Title:  Tensions in the Life of Tenant Farmers in Rural South India


A small mixed-methods research project led by Wendy Olsen under the Global Poverty Research Group.


Oct 2006 – April 2007


Research Question:  What are the variations in the mixture of choice and constraint factors that are felt to influence people’s decisions about their work, occupational status, and land management, among the people in rural Andhra Pradesh villages who are renting in land?


Background:  In India a shrinking but still substantial number of rural people rent land (Olsen, 2006 forthcoming).  As tenants, their families divide their time between unpaid domestic work, tenancy work, other informal sector work and paid work.  Young people also have to decide whether to stay in formal education or to exit schooling in order to work more hours. Poverty influences many of these people, although not all tenants are poor.  Approximately 15% of rural people are in tenant households, and 8% of the land is rented at any one time (1999 national data). A recent policy shift establishes Rural Employment Guarantee Schemes for 200 districts (later to be rolled out to all rural districts), but this change has not taken into account alternative forms of employment including tenancy as a potential source of income for poor people.  Arguments among specialists (e.g. Agarwal, 2003; Jackson, 2003) suggest that women may be a target group deserving to be encouraged to grow crops on rented land, and this debate will be competing with the Rural Employment Guarantee Scheme discussions over the next 5 years in India.  Other countries with large rural sectors will be watching to see which policy (labour guarantee, or encouragement to increase access to land) is more effective in the Indian test case.


Decisions made by tenants are bound to reflect a mixture of background factors.  A realist model of the causal factors would note the contextual factors, major mechanisms (including choices) and the outcomes that result. This project is not about choosing to rent land, but rather about a variety of other choices that arise in the context of doing farming on the land.  Some respondents will own a little land, whilst others will be landless.  The selection of respondents will be aimed at maximizing contrast between them – jobless vs. employed; male vs. female; old vs. young – whilst also sticking to the original focus. 


Theoretical models of decisionmaking for such people are of three main types. In economics the models tend to portray ideal-typical households with idealized people in them, and then to fit survey data to these models.  The new home economics is a typical approach used in economics.  In political economy the models are more class-based and broad-brush, and tend to assume that the people of one social class act in ways consistent with the motives and interests of that class.  It is important, in political economy models, to recognize the structural constraints that exist for people in the poorer social classes.  Already we can see in this literature a diversity ranging from the methodological individualist to the collectivist.  A third set of models promoted by some feminists see the household’s outcomes as a result of interpersonal bargaining.


Innovative Methodology Proposed:  These models are not meant to be mutually exclusive, since they may all really apply in a given situation.  In the present research however a more retroductive question is being asked:  what are the mental models that people use when they actually make decisions?  How do these cognitive or habitual approaches vary from person to person; are there patterns or types or styles?  Do people’s practices reflect their habitus and practical habits, as Bourdieu would suggest (Bourdieu, 1999; Bourdieu and Nice apply this framework to poverty in France), or is there a point where conscious diversion from habit takes over?  These sorts of retroductive questions were asked by Margaret Archer in a qualitative realist study of 20 British respondents (2003; see also Archer, 2000).  Archer depicts the respondents as falling into 3 ideal types.  These range from the communicative reflexive people to the meta-reflexive people, and huge personality differences were summarized through archer’s use of ideal types.  Her research challenges others to ask whether there is any cross-cultural applicability of the notion of having a small set of major types of people.  Archer’s work deals mainly with work and occupational choice but can be applied in the rural tenancy setting, too.  In India, the tenant often combines working on rented land with several other occupations, mostly informal but also including paid casual labour or formal salaried employment.


A debate about reverse tenancy notes the existence of well-off tenants in India who are not poor and do not do casual paid labour.  A background paper for this research will describe the frequency of tenancy among each quintile of India’s population.  (See Olsen, 1996 and 2004, for further background.)  More importantly, research on work outcomes (ie decisions) has shown that poverty is a major factor among others that appear to influence the work—self-employed, domestic work, or paid labour – that people do (Olsen 2005; Olsen and Mehta, mimeos, 2005).


By looking at the work outcomes and asking ‘why and how do these outcomes emerge’, I want to look into the processes that lead to the work outcomes.  Retroduction is used to ask ‘why and how have these observed data come about’ (Danermark, 2001; Olsen, 2004).  Retroducing from the secondary data will require qualitative data, and a small scope is planned for this initial foray.  35 interviews covering a range of types of people, often in the same household (see methods below) will be transcribed and analysed in detail.   Using a translator and sub-contractor with substantial research experience in Telugu language, I will ask these sorts of questions:


Were there any doubts about decisions that were made about the renting of land?


How and when was the crop choice made last time, for putting crops on the tenanted land?


Basically why do you rent land?  (or why do you not want to, if you have doubts about renting it?)


Do you have discussions about the use of the water on the rented land?  With whom; what issues arise?  What is your view about it?


Etc.  (see appendix 1 containing the first draft of the interview plan)


It is through the lens of these explicit discussions that we can hope to gain insight into what is actually, really, but often privately happening in families when their work strategies are decided upon.


Literature Review (in Brief)


In the three models mentioned earlier, considerable assumptions are made about people working in their own best interest and how this interacts with the nuclear or joint household’s best interests.  Most models don’t explore the detail of how the actual decisions emerge.  In this study, I want to stress the complexity of the routes to a given outcome (Byrne, 2002 and 2004).  More than one causal mechanism can lead to the same outcome.  Secondly I would urge that gender differences in interests within a household can lead to conflict and tension.  The individual’s interest would be seen (by an outsider) as different from the household’s best interest, and people inside have to navigate these conflicting interests as best they can.  Longterm planning enters indirectly into the causal framing of a current (apparently short-term decision).  Gender difference in the dowry and exogamy behaviour, with women likely to leave home for marriage, creates a difference in the factors shaping girls’ formal schooling and hence their work patterns.  A full study of occupational choice isn’t feasible here, but a review of the literature on occupational choice of young people in India will be conducted.



The research begins with its literature review and the analysis of secondary data (NSS 55th round) on tenants’ declared range of work patterns.  These data are notoriously inaccurate and badly recorded in the area of tenancy itself.  Labour patterns of those declaring tenancy will not give an accurate picture because there are many tenants who probably have not given accurate records of their land use to the NSS interviewers.  Nevertheless where tenancy is recorded, it is likely to be associated with valid records of the labour use of real tenants.  In causal terms the NSS data are instances, and can instantiate arguments about the causal mechanisms leading to tenants doing paid, informal, domestic or no work.  What we are missing is the full range of such cases, since there is very likely to be underreporting.


Then using realist assumptions the research moves into a qualitative stage.  35 interviews are planned to occur in Telugu and to be recorded.   Selection of respondents from an existing village survey in which 1994-95 household income details were recorded gives a detailed background from which to work in framing the questions.  (It also increases rapport and trust, since there were no negative consequences of being in the 1994-5 survey.)  After translating each interview into English, an analysis of discourses of choice will be conducted. 


The village venue is two contiguous villages of Ramasamudram Taluk of Chittoor District, southern Andhra Pradesh, India.  These villages were surveyed previously, and in each village 60 households were selected randomly for 1994-1996 research which used both interviews and a questionnaire survey (ESRC data archive study number 3927).  From these 120 households there were at that time 35 tenant households.  Of these, about half held some land of their own.  Many were poor, although not as poor as some landless non-tenant families.


In these villages, the sampling will include 4 women from households that own some land (besides the house plot) and who do paid labour; 4 women from households that own some land but who do not do paid labour; 4 women from households that are landless, who do paid labour; and 4 women from households that are landless, and who do not do paid labour (e.g. shopkeepers).  Furthermore 4 men of each group will be interviewed.  A wide range of ages 14-50 will be included amongst these people.  A further 3 interviews will be with interested parties who want to tell their stories to the researchers.


This quota sampling method creates great contrasts between respondents.  Some can come from the same household however, and couples may, if they wish, be interviewed together.  A revisit to each household to check on the data and to augment the initial interview is planned, subject to funds (Hollway and Jefferson, 2000).  Some tenants are not at all poor, and others are very poor, so the study will lead to commentaries on how poverty appears to be affecting people (and whether they mention that poverty affects them).  This case study illustrates the co-incidence of ‘causes’ and ‘reasons’ as mechanisms.  Furthermore, baseline data gives an idea of the ‘liabilities’ and ‘capabilities’ that are enabled by the assets held by each household or each person.  Agarwal would argue that a person’s fallback position affects their bargaining position, but this is quite hard to measure and we are not trying to fit the data into a bargaining model at this time.


After the interviews are translated into English they will be typed, with some words being typed in transliterated Telugu (e.g. meemu cheetu kooruku tiisukonaamu, we took the land as sharecropping, or ee  uuruloo manamu svantabhoomi kaadu.  Bhoomi tiisukoodaniiki guttagaa tiisukoovatsee.  In this hamlet, no one has land.  To get land you have to pay cash rents.)   We anticipate having about 700 pages of text (35 x 20 pages) which will be put into NVIVO as a series of documents.  Both individuals and households will be annotated as ‘cases’ in this data.


Analysing the data will take two procedures iteratively.  Firstly, discourse analysis will be conducted to see what patterns of discourse are dominant in this context (and hence, apparently, what are the rules and norms of discourse about the topics in the interview plan).  Assumptions made about discourse are described by Olsen (in Olsen, 2006, in the chapter in a volume edited by Carling).  Since the data are oral products of a remote village population it is unlikely that these discourses can be compared with other primary sources, but we will try.  (E.g. oral history books from Telangana can be compared; interviews from my previous research will be compared with these.)  Secondly, the specific and rather Western discourse of choice vs. constraint, found in the three models that were described in the review of literature, will be explored by looking for parallels, analogies, metaphors and similarities in the interview texts.  In other words we will test the hypothesis that a choice discourse is present among the tenants of this part of rural India. 


Having run this test once, we then iterate back to discourse analysis to see if a choice-centred discourse per se can be discerned. The same process, run once again, allows us to look for evidence of constraints as described by Marxist writers (e.g. Athreya, 1990; Bhaduri, 1983) or as described by feminists (Folbre, 1995; Swaminathan, 2002). 


The central hypothesis of the research design is that these discourses can be co-present in the decisions that are made.  Processes of decision making may use habitual discourses and habits of work so that without mentioning choice, there is nevertheless a sense of freedom whilst strategic decisions are made.  Furthermore, however, choice is made only in conditions set by history and context, e.g. family history; dowry and the assets owned.   As a researcher I cede much ground to both the choice school and the constraints schools, but I want to find out which decisions reflect either one or both.  In Fairclough’s terms, we can expect to find intertextuality.  However we want to explore the overtones and negative/positive connotations that are associated with each co-incidence of choice/constraint.  Are people angry about constraints that bind them?  Which ones, and who feels that way?


A diagram helps illustrate the hypothesis that I’ll be exploring:


Venn Diagram



By looking at the process of decision making, as described retrospectively, it will be possible to reorganize the evidence into a fresh set of cases.  This set of cases is the incidents, each with its characteristics.  It is useful for more than one respondent to talk about a certain incident or type of incident.  A series of incidents can be described and coded with attributes which will indicate the extent of the presence of a choice discourse, constraint discourse, or bargaining situation/discourse.  Considerable revision of this framework of analysis is expected since the project deliberately has a partially inductive mode of reasoning.


Finally to present the findings a summary will place the cases into contrastive configurations.  For instance, there may emerge three types of people (as Archer described), or there may emerge gendered moral rationalities (as Duncan and Edwards described, again for the British case, 1999).  They may also (or separately) emerge several types of incidents:  those which typically occur within a sense of free choice; those which typically are associated with feelings of unfreedom, coercive situations, or various constraints.  Furthermore some situations may be associated with both in the sense that the constraints may be considered desirable prior decisions of the same person or household. 


The project thus explores not only discourses in use, as presented in the interview’s rather formal setting.  It also explores the notion of the rural person as an agent versus the household as a (metaphorical) decisionmaker.  Economists routinely anthropomorphise the household, but this research tries to clarify the nature of a household strategy as an emergent property based upon the individual agency and synergistic character of discussions and behaviour within the family.[1]  In this way the research is consistent with institutionalism (Hodgson, 2005). Because of this agency focus, the bargaining possibilities will also be taken into account and we will look out for evidence of conflict and how it is resolves.  Conflict both within and between households is likely to arise in some cases.  Existing interviews conducted in 1995 showed the rural people to be willing to expose their disagreements over a number of issues. 


The role of caste, status expectations, stereotypes, obligations, traditions, rituals and duties will be allowed for. 


The interviews are grounded on previous ethnographic work by both Olsen and the Indian sub-contractor (likely to be Davuluri. Venkateswarlu; see Dacorta and Venkateswarlu, 1999, and Venkateswarlu and Dacorta, 2001). The results will have implications for institutional economics (Hodgson, 2005) in the specific sense that the evolution of labour-market and land-market institutions rests upon a foundation of human agency.  This project will help to describe  how human agents interact with each other, what assumptions they make, and what they perceive to be problematic in the household-level strategies that result from their habitus.


Using NVIVO a summary diagram can be created for each main finding. 


Ethical Issues:  It is intended that the interviews would be anonymised and put into the ESRC Data Archive.  Verbal consent will be obtained for this use of the data, and for writing up the results by quoting respondents using pseudonyms.  Local help will be obtained for setting up the pseudonyms.  The qualitative data will be linked by ID number to existing survey data (ESRC Study Number 3927). 


Appendix 2 estimates the time costs of doing this research project.



Appendix 1


Were there any doubts about decisions that were made about the renting of land?


How and when was the crop choice made last time, for putting crops on the tenanted land?


Basically why do you rent land?  (or why do you not want to, if you have doubts about renting it?)


Do you have discussions about the use of the water on the rented land?  With whom; what issues arise?  What is your view about it?


What payment is made for the water and fertilizer for the rented land?  Who decides?  Comment please


In your household do you have discussions about who goes out to work and when?  Describe these


Is anyone doing regular unpaid work for the landlord, and please describe the situation.


Why do they do this work?


Does anyone do irregular work, just on festival days, or otherwise, for the landlord?


Which landlord, and why?


Do you also do this for other employers?  Who does?  Why?


Think of a situation when someone wanted to do kuulie [casual paid] work, and there was a disagreement about it.  Tell me about that.


Think of a situation where it is routine to do kuulie work.  Tell us who decides about that.


Describe an argument someone had about the payment for either kuulie work, or the land rental share or ‘gutta’ rate. 


Why?  Why?  Who?  Where?  How was it resolved?


Describe another please.


If you are doing some small informal work, please describe who does it and when.


Who decides when to do it; {if difficult} when did you start doing it?  What discussion was held about deciding to do it?


Did you recently drop any other activity?  Why?  Who decided?  Explain how that change happened.


Etc.  prompting till some disagreements are described, and some household-level agreements are described.




What is the most outstanding legal case you can think of?  -- especially in your family if any.  Was there ever a threat of a legal case in your family?  Who promoted the idea of a legal case, and what happened?  What did wife/husband/parents think of the idea.  Discuss.


When did your family last have a quarrel.  What was it about.  Tell us who took what position.


When did you decide to have a child in this family leave school [the most recent departure from school – it could be respondent her/himself]. Who took what position in making this decision?  Was there any disagreement, and who said what?


Describe the view of the child about leaving school at that age.  (What age?  To do what?  Is the child available to comment?)  Do your children work?  Why?  Why not?  Doing what?  Explain the pros and cons of having them do specific kinds of work.








Appendix 2:  Time Costs of the Work


Setting up interviews by relocating respondents from the 1994-5 work:  5 days full-time equivalent.


Each interview:  1 hour plus preparation, 2 hours total.

35 interviews == 70 hours.


Total days for 35 interviews:  15 days full-time equivalent.


Brief introductory questionnaire for each household’s background details, landholding, and opinions about different employment relations, 40 questionnaires X 1 hour = 40 hours. (Neff with two assistants Tejo and Actawallah)


Coding in Excel and putting into SPSS and then into NVIVO as case attributes, 3 hours (Olsen).


Transcription into English:  handwritten version, 5 hours per interview, 175 hours == 21 days.


Creating the typed version, 5 hours per interview, 20 pages per interview, 175 hours === 21 days.

NVIVO software – site license for NVIVO 2 at Univ. of Manchester; NVIVO 7 purchased from Pugh Software on academic staff discount, £141.

Coding 700 pages of text in NVIVO, each page about 5 codes plus some attributes on cases which are put into sets of three types, 20 hours of coding work == 3 days.

Analytical work and iconic modeling, 3 days.

Writing – minimum of 3 weeks.


Total work time:  researcher 11 days (plus 3 weeks writing).

Sub-contractor 70 hours, which is 15 days full-time equivalent.

Typist and translator 42 days in total.








Agarwal, B. (1997). "Bargaining and Gender Relations:  Within and Beyond the Household." Feminist Economics 3(1): 1-50.

Agarwal, B. (2003). "Gender and Land Rights Revisited:  Exploring New Prospects via the State, Family and Market." Journal of Agrarian Change 3(1 and 2): 184-224.

Archer, M. (2003). Structure, agency and the internal conversation. Cambridge, Cambridge University Press.

Archer, M. S. (2000). Being human: the problem of agency. Cambridge, U.K.; New York, Cambridge University Press.

Athreya, B. V., G. Djurfeldt, et al. (1990). Barriers Broken:  Production Relations and Agrarian Change in Tamil Nadu. New Delhi, Sage.

Bhaduri, A. (1983). "Cropsharing as a Labour Process." Journal of Peasant Studies 10(2-3): 88-93.

Bourdieu, P. (1999). The weight of the world: social suffering in contemporary society. Stanford, Calif., Stanford University Press.

Bourdieu, P. and R. Nice (1977). Outline of a theory of practice. Cambridge; New York, Cambridge University Press.

Byrne, D. (2002). Interpreting quantitative data. London, Sage.

Byrne, D. (2004). Complex and Contingent Causation:  The Implications of Complex Realism for Quantitative Modelling. Making Realism Work:  Realist Social Theory and Empirical Research. B. Carter and C. New. London, Routledge.

DaCorta, L., and Davuluri Venkateswarlu (1999). "Unfree Relations and the Feminisation of Agricultural Labour in Andhra Pradesh, 1970-95." Journal of Peasant Studies 26(2-3): 73-139.

Danermark, B. (2001). Explaining society: an introduction to critical realism in the social sciences. London, New York, Routledge.

Duncan, S. and R. Edwards (1999). Lone mothers, paid work and gendered moral rationalities. Basingstoke; New York, Macmillan: St. Martin's Press.

Folbre, N. (1995). Who pays for the kids? gender and the structures of constraint. London, Routledge.

Hodgson, G. M. (2004). The Evolution of Institutional Economics:  Agency, Structure and Darwinism in American Institutionalism. London, Routledge.

Hollway, W. and T. Jefferson (2000). Doing qualitative research differently: free association, narrative and the interview method. London; Thousand Oaks, CA, Sage.

Jackson, C. (2003). "Gender Analysis of Land:  Beyond Land Rights for Women?" Journal of Agrarian Change 2003(3): 4.

Olsen, W.K.  (1997) “Crops, Debt and Labour in Two South Indian Villages, 1994-1995”, ESRC Data Archive Study Number 3927.  120 households and the members’ data on the assets held; the people in each household; the land owned and operated; irrigation; the crops sown; crops harvested; loans taken to buy assets; loans for various production and consumption purposes; and interlinkage of debt with labour and tenancy. One file, with one case per loan, has details of interest rates and all other terms and conditions of each loan.

Olsen, W. K. (1996). Rural Indian Social Relations. Delhi, Oxford University Press.

Olsen, W. K. (2004). Methodological Triangulation and Realist Research:  An Indian Exemplar. Realism and Empirical Research. B. Carter, and Caroline New. London, Routledge (Taylor & Francis): Chapter 6.

Olsen, W. K. (2005). Globalisation, Liberalisation And A Paradox Of Social Exclusion In Sri Lanka. Globalisation:  Identity, Polity and Economy. A. H. Carling. London, I.B. Tauris.

Olsen, Wendy (2005) Complex Causal Mechanisms, Epistemology, and Poverty, mimeo (on request from the author).

Olsen, W. K. (2006). "Pluralism, Poverty, and Sharecropping: Cultivating Open-Mindedness in Poverty Studies." Journal Of Development Studies forthcoming.

Olsen, Wendy, and Smita Mehta, 2005. Methodological Pluralism and Indian Labour Markets, paper for the Development Studies Association conference, September, available online at:  


Swaminathan, P. (2002). The Violence of Gender-Biased Development:  Going Beyond Social and Demographic Indicators. The Violence of Development:  The Politics of Identity, Gender and Social Inequalities in India. K. Kapadia. London, Delhi and NY, Kali Books and Zed Press: 69-141.

Venkateswarlu, D. and L. Dacorta (2001). "Transformations in the Age and Gender of Unfree Workers on Hybrid Cotton Seed Farms in Andhra Pradesh." Journal of Peasant Studies 28(3): 1-36.


[1] Note that the family extends beyond the household and its discussions are relevant.  However the household is the main unit of analysis for selecting individuals to talk to.