact on agriculture is already available.2However, a cursory look at some of the banking indicators for scheduled commercial banks (SCBs) shows that more than 3,000 rural bank branches were branded “inefficient” and eventually closed down between 1991 and 2005 (Table 1). The percentage of rural branches to total branches started decreasing post-1991; by 2005, the figure was much lower than that in 1981. Similarly, the percentage of credit advanced in rural areas to total credit increased from 11.9 to 14.7 between 1981 and 1991 and dipped sharply to 9.5% by 2005. While the credit-deposit ratio remained more or less unchanged in the 1980s, it sizeably declined from 59.4% in 1991 to 39% in 2001. After 2001, this ratio shows a steep increase but this offers no solace as it is more due to a slower growth of rural depos-its (in nominal terms) than rural credit for SCBs; a joint outcome of aggressive financial reforms and the precipitating agrarian crisis. A systematic erosion of institutional banking services in rural areas and unfavourable returns from cultivation have worked in tandem to enhance the vulnerability of rural households to local moneylenders manifold in several parts of the country. Keeping these two broad macroeconomic processes in mind, we will ana-lyse the production outcomes of agricultural activity using our primary field enquiry data.Income Depression and Indebtedness: A Class AnalysisThe field enquiry was conducted between February and August in 2006 collecting household-level disaggregated crop-wise and farm operation-wise data on labour-use and input costs, output volumes and prices, outstanding loan details, owned and opera-tional landholdings and asset ownership. Data was also collected on paid-employment, both farm and non-farm, regular and casual wages and income generated from self-employed activities. This enabled us to estimate the farm incomes as well as total household incomes taking into account the various non-farm income sources.The survey covered three diverse agro-climatic regions with varying economic and social structures allowing us to assess the situations of rural indebtedness under dissimilar production con-ditions in agriculture. The two villages surveyed in West Bengal, situated in the Raina-II (WB-rn) block in Bardhaman district, lies in the fertile Gangetic plains. Owing to abundance of groundwater in the region, the extent of irrigation is high and facilitates multiple cropping seasons in a year. Paddy (amanand to some extent boro) and potato are the main crops cultivated. The two regions surveyed in Andhra Pradesh are Saidapur mandal (AP-sdp) in the Karimnagar district and Anantapur (rural) mandal (AP-atp) in Anantapur district. The three villages surveyed in the Saidapur mandal consist of cotton-growing households and around half of their operated area is irrigated. Compared to the West Bengal region, the extent of irrigation is lower in this region which mainly has one cultiva-tion season. On the other hand, the Anantapur region is agricul-turally the most backward among the three regions, has negligible irrigation and is characterised by dryland groundnut cultivation. Henceforth, we will refer to the three regions asWB-rn,AP-sdp andAP-atp for convenience.A stratified random sampling methodology was used to draw samples for the survey. The stratification is done on the basis of operated area in order to prevent an exclusion of households with large landholdings, which is normally a thin stratum in the popu-lation. InWB-rn, the two strata are defined by households below and above five acres of operated area. InAP-sdp andAP-atp, the Table 1: Trends in Number of Branches, Credits and Deposits of SCBs in Rural IndiaYear No of Bank Branches Credit Advanced Deposits Credit-Deposit Ratio (%) Rural % to Rural % to Rural % to Rural All Areas (Number) Total (in Rs Crore) Total (in Rs Crore) Total1981 19,45351.23,60011.95,93913.460.668.11991 35,21658.119,68814.733,16315.159.460.92000 32,67348.7 48,753 10.61,20,539 14.7 40.056.02001 32,64048.354,43110.11,39,43114.739.056.72002 32,44347.866,68210.21,59,42314.241.858.42003 32,28347.477,15310.21,76,50213.843.759.22004 32,10746.885,0219.71,95,08212.943.658.22005 31,967 45.7 1,09,976 9.52,13,10412.251.666.0Sources: Ramachandran and Swaminathan (2002) for the figures pertaining to 1981, 1991 and 2000 andBanking Statistics: Basic Statistical Returns for 2001-2005.
SPECIAL ARTICLEapril 11, 2009 vol xliv no 15 EPW Economic & Political Weekly52same is determined by households below and above 12 acres. The precise reason for using different criteria for stratification across the states is due to the significant difference in the average size of landholding. The strata were arrived at after analysing the house-listing data and demarcating the operational holding size above which the top 10% households were placed. The selected sample size inWB-rn is 77 households while that inAP-sdp andAP-atp is 60 households in each region.We have used Patnaik’s Labour Exploitation Index (E-Crite-rion) to classify the households into six economic classes. The calculation of the Labour Exploitation Index (E) for each house-hold (see equation 1) has been done considering both direct labour exploitation through hiring-in and hiring-out as well as indirect exploitation of labour through leasing-out and leasing-in of land using the following equation. E = X/F = [(Hi – Ho) + (Lo – Li)]/F ...(1)where Hi, Ho are labour days hired-in and hired-out respectively, Lo, Li are the total labour days on land leased-out and land leased-in, respectively, and F is the family labour on the operated farm. E represents the exploitation of others’ labour relative to the exploi-tation of one’s own labour. The economic classes into which the households have been classified are rural labour (RL), poor peas-ant (PP), small peasant (SP), middle peasant (MP), rich peasant (RP) and landlords (LLD), each class determined by the range in which the value of E lies (Table 2). The poor and small peasants are the labour hiring-out classes, constituting the lower peasantry, while the middle and rich peasant are the labour exploiting classes and belong to the upper strata of the peasantry. At the two ends of the classification, we have the rural labour and the landlord class. Both these classes do not toil on their fields and have a nil F. The rural labour hires out large quantum of labour in the labour market due to lack of any operational holdings while the landlords employ large amount of hired labour for cultivation on his operated area or leases out land to tenants in exchange of ground rent. In the his-torical time frame of a capitalist transition of agriculture, it is the upper peasantry classes along with the proto-capitalist landlords3 which appropriate surplus labour directly through an “unequal exchange” in the labour market or indirectly through other exploit-ative relations and lead the process of capital accumulation.The households have also been alternatively grouped into six land size-classes starting with zero or nil operated area up to the size-class operating more than 10 acres. The classification according to operated area size-class is based on the primary data collected on Owned Area, Area Leased-in and Area Leased-out. The Operated Area is the sum of Owned Area and the Net Area Leased-in. Some basic characteristics of the households in the three regions are revealed by their distribution across economic and size-classes (Table 3). InWB-rn, a significant feature of the sample is the small size of landholdings. A large 77.9% of the households operate less than 2.5 acres and another 11.7% of the households lie in the 2.51-5.0 acres bracket. While landhold-ings have been historically smaller in West Bengal, a persistent land reforms pro-gramme has also ensured that large landholdings above the ceiling are not very common. The other strik-ing feature of this region is that the rich peasant class forms the largest economic class, close to around 39%, much higher than the other two regions. This is due to the high demand for hired-labour in this region that originates from multiple cropping seasons in a year. Significant migration of labour from other parts of the state is required to ensure that this high demand suffices. A sig-nificant point to note is that the rich peasant households are dis-tributed across all size-classes, which means that the high labour demand is generated from both small and large landholdings.InAP-sdp, the distribution of the households across land size-classes is more spread out. The percentage of households operat-ing five acres (two hectares) and above in this region is a signifi-cant 25%. The small peasant class is the largest class in the region with 41.67% of the households. Unlike WB-rn, the rich peasant class is not only thinner (13.33%) but is also confined to the two highest size-class brackets (five acres and above). The small and middle peasant classes jointly comprise more than 60% of the households indicating that the use of family labour is at least as important as hired labour for a majority of households.Finally in AP-atp, we find a large size of net labour-selling lower peasantry classes (poor peasant and small peasant classes together included more than half of the households) typical of an agricultur-ally backward region. Cross-classifying the households between economic and size-classes reveals that as much as 28% and 60% of the poor peasant and small peasant class, respectively, are operating more than five acres. In contrast, the rich peasants and the land-lord households are concentrated only in the higher size-classes. The average landholdings are also larger in this dry-land region and around 45% of the households operate more than five acres.From the disaggregated cost of cultivation data and output quantity and prices collected through the primary field enquiry, we have estimated the average farm labour income (FLI) for the households in each class. TheFLI is determined according to the following relation:Farm labour income (FLI) = Gross value of output (GVO) – [Total material input costs + Paid-out labour costs] ...(2)4Total household income′(THI′) = [FLI – rent payments] + Income from paid-employment + Income from self-employment ...(3)5Table 2: Details of Economic ClassificationEconomic Class Value of E CommentsRural labour E = -∞ X < 0 and |X| is very large, F = 0.Poor peasant E ≤ - 1 X < 0, F > 0, |X| is large and |X| ≥ F.Small peasant 0 ≥ E > - 1 X ≤ 0, F > 0 and |X| < F.Middle peasant 1 > E > 0 X > 0, F > 0 and X < F.Rich peasant E≥ 1 X > 0, F > 0, X is large and X ≥ F.Landlord E = ∞ X > 0 and X is very large, F = 0.Source: Adopted from Patnaik (1976).Table 3: Percentage Distribution of Households across Classes in the Three RegionsClass WB-rnAP-sdpAP-atpRL 2.611.73.3PP 32.511.7 41.7SP 5.2 41.7 16.7MP 20.820.020.0RP 39.013.315.0LLD 0.0 1.7 3.3Total 100.0100.0100.00 2.6 10.00.00.01-1.0 53.218.31.71.01-2.5 24.7 16.7 21.72.51-5.0 11.730.031.75.1-10.0 3.916.721.710.1 and above 3.9 8.3 23.3Total 100.0100.0100.0Source: Primary Field Enquiry.
SPECIAL ARTICLEEconomic & Political Weekly EPW april 11, 2009 vol xliv no 1553Total household income (THI) = [FLI – rent payments – Interest payments (OIP)] + Income from paid-employment + Income from self-employment ...(4)The FLI represents the income of the farm households from cultivation in an economic situation where there is a complete absence of monopoly over land and credit, which renders both the ground rent and interest payment as zero. For each economic and size-class, we have an average household FLI. In order to assess the impact of credit monopoly on the production process under the real situation, we look at the share of the FLI that is required to clear all outstanding interest payments (OIP) for the average household in each class across the regions (Tables 4.1 to 4.3). The OIP is the accumulated interest from all unpaid loans including those taken in previous production cycle(s). The pay-ment of ground rent due to land monopoly also critically influ-enced the income situation of some lower peasantry classes in the Andhra Pradesh regions but we shall not go into the details of that within the scope of this paper. We have also estimated the total household income, prior to making interest payments (THI′), and post interest payments (THI) for the average household in each class (see equations 3 and 4).From Tables 4.1 to 4.3, we can observe the diverse production situations across the regions. While the share of FLI required to clear all OIP is around 35% inWB-rn, the same figure is a nearly 74% in the cotton-growing AP-sdp. In contrast, the situation is prohibitive inAP-atp, where the OIP exceeds theFLI, for the region as a whole, by more than 60% of the FLI. The constraint of accu-mulated indebtedness of the households on their agricultural production is of a varying degree across regions and more strin-gent or even prohibitive for the commercial crop growers in the Andhra Pradesh regions. One can argue that a portion of the credit taken is often diverted for non-agricultural purposes and it is not necessary that the entire interest payments have to be made from theFLI alone. However, taking into account the income from other sources apart from cultivation, the situation still remains tight with the OIP constituting a high 61.53% and 75.5% of THI′inAP-sdp andAP-atp, respectively.Assessing this constraint from a class perspective, we find that both the lower peasantry classes in WB-rn generate an FLI which is short of meeting interest payments (Table 4.1). However, the poor peasants in this region are closer to the rural proletariat in their character with labour-selling as their primary income source. Con-sequently, the OIP for this class forms a small 8% of their THI′. On the other hand, the small peasant class is beleaguered with a nega-tive FLI as their average GVO does not cover the material and labour input costs incurred in cultivation. This is primarily due to unusu-ally low yields of both paddy and potato for this class. This some-what indicates that the returns in cultivation for this class are even lower than the casual daily wage. The labour-exploiting upper peasantry classes are comparatively in a much more comfortable situation as far as their pending interest obligations are concerned.InAP-sdp (Table 4.2), the small peasants and the middle peas-ants (the latter not even by using their non-farm incomes) are unable to cover their OIP from theFLI. The share of the FLI depleted by the OIP is high not only for the poor peasant class but disturbingly so even for the rich peasant class (71.5%). Unlike WB-rn, a tight constraint on production emerging out of credit relations adversely affects the upper peasantry also. In fact, the rich peasant class are left with a disposable income post-interest payments that barely allows them to meet minimum consumption Table 4.1: Income and Interest Payments Situation across Classes: WB-rnClass FLI Outstanding OIP as % THI’ OIP as % THI (in Rs) Interest of FLI (in Rs) of THI’ (in Rs) Payments (OIP) (in Rs) RL 400400100.011,8003.3911,400PP 1,2531,264OIP>FLI15,6508.0814,385SP -4,1468,669OIP>FLI5,904OIP>THI’-2,765MP 12,9903,01423.212,51024.09 9,496RP 28,9648,98031.046,51019.3137,531LLD 0 0 0.0 0 0.0 0Total 14,1864,996 35.226,41518.91 21,4190 400400100.011,8003.3911,4000.01-1.0 3,787 4,320 OIP>FLI 13,08133.03 8,7601.01-2.5 15,7594,379 27.817,31125.30 12,9322.51-5.0 23,1516,89929.823,62729.2016,7275.1-10.0 55,0563,027 5.5 61,7234.9058,69510.1 and above 87,765 17,456 19.9 2,49,098 7.01 2,31,643Total 14,1864,996 35.226,41518.91 21,419Source: Same as of Table 3.Table 4.2: Income and Interest Payments Situation across Classes: AP-sdpClass FLI Outstanding OIP as % THI’ OIP as % THI (in Rs) Interest of FLI (in Rs) of THI’ (in Rs) Payments (OIP) (in Rs) RL -65 7,336 OIP>FLI 12,428 59.03 5,092PP 8,3305,44965.48,58863.453,139SP 2,776 4,354 OIP>FLI 7,657 56.86 3,303MP 10,41112,764OIP>FLI12,277OIP>THI’-487RP 1,06,96876,43171.51,07,81870.8931,387LLD 2,46,49738,00015.42,78,49713.642,40,497Total 22,57416,68373.927,11561.5310,4320 -76 8,333OIP>FLI 12,32467.62 3,9910.01-1.0 -1,3405,496OIP>FLI6,39385.988961.01-2.5 -2,1587,239OIP>FLI3,482OIP>THI’-3,7572.51-5.0 7,6984,22554.911,13437.956,9095.1-10.0 14,397 12,99590.39,991 OIP>THI’-3,00410.1 and above 2,21,736 1,22,420 55.2 2,29,496 53.34 1,07,076Total 22,57416,68373.927,11561.5310,432Source: Same as of Table 3.Table 4.3: Income and Interest Payments Situation across Classes: AP-atpClass FLI Outstanding OIP as % THI’ OIP as % THI (in Rs) Interest of FLI (in Rs) of THI’ (in Rs) Payments (OIP) (in Rs) RL -2,9003,600 OIP>FLI 22,45016.04 18,850PP 4,041 15,518OIP>FLI 20,449 75.89 4,931SP 8,122 3,513 43.310,74132.71 7,228MP 2,41112,608OIP>FLI10,569OIP>THI’-2,040RP 32,27827,93486.537,16775.169,233LLD 28,69735,100OIP>FLI35,69798.33597Total 9,22115,053OIP>FLI19,93875.504,8840 0 0 0.0 0 0.0000.01-1.0 3901,54,080OIP>FLI14,390OIP>THI’-1,39,6901.01-2.5 2,87113,047OIP>FLI16,79477.693,7462.51-5.0 13,8974,90435.326,01318.8521,1085.1-10.0 19,2928,538 44.328,71829.73 20,18110.1 and above 52 26,809 OIP>FLI 6,855 OIP>THI’ -19,954Total 9,221 15,053 OIP>FLI 19,93875.50 4,884Source: Same as of Table 3.
SPECIAL ARTICLEapril 11, 2009 vol xliv no 15 EPW Economic & Political Weekly54norms, leave alone their role in capital accumulation through new investments. The sole landlord household seems to be the only household that is at ease in the credit market. An investiga-tion into the credit market structure will throw more light as to why the labour-exploiting classes are critically vulnerable with regard to credit exchange relations.We come across a worse scenario in agricultural production in the unirrigated dryland region of AP-atp. The interest obligations for the poor and middle peasants are stunningly high at four to five times their FLI pointing towards their depressed returns in culti-vation and consequent distress. In this region too, a prohibitively high portion (86.5%) of the FLI is required to clear the OIP for the rich peasant class. The fact that the proto-capitalist landlords also have theirFLI exceeded by theOIP indicates that the debt-driven infirmity of agricultural production is more intensive and wide-spread in this region. The field investigations revealed the exist-ence of bonded labour, which is used as a mode of payment in the credit market for past indebtedness. In fact, the astronomicalOIP figure observed for the sole household in the 0.01-1.00 acre bracket is only the money valuation of the outstanding interest while in reality the unpaid debts are being settled by regular pay-ments in the form of dedicated labour-hours to the creditor.There are two distinct impacts that can be traced from the point of view of the larger agrarian question. The majority of the classes undertaking cultivation, including some from the upper peasantry, are unable to carry out even a simple reproduction of the peasant economy. The THI of these classes are abysmally low and clearly not adequate to meet the minimum consumption requirements of the households. A simultaneous default in interest payments and depression of own-consumption to sub-standard nutrition levels are really the strategies of survival (or destruc-tion) for these classes. The intensity of this income depression is illustrated by the spread of this infirmity amongst the labour- exploiting upper peasantry classes in some cases; the middle peasants in both regions in Andhra Pradesh actually end up with annual losses. This non-viability of a simple reproduction of the peasant economy is reflected in the mass farmer suicides that we have witnessed in the country for a long time now.On the other hand, the rare classes that retain some notable surplus from agricultural production are the rich peasant classes inWB-rn and AP-sdp and the sole landlord household inAP-sdp. These classes appropriate significant volumes of surplus labour through the use of modern and capitalist modes of production. However, the income that they generate from agriculture and allied activities does not allow them to undertake large investments and upgrade cultivation techniques without cutting down their consumption requirements to a certain extent. The following simple exercise illustrates this fact.We calculate the monthly per capita income (MPCI),6 from the annual THI, for the average households in the rich peasant classes in WB-rn and AP-sdp and compare them with the monthly per capita expenditure (MPCE) required to attain certain minimum consump-tion standards. The average family size of households in these classes is 7.03 and 4.75 in WB-rn and AP-sdp, respectively (primary field enquiry). Using these two figures, the MPCI for the average household in WB-rn comes to Rs 445 and for AP-sdp, it is Rs 551. On the other hand, the MPCE required for attaining a minimum per capita per diem calorie intake of 2,200 kcal, derived using NSS data on nutritional intake, is Rs 573 and Rs 704 for West Bengal and Andhra Pradesh, respectively.7 The average MPCI of the rich peasants clearly falls short of the MPCE required for minimum consumption norms in both regions, which implies that new investments in production techniques are feasible for these house-holds not only by depress-ing their consumption below the required levels but also by defaulting to some extent in their inter-est or rent payments, mainly the former.8 This largely explains the stag-nation in investment and growth in the agricultural sector that has now become a persisting feature in the post-liberalisation period.The adverse returns in agriculture and a retreat of the formal credit institu-tions from rural areas in the post-liberalisation period, have not only pushed a large section of the peasantry to abysmally low standards of living but has also depleted the capa-city oflabour-exploiting dominant peasant classes to further the process ofcapi-talist development. While intra-peasantry relations of exploitation continue to operate through unequal exchange in the labour or land-lease markets, these two macroeconomic pro-cesses have nearly stalled the capitalist development ofagriculture as a whole. It is precisely due to this that a debt-relief programme is of immense necessity and holds high significance for further agrarian develop-ment. A look into the char-acteristics and conditions of indebtedness across the regions and classes further elaborates this situation.Table 5.2: Household Debt Situation across Classes: AP-sdpClass OutstandingDebt-AssetsOIP-GVO DebtperRatioRatio Household RL 18,4791.060.000PP 22,0200.610.203SP 26,4340.530.141MP 69,0140.700.211RP 1,85,1810.810.323LLD 3,88,0000.430.064Total 60,6990.660.2400 20,8333.160.0000.01-1.0 14,7690.380.8531.01-2.5 30,0390.500.4602.51-5.0 38,8920.710.1005.1-10.0 77,1950.600.12610.1 and above 3,16,420 0.72 0.285Total 60,6990.660.240Source: Same as of Table 3.Table 5.3: Household Debt Situation across Classes: AP-atpClass OutstandingDebt-AssetsOIP-GVO DebtperRatioRatio Household RL 6,6000.140.000PP 50,7380.870.575SP 45,663 0.320.071MP 69,3450.330.183RP 1,19,1230.420.217LLD 2,05,1000.160.086Total 67,5460.380.2270 00.000.0000.01-1.0 1,98,080198.0848.4531.01-2.5 38,2400.610.9032.51-5.0 34,6250.420.1135.1-10.0 69,7300.250.07910.1 and above 1,28,083 0.38 0.242Total 67,5460.380.227Source: Same as of Table 3.Table 5.1: Household Debt Situation across Classes: WB-rnClass OutstandingDebt-AssetsOIP-GVO DebtperRatioRatio Household RL 8,4000.700.103PP 5,4840.150.105SP 44,669 1.250.297MP 18,5640.300.064RP 58,2550.210.067LLD 00.000.000Total 30,8730.220.0740 8,4000.700.1030.01-1.0 13,5470.340.2131.01-2.5 26,2740.340.0632.51-5.0 49,8990.340.0545.1-10.0 42,6940.090.01410.1 and above 2,42,872 0.16 0.041Total 30,8730.220.074Source: Same as of Table 3.
SPECIAL ARTICLEEconomic & Political Weekly EPW april 11, 2009 vol xliv no 1555The debt situation of the households in the three regions is pre-sented in the Tables 5.1 to 5.3 (p 54). The overall outstanding debt perhousehold (sum of the outstanding principle and outstand-ing interest) is roughly double in the Andhra Pradesh regions compared toWB-rn.Aninter-regionalcomparison of the debt-as-sets ratio9 and the OIP-GVO ratio gives a stronger illustration of the disparity in the debt situation across regions. The average debt-assets ratio and the OIP-GVO ratio is 0.22 and 0.074 inWB-rn, muchlower than the ratios inAP-sdp andAP-atp. This emphati-cally points out the high volumes of debt-burdens that have come to exist in the commercial crop-growing regions under the neoliberal policy regime.Inspecting the situation through the lens of the different peas-ant classes, we find that the small peasant and the rich peasant classes inWB-rn have a higher than average outstanding debt (Table 5.1). For the rich peasant class, the debt-assets ratio and OIP-GVO ratio are less than the regional average indicating that the class is carrying a debt-burden that is commensurate with its assets and production. However, the small peasant class bearsan unusually high debt-assets ratio andOIP-GVO ratio. This thin class is definitely ailing from high levels of accumulated debts. Amongthe size-classes, the marginal farmers (0.01 to 1.0 acres), the majorityin the region, also have a high interest-GVOratio of 0.213. Although the debt indicators donotrevealthepresence of any acute indebted-ness problem at the regional level, there are definitely traces of debt-driven vulnerability among the marginal cultivators and the small peasants.We have already observed earlier that in AP-sdp, the upper peas-antry classes are also facing tight constraints in the domain of credit exchange relations. Both the debt-assets ratio and OIP-GVO ratio figures for the different classes in this region authenticate this obser-vation (Table 5.2). The debt-assets ratio was not less than 0.5 for any of the four peasant classes. Similarly in AP-atp, a marked distinc-tion between the lower and upper peasantry classes is absent as far as their debt-assets and OIP-GVO ratios are concerned (Table 5.3).Interestingly, unlike WB-rn, the susceptibility due to huge debt burdens in both these regions is not restricted to the lower size-classes (below two hectare or five acres). The average household in the two top most size-classes in AP-sdp have quite high debt-assets ratios.Also, the high interest-GVO ratio of the cultivators with oper-ated area ofmore than 10 acres in AP-atp at 0.242 is not surprising given that around 21% of these households actually belonged to lower peasantry classes. Glancing back at the income situations (Tables 4.2 and 4.3), we can observe that the OIP depleted a high share of or exceeded the FLI/THI’ for the two top most size-classes in these regions. From this point of view, the unilateral differentia-tion that the ADWDRS adopts between farmers cultivating less than and more than two hectares is irrational. Typically in dryland areas with more commercial cultivation, the landholding size is larger. Moreover, as the current crisis in agriculture is emanating from high output market risks, households with higher operated area and larger expenditure are not necessarily always better-off com-pared to the smaller cultivators. In such a situation, the provision of a “one time settlement” in the ADWDRS for farmers operating more than two hectares may not be enough to salvage their sinking production systems. The general practice prevalent in the credit market, especially in case of institutional credit, is to repay the interests annually on a loan and renew the loan for the next year. A default in interest pay-ment in a year forces a household to take new loans from newer sources (often from private ones) with the objective of both con-tinuing production as well as to partially clear the old debts. A repeated failure of this nature for a few production cycles leads to the accumulation of debts to an extent such that the annual inter-est costs become unmanageably large. The share of the interest component in the total outstanding debt also exceeds the interest rate at which the credit was originally available. Unless there is a marked jump in the returns to production in some period, the interest costs continue to deplete a larger share of the output every year. The interest rate structure of credit itself exhibits large differ-ences across regions. The average interest rate is extremely high at 19.73% per annum inAP-sdp compared to the other two regions (Table 6). The obvious reason is the higher share of non-institu-tional loans in this region. The higher interest rate in this region vis-à-vis the other regions indirectly reflects the interest rate gap that exists between the institutional and non-institutional credit market. The interest rate on loans from private sources in all the regions is found to be at least 24% per annum much higher than that on institutional loans, which hovered between 9% and 12% pa.A comparison of the percentage share of interest in total out-standing cash dues of the households with the average interest rate prevailing in a region reveals the extent of default in interest payments. From Table 6, we observe that the interest component of the average outstanding household debt inWB-rn at 16.18% is marginally higher than the average rate of interest (14.64%) of this region. A drastic build up of household interest costs with time has been absent in this region. The contrary is, of course true for the other two regions, where the interest component of the outstanding debt exceeds the average interest rate by 7.75 and 8.47 percentage points (forAP-sdp andAP-atp, respectively). This brings us to the issue of source of credit and the differential access to formal credit across classes. The distribution of household debt by source of credit presented in Table 7 (p 56) reveals the struc-ture of credit markets in the three regions. The share of institutional credit in AP-sdp is woefully low at only 23.63% of the total debt. The access to institutional credit is quite poor for all peasant classes though it increases for the upper classes. The high pressure of inter-est repayments on household income that we earlier noted can be significantly reduced by greater access to the formal credit market.This region comes in Telangana, which has been one of the earliest casualties of trade liberalisation. The volatile world cotton prices in the mid-1990s adversely affected the cultivators. The consequent high default in interest payments by the cultivators prompted the banks to stop giving credit to farmers. This facilitated Table 6: Number of Loans, Average Interest Rate and Interest Component of Outstanding Dues: All Regions WB-rnAP-sdpAP-atpNo of loans, of which 92 71 104Institutional 752778Non-institutional 174426Average interest rate (pa) 14.64 19.73 13.82Share of interest in total outstanding cash dues (%) 16.18 27.48 22.29Source: Same as of Table 3.
3.0 2.1 6.8 1.0 14 5 15.7 12.7 10.7 38.6 40.4 17 6 30.0 6.7 0.1 
11.9 21.1 8.2 13 8 49.9 44.3 12 3 6.6 0.8 0.6 6.5 11 4 10 4 2.5 
SPECIAL ARTICLEEconomic & Political Weekly EPW april 11, 2009 vol xliv no 1557The clear evidence, both from primary and secondary data sources, of high debt burdens, increasing share of informal credit and a stringent and rising interest rate structure in rural areas is critical for the question of further agrarian development. The sec-ondary trends from the NSS all the more strengthens the case for a simultaneous debt-relief programme and public interventions to make economic returns more favourable for agricultural producers. ConclusionsThe class analysis of the agricultural situation reveals the differ-ential impact of the neoliberal economic policies on the Indian peasantry. In all regions of our study, including a primarily food growing region like WB-rn, we find that the lower peasantry classes face a depression of their incomes to drastically low levels which can only satisfy consumption levels far below the required subsistence norms. In the cash crop cultivating regions of Andhra Pradesh, income depression pervades the production by some upper peasantry classes as well. The depression is also of such intensity where even non-payment of debts barely allows attain-ment of the required consumption levels for a good number of households in these classes. In contrast, some rare upper peas-antry classes retain a sizeable surplus from cultivation but they are just about able to manage and cover their required consump-tion expenditures, i e, they merely accomplish a simple reproduc-tion of the peasant economy. Although these peasant classes appropriate surplus labour, both directly in the labour market and indirectly through exchange relations in the land lease market, the lack of effective demand and consequent low output prices do not allow the realisation of that surplus. Coupled by the simultaneous scaling down of government operations with regard to inputs and infrastructure required by agriculture over the reform period, this has hindered the process of capital accu-mulation that is so essential for furthering agrarian transition in the country. The prolonged stagnation and low investments in the primary sector are precisely due to these reasons.Given this specific nature of stagnation, there is a strong case for stabilisation of output prices for primary and food commodi-ties in the economy. For this purpose, there is an urgent need to enhance public procurement operations, for both food and non-food crops, and rejuvenating the public distribution system to pre-TPDS levels. Given that an overwhelming majority of the peasantry in the country are net food-buyers, robust procure-ment and distribution operations are of central importance if the real output prices faced by peasant production has to come out of their deflationary trends.The magnitude of the constraint that has emerged for primary commodity producers in the sphere of credit exchange relations also call for a comprehensive debt-relief policy. The ADWDRS has been a welcome step towards that but there is need for more effective measures that will negotiate the huge debt burden of farmers to informal sources. The adoption of unilateral and mis-placed targeting rules based on size of landholdings also limits the effectiveness of such a policy. A more inclusive debt-relief policy that also constitutes a debt-relief commission, expansion of rural institutional credit facilities and improved real returns for agriculture can effectively release the primary sector from the clutches of deflation and indebtedness in which it has come to be ensnared under the neoliberal regime.Notes 1 Nagaraj (2008) has made detailed state-wise estimates of farmer suicides using the National Crime Records Bureau (NCRB) data. 2 The shrinkage of bank operations in rural areas as a result of the Narasimham Committee recom-mendations and in particular, the decline of pri-ority sector lending to agriculture, has been dis-cussed and documented comprehensively in Ramachandran and Swaminathan (2002) and Chandrasekhar (2004). 3 The proto-capitalist landlords are those who pri-marily undertake cultivation on their land by employing large quantum of hired labour and farm-servants. The other category of landlords would be the ones primarily leasing out land in exchange of ground rent. 4 The GVO is the sum of Crop Output, By-Product and Livestock Product. The Total Material Input costs is determined by the following relation- Total Material Input Costs = [Sf+ Livestock Feedf + Mf] + [Sp + Livestock Feedp + Livestock Main-tenancep+ Mp + Fp+ Fuelsp + Irrigation chargesp + Pp + Service chargesp] + Amortisation cost. where ‘S’, ‘M’, ‘F’ and ‘P’ denotes seed, manure, fertilisers and pesticides respectively. The suf-fixes ‘f’ and ‘p’ denotes farm-produced inputs and purchased inputs, respectively. The total paid-out labour costs comprise of wages paid in both cash and kind. The yearly amortisation cost for pro-ductive assets has been derived by a straight-line depreciation exercise based on the data on expected life of these assets. 5 As very few households were accruing ground rent by leasing out land and collecting service charges by renting-out their productive capital, we have mergedboth these in the category of income from self-employment for the sake of convenience. Both rent and service charges are returns to the ownership right ofland and productive capital respectively and are usually not placed under the category of income. 6 MPCI for ith class = [(Annual THI for ith class/12)/Average Family size for ith class]. 7 The NSS data on Nutritional Intake gives the per capita per diem calorie intake by the different MPCE classes. The per capita calorie intake (y-axis) plotted against the Average Per Capita Expenditure or APCE (x-axis) for each MPCE class reveals a direct relationship between the two. This relationship allows us to determine the expenditure level that achieves any particular calorie intake norm. By doing this exercise using the 61st NSS round data on nutritional intake (Report No 513) and applying a norm of 2,200 kcal, we arrive at these figures for West Bengal and Andhra Pradesh. A calorie intake norm of 2,200 kcal is chosen as that was the norm used for estimating the official poverty line of 1973-74, post the recommendations of the 1979 Task Force. For further insight on the issue, see Patnaik (2007). 8 Rent payments for this class in both WB-rn and AP-sdp are small and insignificant.9 Assets comprise buildings, livestock, means of production and transport and consumer durables of the households. The land assets and other financial assets are not included while computing the debt-asset ratio.ReferencesChandrasekhar, C P (2004): “Bank Reform and the Rural Sector”, available on www.macroscan.com.GOI (2007): Report of the Expert Group on Agricultural Indebtedness, Ministry of Finance. – (2008): Agricultural Debt Waiver and Debt Relief Scheme. – Reports of the Commission for Agricultural Costs and Prices for the Years 1997, 2002 and 2008.Government of Kerala (2007): “The Kerala Farmers’ Debt Relief Commission Act”.Labour Bureau, GOI,Consumer Price Index for Agricul-tural Labourers (Base: 1986-87=100).Nagaraj, K (2008): “Farmers’ Suicides in India: Magni-tudes, Trends and Spatial Patterns”, Macroscan, March, available on http://www.macroscan.com/ NSSO, Report No 420, Indebtedness of Rural House-holds as on 30.6.1991, Debt and Investment Survey, 1992.– Report No 498,Indebtedness of Farmer House-holds, Situation Assessment Survey of Farmers, 2002-03. – Report No 501,Indebtedness of Rural Households as on 30.6.2002, Debt and Investment Survey, 2002-03.– Report No 513,Nutritional Intake in India, 2004-05.Patnaik, U (1976): “Class Differentiation within the Peasantry: An Approach to Analysis of Indian Agriculture”, Economic & Political Weekly, Vol11(39),Review of Agriculture, 82-101.– (2007): “Neoliberalism and Rural Poverty in India”, Economic & Political Weekly, Vol 42(30): 3132-50.Ramachandran, V K and M Swaminathan (2002): “Rural Banking and Landless Labour Households: Institutional Reform and Rural Credit Markets in India”, Journal of Agrarian Change, Vol 2(4): 502-44.RBI, Banking Statistics, Basic Statistical Returns, 2001-06 (Mumbai: Reserve Bank of India). – (2007): Handbook of Statistics on Indian Economy.
