Economic and Political Weekly April 22, 20061538

Farmers’ suicides have been receiving a lot of social andpublic policy attention, particularly in the states of AndhraPradesh, Karnataka, Kerala, Punjab and Maharashtra, in

recent years.1 A brief review of issues indicates that the depen-dence on agriculture is largely among marginal and small farmersand agricultural labourers. These groups also bear the brunt ofthe unavailability of water and its associated yield uncertainty.Linking of the national market with international markets hasalso increased price uncertainty, particularly in crops like cotton.The absence of a formal rural financial market also leads to adependence on informal sources with a greater interest burden.The marginal and small farmer is, however, willing to experimentand take risks. But, the farmer is not able to visualise that a badmonsoon leading to a crop failure or a glut in the market canpush him into indebtedness and a crashing of dreams. This paperdiscusses farmers’ suicides in Maharashtra, which is relativelyhigher in certain districts of Vidarbha.

Suicide is a complex and multifaceted phenomenon. Someimportant correlates can be identified either in the neurobiologi-cal domain [Mann 2002] or in the socio-economic domain(Durkheim 2002 (1897)]. The former are predisposing in nature.They are internal factors that exist in the individual. All thoseidentified with these factors do not commit suicide. The presenceof additional factors that are external to the individual becomescrucial. This takes us to the socio-economic factors that areprecipitating in nature – they can act as a trigger. These can beeither systemic or idiosyncratic. The current exercise identifiesimportant socio-economic risk factors.

The rest of the paper is organised as follows. The suicidescenario and agrarian situation in Maharashtra are discussed inSections I and II respectively. Micro level analysis based on fieldsurvey in the districts of Wardha, Washim and Yavatmal isdiscussed in Section III. Concluding remarks are in Section IV.

ISuicide Scenario in Maharashtra

Suicide data in India is compiled through police records.2 In2001, Maharashtra constituted about 9.4 per cent of the all Indiapopulation but accounted for 13.5 per cent of the total suicidedeaths in the country. Suicides accounted for 2 per cent of thetotal deaths in the state whereas at the all India level suicides,

accounted for 1.3 per cent of the total deaths. Suicide is a socialphenomenon that differs across gender groups and it is appro-priate to discuss patterns in males and females separately. In 2001,the age-adjusted suicide mortality rate (SMR, suicide deaths per1,00,000 persons) for Maharashtra was 20.6 for males and 12.6for females whereas for India it was 14.0 for males and 9.5 forfemales.3

In Maharashtra, the age-adjusted SMR for males increasedfrom 17.4 in 1995 to 20.3 in 2004 and that for females decreasedfrom 13.6 in 1995 to 10.8 in 2004 (Table 1). The absolutenumbers of male suicides decreased in 1996, but thereafter theyhave been increasing. For females, the absolute number ofsuicides decreased in 1996 and then increased in the next twoyears, but have been declining since 1999. The age-adjustedSMR for males has not always been increasing indicating thatthe increase in the number of suicides has not been commen-surate with the increase in population. Between 2001 and 2004the age-adjusted SMR for males has been in the range of 20-21. The decline in absolute female suicides when the populationhas been rising explains the declining age-adjusted SMR forfemales. Across divisions, Amravati and Nagpur have relativelyhigher SMRs.

The total number of farmer suicides in Maharashtra increasedfrom 1,083 in 1995 to 4,147 in 2004. The increase was largelybecause of a 288 per cent increase in male farmer suicides from978 to 3,799. During this period, male farmer suicides as aproportion of total male suicides in Maharashtra increased from14 per cent to 38 per cent. We can arrive at the SMR for farmersby normalising it with the population of cultivators. The SMRfor male farmers increased by nearly four times from 15 in 1995to 57 in 2004, but for females the trend fluctuated to reach apeak of 12 in 2001, but thereafter it has been declining. Formales, the ratio of SMR for farmers to age-adjusted SMR waslower than unity in 1995, but thereafter it has been greater thanunity and has been steadily increasing, and in 2004 the SMRfor farmers was 2.8 times greater than that for the generalpopulation.

The SMR for farmers across divisions of Maharashtra during2001-04 indicates relatively higher suicide deaths than the stateaverage for males in Amravati and Nagpur divisions and forfemales in Amravati, Aurangabad and Nashik divisions (Table 2).Across divisions, the ratio of farmers-to-age-adjusted SMR is

Farmers’ Suicides in MaharashtraAn agrarian crisis has precipitated a spate of suicides in Maharashtra. The suicide mortality

rate for farmers in the state has increased from 15 in 1995 to 57 in 2004. The rain-dependentcotton growing farmers of Vidarbha are faced with declining profitability because of dumping

in the global market by the US, low import tariffs, failure of the Monopoly CottonProcurement Scheme and withdrawal of the state (resulting in declining public investment in

agriculture, poor government agriculture extension services and the diminishing role of formalcredit institutions). The farmer now depends on the input dealer for advice, leading to

supplier-induced demand, and on informal sources of credit, which result in a greater interestburden. In short, the farmer is faced with yield, price, credit, income and weather uncertainties.

The way out is to merge bold public policy initiatives with civil society engagement.

SRIJIT MISHRA

Economic and Political Weekly April 22, 2006 1539

the maximum in Amravati among males and in Nashik amongfemales. Nagpur has a higher SMR for farmers, which is loweronly than in Amravati, but the ratio of farmers to age-adjustedSMR is much lower. This indicates that in a relative sense theSMR in Nagpur is also high across other subgroups of populationwhich are not self-employed in farming.

IIAgrarian Scenario in Maharashtra

In Maharashtra, the contribution of agriculture and alliedactivities to the net state domestic product in current prices camedown from 40 per cent in 1960-61 to 13 per cent in 2004-05,whereas as per the 2001 Census 55 per cent of the total workersare either cultivators or agricultural labourers. Between1993-94and 2003-04, the linear trend growth rate of Maharashtra’s grossstate domestic product (GSDP) at 4.8 per cent per annum waslower than that of India’s gross domestic product (GDP) at 5.8per cent per annum.4

The GSDP of agriculture’s in Maharashtra at 1993-94 prices,increased from Rs 21,750 crore for the triennium ending (TE)1995-96 to Rs 24,988 crore for TE 2002-03. The eight majorcrop groups in terms of gross area under cultivation are cerealswith nearly half of it under jowar, pulses (nearly three-tenthsunder tur), fibres (most of it cotton), oilseeds (nearly half of itunder soyabean), sugar cane, fruits and vegetables, condimentsand spices and drugs and narcotics.

The share of estimated gross value added is lower comparedto the share of area for cereals (particularly low for jowar), pulses,fibres (mostly cotton) and oilseeds (Table 3). In TE 2002-03,cereals and pulses accounted for 64 per cent of the gross areaunder cultivation, but contributed to only 27 of the gross valueadded. Cotton accounts for 15 per cent area, but contributes toonly 9 per cent of the gross value added in agriculture. Oilseedsaccount for 12 per cent share of area and 10 per cent of the grossvalue added but for soyabean the share of gross value added isslightly higher (5.7 per cent) than its share of area. Sugar caneaccounts for less than 3 per cent of the area under cultivation,but 17 per cent of the gross value added. Fruits and vegetablesaccount for less than 5 per cent of the area under cultivation,but nearly 36 per cent of the gross value added.

Between the annual average of TE 1995-96 and TE 2002-03,the incremental gross value added is positive for pulses, soyabean,sugar cane and fruits and vegetables; it has declined for cereals,cotton, condiments and spices and drugs and narcotics. Theincrease in incremental value for pulses, soyabean, sugar caneand fruits and vegetables is largely explained by an increase inarea under cultivation, but for specific crops like tur, soyabean,banana, grape and tomato the share of other factors like yield,or advantages in relative prices also contribute.

The decline in incremental value for cereals is contributed bya decline in area as well as other factors. Condiments/spices showa decline despite a favourable impact of other factors becausesuch areas might have shifted to more value addition crops underfruits and vegetables. The decline of drugs and narcotics couldbe indicative of the legal hassles associated with its cultivation.

In cotton, the incremental value has declined, but there hasbeen an increase in area. With increasing yield, it suggests thatthe positive effect of an increase in area has largely been offsetby unfavourable relative prices. In other words, the profitabilityof cotton cultivation is declining.

Table 1: Age-Adjusted Suicide Mortality Rates for Farmersin Maharashtra, 1995-2004

Year Age Adjusted (5+) SMR for Farmers Ratio of SMR forSMR Farmers to

Age-Adjusted SMRMales Females Males Females Males Females

1995 17.4 13.6 14.7 1.8 0.84 0.131996 16.0 12.6 23.5 7.1 1.47 0.561997 17.7 13.8 23.9 5.6 1.35 0.401998 18.9 14.4 29.0 8.5 1.53 0.591999 18.5 13.9 30.6 6.9 1.66 0.502000 19.6 12.9 37.3 10.1 1.90 0.782001 20.6 12.6 44.1 11.5 2.14 0.912002 20.3 11.9 47.3 10.8 2.33 0.912003 20.6 11.3 50.8 9.4 2.46 0.832004 20.3 10.8 57.2 7.4 2.81 0.68

Notes: SMR = suicide mortality rate; age-adjusted SMR takes intoconsideration 5+ population only. SMR for farmers is calculated bynormalising suicides for those self-employed in agriculture on thepopulation for cultivators. Year-wise population for each subgroupwas interpolated/extrapolated and adjusted to give estimates that aresubgroup consistent.

Sources: (i) Census of India, 1991 and 2001; (ii) Accidental Deaths andSuicides in India, 1995, 1996, 1997, 1998 and 1999, National CrimeRecords Bureau, New Delhi; Communication from CID, Pune; andwww.indiastat.com (accessed October 5, 2005).

Table 2: Suicide Mortality Rate for Farmers acrossDivisions in Maharashtra, 2001-04

Divisions/ SMR for Farmers Ratio of SMR for Farmers to SMRDistricts for Age-Adjusted Population

Males Females Males Females

Amravati division 115.6 24.3 2.8 1.1Aurangabad division 47.6 9.2 2.5 0.8Konkan division 25.1 7.7 2.0 0.8Nagpur division 55.5 8.5 2.0 0.6Nashik division 36.6 12.0 2.2 1.4Pune division 34.7 3.6 1.6 0.4Maharashtra state 48.1 9.1 2.4 0.8

Note: The SMR for farmers are calculated with the assumption that cultivatorsas a proportion of 5+ years population is the same as it is in 2001. Thisis likely to underestimate SMR for farmers.

Source: As in Table 1.

In Vidarbha, particularly in Amravati division and in theselected districts, the cropping pattern indicates a shift in areafrom cereals (particularly jowar) and cotton towards soyabeanand to a lesser extent pulses. There is a shift, but with the absenceof viable alternatives, Cotton continues to be a major cash cropaccounting for 33 per cent of area in Wardha, 21 per cent ofthe area in Washim and 46 per cent of the area in Yavatmal intriennium ending (TE) 2002-03. Cotton and soyabean are the twoimportant cash crops of the region.

In 2004-05, the year relevant for our primary survey, produc-tion of cotton was a record high worldwide as also in India.Maharashtra’s production at 52 lakh bales showed a 68 per centincrease over the previous year’s 31 lakh bales. This is largelybecause of a record yield of 297 kg/hectare in the state. Regardlessof this, Maharashtra’s productivity at 64 per cent of the nationalaverage continues to be among the lowest. This growth wouldhave bypassed Amravati division where the monsoon in May-October 2004 was largely deficient. In the selected districts,rainfall fell short of the normal by 22 per cent in Wardha, 31per cent in Washim and 51 per cent in Yavatmal. The cottonfarmer in this region faced both price as well as yield shockssimultaneously.

On cotton prices, there are a number of other relevant factors.

Economic and Political Weekly April 22, 20061540

Excess international supply at a lower price is also because ofdirect and indirect subsidies leading to dumping by the US.During the period 1998 to 2003, cotton export prices for the USwere lower than their cost of production by more than 50 percent on average [Murphy, Lilliston and Lake 2005]. Domesticpolicies in India have led to removal of quantitative restrictionsand subsequently the reduction of import tariff from 35 per centin 2001-02 to 5 per cent in 2002-03. All these exposed domesticprices to the volatility of international prices that has beenadversely affecting the cotton farmer. Similarly, excessive cottonexports leading to an increase in yarn prices can adversely affectthe handloom and powerloom weavers [Ghosh 2005]. TheMonopoly Cotton Procurement Scheme (MCPS) in operationsince 1972-73 in Maharashtra was meant to stabilise prices, butover time a plethora of problems and cumulative losses haverendered it non-functional [see also Shroff 2006]. In short, whenthe farmer is being exposed to the global market, there is nomechanism that will guard him/her against price volatility.

In 2004-05, the cost of production for cotton in Maharashtraaccording to the Commission on Agricultural Costs and Priceswas Rs 2,216 per quintal, but the Minimum Support Price (MSP)was only Rs 1,960.5 The MSP indicated is for long staple varietyof fair average quality and is assured only when the crop is soldin authorised centres. Thus, even if one includes the additionaladvance price of Rs 500 paid in Maharashtra in that year, whichwas discontinued in 2005-06, the MSP is still likely to fall shortof costs or at most break even.6 The MSP for jowar also fallsshort of the cost of production in Maharashtra, but the MSP forsoyabean and tur are greater than the cost of production. Theseobservations are in line with shifts in cropping patterns and alsoto the changes in incremental value additions indicated above.

The state’s role as a facilitator of agricultural growth is ex-amined by looking at public investment in agriculture, the ruralfinancial market and agricultural extension, among others. In theabsence of data for Maharashtra, a perusal of the all India situationsuggests that gross fixed capital formation in agriculture as aproportion of the GDP declined from 3.1 per cent during 1980-85 (Sixth Plan) to 1.6 per cent during 1997-2002 (Ninth Plan).During the same period, gross fixed capital formation in agri-culture as a proportion of total gross fixed capital formationdeclined from 13.1 per cent to 7.4 per cent in India, the proportionof plan expenditure towards agriculture and allied activitiesdeclined from 6.1 per cent to 4.5 per cent in India and from 6per cent to 3.3 per cent in Maharashtra.

Recent trends with regard to agricultural credit in Maharashtrafrom 1991 to 2004 indicate the following. Credit utilisation toagriculture as a proportion of total credit utilisation in the statehas declined from 20.2 per cent to 11.2 per cent – this is largelyoffset by an increase in personal loans. Agricultural credit utilisationis shifting from rural regions to urban areas, with Mumbai’s sharehaving increased from 5 per cent to 48 per cent. Within agri-culture, the share of direct finance reduced from 79 per cent to51 per cent. Even after excluding Mumbai, division-wise dis-tribution shows a decline in the share of both direct and indirectfinance components of agricultural credit in Vidarbha.

In Maharashtra, the cooperative banks have been an importantsource of credit, particularly for agricultural purposes, in ruralareas. A recent survey (59th round National Sample Survey,January-December 2003) indicates that in Maharashtra nearlyhalf of the loans are from cooperative societies compared to aboutone-fifth for the all India average. However, much of these loans

are likely to be outstanding debts, not current loans. The perhectare loan in the selected districts is relatively lower [Shah2006]. A recent study of Yavatmal indicates that more than halfthe members are defaulters with their credit lines choked fromone to many years [Sarangi 2004]. This is so because over theyears, the cooperative credit institutions were faced with a numberof problems – high interest rates, accounting practices were notrationalised and no professional management to mention a few[Government of India 2004].

An important issue raised in our focus group discussions(FGDs) conducted in 98 villages with an average participant sizeof six to seven (minimum-two, maximum-nine) is that currentoperational loans are likely to be from moneylenders. In 70 percent of the FGDs, the availability of the informal loans in thevillage was mentioned. One participant’s remark during an FGDwill elucidate the socio-economic dominance of the moneylender.The participant said that: “Gentleman, you will go away afterthis discussion. It is we who have to stay in the village. Pleasedo not probe further into the details. Further revelation by uswill make our stay in the village difficult.”

Informal loan transactions could be in ‘dedhi’. The debtor hasto return the loan around harvest (within four to six months) andpay Rs 150 for a loan of Rs 100. Similarly, there is ‘sawai’(payment of Rs 125 for a loan of Rs 100). Another popular formof loans for agricultural and social purposes is at an interest rateof Rs 10/Rs 5 per month. Non-payment of loan leads to rewritingof a fresh loan with some additional credit being given duringthe start of the next agricultural season.

A conventional form of collateral is land. Creditors now considerit risky because suicides can lead to cancellation of such contracts,and hence, insist on sale of land with a verbal (not legal) promisethat it will be sold back to the debtor after the loan is repaid.If required, legal registration expenses on both counts are borneby the debtor. Land seizure/mortgage was mentioned in 17 percent of FGDs.

Some of the moneylenders would also be traders. The loan takencould be for purchase of an input and repayment through saleof produce. Interlocking of credit, input and output markets arenot necessarily enforced by the trader-moneylender, but

Table 3: Crop-wise Share of Estimated Gross Value Added(EGVA) to Share of Area (TE 2002-03) and Incremental EGVA

(TE 2002-03 over TE 1995-96)

Crop Group Share of EGVA Incremental EGVA, TE 2002-03 overto Share of Area TE 1995-96

TE 2002-03 Level Area Effect Other Effect(Lakh Rupees) (Per Cent) (Per Cent)

Cereals 0.38 -61868.9 -53.0 -47.0Pulses 0.49 5743.1 98.3 1.7Fibres 0.61 -8230.5 177.5 -277.5Oilseeds 0.85 5374.6 -358.5 458.5Sugar cane 6.05 54313.9 130.5 -30.5Fruits/vegetables 7.28 200564.0 205.2 -105.2Condiments/spices 1.41 -402.9 -491.0 391.0Drugs/narcotics 2.78 -359.7 -82.0 -18.0Total 1.00 195133.6 -8.8 108.8

Notes: TE denotes triennium ending for three year annual average,Crop specific area was estimated from EGVA for crop and EGVA forcrop per hectare. Incremental EGVA comprises area and othereffects (Et-Et-k)=((At-At-k)Ot-k)+((Ot-Ot-k)At) where positive/negativesign for shares of area/other indicate the proportion contribution toincrease/decrease in the level and E = EGVA, A = area and O = other.

Source: State Income of Maharashtra (Soft copy), Directorate of Economicsand Statistics, Mumbai, 2005.

Economic and Political Weekly April 22, 2006 1541

operating with a single trader-moneylender, would save trans-action costs to the farmer.

Nearly 50 per cent of FGDs discussed the paucity of water.Despite delay and deficient rain, there were instances of people(whole villages) opting for a second or third sowing without anygroundwater dependence. With seed replacement being almostcomplete, it contributed to additional expenses for seed. In thelast 5-10 years there has been an increase in the number ofsprayings for insecticides/pesticides and an increase in the needand cost for fertilisers. All these added to the cost. The issueof spurious quality of inputs also came up in the discussions.This brings forth some important points. First, the absence ofan extension service which could have advised the farmers againstlate sowing or improper use of other inputs. Second, with newtechnology on the anvil, there is deskilling and the farmer’sexperience becomes redundant. The private traders selling farminputs advise the farmer on extension service. This leads tosupplier-induced demand.

We also came across villages from where many people havemigrated in search of jobs. In 14 per cent of the FGDs, difficultyin getting employment or availability of work at low wages wasmentioned. These not only indicate the unavailability of non-farm jobs in the study regions, but also indicate poor publicinterventions.

One of the important social welfare measures in the state isthe Maharashtra Employment Guarantee Scheme (MEGS), whichhas been in operation since 1970s. In the inland eastern region(consisting of all the districts of Amravati division and Wardhaand Nagpur), the share of MEGS expenditure from the state’stotal MEGS expenditure decreased from 14.4 per cent in 2000-01to 6 per cent in 2003-04. This is much lower than the region’sshare of rural poor (23 per cent) or its share of rural population(17 per cent). The region’s share of item-wise expenditure between2000-01 and 2003-04 was 4.2 per cent for agriculture, 6.5 percent for irrigation and 13.0 per cent for horticulture. In particular,it reflects poor intervention in works associated with developingagriculture either directly or indirectly through interventions inirrigation and horticulture.

What is intriguing is that in Yavatmal, the expenditure on esta-blishment as a share of total expenditure in the state is greater thanthe district’s share of rural population. In Wardha the total expen-diture under MEGS in 2003-04 as also 2004-05, a deficient rainfallyear, was either on establishment or on miscellaneous accountsonly. These observation are in line with the findings mentioned ina recent study that MEGS has been successful as a relief measurelargely concentrated in the drought-prone areas of Marathwadaand western Maharashtra divisions of the state and has had alimited success as a poverty eradication measure [Vatsa 2005].

IIIMicro Level Analysis

This is based on a field survey largely conducted in Wardha,Washim and Yavatmal districts. The village were identified basedon a list of farmers’ suicides for 2004 and a few cases for January2005.7 Data from 111 suicide case (one Amravati, 21 Wardha,29 Washim and 60 Yavatmal) and 106 non-suicide controlhouseholds spread across 105 villages form the basis of ouranalysis. In our sample from suicide cases, 91 per cent were males,55 per cent in the age group of 31-50 years, and 80 per centwere currently married.

On educational status, 21 per cent were illiterate, 14 per centwere literate but below primary level, 26 per cent completedprimary education but were below matriculation and the rest werewith higher education. On experience in farming, 24 per centhad less than five years, 18 per cent had about 6-10 years andthe remains 58 per cent had more than 10 years of experience.

The three main castes are Kunbis (31 per cent), Banjaras (16per cent) and Baudh (9 per cent). Kunbis are the predominantpeasant community in the selected districts and Banjaras havea substantial presence in Yavatmal and Karanja subdivision ofWashim. The size-class of land shows that 14 per cent aremarginal (0-2.5 acres), 39 per cent are small (2.5-5 acres), 21per cent semi-medium (5-10 acres), 15 per cent medium (10-20acres), 4 per cent are large (20 + acres) and 7 per cent have notgiven information on their landownership.

Superimposing aggregate caste and land-size categories on theagricultural census of 1995-96, the differences in the ratios were notstatistically significant. Nevertheless, we elucidate the followingpatterns. For all size-class of land, the ratios indicate relativelyhigher suicides among the scheduled castes (SCs) and scheduledtribes (STs) and lower among “other” castes. For all castes, theratio of sample to population proportion indicates relatively highersuicides among the marginal size-class and small size-class offarmers and lower suicides among the “other” size-class of farmers.

In 79 per cent of the cases, suicides were committed byconsuming insecticides. These proportions are higher than thatindicated for the overall population in the selected districts. Thisis so because of the easy accessibility in farming households(particularly, those cultivating cotton) to pesticides.8 A hospitalthat can treat emergencies like poisoning is on an average morethan 20 km away (particularly, in Washim and Yavatmal). Thismeans that the time taken to reach a treatment centre in thesehilly regions can easily be more than an hour. This delay canprove fatal. Restriction on availability and toxic content ofpesticides and access to early treatment are important policyparameters. Important policy lessons can be taken from SriLanka’s experience [Gunnell and Eddleston 2003].9

Table 4: Risk Factors Identified with Deceased Individual

Risk Factors N=111 Per Cent

Was the deceased indebted? 96 86.5Did his economic status deteriorate before the incident? 82 73.9Did the deceased not share problems with otherfamily members? 61 55.0

Was there a crop failure? 45 40.5Was there a change in his social position beforethe incident? 40 36.0

Did the deceased have a daughter/sister ofmarriageable age? 38 34.2

Was there any suicide in the nearby villages recently? 36 32.4Did the deceased have any addictions? 31 27.9Was there a change in the deceased’s behaviourbefore the incident? 29 26.1

Did the deceased have disputes with neighbours or others? 26 23.4Did the deceased have some health problem? 23 20.7Did any death occur in the family recently beforethe incident? 11 9.9

Has there been any suicide previously in the family? 7 6.3Are some other family members chronically ill/handicapped? 4 3.6Average number of risk factors 4.8Minimum number of risk factors 2Maximum number of risk factors 9

Note: N indicates number of households. The risk factors are not mutuallyexclusive, and hence, will not add up to 100 per cent.

Source: Field Survey.

Economic and Political Weekly April 22, 20061542

A suicide is the complex interplay of multiple factors. Anumber of risk factors can coexist and one particular individualcan come across all or none of the risk factors identified by us.In our sample, the minimum number of risk factors is two andthe maximum is nine. The most common was indebtedness (86per cent). From all those indebted, 44 per cent were harassedfor repayment of loans and in 33 per cent of cases the creditorinsisted on immediate repayment (Table 4). Next in importanceis the fall in economic position (74 per cent). Indebtedness perse will not lead to a fall in economic position, but if it reachesa stage that will lead to a sale of assets then it can be associatedwith a fall in economic position. Similarly, a fall in economicposition can also lead to greater reliance on credit and therebyincreasing the debt burden.

In 55 per cent of cases, it was observed that the individualconcerned had not discussed his/her problem with other familymembers. He/she was shouldering the entire burden that wastroubling him and was not sharing his/her difficulties with others.An avenue for letting out one’s pent up feelings and frustrationwas closed.

Crop failure was mentioned in 40 per cent of the cases andmost of these also mentioned a loss in second or third sowingdue to delay in rainfall. There were a few cases which mentionfire or theft. Crop loss can also happen due to excessive untimelyrain, say, during the harvest. Crop failure can lead to economicdownfall and make it difficult to repay existing loans. This willalso increase the need for additional credit. Crop failure leadingto a fall in economic position is quite straight forward, but thecausal links can also be the other way round. A household facedwith a downfall in economic position or with greater debt burdencould not take additional loans for investing in agriculture (say,during a pest attack) and this can lead to a reduction in yieldor total crop failure.

Change in social status was identified in 36 per cent of thecases. This can be associated with a fall in economic position.Harassment by creditors or their agents due to non-payment ofloans can also lead to social disgrace. Crop failure due tounsuccessful experimentation by a farmer who was recognisedas successful entrepreneur may find a change in his social status– people who earlier came for advice are now providing solace.

A socially important role of a brother/father is to get one’ssister/daughter married. Communities have norms in terms of ageand expenditure.10 A farmer is largely dependent on a good returnfrom his produce to fulfil this obligation. Thus, crop failure,greater credit burden or a fall in his economic position can comein his way of fulfilling this obligation. Inability to conduct asister’s/daughter’s marriage can be socially humiliating. It canalso increase intra-household conflicts. To complete this socialobligation a farmer may also take loans thinking that he can repaythe amount after the harvest. The recent marriage of a sister/daughter or inability to get one’s sister/daughter married has beenidentified as a risk factor in 34 per cent of the cases.

We have also taken note of recent suicides in a nearby villageand identified them as an additional risk factor in 32 per centof the cases. This was done because an individual who is facingsome similar socio-economic problem can relate to the earlierincident and contemplate suicide. There could be an imitationeffect.11 Addiction (particularly, alcohol) was identified in 28per cent of the cases. It is said that under intoxication an individualmay indulge in an act of self-harm without being aware of theconsequences. Alternatively, getting intoxicated could itself be

a reaction to get out of depression that can be associated withsome socio-economic problem.

Change in the individual’s behaviour was identified in 26 percent of the cases. These are symptoms and indicate that the indi-vidual needs some psycho-social help. Dispute with neighbours/others in the villages was identified in 24 per cent of the cases.This could be related with property disputes or an altercationleading to a social humiliation. Or, it could be a part of hischanged behaviour indicating that he needs some help.

The personal health problem of the deceased was identifiedin 21 per cent of the cases. From these, 26 per cent (six cases)were those who were perceived by others’ with some mentalhealth problem. Illness gets aggravated due to a poor economiccondition because it makes care seeking difficult. Similarly, illhealth can lead to a loan to meet medical expenses and alsoreduce the ability to work aggravating the economic condition.If the sick person is some other member (3 per cent of the cases)then the breadwinner has the added frustration and helplessnessin not being able to provide appropriate care for an ailing parent/spouse/child. Death of another member in the family before theincident was identified in 10 per cent of the cases. The nearone’s death could have been because of not receiving appropriatehealthcare. Inability to provide care is largely because of thepoor economic condition rooted in the larger agrarian crisis.

Suicide history in the family could be identified in 6 per cent ofthe cases. This could be indicative of a genetic factor. However, asmentioned earlier such individual factors are predisposing in natureand they can be intensified with some additional risk factors.

Being indebted has been one of the most important risk factors,but then this may not reveal much about the extent of debt burden.Comparing suicide case with non-suicide control householdsshows that the average outstanding debt is higher in the formerby 3.5 times and after normalising for family size or land sizeit is higher by three times and all these differences are statisticallysignificant (Table 5).

Table 5: Average Outstanding Debt per Household

Household Suicide Cases Non-Suicide ControlsDebt Amount N Amount N

All * 38444 101 10910 95Per person* 7224 101 2405 95Per acre* 7079 94 2365 90

Notes: N=number of households. * Difference between suicide cases andnon-suicide controls is statistically significant at 95 per cent CI.

Source: Field Survey.

Table 6: Average Outstanding Debt per Transactionacross Source

Source Suicide Cases Non-Suicide ControlsAmount N Amount N

Cooperative bank 30466 50 13756 26Rural bank 11314 7 40800 4Commercial bank 95000 4 11500 1Moneylender 22080 60 11973 26Friends/relatives 16268 28 2000 2Self-help group 4500 2 14000 1Trader 5000 2 20000 1Landlord/employer 4000 2 10000 1Others 12571 7 16089 9Not available 5000 1 - -Total 23821 163 14598 71

Note: N indicates the number of transactions with outstanding debt. Thetransactions are from 93 suicide cases and 55 non-suicide controls.

Source: Field Survey.

Economic and Political Weekly April 22, 2006 1543

The 95 suicide cases had outstanding debt from 163 loantransactions and 55 non-suicide controls had outstanding debtfrom 71 loan transactions. Thus, the average number of loanswith outstanding debt in suicide cases (1.8) was higher whencompared with non-suicide controls (1.3). Analysis of the sourceof loan indicates a greater reliance on cooperatives in the formalsector and moneylenders in the informal sector (Table 6).

The reliance on moneylenders and friends/relatives is higherfor suicide cases (54 per cent of 163 transactions) than non-suicidecontrols (39 per cent of 71 transactions). A very high amountis indicated for suicide case households under commercial bankbecause a large farmer (owning 28 acres) having an outstandingloan of Rs 2.5 lakh which was incurred for marriage in the family(in fact, the individual had taken a loan of Rs 5 lakh and hadalready returned Rs 2.5 lakh). After excluding this extreme case,the distribution of total outstanding debt indicates that 42 percent is from cooperative banks, 36 per cent is from moneylendersand 13 per cent is from friends/relatives. In non-suicide controlhouseholds, after excluding a loan transaction with outstandingdebt of Rs 98,200 from a rural bank, the distribution of totaloutstanding debt indicates that 38 per cent is from cooperativebanks, 33 per cent is from moneylenders and 15 per cent is fromother unspecified informal sources.

The purpose of loan is given in Table 7. After excluding thetransactions where the purpose is not available, 67 per cent in thesuicide cases and 89 per cent in the non-suicide controls are foragricultural purposes only. This proportion further increases if wetake into consideration transactions from formal sources only.

Next to agriculture is marriage, which is mostly from informalsources. For each specific purpose the number of transactionswith outstanding debt and the average outstanding debt pertransaction is higher among suicide case households when com-pared with non-suicide control households. The average amountof outstanding debt per transaction for agricultural purposes isgreater than Rs 10,000. For marriage, after excluding an extremeobservation with outstanding debt of Rs 2,50,000 from a com-mercial bank by a suicide case household, the gap is Rs 2,324only. There was one suicide case household where a loan forhealth expenditure was to the tune of Rs 1.5 lakh. From totaloutstanding debt (including those where the purpose is notavailable), agriculture being the sole purpose accounts for 65per cent of the outstanding debt in suicide case households and79 per cent in non-suicide control households. Marriage beingthe sole purpose accounts for 18 per cent of the total outstandingdebt in suicide case households (reduces to 13 per cent if we

exclude the extreme case of Rs 2.5 lakh outstanding debt froma commercial bank); it is 10 per cent of the total outstandingdebt among non-suicide control households.

The total outstanding debt from suicide cases is 3.7 times morethan the total outstanding debt from non-suicide controls. Com-pared to non-suicide controls, the suicide cases have a greaterproportion of outstanding debt that is more than one year old(74 per cent and 65 per cent respectively) and for 2004, the mostrecent year, a greater proportion of outstanding debt is frominformal sources (72 per cent for suicide cases and 38 per centfor non-suicide controls).

Comparison of suicide cases with non-suicide controls alsoindicate that on average the former have a lower proportion thanthat which owns bullocks (a productive and liquid asset), rela-tively greater family size (particularly, female members) andlower value of produce (Table 8). The relevance of bullocks tothe agrarian economy as a productive asset in Indian agricultureis also well known [see Vaidyanathan 1988 among others].Bullocks are the major means of ploughing, an act that dependson rain and has to be done within a short span before sowing.Hiring of bullocks or tractors will increase costs and the lattermay not be as effective in these dry land rain-dependent con-ditions. Bullocks are also used as a liquid asset that is sold underdistress conditions. Thus, the absence of bullocks may be areflection of the hardship that the household has been facing.It was observed that suicide case households had on average alower ownership of bullocks as well as other livestock, agricul-tural implements, consumer durables and access to basic ameni-ties. These get further compounded with a crop failure or pooryield leading to a lower value of produce. If this also happensto be a family with a daughter or sister of marriageable age thenthe social responsibility would further add to the burden.

A statistical exercise is done to compare case-control house-holds. Households suicide status is a binary dependent variable,Y; 1=case and 0=control. The independent variables, Xi’s, areoutstanding debt in rupees, a yes/no binary variable on ownershipof bullocks, family size, value of produce in rupees and valueof produce per acre of landowned in rupees. Using these, weestimate a step-wise logistic regression,12

ln[p/(1-p)]=α+βiXi+u; i=1,…5.where ln is natural logarithm, p is probability of obtaining asuicide case household, ln[p/(1-p)] is the log odds ratio of asuicide case household, α is a coefficient on the constant term,βi’s are the coefficients of the five independent variables, Xi’s,and u is error term.

Table 7: Average Outstanding Debt per Transaction by Source across Purpose/Year of Loan

Purpose/Year Suicide Cases Non-Suicide ControlsFormal N Informal N Total N Formal N Informal N Total N

PurposeAgriculture 28861 49 20660 53 24600 102 18475 26 11353 30 14660 56Marriage 250000 1 27324 17 39694 18 - - 25000 4 25000 4Others 42261 5 12492 24 17624 29 6000 2 21000 5 16714 7Not available 17833 6 5125 8 10571 14 13333 3 11700 5 12313 8Year2005, January-March - - 7760 5 7760 5 - - 50000 1 50000 12004 18427 15 16273 44 16821 59 17562 13 9629 14 13448 272003 29333 18 15722 27 21167 45 10080 7 14955 11 13059 182002 18810 11 22588 17 21104 28 17083 6 9614 7 13062 132001 38560 5 21000 4 30756 9 22500 2 15000 5 17143 72000 and earlier 64050 8 200000 1 79156 9 28667 3 - - 28667 3Not available 66500 4 13250 4 39875 8 - - 6250 2 6250 2Total 32500 61 18630 102 23821 163 17173 31 12603 40 14598 71

Note: N indicates the number of transactions with outstanding debt.Source: Field Survey.

Economic and Political Weekly April 22, 20061544

���

While discussing results, instead of coefficients, odds ratio,eβi, are given because the interpretation is more intuitive – fora unit increase in the independent variable there would be acorresponding change in the odds ratio (probability of a suicidecase/probability of a non-suicide control).

The result for complete case-control analysis of the 68 pairsof observation is as follows. It gives outstanding debt and absenceof bullocks as statistically significant variables that differentiatesuicide case from non-suicide control households (Table 9). Itsuggests that if outstanding debt increases by Rs 1,000 then theodds that the household is one with a suicide victim increasesby 6 per cent and if the household owns bullocks then the oddsthat it is a household with a suicide victim decreases by 65 percent. When we estimate the results for the three selected districtsindependently, then under different scenarios family size, valueof produce, and land size turn out to be statistically significantin explaining the differences.

IVConcluding Remarks

Maharashtra’s age-adjusted suicide mortality rate for males hasstabilised in the range of 20-21 from 2001 onwards and that forfemales has been declining after 1999. The SMR for farmers haveincreased by nearly four times from 15 in 1995 to 57 in 2004.It is the highest in Amravati division where SMR for male farmerswas 116 during 2001-04, which was more than eight times theage-adjusted SMR for males in India in 2001.

Gross value addition is relatively higher for fruits and vege-tables and sugar cane, but these crops are not grown in theselected districts. Traditionally, cotton has been the cash cropfor farmers of the districts. Over the years, profitability fromcotton has declined. Some of the reasons are: high subsidies bythe US leading to price distortions, low import tariffs inIndia, and failure of the MCPS in Maharashtra. The withdrawal ofthe state is evident from declining public investment in agricul-ture, poor government agricultural extension service, and a dimin-ishing role of formal institutions in rural financial market amongothers. The farmer now depends on the input dealer for adviceleading to supplier-induced demand and on informal sources ofcredit with a greater interest burden. To add to this, 2004 wasa rain deficient year that affected yield in at least some pocketsof the selected districts, but the overall macro supply scenariobeing good, market prices were low. The farmer was exposedto yield and price shocks simultaneously. In short, the systemicrisk factors indicate a larger socio-economic and agrarian crisis.

In the micro-level analysis, the inter-related and coexistingsocio-economic risk factors identified are indebtedness, deterio-ration of economic status, conflict with other members in thefamily, crop failure, a decline in social position, burden ofdaughter’s/sister’s marriage, suicide in a nearby village, addictions,change in behaviour of the deceased, dispute with neighbours/others,health problems, a recent death in the family, history of suicidein the family and other family members being ill. Comparingsuicide cases and non-suicide controls, one observes that on anaverage the former have a higher outstanding debt, a relativelylower owner ownership of assets (particularly, bullocks whichis a productive and liquid asset) and access to basic amenities,a larger family size (particularly, female members), and a lowervalue of produce. These indicate that the idiosyncratic factorsdo not occur in isolation – they are exacerbated because of thelarger socio-economic and agrarian crisis.

The policy implication from the above discussion calls for anemphasis on the larger agrarian crisis. Availability of affordablecredit requires revitalisation of the rural credit market. Riskmanagement should address yield, price, credit, income or weatherrelated uncertainties among others. Improving water availabilitywill facilitate diversification of the cropping pattern, but thisshould go hand in hand with policies that increase non-farmemployment. Improving agricultural extension that addressesdeskilling because of technological changes and also facilitatesappropriate technical know-how for alternative forms of culti-vation such as organic farming will be of help. There is a strongcase for regulating private credit and input markets. Public healthinterventions would require improving access to treatment ofpoisoning cases and psychosocial care. This should be comple-mented with policies that reduce the access to organo-phosphorouspoisons. Besides, support from public institutions, there is needfor a greater involvement from civil society.

Email: srijit@igidr.ac.in

Notes[This paper draws largely on the report, Suicide of Farmers in Maharashtrafunded by the government of Maharashtra. Many people helped in the studyand the detailed acknowledgements are in the report, but the cooperation ofthe individuals interviewed (particularly, the members of the suicide casehouseholds) needs special mention. The author thanks R Radhakrishna forgiving him this responsibility and, more importantly, for his comments andsuggestions at various stages of the work and D Narasimha Reddy forcomments on an earlier version of this paper.]

Table 8: Comparing Control Cases by AverageHousehold Characteristics

Average Characteristic Suicide Cases Non-Suicide ControlsValue N Value N

Own bullocks ( per cent)* 42.9 105 63.5 104Value of produce (Rs 000) 22.9 93 41.4 90Value of produce per Acre (Rs 000) 3.4 93 6.5 90Family size 5.53 111 5.08 106

Notes: N=number of households. * Difference between suicide cases andnon-suicide controls is statistically significant at 95 per cent confidenceinterval.

Source: Field Survey.

Table 9: Results (Odds Ratio) of Step-wise LogisticRegression Analysis

Complete Wardha Washim YavatmalCase-Control Analysis

N 136 24 32 80Debt 1.000061 1.000237 1.00009 1.000055

(0.0000138) (0.0000974) (0.0000383) (0.0000176)[0.000] [0.015] [0.019] [0.002]

Own bullocks 0.3462934 0.3084751(0.1403603) (0.1685215)

[0.009] [0.031]Family size 2.124579

(0.8003213)[0.045]

Value produce 0.9997406(0.0001221)

[0.034]Value per acre 0.9997575

(0.0001234)[0.049]

Log likelihood -74.6497 -5.4993989 -16.388205 -42.176024LR Chi2 39.24 22.27 11.59 26.55Prob >Chi2 0.0000 0.0000 0.0031 0.0000PseudoR2 0.2081 0.6694 0.2612 0.2394

Notes: Round brackets give standard error, square brackets give prob > |z|.The variables are indicated in the order in which they were selected inthe step-wise logistic regression. In Wardha, estimation was possibleafter excluding the variable value of produce per acre.

Economic and Political Weekly April 22, 2006 1545

1 Some of the recent studies include Assadi (1998), Bhalla et al (1998),Chowdry et al (undated), Dandekar et al (2005), Deshpande (2002),Government of Andhra Pradesh (2004), Government of Maharashtra(1998), Iyer and Manick (2000), Mishra (2006a, 2006b), Mohan Rao(2004), Mohanty (2001, 2005), Mohanty and Shroff (2004), Parthasarathyand Shameem (1998), Purendra Prasad (2003), Revathi (1998), ShambhuPrasad (1999), Shiva et al (2000), Stone (2002) and Vasavi (1999) amongothers.

2 There is under-reporting because of legal hassles and shame identifiedwith the act.

3 Medically, suicide is not defined for the age group of 0-4 years. Hence,5+ population is used to compute the age-adjusted SMR.

4 For a recent discussion on Maharashtra’s growth and poverty see Mishraand Panda (2006). For Maharashtra’s agricultural development till early1990s see Sawant et al (1999). For agrarian scenario in selected districtsof Vidarbha see Kulkarni and Deshpande (2006). On public interventionbacklogs, a dated but comprehensive analysis is the Dandekar committeereport on Regional Imbalance [Government of Maharashtra 1984].

5 The MSP is declared by ministry of agriculture, government of India.6 The additional advance price has been discontinued from 2005-06 leading

to drastic fall in the prices received by the farmer when costs are likelyto have further increased.

7 Occurrence of suicide, a rare event, was the basis for selecting a village.In the village we surveyed the suicide case household, a control household,conducted a focus group discussion and also collected some village levelinformation before moving over to another village. The distance fromone village to another, more often than not, was 20-30 km.

8 The greater incidence of poisoning in farming households, particularlyduring the peak agricultural season, has been mentioned in Bhatkule(2006) among others.

9 A recent study suggests that pesticides are not only agents for suicide,but are also part of the causal pathway [London et al 2005]. Exposureto organophosphorous pesticides can affect the central nervous systemwhich in turn can lead to depression and subsequently suicide.

10 In our selected region, instances of females getting married before 18years of age are prevalent. Marriage expenditure would depend on theeconomic position of the household and FGDs indicate it to be aroundRs 20,000-Rs 40,000 only for marginal/small farmers. With such a smallamount, it is not the social custom of marriage but the poor returns fromagriculture that becomes relevant.

11 Suicide contagion also increases the responsibility of media and guidelinesby the World Health Organisation on suicide reportage would be of help[Mishra 2006b]. This is not to deny the important role that media playedin highlighting the issue of farmers’ suicides. For a media perspective,see Deshpande (2006) among others.

12 In the step-wise procedure, a variable is added if it increases chi-squaresignificance by 0.05 and it is dropped if it increases chi-square significanceby 0.1.

ReferencesAssadi, M (1998): ‘Farmers’ Suicides: Signs of Distress in Rural Economy’,

Economic and Political Weekly, 33 (14): 747-48.Bhalla, G S, SL Sharma, N N Wig, S Mehta, S and P Kumar (1998): Suicides

in Rural Punjab, Monograph Series-5, Institute for Development andCommunication, Chandigarh.

Bhatkule, P R (2006): ‘Poisoning Cases in Yavatmal Medical College, July2004-June 2005’, Suicide of Farmers in Maharashtra, Background PapersSubmitted to the Government of Maharashtra), Indira Gandhi Instituteof Development Research (IGIDR), Mumbai.

Chowdry, K R, R R Prasad, P Manikyamba, S Indrakant and T N Rao(Undated): ‘Farmers’ Suicides in Andhra Pradesh’, mimeo.

Dandekar A, S Narawades, R Rathod, R Ingle, V Kulkarni and Y D Sateppa(2005): Causes of Farmer Suicides in Maharashtra: An Enquiry, FinalReport Submitted to the Mumbai High Court, Tata Institute of SocialSciences, Rural Campus, Tuljapur.

Deshpande, R S (2002): ‘Sucide by farmers in Karnataka: Agrarian Distressand Possible Alleviatory Steps’, Economic and Political Weekly, 37 (25):2601-10.

Deshpande, V (2006): ‘Farmers’ Suicides: A Media Perspective’, Suicideof Farmers in Maharashtra, Background Papers (Submitted to theGovernment of Maharashtra), IGIDR, Mumbai.

Durkheim, E (2002, [1897]): Suicide, Routledge Classics, London and NewYork.

Ghosh, J (2005): ‘Trade Liberalisation in Agriculture: An Examination ofImpact and Policy Strategies with Special Reference to India’, BackgroundPaper for Human Development Report 2005, Human Development ReportOffice, Occasional Paper: 2005/12.

Government of Andhra Pradesh (2004): Report of the Commission on Farmers’Welfare, Hyderabad.

Government of India (2004): Draft Final Report of the Task Force on Revivalof Cooperative Credit Institutions, (chairperson of the committee: AVaidyanathan).

Government of Maharashtra (1984): Report of the Fact Finding Committeeon Regional Imbalance in Maharashtra, Planning Department,Bombay, April.

– (1998): Suicides of Farmers in Maharashtra: A Socio-Economic Survey,Commissioner Agriculture, Maharashtra State, Pune, Special Study, August.

Gunnell, D, and M Eddleston (2003): ‘Suicide by Intentional Ingestion ofPesticide: A Continuing Tragedy in Developing Countries’, InternationalJournal of Epidemiology, 32 (6): 902-09.

Iyer, K G and M S Manick (2000): Impoverishment and Suicides in RuralPunjab, Indian Publishers Distributors, New Delhi.

Kulkarni, AP, V S Deshpande (2006): ‘Agrarian Scenario in Yavatmal,Washim and Wardha Districts’, Suicide of Farmers in MaharashtraBackground Papers (Submitted to the Government of Maharashtra),IGIDR, Mumbai.

London, L, A J Flisher, C Wesseling, D Mergler and H Kromhout (2005):‘Suicide and Exposure to Organophosphate Insecticides: Cause or Effect?’American Journal of Industrial Medicine, 47 (4): 308-21.

Mann, J (2002): ‘A Current Perspective of Suicide and Attempted Suicide’,Annals of Internal Medicine, 136: 302-11.

Mishra, S (2006a): ‘Suicide of Farmers in Maharashtra’, report submittedto the Government of Maharashtra, IGIDR, Mumbai. This is availableat http://www.igidr.ac.in/suicide/suicide.htm.

– (2006b): ‘Farmers’ Suicides in Maharashtra: Content Analysis of MediaReports’, Suicide of Farmers in Maharashtra, Background Papers(Submitted to the Government of Maharashtra), IGIDR, Mumbai.

Mishra, S and M Panda (2006): ‘Growth and Poverty in Maharashtra’, WP-2006-001, IGIDR, Mumbai.

Mohan Rao, R M (2004): Suicides among Farmers: A Study of CottonGrowers, Concept Publishing Company, New Delhi.

Mohanty, B B (2001): ‘Sucides of Farmers in Maharashtra’, Review ofDevelopment and Change, 6 (2): 146-88.

– (2005): ‘We are Like the Living Dead: Farmer Suicide in Maharashtra,Western India’, Journal of Peasant Studies, 32 (2): 243-76.

Mohanty, B B and S Shroff (2004): ‘Farmers’ Suicides in Maharashtra’,Economic and Political Weekly, 39 (52): 599-606.

Murphy, S, B Lilliston and M B Lake (2005): WTO Agreement on Agriculture:A Decade of Dumping (United States Dumping on Agricultural Markets),Institute for Agriculture and Trade Policy, Minneapolis.

Parthasarthy, G and Shameem (1998): ‘Sucides of Cotton Farmers in AndhraPradesh: An Exploratory Study’, Economic and Political Weekly, 33 (13):720-26.

Purendra Prasad, N (2003): ‘Rights of Farmers in the Context of Suicidesin Andhra Pradesh’, Indian Journal of Human Rights, 7 (1-2): 101-19.

Revathi, E (1998): ‘Farmers’ Suicides: Missing Issues’, Economic and PoliticalWeekly, 33 (16): 1207.

Sarangi, U C (2004): ‘Report on Cropping Pattern Optimisation in the Districtsof Yeotmal, Jalna and Nandurbar to Augment Livelihood Opportunities’,Study instituted by IGIDR on behalf of UNDP and Government ofMaharashtra.

Sawant, S D, B N Kulkarni, C V Achuthan and K J S Satyasai (1999):Agricultural Development in Maharashtra Problems and Prospects,National Bank for Agriculture and Rural Development, Mumbai, OccasionalPaper: 7.

Shah, D (2006): ‘Resurrection of Rural Credit Delivery System in Maharashtra’,Suicide of Farmers in Maharashtra, Background Papers (Submitted tothe Government of Maharashtra), IGIDR, Mumbai.

Shroff, S (2006): ‘Cotton Sector in Maharashtra’, Suicide of Farmers inMaharashtra, Background Papers (Submitted to the Government ofMaharashtra), IGIDR, Mumbai.

Shambhu Prasad, C (1999): ‘Suicide Deaths and Quality of Indian Cotton– Perspectives from History of Technology and Khadi Movement’,Economic and Political Weekly, 34 (5): PE12-20.

Shiva, V and A H Jafri, A Emani and M Pande (2000): Seeds of Suicide:The Ecological and Human Costs of Globalisation of Agriculture, ResearchFoundation for Science, Technology and Ecology, New Delhi.

Stone, G D (2002): ‘Biotechnology and Suicide in India’, http://artsci.wustl.edu/~anthro/research/biotech_suicide.html (accessed June 20, 2005).

Vasavi, A R (1999): ‘Agrarian Distress in Bidar’, Economic and PoliticalWeekly, 34 (32): 2263-68.

Vaidyanathan, A (1988): Bovine Economy in India, Oxford and IBH, NewDelhi.

Vatsa, K S (2005): ‘Employment Guarantee Scheme in Maharashtra: ItsImpact on Drought, Poverty and Vulnerability’, Mumbai, mimeo.