University of Southern California Dissertations and Theses

Patronage, dependence and debt in peasant agriculture

(USC Thesis Other)

Patronage, dependence and debt in peasant agriculture

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content PATRONAGE, DEPENDENCE AND DEBT IN PEASANT AGRICULTURE by Gautam Bose A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (Economics) December 1990 UMI Number: DP23368 All rights reserved INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted. In the unlikely event that the author did not send a com plete manuscript and there are missing pages, th ese will be noted. Also, if material had to be removed, a note will indicate the deletion. Dissertation Publishing UMI DP23368 Published by ProQuest LLC (2014). Copyright in the Dissertation held by the Author. Microform Edition © ProQuest LLC. All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code ProQuest LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 4 8 1 0 6 - 1346 UNIVERSITY OF SOUTHERN CALIFORNIA THE GRADUATE SCHOOL UNIVERSITY PARK under the direction of h is Dissertation Committee, and approved by all its members, has been presented to and accepted by The Graduate School, in partial fulfillm ent of re quirem ents for the degree of LOS ANGELES, CAUFORNIA 90089 PhX > t t ' This dissertation, written by Gautam Bose D O C TO R OF PH ILOSOPH Y Dean of Graduate Studies Date 30 M AY 1988 DISSERTATION COMMITTEE Chairperson ii Acknowledgements My interest in the problem explored in this dissertation dates back to at least 1979, when I was a first-year graduate student at Cornell. There the ‘indebted peasant’ and the ‘exploitative landlord’ provided the staple for numerous heated arguments, in which the participants, in various combinations, consisted of Debraj Ray, Shubhashis Gangopadhyay, Kunal Sengupta, Ronnie Dasgupta, and Haider Ali Khan. A lasting interest in economic problems was generated much earlier by our teachers at Presidency College, Calcutta, where many of us had been students together. Later, at the University of Southern California, Professor Jeffrey B. Nugent provided the ideal mix of freedom and prodding encouragement which was needed to think things out, and in particular to write them down in an organized fashion. He did for me far more than was his responsibility as research guide. Tim ur K uran commented extensively on several drafts. Professor Richard Day helped me see around my dogmatism, and opened many avenues of thought. To all of them I owe a debt of gratitude. Most of all, through the often discouraging process, it was the unwavering faith of my parents which kept me going. It is they th at believed th at I was capable, and they, perhaps more than I, th at wanted me to complete the Ph.D. To them , therefore, this dissertation is dedicated. May 1988 Los Angeles ’1 1 1 C o n t e n t s In tro d u c tio n 1 1 T h e S tr u c tu r e o f A g ra ria n R e la tio n s 9 2 In te rte m p o ra lly O p tim a l C o n tra c ts 26 3 In te rlin k e d C re d it a n d O u tp u t M a rk e ts 42 4 S u p p o rt M e ch a n ism s 59 5 C o n clu sio n s 69 R eferences 82 1 In trod u ction Dependence Relations in Agriculture i This dissertation analyzes, primarily in general theoretical term s, a specific kind of relationship between two agents which has been widely observed in agrar ian regions of underdeveloped countries, especially in parts of eastern India. The relationship, which has sometimes been term ed ‘semi-feudal’ (Bhaduri 1973) or ‘patron-client’ (e.g. Rudra, 1984), occurs most frequently between two agents of whom one (the ‘patron’) is relatively powerful in economic term s, and the other (the ‘client’) is poor both in comparison as well as in absolute term s, the rela tionship is highly personalized, and incorporates substantial economic dependence of the weaker agent on the stronger one. For convenience of exposition, we shall freely use the terms ‘patron’ and ‘client’ to designate the stronger and weaker agents respectively.1 Our intention is to develop a consistent theoretical framework which incorpo rates the most general traits commonly observed in these relationships. Thus, on the one hand, it is not claimed that the models presented here are completely consistent with survey data from eastern India; on the other hand they may have some explanatory power in widely different situations, both contemporary and historical.2 1 Several studies over the last decade contribute towards a general picture of agrarian rela tionships, though that picture may not be an accurate representation of relationships in any one specific area. I used the large-scale surveys of Bardhan and Rudra (1978, 1980a,b,c) as the primary empirical resource. Interesting supplementary material is found in Bardhan and Rudra (1986). It seems to me, however, that the interpretation presented at the conclusion of the latter paper con tradicts the impression created in the rest of the paper. For somewhat different conjectures by the same authors, the reader may consult the particularly important work of Rudra (1982), and also Rudra (1984) and Bardhan (1980). For related evidence and discussion, see Sarap (1987), and Binswanger and Rosenzweig (1986). See also parts of Brass (1986) and the references therein. 2As for example in the post-bellum American south (see Alston and Ferrie), or nineteenth- century Russia (see footnote 15 to chapter l). 2 Several aspects of the general relationship analyzed here have received attention in the literature since early in the last decade, and some may be considered to have achieved the status of stylized facts. Among these there are two which are most prominent : the prevalence of sharecropping arrangements between landowners and cultivators as a form of land tenure, and the widespread existence of interlinked contracts in land, labour, credit and output between a landlord and a peasant. Sharecropping arrangements were subjected to a thorough analysis by Cheung (1969). Substantial additions have since been m ade to the analysis. Currently, sharecropping is viewed as a mechanism to distribute production risk between the two agents, and also as a means to provide each agent with the proper incentives in order to bring forth optim al amounts of resources - the two reasons either working separately or together.3 Bhaduri (1973) noted th at sharecropping arrangements in eastern India are fre quently accompanied by recurrent indebtedness of the cultivator to the landlord. He viewed this compound arrangement as an institutional mechanism whereby the landlord extracts the maximum possible surplus from production (which is defined as the excess of production over the peasant’s minimum subsistence consumption needs). While he did not make it clear why such a twofold process is necessary, he did show th at under certain conditions it was sufficient for this end. He es tablished two consequences of this process which are of particular interest. First, the cultivator would remain perpetually indebted to the landlord. Secondly, the existence of this twofold arrangem ent would lead the landlord to resist small tech nological improvements in the production process. Bhaduri called this structure of relationships ‘semi- feudal’, since the peasant is rendered completely into the landlord’s control and becomes dependent upon him for livelihood. A barrage of papers and comments followed Bhaduri’ s article. Srinivasan (1979) argued that his results were based on the implicit assum ption th at credit 3see Braverman and Stiglitz (1982), Eswaran and Kotwal (1985b). 3 was an inferior good for the cultivator, and if peasants are proper neoclassical consumers then the second result would not necessarily hold. Ghosh and Saith (1976) showed th at the specific steady-state solution which Bhaduri had obtained was a ‘knife-edge’ which depended on the rate of interest charged on loans, and if this specific rate did not obtain then the landlord may delay the adoption of an in novation but would not necessarily eschew it. As long as the rate was higher than the knife-edge, however, the peasant would continue to be perpetually indebted. As we mentioned above, it was not clear why Bhaduri’ s landlord needed both a share-tenancy arrangement as well as a credit arrangement to extract surplus from the peasant. If he has sufficient monopoly power vis-a-vis the peasant (which is not an unjustified assumption), then maximum surplus could be obtained by simply m anipulating the term s of one or the other contract (Newbery, 1975). If the quantity of surplus itself is sensitive to the specification of the arrangem ent, however, then a more complicated arrangement may become necessary. This would also hold true, of course, if the dual arrangement was not aimed exclusively at realising maximum surplus, but served some other purpose besides. Braverman and Stiglitz developed the first of these two possibilities in a par ticularly im portant contribution (1982). They observed th at ‘...the worker has considerable discretion over his own actions, and th at, because of the nature of the contractual arrangement between the worker and the landlord, the worker’ s actions have an im portant effect on the landlord’s expected profits...’4 It is therefore necessary for the landlord to introduce mechanisms into the contractual arrangement which will ‘induce workers to behave in the way he would like them to behave.’ Braverman and Stiglitz showed th at such inducements can indeed be created by suitably interlinking share and credit contracts. Further, such interlinking results in an outward shift of the utility-possibility frontier between 4 In general form, this problem has been well known for some time in the game theory literature as the ” principal-agent” problem. the two agents, so th at an outcome is feasible which is strictly pareto-superior to the one in the case w ithout interlinking. We thus have an explanation for the landlord’s use of both credit and share- tenancy, though one which is far different from Bhaduri’s original suggestion. In ; the process of this explanation, however, some of the early questions which were ! raised fall out of sight, two of which seem to be of particular interest. Questions, Arguments, and Assertions The first question relates to the observation th at a peasant normally works for the same landlord for long periods of time - a characteristic which Bhaduri considered a defining feature of semi-feudal agriculture. Bardhan and R udra found this phenomenon to be quite widespread (1978,1980a). A priori, one would expect this in a situation where the landlord is also the normal source of credit for the peasant, because this would make it easier for him to monitor the loan, and also because he would have a selfish interest in the peasant’s solvency while the latter is indebted. Thus while credit may be extended in the first place to provide incentives for the peasant to work, it may in turn become the rationale for a long-term association between landlord and peasant. This explanation is viable, of course, if increasing the length of the contract does not reduce its value to the landlord. Once we explicitly adm it multiple-period arrangements, however, further con siderations m ust be taken account of, which relate to the fact th at ‘the worker has considerable discretion over his own actions.... (which) have an im portant effect on the landlord’s profits.’ In the one-period case, where the peasant’s payoff is defined over current effort and current output, this discretion allows him to choose a particular rate of tradeoff between the two - a choice, as we have seen, th at can be affected by offering him credit. 5 It should be noticed, however, th at the one-period decisions which are made will in general affect not only current output but future output as well, since they determine the quality of land and other durable assets th at rem ain after the exercise.5 If some effort is needed purely to m aintain the future productivity of the land, or if this productivity is in part determined by the specific allocation of effort between tasks, then the peasant who is on a one period contract is unlikely to apply such effort appropriately. The worker, in other words, may have discretion not only over production decisions but over investment decisions as well — in fact the two sets of decisions may in many cases have to be made jointly. If the peasant’s rem uneration is not dependent on the investment decisions at all, then he is hardly likely to make and implement such decisions in a way which the landlord considers desirable .6 If the peasant must be trusted with these decisions, and is to be offered proper inducements to make and implement them responsibly, then the future productiv ity of land and associated capital m ust enter his objective function in a non-trivial way. Thus his payoff m ust depend not on the current output alone, but on the stream of future outputs. This would be true, in particular, if he were engaged to work the land for a multiple (possibly indefinite) number of periods. As in the one-period case, however, this does not in itself ensure th at the peasant makes his decisions according to the landlord’ s preference. The two agents may well discount future incomes at widely different rates, which is quite likely in view of the usual differences in asset-ownership and creditworthiness of such 5Part of this problem can be resolved by enforcing an input-sharing contract in which the landlord provides land-augmenting inputs (as long as the peasant cannot remarket these), see Eswaran and Kotwal 1985b. But part of the tradeoff between present and future output is certainly determined by the allocation of effort between different tasks. 6The inseparability of production and investment decisions, on casual reflection at least, seems more likely in labour-intensive, small- farm agriculture than in large-scale mechanised ventures, and least likely in modern industry. It would be interesting to determine how far this specific char acteristic of the production process is responsible for institutional and organizational differences between regions dependent on one or the other type of venture. 6 individuals. The following question therefore arises, is it possible for the landlord to offer inducements in order to affect the peasant’s investment choices and bring them in line with his own preferences? This question is addressed in chapter 2. We show th at the answer is in the , affirmative, and th a t the instrum ent used to affect the peasant’s choices is in fact the credit contract duly interlinked with the cultivation contract. A second problem arises at this point. Suppose th at, if the peasant did work for multiple periods and was offered the proper inducements, then he would in deed make the ‘right’ investment decisions. However, the landlord has no way of knowing th at the peasant actually intends to work for a sufficient length of time, since a more attractive income opportunity may present itself to him at any point. Such opportunities may include migration to a nearby booming urban area, or the possibility of utilizing newly available technology to multiply production on the peasant’s own plot of land (which was perhaps hitherto insufficient to support his family). Indeed, the peasant may himself be aware of such opportunities, and plan to leave the landlord’s employ at an appropriate time, in which case he would obviously be unconcerned with the future productivity of the landlord’s land and capital. The problem then is to find a mechanism which would ensure th at the peasant is in fact unable to make use of such opportunities, whether expected or not, thus effectively enforcing a long-term contract. In chapters 3 and 4 we show that such a mechanism does indeed exist, and the credit arrangem ent is its prim ary vehicle. Its effectiveness, however, depends more on specific institutional characteristics - such as the timing of transactions and the particular goods transacted - than on market variables such as prices' and rates of interest. Further, it is im portant whether the peasant has access to an anonymous (‘open’) market, and the the eligibility to trade on it. Alternatively, the implementation of the mechanism requires constraints to be placed on the 7 specific commodities (especially money and durable assets) to which the peasant has access, which in turn restrict his economic options. If some agents systematically acquire the power to make certain economic op tions inaccessible to some other agents, then this may serve some or even most of the purposes of ‘extra-economic’ power. In a working equilibrium, if coercion does not in fact have to be used, then the functioning of this system may display fundamental resemblances with th at of social structures based on extra-economic coercion. Also, at least in a static context, it may be meaningful to designate agents as ‘stronger’ and ‘weaker’, irrespective of whether the categories are sus tained by economic or extra-economic means. We are concerned here with the economic characteristics of the landlord- peasant relationship in question, and the functioning of the working equilibrium. We find th at the equilibrium is sustained by an institutionalized pattern of in teractions which emphasizes the systematic exercise of personal control by some agents over others, where person specific constraints are placed on the second type of agents. Such agents find their access to specific markets and goods restricted as a result, even where the lack of means may not have restricted them . It is an interesting question whether the functioning of this arrangement nevertheless is most adequately explainable in terms of categories developed in the context of other socio-economic structures, ones in which agents customarily face different kinds of restrictions, determined through more standardized transactions. In its most comprehensive form, the question is one of determining the mode of production in this type of agriculture. The analysis presented here does not allow us to address this question, since our investigations are not cast in term s of class-relations. However, the question cannot be avoided entirely, since it is intim ately connected with some of the problems we do address. We have therefore entered into the discussion and offered comments which seem relevant given the rest of our analysis, though these should be considered tentative and subject to 8 revision in light of more comprehensive study. This discussion is contained in parts of chapters 1 and 5, and is centered around the designation ‘semi-feudal’ which Bhaduri applied to this type of agriculture. C hapter 1 also contains a more general and detailed introduction to the con- . text under discussion, and presents the hypotheses proposed in this dissertation. Chapters 2 and 3 present analytical models intended to illustrate some of the propositions, and chapter 4 discusses the nature, purpose, and sufficiency of the institutional constraints referred to above. Chapter 5 summarizes the argument, and derives certain consequences regarding development processes. 9 C h a p t e r 1 T h e S t r u c t u r e of A g r a r ia n R e l a t i o n s 1.1 Economic Transactions and Personal Dependence There is no one rigid mold to which all patron-client relations conform. The specific characteristics differ widely over time and place. However, some character istics are sufficiently widespread that we can use them to define a prototype, and conditions actually obtaining in empirical situations may be considered variations on this prototype. Nor do these variations between them exhaust all economic relationships to be found in underdeveloped agriculture, even in the specific areas we have mentioned. But they do cover a significant proportion of interactions to m erit analysis as a category.1 Some aspects of this prototype have attracted substantial analytical attention in the literature. Primary among these is the interlinking of market transactions and the prevalence of share- tenancy. In share arrangements, the landlord often fills the role of patron, while the share-tenant conforms to the image of the client or weaker agent, though the contrary has also been observed. Along w ith the phenomenon of interlinked markets, attention has focused on the fact th at the i For interlocking markets see Bardhan (1980). For patron-client relationships see Rudra (1984). See also Rudra (1981). 10 peasant tends to be perpetually or recurrently indebted to the landlord. It has also been widely noted th at the pattern of economic transactions between the same two agents tends to be repeated for long period of time, often indefinitely, but less analytical attention has been paid to this fact. The existence of relationships which display these characteristics is often found to coincide with a certain profile of economic and technological development. It is a common observation th at, in these regions, the positive fruits of economic de velopment flow entirely to the more powerful agents, bypassing the weaker section of the agrarian population. Technological innovations, where adopted, are found mainly on the self-cultivated lands of richer landlords, who effectively corner all increases in rural income. As a result, while income expands, it makes no inroads on poverty, and inequality of income increases steadily. This, for example, has been the experience in eastern India following the promotion of ‘Green Revolu tion’ - type technologies.2 It is the ‘interlinking’ aspect of economic transactions th at has perhaps attracted the most attention in recent times. Interlinking is said to occur when a transaction between two agents in one m arket (say land- lease) is concluded on the condition th at another transaction in a separate market (say credit) be also concluded between the same two agents, simultaneously or at another time. Thus a landlord may rent land to a peasant only on condition th at the latter also acquire necessary credit from him at pre-specified term s, or a moneylender-cum-trader may extend credit to a cultivator only if the latter agrees to market his output through the former.3 Indeed, it may be a fairly common 2See the article by Ajit K. Ghosh on West Bengal in Ghosh (ed.1983), and the Article by Bardhan in Mellor and Desai (ed. 1985). 3See Bardhan and Rudra (1978, 1980a), Sarap (1987) p.95-97, Brass (1986). While the last paper deals primarily with the state of Punjab, the author quotes several studies from other parts of the country which report interlinking of credit and labour contracts. A similar pattern of interlinked transactions was prevalent between pearl-divers and ship- captains in the Bahrain Gulf, until the pearling industry declined in the 1930’s. The divers were assured of credit at the beginning of the fishing season by captains who were familiar with their skills and wanted to ensure the use of their services, and the captains were able to retain a lien on the divers’ share of the pearl-harvest (see Datta and Nugent). 11 occurrence th at all of the external economic transactions of the client’s household are conducted with or overseen by the patron (Rudra 1984).This last phenomenon is not unique to eastern India, it has, for example, been noted in the American 1 south (Alston and Ferrie). A possible motive behind such arrangements on the part of the stronger agent is immediately apparent, particularly in view of the unequal economic power of the two agents. The weaker agent may be reduced to a state of complete personal dependency on the stronger, thus enabling the latter to exploit his labour and other productive resources to the full (Bhaduri 1983, chapter 2). In a seminal paper, Amit Bhaduri presented a model of interlinked transactions which pro duced exactly this result, and induced the patron to resist technological change in agriculture (Bhaduri 1973). B haduri’s thesis has come under much criticism, for reasons which we shall briefly recount in chapter 3. An alternative track along which explanations have been sought is th at of incentive compatibility between labour supply and payoffs. In an im portant theo retical contribution, Braverman and Stiglitz (1982) argued th at certain patterns of interlinking between land- tenure and credit arrangements produce greater work- incentives for the cultivator. Thus such arrangements reduce the probability that he will shirk on the job, and eliminate the need for the patron to oversee his duties. This leads to higher effort and a pareto- superior outcome overall. Incentive compatibility, along with risk aversion, has born the brunt of ex planation for share-tenancy arrangements. Share contracts distribute production risk between the landlord and cultivator, and provides a positive link between the cultivator’s effort and his payoff.4 In the above explanations, the incentive compatibility problems are phrased in the context of the cultivator’s labour-leisure tradeoff in single periods. The problem involved relates to the possibility th at he may shirk work in the current 4Cheung 1969, Newbery 1977, Kotwal 1985. For a survey see Kaushik Basu 1984 ch.10. 12 period if the structure of the current contract does not incorporate an adequate penalty. Conversely, the purpose of the contract is satisfied even if the dealings between landlord and cultivator are term inated after the current period. Repeti- . tions of the arrangement constitute a concatenation of self-contained one-period contracts, and are not linked over tim e in any essential way. The explanations for share- tenancy and interlinking therefore do not provide an insight into the other prom inent observations about patron-client relationships - namely, perpet ual indebtedness, and long-term associations between the same pair of agents. In other words, these explanations do not contribute to an understanding of the personalized nature of interactions. (But see Eswaran and Kotwal 1985b). One hypothesis is th at the patron usually has close personal ties with the client, as well as private information on his economic condition, and hence incurs fewer costs in the monitoring and enforcement of contracts. The personal nature of ties is thus quoted as a cause of the long-term nature of contracts. On reflection this seems unsatisfactory. W ithin the close society of an agrarian village, prominent villagers usually have intim ate information on all of the lesser residents, or can acquire such information easily. Besides, reputation plays an im portant role in acquiring employment, hence shirkers and other offenders can be punished even if they do not repeatedly seek work with the same employer. Thus, when it is routinely observed th at segments of the village workforce remains primarily associated with the same employer, and develop strong personal ties with him, such partitioning needs to be explained as the result of economic optim ization at some level, and not quoted as a cause of other phenomena ex ante. Indeed, a satisfactory explanation may also contribute to our understanding of the resilience ' of the village socio-economic structure, and its slow pace of change in response to rapidly changing conditions outside. 13 One explanation which has been suggested in passing by various authors seems to us worthy of further investigation. It is th at, in a world of high uncertainty in both production and exchange, long term associations are capable of providing insurance to the agents concerned. Thus a peasant may be assured of finding employment with a landlord year after year, and be able to further smooth out his consumption stream through occasional borrowing in lean years. The landlord, on the other hand, is assured of appropriate supplies of labour at appropriate times, particularly in peak seasons when the labour market may be exceptionally tight. (Bardhan and Rudra 1980a, Bardhan 1985). Such an association, and consequent necessary allegiance developed by the client towards the patron, may produce further benefits for the latter - not least of which is the status of such patrons in village society. It also puts in his hands substantial ‘extra- economic’ power to coerce and dominate other members of th at society. Influential members of the village elite are thus able to restrict the impact of outside agents on the village by ensuring that local people do not cooperate with them , and also to enforce cooperation in other m atters of local policy which is also determined by th at elite. Use of such compulsion is a widespread phenomenon in Indian villages, and is repeatedly reported in the press. An im portant aspect of such long-term associations, however, has been almost entirely neglected in the relevant literature. It relates to the question of moral hazard in the use of durable productive resources. In a 1985 paper, Eswaran and Kotwal argued that owners of productive resources (capital) are likely to hire workers, rather than the reverse, owing to the existence of limited liability which limits the extent to which workers can be held accountable for damages to the capital. The owner can circumvent this moral hazard problem by retaining supervisory and managerial control of the production process, and consequently of the use of the resources (Eswaran and Kotwal, 1985c). 14 In the agrarian arrangements which are our concern, however, the peasant often works w ithout supervision, makes decisions - or at least implements them - independently, and in several ways enjoys the confidence of his patron. And yet production is overwhelmingly carried out on the patron’s land (the most valuable productive resource), often using the latter’s livestock and implements, and in many cases with inputs such as seedgrain and fertilizer also being provided by him. 5 This, moreover, is in a context of high production uncertainty, which may well make it impossible to judge the extent of input use from the output generated. The incentive problem mentioned earlier stems from the fact th a t the worker uses the owner’s property in his productive activity. W ith a limited-time involve m ent, the worker therefore is concerned only with the current productivity of th at capital, and not in the future returns after th at involvement is term inated. The pattern of capital use which is optimal from the worker’s point of view, therefore turns out to be suboptim al when viewed from the owner’s perspective.6 An agent who has a perpetual right to a share of the residual returns from a productive resource will behave as if he had a lien on the ownership of th at resource. It is our conjecture, as we shall argue in detail in chapter 2, th at an interlinked share and credit arrangements with perpetual renewal of contracts produces exactly this result-that the peasant treats the land and associated capital as if it were his property,and as if his subjective disposition towards the tradeoff between present and future output were the same as th at of the landlord. Hence he does not need to be supervised in order to circumvent the inoptim al use of resources. While we will not pursue it explicitly, this track may also shed some analytical light on the special structure of property rights in land which has existed in India 5 For details on sharecropping arrangements see Bardhan and Rudra (1980c). 6Eswaran and Kotwal (1985c). See Brock (1971) for a theoretical framework and results. Mitra (1983) presents a generalizarion of this framework to the case of irreversible investment. See Alston, D atta and Nugent (1986) for a discussion of the connection between present effort and future output in agriculture. 15 for centuries. It has been observed th at exclusive private property in land does not exist either in principle or in practice, and uniform efforts to impose such property rights has consistently m et with failure since early British times. Eric Stokes has called the structure of property rights ‘m ultiplex’, with different agents, including the state, holding different rights to the land. Nor can any of these rights be revoked by an ‘owner’, and in many cases the holder cannot alienate the rights either.7 If we take this intertem poral incentive problem into account, then an indef initely repeated personalized patron-client arrangem ent turns out to be pareto- superior to limited-time anonymous contractual arrangements. This is in the sense th at, under such an arrangem ent, the landlord makes a larger payoff than under any alternative arrangements, while the peasants payoff is at least as large as it would have been otherwise. The comparisons are made between the capitalised income stream s accruing to the two agents over the durations of the alternative arrangem ents. However, the solution is tenable only if one or both agents can guarantee th at the arrangem ent will in fact persist over a long time. In partic ular, if there is a significant probability th at the poverty and uncertainty of the peasant’s condition will be m itigated at some future tim e, then the solution would not be enforceable. This is because the peasant will foresee the term ination of the arrangem ent and switch to a short-term strategy, and the employer (assum ing he has knowledge of the peasant’s prospects) will be forced to revert to an arrangem ent incorporating supervision. A similar result would of course obtain if there is a significant probability th at the employer may want to term inate the arrangem ent in the near future. Such a turn of events would be prom pted primarily by technological and organizational changes in production. For example, a consolidation of small plots into large farm s, together with the introduction of a cultivation pattern involving team s of 7Stokes (1978). See also Neale’s paper in Fryckenberg (1969) (ed.). 16 workers, perhaps incorporating some degree of mechanization, would likely render unnecessary the arrangem ent discussed above. However, the feasibility of such changes in the Indian context may be severely restricted by historically created conditions. Where the landlord’s holdings are highly fragmented and isolated from . each other-as a result of unsystem atic acquisitions and subdivided inheritance- : consolidation may turn out excessively costly. Further, if the landlords are not resident in the village and not closely connected with agricultural activity, they may well be incapable of conceiving and executing an extensive reorganization of the production process.8 If there is a possibility th at one or the other party may want to opt out of the arrangem ent if certain developments occur, then the continuation of the arrange ment would be contingent on the restriction of such developments. Generally, the stability of any socio-economic structure is ensured by institutionalized sanctions on deviant behaviour, backed by a threat of punishment. The most basic sanction, of course, is the threat of violence (police, military, ruffians, etc.), but often the sanctions are encoded in social norms (Moore,1978). In the Indian context, a detailed specification of proprieties, largely encoded in taboos and the caste-system, plays a dominant role in keeping socio-economic agents in their ‘proper’ places. Yet, as traditional society is growing weaker- especially in urban areas, but also to a great extent in the villages-it is a puzzle th at the patron-client arrangement has shown itself so resilient to change. W ith the introduction and spread of markets and mobility, a rapid and complete trans form ation of the socio-economic structure would not have been surprising (if we 8I learnt in a conversation with Dr. Amit Mitra in 1986 that, in parts of Bihar, highest caste landowners often live away from the village and depend on their lower-caste clients to carry on cultivation on their behalf. The intermediate- caste farmers, who are comfortable self-cultivators looking to expand their landholding, often come into violent conflict with the lower- caste retain ers in course of their consistent efforts to acquire land and dimmish the control of the high-caste absentees. In such a situation, the allegience of clients would be highly valuable to the absentee landlords, while it is unlikely that they would be in a position to institute fundamental organiza tional change. 17 are to believe the neo-classical formulation of historical change); it is remarkable th at there are yet large areas which, at best, show only superficial adaptation to a more impersonal market economy, and where the traditional allegiances manage to hold together in rather robust fashion. It is interesting to seek the cause of this resilience in the pattern of transactions and institutions th at prevail in the areas of our concern. Our suspicion is th at it is often the case th at an institutional arrangement, formally similar to (and indicative of) a simple m arket relation, in certain contexts actually serves the purpose of resisting the development of such relations, and instead helps perpetuate a set of pre-capitalist relations of production and exchange.9 In chapters 3 and 4 we have investigated this suspicion in more detail. The contention is th at the institution of indebtedness, recurrent or perpetual, serves the patron as a device to monitor and control the client, and to ensure th at the latter cannot avail of alternative income and employment opportunities which are superior to the arrangem ent discussed earlier. This is achieved not so much by the bonding of labour on account of past debt, though this also occurs frequently, b ut rather by ensuring th at the client does not accumulate sufficient resources to pay the costs of the move. Indebtedness can thus serve to control the economic options of the client even though usurious rates of interest are not charged, but cannot be easily substituted for by other transactions. We are thus able to explain indebtedness and dependence in term s of the asset- distribution and the production conditions that describe the agrarian structure, in addition to preferences and agents’ objectives. Consequently, this may also be construed as an explanation of why such relations are not frequently observed in other contem porary economic structures. 9This assertion is not too far removed from that of the ‘development of underdevelopment’ school. See Aidan Foster-Carter (1978). See also the discussion in Braudel on W itold Kula’s characterization of eastern Europe during the 17th and 18th centuries, in Braudel (1979) pp.270-1. 18 1.2 Semi-Feudalism: W hat’s in a Name? The question of production and exchange relations is closely connected with the ongoing debate on the ‘mode of production’ in specific regions of rural India. A phase of this discussion was inaugurated when Bhaduri designated the system in Eastern India as ‘semi-feudal’. (Bhaduri 1973).1 0 His definition of semi-feudalism comprised three major characteristics, which together led to close personal depen dence of peasant on landlord. This dependence was created through a combination of share-tenancy arrangements, perpetual indebtedness of tenant/peasant to land lord, and a rational reluctance on the part of the landlord to adopt technological innovation. Essentially, by ensuring th at the tenant remained at a low income level, the landlord could use both rent and usury to extract the entire surplus produced by the peasant. Since he would be unable to do this (and hence suffer a drop in income) if the peasant’s earnings increased, the landlord was opposed to technological innovation. The formulation of B haduri’s model has been discussed and debated extensively on technical and empirical grounds. Only a small handful of authors, however, have addressed the im portant theoretical question posed by him -is the structure of landlord-tenant relations semi-feudal by some meaningful definition of the term? While this may at first sight seem a mere terminological quibble, it is not so for two reasons. First, if the designation is meaningful, th at is, if the economic structure under consideration is significantly different from the one analysed by standard economic theory, then a separate analysis must be developed in order to produce effective policy recommendations. Secondly, if the relationship under discussion is suffi ciently sim ilar to a feudal one, then our knowledge of feudal agrarian relations may provide an initial framework to orient th at alternative analytical structure. 10For a summary of an earlier debate on the nature of the mode of production in Indian agri culture see D. McEachern (1976). 19 < Bardhan, who is a consistent critic of B haduri’s hypothesis, has raised ob jections to his formulation which cover both grounds. He has argued th at the contractual arrangements th at have been observed are sufficiently similar to trans actions in full fledged market economies, and are driven by identical optim ization behaviour, to be susceptible to analysis by neoclassical theory. Thus long-term contracts can be explained as optim al behaviour in the presence of uncertainty, share-tenancy by incentive compatibility, personalization as a means of reducing information costs etc. He also points out th at while transactions are impersonal in principle in capitalist society, in fact personal ties often exist between interacting economic agents, but this does not render those transactions feudal.1 1 Besides, Bardhan argues th at the personal dependence which is under discussion is not m aintained by extra-economic coercion, as under feudalism, but rather by eco nomic necessity which is recognized by both the participating agents. Since the political system th at supports the economic relationships is not feudal, the analogy w ith a feudal mode of production is not fruitful in analytical endeavours. There are, of course, several problems even in defining the term s of the ques tion. Feudalism as a mode of production is not defined by individual relationships between lord and serf-though this relationship is basic to it-b u t by a production relation between classes, and a specific organization of state power which supports th at relation. To justify the designation in a different context, one m ust inves tigate whether im portant consequences of that individual relation rem ain valid even when the context is changed, and if alternative mechanisms to support it are available in the new context. Presumably Bhaduri had this in m ind when he added a prefix to ‘feudalism’, for it would not be argued even by the most myopic th at a feudal mode of production actually obtains in India. One could, however, meaningfully argue th at the individual relations are semi-feudal, if it can be shown th at they bear significant similarity to the individual relationship between feudal n See Bardhan (1985), in particular chapter 9. See also Bardhan and Rudra (1978, 1980a). 20 lords and serfs, th at significant consequences of th at relationship is preserved in the present one, and th at this is in spite of the fact th at today’s landlords do not norm ally and legitimately wield extra-economic powers of coercion. It may be noted th at the last clause often contributes more to analytical clarity than to empirical realism. The m ost prominent aspect of feudal structure was its basis in relations of personal dependence. Marc Bloch considered this a central defining feature of feudalism - one which pervaded every level of society. Moreover, dependence and correspondingly lordship was not a voluntary relationship, nor could it be term inated at will. While a lord could, and occasionally did, grant freeman status to a serf, a serf could not in general leave his lord. It was only during the later Middle Ages, when feudalism was in decay, th at serfs could buy their freedom with money. The institutional structure which supported these relations of personal depen dence mainly comprised legal, political and m ilitary elements, though it was clearly not w ithout economic consequences. In this sense the institution of serfdom was ‘coercive’ and ‘extra- economic’ (as B ardhan notes). However, it may be useful to retain in m ind th at the ‘extra-economic’ nature belonged to the institutions which supported serfdom, not to the dependence relations themselves. The latter had economic manifestations which were hardly less im portant than the social and political symptoms. Further, it is im portant to note th at, while the enforcement mechanism was potentially coercive, this does not necessarily imply th at the serfs were always unwilling to m aintain their status. For given the feudal structure, during most of the feudal period, it was certainly preferable for a serf to stay in his estate rather than to defect. W ithin an operating feudal economy, the lord was the natural provider of protection and security to the serf, if also at the same tim e his exploiter. In England, a serf was generally better off than a squatter or cottager, 21; which would be his alternative status were he to flee his m aster’s land. In addition, the serf had certain inalienable rights, prominently the right to a plot of land for cultivation and the right to his m aster’s protection. Even when the feudal order was eroding and a capitalist economy gradually gaining ground, for many of the serfs the static security of the old regime looked more attractive than the ravages of the new. They formed the most conservative group, though they were powerless to stem the tide of change (Moore 1966, ch.l). It may have been beneficial for the serfs as a group to be em ancipated, if such a change were accompanied by appropriate restructuring of economic and political power, in particular if the lower orders seized power themselves. However, the question is counterfactual since corresponding class awareness, organization or action was never present on a wide scale. In their absence, on an individual level, it was not necessarily rational for serfs to seek freedom from their estate, however exploited they may have been. Indeed there is no a priori reason to think th at the serf always and everywhere perceived himself as being exploited-in view of the alternatives open to him within the existing system, he may well have judged his position as relatively comfortable.1 2 1 3 The relation between lord and serf in a feudal context, then, was characterized by a perm anent condition of personal dependence of the serf on the lord, from which the serf did not have a legal option to unilaterally extricate himself. This dependence allowed the lord to set the conditions under which the serf could engage in economic activity and in transactions with other agents. 12Nor does an individual or group’s own perception of the extent to which they are exploited bear direct correlation to assessment made according to an external (perhaps ‘objective’) criterion. See Moore (1978). 13Indeed the position of the serfs fluctuated in response to changes in economic, technological, and demographic variables, though the pattern of this response was determined by the specific structure upon which these forces operated. See Postan (1937), Brenner (1976), Hatcher (1981), and the symposium on Brenner’s paper (Past and Present 1976-79). 22 This structure of dependence relations was itself supported by a set of legal, political and m ilitary institutions, which provided for mechanisms which discour aged defection and punished offenders. These mechanisms were indeed coercive and extra-economic, in the sense th at they imposed physical restraints on the agents. However, the stability of the feudal system did not require the constant ' deployment of these extra-economic mechanisms. Rather, stability was a conse quence of the fact th at, for much of the tim e for an individual serf, his present estate was also the best he could aspire to, given the options open to him. This was because serfs did not have property except by the lord’s grace, and could not find work independently with another property owner as a m atter of course. The permanence and stability of the dependence relations were thus postu lated, in large part, on a restriction of the economic alternatives open to the serf. It follows th at such restrictions could not be removed within the existing socio economic structure as supported by the extra-economic institutional apparatus. * * * 9 We can now pose two questions. First, is there reason to believe th at perm a nent dependence relations benefit the stronger agent (patron, lord) in a variety of institutional structures? In other words, are the benefits derivable from such rela tions valid for a class of economic situations which is more general than the classic case of feudalism? Secondly, if so, w hat are the general conditions under which such relations are profitable, and w hat mechanisms-economic or non-economic- support such relations in the alternative situations? If there exist other contexts, distinct from the historical phase of feudalism, in which perm anent dependence relations are nevertheless valuable, and if, further, there exist mechanisms to support such relations, then the designation ‘semi- feudal’ may be appropriately applied to th at context. We could then investigate whether as a result the operation of economic rationality implies consequences 23 different from those we would expect in situations traditionally considered distinct from feudalism. In particular we could ask whether, as a result, specific constraints or qualifications are imposed on the direction of economic development. The first question is particularly relevant in contexts where a powerful agent is not required - by custom, law or considerations of status - to m aintain dependents. It is known th at, centuries after serfdom had become obsolete in western Europe, the system continued in Europe east of the Elbe, not as a custom ary process, but supported by the constant deployment of force and the conscious segregation of serfs from alternative opportunities and contact w ith the m arket. 1 4 In Russia, peasants who were not formally serfs were held in a position of virtual serfdom by means of debt-bondage.1 5 And in India in contemporary times, several authors note the operation of a patron-client relationship which ties a peasant’s survival and well-being to the perpetual patronage of a stronger agent - usually a larger landowner or moneylender. 14On this we have the opinion of no less an authority than Fernand Braudel. He writes, An economic conjuncture with multiple effects was pushing eastern Europe back, in the early sixteenth century, towards a colonial destiny as a producer of raw ma terials, a development of which the second serfdom was only the most visible sign. Everywhere, though with variations according to time and place, the peasant was being even more firmly attached to the land.... (A)lthough it might be organized differently or m itigated here and there, the rule of six days a week unpaid labour was tending to become established everywhere without exception.... The decline of the cities and the urban markets and the weakness of the state corresponded to the takeover of the labour force (and also of the best land ) which formed the drive behind the success of the feudal lords... (T)he landlords had succeeded in integrating the peasantry into closed economic units, sometimes very large ones.... These economic units were self-contained. The peasants had virtually no access any longer to the urban markets, which were in any case much reduced. When he did come to town, it was only for small scale economic transactions... Braudel (1979) pp.267-270. Also see Brenner (1976, 1982) and Klima (1979). 15<In Russia, it was the indebtedness of the peasantry which made it possible for the nobles to draw up contracts for their tenants pinning them down to one estate, a sort of “voluntary enserfment” as it has been described, which was later legalized.’ Braudel (1982) p.267. 24 The conditions which make such relations profitable of course vary from context to context. In Europe, during the sixteenth century and later, the grain trade flourished continuously because of the increases in wealth and property which accompanied the expansion in urbanization, industrial production and the market. But this same process also opened up potential opportunities of migration and alternative employment for the rural population. In order to reap the profits from the grain trade, the landowners east of the Elbe therefore had to devise non-market means to contain their workforce. The Indian example does not fit into the explicitly coercive mold. The peasants here are not forced into dependence by physical restraints, but by economic neces sity. Given th at they own negligible amounts of assets, and face a subsistence-level income stream which is, moreover, subject to high production risk, it is considera tions of survival which lead them to seek the patronage of a prosperous benefactor. It is im portant th at benefits of such patronage be always available in times of need, irrespective of the peasant’s recent credit history and asset position. It has been rightly observed th at in most cases the only collateral th at the peasant can borrow against is his future capacity to work. In such a situation it is rational for him to provide assurance to his patron that he will not default in the future. Such assurance may be a function of the reputation which he has built up with his benefactor, or a consequence of conditions which restrict the feasibility of default. The precariousness of the peasant’s condition, and his corresponding need for patronage, is crucial in determining whether coercion is or is not necessary to cement a relation of dependence. If he can indeed not survive w ithout patronage, then coercion is unnecessaxy, even for a patron who derives positive benefits from dependence. Since it is rational for the peasant to prove his loyalty, he need not be supervised at work, as long as there are observable indicators which reflect his diligence, even though these indicators may only be observed with substantial lag. 25 The peasant would thus work with the diligence appropriate to a member of a p atron’s ‘family’, with obvious payoffs for the patron. It would not be surprising, then, if we found th at the landlord or patron has ' available to him a mechanism which enables him to m aintain the peasant in a precarious condition, one in which the peasant necessarily continues to remain dependent on the lord. We indicated earlier th at recurrent indebtedness provides precisely such a mechanism. The effectiveness of this mechanism, however, de pends on the specific physical form of the dealings between the two agents, not merely the values th at are transferred between them . It would be a fallacy to analyze this economy under the implicit assumptions of generalized exchange and commodity production-it is the restriction of these characteristics th a t give rise to the specific relations of power and subordination between patron and client. In the next chapter, we present a formal model which illustrates the optim ality properties of a long-term interlinked contract. The question of debt as a controlling mechanism is taken up in chapters 3 and 4. The concluding chapter discusses some possible consequences such arrangements may have on the direction of economic change. 26 C h a p t e r 2 I n t e r t e m p orally O p tim al C o n t r a c t s In the type of production arrangements discussed in the introduction, it is usually the rule th at the peasant works on the landlord’ s land, often using the latter’s livestock and other ’ ’capital” . This creates a possibility of potential moral hazard, since the peasant may m aintain a high productivity at the cost of exces sive depreciation of productive equipment and land. If, however, he expects his future stream of income to be exclusively associated with the time-profile of the productivity of this stock, then it is to his interest to use this stock in a pattern that is optim al for a long time-horizon, perhaps as long as th a t for which the landlord would himself plan. Essentially, the peasant could be expected to treat the landlord’ s property as if it were his own, insofar as planning and care in use are concerned.1 In the present chapter we shall show, within the context of a schematic model, th at this is indeed generally true, and that the landlord can set the term s of the contract to ensure th at an optimal production pattern is adopted. However, this requires him to interlock the land-tenure contract with a credit contract. The basic model is set out in section 2.1. Some useful standard theorems are quoted in section 2.2. Results are also derived in section 2.2. 1This supposition is theoretically sound, see Eswaran < k Kotwal (1985). 27 2.1 Production and Contracts Our concern is to determine an optim al land-tenure arrangem ent between a landlord and a peasant. The peasant is employed to produce output using the landlord’s land, livestock and implements, which we shall assume can be aggre gated into a homogeneous quantity called ‘stock’. Production is carried out by a single peasant (who perhaps uses family labour). This is widely true in many less-developed agricultural areas, and may be a consequence of existing farm-size and technology. The stock or productive capital may vary in quality. Stock of a given quality x is capable of producing a well-defined maximum output g(x) at the end of the current period. g(x) is stochastic. The current output actually realised may be lower than this maximum, depending on the intensity of cultivation. The intensity of cultivation also affects the quality of the stock. Thus if the maximum output g{x) is elicited, this depreciates the quality of the stock (or sim ply, the stock) by a constant factor d. On the other hand, if the land is allowed to lie fallow and hence no output is produced at all, stock is augmented by some determ inate factor. An intensity of cultivation between zero and the maximum produces a corresponding change in stock. In other words, of the maximum pos sible current output g(x), a part may be sacrificed to counteract depreciation or even to add to stock. The choice of the intensity of cultivation is thus at the same time an investment decision. The difference is th at the investment is not m ade out of current output after it becomes available. Rather, the level of investment in the next period is determined prior to production in the current period. Production and investment are not separated physically, but codetermined by the intensity of cultivation, which is the single choice variable. 28 We shall assume th at, once this decision is made, it fixes the level of investment, while output is stochastic depending on the state of nature realised in th at period. Denoting output by y, investment by 2 , and stock quality by x , we can then write output in period t as a function of the stock in periods t and t + 1, yt = h ( x t- i,x t,dt-i) (2.1) where 9 is a random variable which is independently and identically distributed in each period. Thus y * is also random dependent on 6. x t+i = dxt + zt (2.2) Where d is the rate of depreciation of stock. Clearly the function h(.) implies a transform ation between current output and investment. In w hat follows, we make the simplifying assum ption th at equal amounts of output foregone produce equal augmentations in land quality, inde pendent of the initial productivity of the land. O utput units can be defined such th at this linearity constant is normalised to one. Thus * = (ff(*) - E v) (2-3) The model is a special case of th at used in the theory of aggregative growth w ith irreversibilities in investment (see M itra, 1983). If investment in each period is assumed to be m ade explicitly out of realized output, then it would be observ able to an agent who is not himself making the allocation (e.g. a non-participating landlord). We interpret investment as being m ade implicitly through the decision on the intensity of cropping. Since output is stochastic, the actual level of in vestment or land augm entation is not observable to a non- participating agent, even over several periods.2 This introduces the potential for m oral hazard, and a principal-agent game, since the cultivator can adopt a m ulti-period program which is different from th at which the landlord considers optimal. 2 This is strictly true only if the range of y is unbounded above. Otherwise overcropping may be detected in a very good season, if the value of y realised is too high. 29 P r o d u c tio n Productive activity from any period t yields output in the following period . t + 1. Changes in stock consequent on the output decision also become effective ! in the next period. Gross potential output, which is expected current output E y plus additions z to (depreciated) land quality , is denoted g(x). The following . conditions obtain: (T .l) f(0) = 0 (T.2) g(x) is strictly increasing, continuous, and concave for x > 0, and differentiable for x > 0. (T.3) d ^ hmz— to o Q (^) ^ 1 d hm x —. q g * ^x) — oo. (T.4) 0 < d < 1 Given an initial stock x a feasible program for a T-period horizon is defined by a sequence {< x t,y t+1,z t + 1 > , t = 0, ...,T } which satisfies (F .l) x 0 = x (F.2) g{xt-i) = E yt + zt t = l ,2 ,...,T (F.3) x t+1 = (1 - d)xt + zt+i t = 0 , l , . . , T - l (F.4) (xt,y t+1,zt+i) > 0 t = 0 , l , . . , T - l In w hat follows, we want to focus specifically on the problem of moral hazard associated w ith the unobservability of investment and stock. While this aspect of the production arrangement derives in part from the uncertainty in production, it is convenient for us to abstract from adjustm ents owing to uncertainty and risk aversion per se. In order to do this, we shall make the following assumption: A s s u m p tio n 1 Each agent forms point expectations regarding future streams of output based upon the information at his disposal, and formulates his optimization problem on the basis of these expectations. 30 Henceforth we shall simply write y for E y etc., since these are the only relevant variables in the agents’ plans. The entire argument th at follows is in term s of ex ante plans. Also in order to concentrate on the problem at hand, we want to abstract from the problem of shirking on the part of the peasant. We therefore ! ; assume th at cultivation, if undertaken at all, requires one unit of effort per period, I regardless of the intensity of cultivation. S p ecifica tio n o f C o n tra cts A contract between the landlord and the peasant consists of two parts, (a) the production arrangem ent, which specifies the rules for output- sharing (rent, share or wage) as well as the extent of direct supervision to be undertaken by the landlord, and (b) a credit agreement, including the rate of interest on loans and the rules for repaym ent. In addition, the length of the contract m ust be specified. Thus A contract is defined by the following (superscripts N and L denote the peasant and landlord respectively): (i) A positive integer T which defines the length of the contract. The peasant works in the periods 0 ,1 ,..., T. (ii) A pair {a, ft) which defines output shares. (iii) A supervision sequence s = (st, t = 1 ,2 ,..., T), where st = 1 if supervision is undertaken 0 if not (iv) A rate of interest r at which the peasant may borrow. The peasants share of output in period t is h f = ayt + /? 0 < a < 1 ft £ R (2.4) This includes rent (a = 1, /? < 0), share (0 < a < 1) and wage (a = 0, ft > 0) contracts. Note the restriction to linear schemes.3 The landlord’s share is 3see Eswaran and Kotwal (1985b), and Gangopadhyay and Sengupta. 31 t i = (1 - a)yt - /?. (2.5) A s s u m p tio n 2 The cost of supervision is q in each period t in which st = 1. Thus a contract is completely specified by a pair (Q, T), where Q = {o:,/?,s,r}. The peasant works in the periods 0 ,1 ,2 ,..., T. O utput becomes available and shares are paid at the beginning of periods 1, 2 , T + 1. The peasant may take out loans in periods 0,1, ...,T , and repay them in 1,2, ...,T + 1. Let kt be the amount the peasant borrows in period t. For convenience we shall assume th at this am ount m ust be repaid with interest in period t + 1, so th at his net borrowing in t + 1 is given by lt+i = kt+i — (l + r)hf t = 1,2,..., T + 1. (2-6) lo = k0 (2.7) The following conditions are imposed : (L.l) £*>0 t = 0 ,1 ,...,T + 1. (L.2) kx+x 0 i.e. the peasant cannot lend to the landlord, and he cannot have an outstanding debt at the end of the contract period. Notation: We shall denote the sequence (xo, x u by the letter x (similarly k ,c ,y , etc.). For a given feasible sequence of stocks x and sequence of loans k satisfying (L.l) and (L.2), the peasant’s disposable income in each period t is then given by C o = lo (2-8) ct = h ? + l t t = 1, ...,T + 1 . (2.9) The landlord’ s income is the residual 32 T T o = - l o (2.10) 7 c t = yt - c t t = 1,2,..., T + 1. (2.11) We restrict the peasant to non-negative incomes in each period (implicitly, the peasant is assumed to consume only his disposable income in each period). (C .l) c > 0. This assum ption together w ith (L.2) above eliminates the possibility of bankruptcy. O b se rv a tio n 1 The peasant’ s disposable income is continuous in a and (3. Also, from the constraints on borrowing and consumption it follows that if a = 0, (3 = 0, then the peasant is restricted to a zero consumption in each period. Given the contract (Q, T ) and a pair of feasible vectors x, k, each agent’s payoff sequence is completely determined. The stock sequence, in turn, is determ ined by the initial stock xo and an investment sequence z. We can therefore write the consumption and profit sequences as functions of these, c = c(Q, T, z ,k ) 7 r = 7r(Q, T, z, k ) Given the contract (Q, T), the choices of z and k axe m ade as follows : in each period t such th at st = 1, the landlord chooses zt subject to the feasibility constraints, while this choice is m ade by the peasant in the rem aining periods. These two subsets of {zt, t = 1 ,...,T + 1} are w ritten z L and z N respectively. The vector k is chosen by the peasant subject to the constraints (L .l), (L.2) and '(C). The landlord’s choice variables then consist of the contract and the vector zL and the peasant’s choice variables are z N and k. In order to determine their optim al choices we must now introduce their objective functions and some further constraints. 33 O b jective F u n ction s Let fj, and S represent the discount factors of the landlord and the peasant respectively. The peasant has a utility function tt(.) defined over his disposable income in each period t. His objective function is the discounted sum of the one- period utilities. r+ i U(Q , T ,z ,k ) = ' £ e 'u (c ,(Q ,T ,z,k )) (2.12) t= 0 Standard assumptions are imposed on the one-period utility function u, (U .l) u(c) is continuous, twice differentiable, and strictly concave for c > 0, and strictly increasing for c > 0. (U.2) u'(c) — > oo as c — ► 0. (U.3) The peasant has a reservation utility UT, which is greater than th at afforded by a consum ption of zero in every period. The landlord’s objective function depends on the stream of payoffs he receives over the duration of the contract, as well as on the value of the stock th at remains in the T + 1th period. The reason is obvious - his earnings from the land in future periods depend positively on this stock. Let u (x r+i) represent the valuation of the stock remaining at the end of the contract. The landlord’s objective function is assumed to be the linear discounted sum of the one-period payoffs and the term inal stock, T +1 n [ Q ,T ,z ,k ) = ^ /i* 7 rt(Q ,r,2 ,fc) + » T+1v{xT+1) (2.13) t=o A s s u m p tio n 3 (fi > 6). 4 If z(Q ,T ), k (Q ,T ) are optim al choices m ade by the relevant agents under the contract (Q ,T ), then we write 4 This assumption, as well as assumption 4 are in keeping with the asymmetry of the problem. We would obtain a similar result without these assumptions if we allowed the peasant to lend to the landlord as well. 34 i r ( Q , r ) = n ( Q ,r , *(q , t ), *(<?,r)) (2.14) In order to assign a definite form to u(.), we assume th at, in period T + 1 , the landlord will enter into another T-period contract (Q i, T), spanning the periods T + 1 to 2T + 2, with an identical peasant. This is followed by another contract (Q2, T), and so on. The stock at the beginning of each contract is given by the stock at the end of the previous one. Then v(x T) is given by: OO v(xT) = m a x ^ /A ^ r+1^n(Q,-, T, Zi,kx) (2-15) «=i where variables relating to the *-th contract are superscripted by i, and the max imization instrum ents are the contracts (Q,-,T) and the vectors z L corresponding to the i-th contract. The optim ization problems for the two agents are therefore given: MAX I (Peasant) : choose (zN/Q ,k ) to maximize (2.12). MAX II (Landlord): choose (Q ,z L) to maximize (2.13) such th at U *(Q ,T) > UT. 2.2 An Optimal Contract The plan of the rest of the chapter is as follows. Given a contract (Q, T ), we can determ ine the optim al choices of the two agents, and hence compute the values . attained by their objective functions. Further, if we define a reservation value for the peasant, we can compute the maximum value which the landlord’ s objective function can attain, given the contract and a reservation constraint. We show th at this maximum has a well-defined upper bound. We then define a contract which utilizes both output-sharing and credit, and show th at this contract yields a value ‘close’ to the upper bound. Further, as the length of the contract gets larger, the value of the contract asym ptotically approaches the upper bound. Finally, we argue th at a contract which does not incorporate both output-sharing and credit m ust yield a lower value than the one we defined. 35 S ta n d a rd R esu lts It is convenient, at the beginning of this section, to state a few standard results which hold true in our model. For proofs, the reader may consult M itra (1983). Consider the problem (*): ( * ) Given xq choose x = X j , . . . , x t + i to maximize T Y , tfW V t) 0< p<l t= 1 where (i) the sequence x satisfies feasibility conditions, and (ii) < f > is a continuous, strictly increasing, concave, and twice differentiable (utility) function. We note the following results: T h e o re m 1 Let x * (T ,p ) solve the problem (*). This sequence exists, and is unique. T h e o re m 2 There is a feasible infinite sequence x* (p) starting from x 0 such that, as T — ► oo, for each t, x *t (T ,p) — > x *t (//)• T h e o re m Z As t — * oo, x * < (Xo,p) — > x^ where x^ is defined by b '(^ M ) + (! ~ d)\ = x- This limiting value of the optim al stock is the golden-rule (cf. Solow, 1956) modified for the discount factor p. Note th at the golden rule does not depend on the precise form of the utility function, but only on the discount factor. The infinite sequence x *t [p) which leads to the golden rule is not independent of the precise form of the utility function. We shall call this sequence the golden rule path. At this point we shall make a further assumption which is somewhat restrictive : A s s u m p tio n 4 x 0 < x(p) 36 The sequence of necessary and sufficient Kuhn-Tucker conditions th at characterizes the solution to (*) is given by: 1) <f>'{yt) > n[g'{xt) + (1 — d)] zt > 0, t — 1, ...T. (2.16) < p[sf(xt) + (1 - d)] zt < g{xt- 1), t = 1, ...T. (2.17) < t > ’{yt+ i ) T h e A g e n ts’ O p tim iza tio n P ro b lem s Consider MAX I, the peasant’s maximization problem. The necessary and sufficient conditions which characterize the solution are the following: u'{ct) u'(ct+1) tt'(Cf) < 6 (l + r) kt > 0 i = 0 ,l,..,T . (2.18) «'(Ci+l) u'(ct) > 8\g'(xt) + (1 - d)] 0 < Zt < g(t-i) t / s t = 0. (2.19) < 8[g'{xt) + (1 - d)] 0 < zt < g{xt- 1 ). (2.20) u'(ct+1) and the appropriate complementary slackness conditions. The first determines the optim al loans and the other two determine the choice of investment in those periods in which supervision is not undertaken. O b s e rv a tio n 2 (K.S) and (K.4) are relevant only if the share-rate a > 0. If a — 0, then the output sequence (and hence the investment sequence) does not enter into the peasant’ s objective function. As we noted in the previous section, the landlord’s objective function is an infinite-dimensional one, of which only the first T + 1 periods are relevant to our present problem. Any feasible solution to the infinite- dimensional problem will specify, in particular, the following infinite sequences: a feasible stock sequence {x|} (with an associated output sequence {t/J}, a supervision sequence {s't},and a consumption sequence for the peasant(s) {c|}. This last sequence m ust satisfy the 37 reservation utility constraint for each segment of periods [*(T + 1 ) , (t + 1 )(T + 1)] i = 0 ,1 ,2 ,... Thus the landlord’ s infinite-dimensional profit may be written: 0 ° OO o o n'(Q,,r,a,A:) = Y ^ V t ~ Y ^ ct - Y r f q - s t (2.21) t=o t=o t=o which he maximizes subject to the reservation utility and feasibility constraints. O b se rv a tio n 3 T l'(Q i,T ,z,k ) has an upper bound which is attainable under the feasibility and reservation utility conditions. Proof Consider the three term s in 3 in turn. The infinite discounted sum of outputs y[ has an upper bound which is attained under the program x * (p) described in theorem 2 above. The term involving consumptions is minimized by a unique and well- defined sequence c* which is T-cyclic, cfi(r+i)+*l-=c? * = 0,1,2,.. t = 0 ,1,2, ...T + 1. (2.22) and (cq, c j,..., c f+1) solves (MIN I) M I N £ t T =o1p tct s.t. U {c*0,...,c*T+1) > U T. Given the assumptions on the peasant’s utility function, this has a unique solution, which is characterized by u'{ct)/u '(c t+1) = 6 /p (2.23) and equality in the constraint in (MIN I). Finally, the discounted sum of supervision costs has a minimum of 0, attainable by setting st = 0 for all t. Let y* be the output sequence associated with x*(p). Define T + 1 T+1 n*(T) = £ - £ n ‘c[ + v(x-T+1) (2.24) t= 0 t= 0 i.e., i r ( r ) is the upper bound on the profit from a T-period contract. □ 38 Now consider a contract (Q'0,T ) which has st = 0 t = l , 2 , . . , r - 1 ^r+i = 1 zt+i = m ax{m in[i(/i) — dxT ,g{xT)],Qi) T = (1/^) - 1 with a > 0, (3 > 0. P ro p o s itio n 1 Under {Qq, T ), the peasant chooses kt > 0 in each period t = We shall first prove some simple lemmas. The proof of the proposition will then follow easily. Let c ,z ,x , etc. represent the equilibrium choices under (Q'0,T ). L em m a 1 ; ct > ct+j for t = 0 ,1 ,...,2 \ proof: 6 < p> hence 6(1 + r) < 1, so the claim follows from 2.18. □ L e m m a 2 g(x) — dx is monotonic increasing in x. This is simply a restatem ent of the first inequality in (T.3). L e m m a 3 ; x t < x^ for t = 0 ,1 ,.., T + 1. proof (i) xo < Xft by assumption. (ii) Suppose xt > x „ for some t, 0 < t < T . Then 2.19 applies, and zt — 0 => x t- i > x t > x^. By backward induction we have x 0 > x h, a contradiction. (iii) By (ii) we have x T < x^ => (1 — d)xT < x^ ■ = > zt+i < — (l — d)xT — ^ ^ x^m □ 39 L e m m a 4 : kt — 0 = > x t > for t = 0 , 1 , T . proof Otherwise suppose th at for some t, kt = 0 and x t < x^. Then (2.20) applies hence zt — g{xt~i) = ► ct = (3 - (1 + r)k t - 1 < P < ct+1 = a [ f f ( a : t ) - zt+1) + 0 + A;t + i which contradicts lemma 1. □ Proof of proposition 1 Suppose th at for some t < T we have kt = 0. We shall show th at this contradicts lemma 1. kt = 0 = > • x t > Xp by lemma 4. ®t-i < by lemma 3. Therefore x t > x t~i ct = alg(xt-t) - Z t ] + 0 - ( l + r)kt- i < a[g(xt- 1) - zt] + /3 ct+i = - zt+i] + P + kt+i > cc[<7(xt) - zt+ 1] + (3 B ut then by lemma 2, ct+i > ct contradicting lemma 1. □ P ro p o s itio n 2 Under (Qo,T), the golden-rule path is adopted in periods 1, ...,T + 1. Proof By proposition 1, the constraint in (2.18) is m et with equality in each period t = 1,2, ...,2 \ By definition of r in the contract, we have 1/(1 + r) — p, , so th at (2.19) defines the golden-rule stock if zt > 0 (see theorem 3). But since our initial stock is no greater than the golden- rule stock, and the depreciation rate is strictly positive, so this constraint is not binding, hence the peasant chooses the golden- rule in each period 1 to T. In the final period the landlord chooses the golden- rule, which is optim al for him and clearly feasible. □ 40 Recall th at the peasant’s disposable income in each period is continuous in a and {3 - the share param eters. By (U .l) and the definition of U(.), his objective function is also continuous in these variables. By observation 1 and (U.2), for any value of a between 0 and 1, there is some /3 = /3(a) such th at the value of the peasant’s objective function under (Q'0,T ) is reduced to his reservation utility UT. T h e O p tim a l C o n tract Define the contract (Q*, T ) with s and r as before, 0 < a < 1, and (3 = /3(a). P ro p o s itio n 3 Under (Q *,T), the peasant’ s optimal consumption sequence is (cj!, t = 0 ,1 ,.., T + 1 ) , which solves (M IN I). Proof By proposition 1, (2.18) holds with equality, and by definition of /3(a) we have U(c*) = UT, hence the result follows. □ P ro p o s itio n 4 II* ( r ) - U*(Q*,T) = p T+1q. This is obvious, since the only difference between the two is the cost of supervision in the last period in the second contract. P ro p o s itio n 5 A s T ^ oo, ir(<3%!T) — ► IP (T ). This follows from proposition 4, since 0 < p, < 1. Thus (Q*, T) defines a contract which yields the landlord a profit ‘close’ to the maximum possible profit, which tends asym ptotically to the upper bound as the contract length T gets large. Further, the equilibrium outcome of this contract has the peasant perpetually indebted to the landlord, and working under a share- cropping arrangem ent. These were the characteristics of observed arrangem ents which we were attem pting to explain. 41 It now remains for us to show th at a contract w ithout a positive share-rate or a credit arrangem ent would not yield this result. Suppose a = 0. Then the peasant’s income is insensitive to the current output, hence any choice of investments is optim al for him. The golden-rule, in particular, is also optim al, hence this choice would be an equilibrium response (in the Nash sense) to any contract. But this equilibrium is very ‘weak’. For example, if we assign equal probabilities to each of his optim al strategies (since ‘if it doesn’t m atter then it doesn’t m atter’), then the probability of the golden-rule being adopted reduces to zero. W ith the contract we defined, the golden-rule is the unique optimizer, which makes it much more likely to be adopted. Secondly, suppose no credit is offered. Then the peasant reverts to an invest ment sequence which solves (2.19), with consumptions being defined strictly by (oc,j3) and y. The solution to this is the golden-rule path corresponding to the peasant’s discount factor, which clearly diverges from th at corresponding to the landlord’s (the limit of each path is the corresponding golden-rule stock, and by theorem 3 the two stocks are distinct). Since the golden- rule path is unique, the result m ust be suboptim al for the landlord. The alternative in each of the above two cases is, of course, for the landlord to undertake direct supervision in each period (or in a number of periods which is determined by comparing cost of supervision with the possible loss), bu t th at clearly leads to a lower profit as well. Hence a contract which maximizes the landlord’s profit subject to the peasant’s reservation utility is infinite-dimensional, incorporates credit and output-sharing, and in the equilibrium outcome the peasant remains perpetually indebted to the landlord. 42 C h a p t e r 3 In te r lin k e d C r e d it a n d O u t p u t M a r k e ts In his 1973 paper, Amit Bhaduri proposed a model of a landlord-tenant relation in ” semi-feudal” agriculture, in which the landlord used the dual mechanism of land-rent and usury to extract from the tenant all of the latter’s output save his minimum consumption needs, in the process reducing the tenant to the status of a virtual debt-serf (Bhaduri 1973). It was argued that in a situation where the landlord is able to interlink the m arkets in land and credit to exploit his tenant, he would be inimical to investment and innovation in production in order to protect his economic interest. Bhaduri applied the designation ‘semi-feudal’ to the agrarian structure in spe cific geographical areas (nominally eastern India) based upon four characteristics which this structure incorporates: (i) landlords engage peasants to cultivate land as share-tenants, (ii) the sharecroppers are perpetually indebted to their landlords, (iii) the landlords use usurious rates of interest to extract surplus from peasants, and (iv) landlords are unwilling to make investment and adopt innovations in production since their net income may actually decrease as a result. 43 The designation is appropriate since (a) the peasant is inextricably tied to the landlord by his debt-som ew hat like a serf, and (b) there is little tendency towards technological progress, an observed characteristic of feudalism.1 The topic of interlinked markets has subsequently aroused much interest. Bha- duri’s thesis itself has been much debated.2 Empirical research shows th at inter linking of m arkets is certainly a widespread phenomenon, at least in eastern India.3 B ut authors hold disparate views as to the exploitative content of such relations, their economic significance, and particularly the importance of such relations in prom oting or inhibiting investment and the adoption of innovations (Newbery 1975, Ghosh and Saith). It has also been argued th at the designation ‘semi- feudal’ is im proper because the serf-like status of the peasant is not enforced by extra-economic means (Bardhan 1984). One of the strongest arguments against B haduri’s thesis is th at, if the stronger agent is indeed absolutely superior in term s of bargaining position, then it is un necessary for him to use the tactic of interlinking to extract surplus from the weaker agent. If the landlord is a monopolist in land, and labour is abundant, for instance, then he could extract the entire surplus merely by an appropriate adjust m ent of the rent- or share-rate, and usury would be redundant (Newbery,1975). In such a situation, the result regarding resistance to innovation would not hold. Alternatively, even if both forms of surplus extraction are used, the landlord only has to adjust the rate of one or the other in order to appropriate the entire gains from innovation. Further, it has been observed th at, in some cases, the landlord often grants loans at a rate of interest lower than th at prevailing in the ‘m arket’, and hence the accusation of usury may be unfounded (Bardhan and R udra 1978). 1 At least in the English version, see Hilton (1975, p .178), Mingay (1963, ch.3), Habakkuk (1953, p. 190). 2Newbery (1975), Ghosh and Saith (1976), Srinivasan (1979), Bardhan (1983, ch.9). 3Bardhan and Rudra (1978, 1980a,c), and Sarap (1987) p.94-95 and Table 7. 44 Indebtedness, however, is not only a vehicle of usurious exploitation in Bha- duri’s model, it is also a substitute for extra-economic coercion, and forces the peasant to rem ain in serf-like dependence.(Bhaduri 1983). If this were true, then the situation in eastern India would call forth interesting analogues in history and in the theory of underdevelopment and dependency. In the present paper we intend to show th at debt can be used in order to interlink m arkets, and in the process to extract the entire surplus of the peasant. This am ount may include superprofits (in the monopoly sense). Further, the lender can restrict the peasant’s access to other traders, thus m aintaining the peasant in a state of dependence. This is only feasible if the peasant does not produce ‘too m uch’ surplus, hence there is reason for the lender to be opposed to innovation. If this were true, then it would add to the accusation th at indebtedness is an ‘underdeveloping’ or progress-inhibiting factor in backward agriculture, and th at it sometimes reduces peasants to serf-like conditions. In the situation th a t is investigated, the peasant retains some surplus product after paying land-rent and meeting his basic consumption needs. He is, however, forced to get recurrently into debt to a moneylender, who is also a commodity trader. Initially the moneylender- trader is a monopolist in the credit market, and is able to interlink the markets in a way th at allows him to extract the entire surplus of the peasant. He however has no incentive to block innovations which the peasant may wish to undertake (and the peasant has no incentive to undertake such innovation either). Com petitors are then allowed to enter one or both of the m arkets in grain and credit. It is shown th at, given the initial situation of recurring indebtedness, the moneylender can continue his hold over the peasant if the latter does not become ‘too productive’. Consequently, he now has an incentive to block innovations which the peasant might want to undertake, in order to be able to protect his exploitative profits. 45 We do not suggest any feasible economic strategy which the moneylender may use to block innovation. However, given the social and political realities in the societies where this analysis may be relevant, the absence of an explicit economic strategy may not be an im portant deterrent. Agriculture in pre-capitalist village societies is usually carried out on the basis of substantial interdependence between peasants. If many of them are dependent for subsistence on the same powerful agent, they may be easily threatened into withdrawing their cooperation from the productive activities of a specific cultivator, thus making it impossible for the latter to carry on production. The schematic model is defined in section 1. It is form ulated along the lines of Bhaduri (1983 ch.2). Section 2 describes a fictional insurance m arket which later serves as a benchmark to identify ‘exploitative’ elements in prices and equilibria in later sections. The following sections establish the results in specific term s. 3.1 The Model There are two agents, a peasant, and a moneylender (M) who is also a trader in grain. Grain is the only produced good, there is also another good - money. The production period of grain, which is also the unit of tim e, is divided into two seasons (or m arket periods)-a pre-harvest season and a post-harvest season. T h e A g e n ts an d P r o d u ctio n The peasant cultivates land in order to produce grain. The land may be his own, or he may lease it from a landlord. We assume this arrangem ent to be given and external to our problem. His net income from the land-after meeting rent, taxes and all other similar obligations-is y units of grain. This is a stochastic variable dependent on the state of nature s. The peasant receives this income in the form of grain, s is independently and identically distributed in each period. 46 In each production period, the peasant has a minimum consumption require m ent of c units of grain. If his consumption falls short of c he dies. The excess of y over c is defined as his surplus, x. The following conditions hold on x(s) : E sx(s) = x > 0 (3-1) Pr([s : x(s) < 0]) > 0 (3-2) Condition 3.1 says th a t on the average the peasant has a positive surplus. 4 The peasant’s income accrues to him in the post-harvest season, and he needs his consumption in the pre-harvest season. He consumes only grain. Condition 3.2 says th at, in any given period, there is a positive probability th at the peasant’s output will fall short of his minimum consumption needs. This is a situation under which the peasant may be forced into debt. The peasant has a utility function defined on consumption in excess of c. M possesses a stock of grain and of money. He acquires grain in the village markets through purchase or as settlem ent of debts, and exports it to a ‘foreign’ m arket. The profits from these transactions form his objective function. M ark ets M arkets meet once in each of the two seasons - pre-harvest and post- harvest. Transactions in both grain and credit are perm itted. All credit is transacted in money, whether the actual loan is extended in money or in grain. Production periods run from harvest to harvest, and are indexed by (subscript) t. The seasons are indicated by ti (post-harvest) and t 2 (the following pre-harvest). A tim e-subscript denotes a quantity actually available and transacted at th at time, thus xt is the surplus available in period t, which is the result of production in period t — 1. For convenience we shall replace (t + l ) x etc. by t\ + 1 etc. 4This may be because the land market is not completely monopolistic, or because the peasant owns some complementary inputs such as a pair of bullocks, or because he has some land of his own. 47 3.2 Insurance Consider the situation where an agent undertakes to insure the p easan ts output against production risk. This is a contract to replace the variable x(s), contingent on the state of nature, by another less risky variable 2(5). Let z(s) be a function of s such th a t x(s) is a mean-preserving spread of z(s), x — z. An insur- ance scheme z is a contract under which an insurer undertakes to replace the surplus x(s) by the quantity z(s) in each period t, contingent on the realized value of s. This is not intended as an exhaustive definition of insurance schemes, but it is sufficiently general to suit our purpose. Since the peasant is risk-averse, it follows th at he would be willing to pay a positive price per period for an insurance plan which reduces the riskiness of his income or, equivalently, his surplus. An insurance contract is usually m ade on the condition th a t a fixed sum (the ‘prem ium ’) be paid by the insured party to the insurer each period as price of the service. This price reflects the cost of providing the service, as well as the relative bargaining positions of the two agents. We assume th a t the cost is given, and is no greater than the peasant’s expected surplus for any contract z. Expressing all costs in term s of units of grain, we have, tf[x(s)] < x (3.3) < 7(.) is the cost of providing full insurance (i.e. the cost of providing the scheme z where z(s) = x for all s). This is a relevant feasibility condition which allows the peasant to potentially insure himself to the full extent of his risk. For convenience we shall only consider full insurance. In order to focus exclusively on m arket strategies, we assume th a t the technol ogy of insurance is available freely to all agents who may compete on the market. This guards against the possibility th at the m oneylender’ s profit is a monopoly rent due to private information or some similar resource. 48 A s s u m p tio n 5 All insurers including M incur the same cost gx in offering insur ance to the peasant. The peasant is not able to insure himself. Of course in reality such an explicit insurance contract is never available to the peasant. In w hat follows, however, we will be considering equilibria w ith recurrent indebtedness which are formally similar to an insurance arrangem ent. The implicit cost of insurance can therefore serve as a benchmark to evaluate the profitability of the credit arrangem ent to the lender. Further, the peasant’s ability to make scheduled repayments on a loan will depend upon his realised income in each period. In order to insure against the risk of default, lending agencies would normally add a risk- premium to the risk free lending rate. In addition, to render the peasant’s debt completely risk-free, his subsistence m ust also be guaranteed for all the periods in which he has to make repayments. In effect, this requires the peasant to buy output insurance such th at his surplus is guaranteed to the amount of the scheduled repayments. We may assume for convenience th at only the peasant’s future stream of surpluses can be used as collateral for loans. Let i" be the rate of interest at which credit is available to a riskless borrower, and let gx be the price of full insurance for the peasant, when both credit and insurance markets are perfectly competitive. If the peasant buys full insurance, then the largest loan he could obtain at the rate i is given by: D m a x — [ ■ £ ffa ] ~ * (^'^) otherwise he borrows at a rate where *' > * + ax) (3.5) The second term on the right-hand side is the minimum possible risk- premium m arkup which is sufficient to cover insurance cost. 49 P u r e In su ran ce v s. L oans The substance of the transactions involved in the full insurance contract de scribed earlier is as follows: (i) In each period t, contingent on the realised state of nature, a surplus x t is pro duced- The insurer then transfers to the peasant (or obtains from him, if negative) a quantity x — x(s) of grain. (ii) The peasant then pays the insurer an agreed upon price or prem ium 9z > 9x- The peasant thus retains in each period a net surplus e’ = x — g 'x > x — gx — e (3-6) e is his net expected surplus when all markets are perfect and accessible. Now consider an alternative arrangem ent whereby the insurer, instead of pro viding pure insurance in this form, agrees to lend to the peasant in each period an amount L (s) = max[c — x(s),0] (3-7) to be repaid in a subsequent period. The peasant can still meet his consumption needs, however, the amount of his net surplus in any period now depends upon his outstanding debt and the condition of its repaym ent. We shall call this a subsistence-loan contract. D e fin itio n 1 A subsistence-loan contract is an undertaking by a lender to lend the peasant an amount L(s) in grain in each period. 3.3 Moneylender Monopolist In this section we shall assume : (a) th at M is a monopolist in the market for credit an d /o r insurance, and (b) th at he aims to maximize expected profits. Our conclusions will follow from considerations of his optim al strategy. The grain m arket is competitive. Thus there is a price at which all agents can buy and sell 50 grain. We assume th at this price is constant over market periods (seasons), and hence can be used as a num erator. This assum ption is m ade for convenience only and does not qualitatively affect our results. In terlo ck in g Let dtj be the debt in money (if any) owed by the peasant to M at the beginning of m arket period tj. M can impose the following conditions on the peasant. A s s u m p tio n 6 If dtj > 0, then in market period tj the peasant must settle this debt to the maximum amount permitted by his current endowment prior to con sumption. A s s u m p tio n 7 If the peasant obtains a subsistence loan from M, then he must \ use it entirely to purchase grain from M. Both these conditions are implementable, the first can be imposed through timely foreclosure, and the second through a refusal to extend the loan. A s s u m p tio n 8 The moneylender can choose a price pt . in market period tj for his grain transactions with the peasant. This is clearly feasible if M is a monopolist, but is true other- wise, even though the peasant may decline to trade at the price pt]. T h e O p tim a l S tra te g y In order to concentrate on situations in which the peasant might become indebted, we shall suppose th at he has no stocks of grain. Thus, if in some period his output falls below his consumption needs, he would be forced in th at period to acquire either insurance or a subsistence- loan, and he would be willing to pay a price as high as x per period in subsequent periods. A higher price cannot be sustained. 51 Since M is a monopolist, he can choose whether to offer insurance or loans, and the period in which to offer it. In what follows, we shall describe a strategy for M, which also includes choosing prices at which to trade in grain, and show th at this is an optim al strategy. Let t —1 be a period such th at xt_x < 0. The probability of such an occurrence is positive by (3.2). The peasant then needs — xt~\ in grain to meet his consumption needs. Let M extend him a subsistence-loan to cover the shortfall. By assumptions 7 and 8, he can then require the peasant to buy grain from him, set a price for the grain sale, and extend exactly the money needed to acquire the necessary am ount of grain. The transaction takes place in the sub-period t2 — I, which is the season when the peasant needs his consumption. Let Pt2- i be the price of grain chosen by M. Then in the following post-harvest period t\ the peasant owes a debt to M given by: dtx = Pt2- i- (~ x t-i) (3.8) In the sub-period h , M is again entitled to choose a price on grain transactions. Let this price be P(ti) = dtl -f yt (3.9) Since the num erator on the right-hand side represents the peasant’s total debt in this period, he is obliged to sell his entire output yt to settle this debt since, by assumption 6, the debt has to be repaid to the extent of his current endowment. This does not hold if the peasant can obtain a better price elsewhere. B ut this objection need not be relevant. O b se rv a tio n 4 By choosing pt2~\ sufficiently high, M can ensure that ptl as de fined in (S.9) is greater than any predetermined number p. Hence in the post harvest market he can offer the peasant a higher price than the market, making it most profitable for the peasant to sell to him. 52 Then in sub-period t 2, the peasant’s endowment is 0, and he needs a loan to cover his entire consumption c. M can now choose another appropriate price p< 2 , and repeat the process in all future periods t + n, n = 1 ,2 ,.... Then in t and all subsequent periods, the net transactions are (i) a transfer of y = t/(s) units of grain from the peasant to M in the post-harvest period, and (ii) a transfer of c units of grain from M to the peasant in the pre-harvest period. It will be noted th at the money transactions are purely ‘notional’ (see Bhaduri 1983 ch.2) and for accounting purposes only, but they are indispensable in order to segregate the peasant from the market. The profit m ade by net transaction is the same as th at which would be entailed by a full insurance contract where the price of insurance is set at x units of grain per period (which is the maximum feasible price). M ’s expected revenue from this transaction in each period is given by R — E s [y(s) — c] = x (3.10) Since the cost of providing full insurance is gx, his net profit is given by (in expected term s): e = x - g x (3.11) which is a monopoly rent. It is easy to check th at this is the maximum profit that can be made. We can now state the following propositions for the case where M is a monop olist in credit and insurance: P ro p o s itio n 6 Provision of a subsistence-loan to the peasant, beginning in an appropriate period, is an optimal strategy for M. 53 P ro p o s itio n 7 Interlocking of grain and credit markets is a feasible and optimal * strategy for M in his transactions with the peasant. P ro p o s itio n 8 M can appropriate the entire surplus of the peasant, net of insur ance cost, as rent using this optimal strategy. P ro p o s itio n 9 Under this strategy, the peasant owes a debt to M which is renewed every period. It should be pointed out th at this is not a unique optim al strategy. In par ticular, M can obtain the same profit by offering a pure insurance arrangem ent at a price x , thus charging a monopoly rent equal to w hat he can obtain with the subsistence-loan arrangem ent. The only difference between the two is th at an insurance scheme can, under proper circumstances, be unilaterally canceled by the peasant, whereas the loan arrangem ent we described above cannot be so canceled. This is because an insurance scheme, at least in principle, requires a payment to be m ade in each period, and lapses or becomes inactive in case of default. In order to discontinue the scheme, therefore, the peasant merely needs to stop m aking payments. In order to cancel the loan arrangem ent, on the other hand, the peasant m ust pay off his outstanding balance which, as we show, he may be unable to do. In any case, the cost of canceling the loan arrangem ent is clearly higher than th at of canceling the insurance arrangem ent. The latter can therefore be viewed as a longer term arrangement, which might lower the cost of providing insurance. The two arrangem ents, however, have very different implications for the mon eylender’s profits if there is potential threat of com petition in the credit/ insurance m arket. These implications are more serious if the peasant potentially has oppor tunities for innovations in productive techniques available to him. In this last case, different reactions can be expected to different types of innovations, according to 54 their effect on the expected output of the peasant, and on the variability of th at output. In the next two sections, respectively, we shall consider the cases of po tential competition and innovation opportunities, the latter both in monopolistic and competitive situations. The problem of differing reactions to different types of innovations will be discussed in somewhat detail in the last section. 3.4 Competition in Credit In this section we assume th at there is perfect competition in all three m arkets - credit, insurance and grain. The peasant can therefore potentially trade in grain at the m arket price, obtain full insurance at a price gx (cost of insurance) per period, and borrow at a rate of interest * as long as his output is insured. The initial condition, however, is one in which he is indebted to M under a subsistence-loan arrangem ent. Thus in each pre-harvest season the peasant obtains a loan, nominally in money, from M in order to cover his consumption requirements. M uses the interlinking conditions stated in assumptions 7 and 8 to force the peasant to buy grain from him against this credit, and recovers the peasant’s entire output in the following post-harvest season as repaym ent for the loan. Our intention is to show th at, if a certain condition on the peasant’s average surplus x holds, then the peasant cannot withdraw from the arrangem ent to take advantage of the more favourable prices which prevail on the open m arket. The conditions under which such indebtedness can persist, of course, are the obverse of those under which the peasant can break out of the debt-bond. In a given production period t he may extricate himself in either the pre-harvest or post-harvest season. If he succeeds in doing this, then he may purchase insurance on the open m arket and therefore be assured of a surplus in all succeeding periods. This future stream of surplus is the only asset against which the peasant can borrow. 55 Let D represent the peasant’s borrowing on the open m arket, and d his out standing debt in any given m arket period. Clearly, this borrowing will be positive, whichever season the peasant attem pts to shift his transactions to the open market. We look at each period in turn. In period t x he needs to settle his outstanding debt dtl w ith M and also retain c for consumption. Therefore, D {tx) = m ax[dtl - (z - gx), 0] (3.12) where D (tx) is the open m arket loan and x — gx is available surplus. Alternatively, he may obtain his consumption loan in period t2 on the open market. In which case, D (t2) = c (3.13) In either case, his surplus x — gx in succeeding periods m ust be sufficient to cover at least the interest payments on the loan. The condition for extrication is therefore either e > i (max[dtl — e, 0]) (3-14) or e > i.c (3.15) It was seen earlier th a t dtl is effectively a choice variable for M, since he can arbitrarily set the price Pt2- i (Observation 4). He can therefore always set the nom inal value of the debt at a level such th at 3.14 can never be fulfilled. 3.15 remains the only effective condition under which the peasant can break the debt-bond. The condition which allows M to continue earning his superprofits is therefore the negation of condition 3.15. e < i.c (3.16) We shall term this the Condition of Indebtedness. 56 This is only an alternative way of stating th at the maximum loan which the peasant can obtain on the open m arket is smaller than his consumption require ment c. e > £ U x (3.17) Note th at, if 3.16 holds, then the peasant pays the moneylender e in each period, in addition to insurance costs and capital repayment. By 3.16, the implicit rate of interest on the periodic loan c is then < i (3.18) This, however, does not translate into a loss for M. As pointed out in Section 3.3, the net transaction between M and the peasant constitutes an insurance ar rangem ent, not a twofold arrangem ent consisting of insurance and a loan. Its cost to M is therefore gx, and not gx + i.e. The periodic paym ent e therefore indeed constitutes superprofits for M. Alternatively, consider the eventuality th at the peasant does extricate himself from the debt-bond, say in t 2. M is then left w ith a quantity c of grain (or equivalently c in money) which the peasant would have otherwise borrowed from him. M can presumably lend this money elsewhere on the m arket at the net (competitive) interest rate i. However, he would now incur a further cost i.c each period in administering the loan; hence he would make a norm al (zero) rate of profit. This is lower than the profit he makes on the subsistence-loan arrangem ent. We can now state the following proposition. P ro p o s itio n 10 If the peasant is indebted to M, and the Condition of Indebted ness (S.16) holds then propositions 6 to 9 continue to hold even if insurance and credit markets are perfectly competitive. It was mentioned at the end of Section 3.3 th at in the absence of com petition, M would be indifferent between extending insurance at a monopoly price x and 57 offering a subsistence-loan. An insurance plan however, can be unilaterally dis continued by the peasant, whereas the loan arrangem ent is binding on him unless he can pay off the debt. It will be obvious from the analysis in this section th at whenever there is a threat of potential com petition in either of the two m arkets, M would opt for the subsistence-loan plan in order to restrict entry. 3.5 Reaction to Innovation In this section we indicate the implications of the preceding analysis for responses to opportunities for innovation, in particular on the p art of the moneylender. There are two kinds of effects an innovation may have on output - it may change (increase or decrease) average output, and it may affect the variability or riskiness in production between different states of nature. Formally, let A be an innovation, and let x(s) + A x(s) be the surplus in state- of-nature s if the innovation is adopted. In other words, the innovation produces a net change of A x(s) in the surplus in state-of-nature s. Let the new average surplus be given by x + A x. If the riskiness of the new surplus is different, it would affect the cost incurred by an insurer to insure it. Let this change in cost of insurance be represented by A gx. A gx is positive if the innovation increases the riskiness in output, and vice versa. The peasant’s periodic surplus after insurance cost, in term s of open- m arket values, now changes to (we drop the hats and consider only expected values of variables) e + A e = x + A x - (gx + Agr*) (3.19) Suppose th at the peasant was initially indebted to M and condition (3.16) was satisfied. Also let the innovation A be available to the peasant at zero cost (this does not qualitatively affect our conclusions). If the m arkets are competitive, and 58 if A x — A gx > i.c — e (3.20) then by adopting the innovation the peasant could raise his income sufficiently to break out of the subsistence-Ioan arrangem ent. As we saw earlier, this reduces the m oneylender’s profits by e per period. It would therefore be rational for M to try to restrain the peasant from adopting the innovation. 3.20 would be satisfied if the innovation raised mean output and/or lowered riskiness sufficiently. We can therefore state the following: P ro p o s itio n 11 It is in the interest of the moneylender to restrain the peasant from adopting an innovation which significantly raises output and/or reduces pro duction risk. It is interesting to note th at if the credit/insurance m arket is monopolized by M, and a subsistence-loan scheme is in effect, then neither increase in production nor a reduction in riskiness need threaten the monopoly profits th a t he earns. The profits in this situation are realised by m anipulating prices at a tim e when the peasant is forced to borrow, and this m anipulation ensures th a t the peasant’s need for credit will be renewed in the next period. Conclusion We have shown th at, if the peasant has an outstanding loan to the moneylender, then this loan can be used to interlock the credit and output transactions of the peasant, and as a result the moneylender can extract the entire surplus produced by the peasant. This holds even in the presence of competition in one or both markets. The crucial condition is th at the peasant does not produced more than a critical am ount of surplus. The moneylender may thus have an incentive to restrain the peasant from adopting technological innovations which raises his productivity above this critical level. C h a p t e r 4 S u p p o r t M e c h a n i s m s 4.1 The Need for Permanence W hen we discussed the benefits of a patron-client relationship in chapter 2, we concentrated explicitly on the problem of different time-preference profiles of the patron and the client. Other benefits may also be considered. For example, it is a sufficiently common observation that, in a patron-client relationship, most or all of the transactions which the client enters into are conducted with the patron. Thus all of the form er’s purchases and sales may be mediated by the patron, who often owns the local store. If the latter has a large number of dependent clients, it is obvious th a t this would augment his bargaining power when dealing with outside traders-w hether in selling agricultural output which is transported to the non-agricultural sector, or in acquiring commodities from th a t sector. We need not limit the discussion exclusively to the economic benefits of a patron-client type relation. There are other gains which follow upon such an arrangement. Especially relevant axe the consequences on the landlord’s position in the local power structure and political apparatus. It is widely accepted that political patronage and local power networks severely condition economic opportunities and performance in third-world agriculture. We shall not argue this assertion here; it is sufficient to note th at such connections 60 are consistently used by substantial agents - to evade land ceiling laws and laws regarding wages and tenure, to avail of (and misdirect) development allocations, and to control insubordination among weaker agents. Access to power networks and patronage, in turn, is often crucially dependent upon the exercise of control over blocks of people. This may enable the landlord to deliver a block of votes in favour of a benevolent politician, or summon forces at times of potential physical conflict, or merely accord him sufficient status to become p art of the rural ruling class. 1 It is obvious th a t the benefits to either party, as discussed above, presuppose a certain organization of production, and specific asset endowments for each agent. The curtailm ent of supervision costs to regulate effort and capital use is likely to be im portant incentives for the landlord in a context where agrarian production is normally carried out by single peasants or households on small plots of land, as opposed to large farm s employing co-ordinated team s of labourers. In the latter case, one would expect substantial economies of scale in supervision, which would render the gains from voluntary compliance insignificant. In mechanized agriculture, for instance, perm anent dependence would not be found as valuable, both because m echanization necessitates a large scale of operations, as well as since the machine dictates, to some extent, a pace and scope of activity which cannot be easily overridden by individual workers. For the peasant, on the other hand, the benefits of patronage are more valuable when his income stream is risky, and he does not have alternative means to stabilize his consumption. In particular, lack of access to other sources of consumption credit (or production loans to facilitate cultivation on his own land) would increase the value of dependence. 1 W hile it is not our intention to pursue this point here, it should be observed that dependence thus performs a function reminiscent of a feudal order. 61 A special case of this arises when the peasant has no accumulated stocks of consumption goods or money. Depending upon their nature, such stocks can serve as collateral to borrow against on the open m arket, or as a consum ption fund to draw upon in bad states of nature. The insurance m otivation in building up such , stocks is clear. It is less obvious th a t the peasant’s ability to avail of alternative income opportunities may also be affected by the existence or non-existence of such stocks (or other assets). One way in which dependence may arise as a consequence of insufficient income, and be sustained, was discussed in chapter 3. The specific result we derived there is th at a moneylender who is a monopolist can get a peasant indebted to him by capitalizing on the event th a t the peasant’s income falls below his consumption needs. He can renew this debt every period by m anipulating prices, and use it to appropriate the entire surplus produced by the peasant in all subsequent periods. Further, if the indebtedness condition holds, then he can protect this profit and m aintain the peasant in a dependent status even if competing sources of credit enter the m arket. It may be noticed th at the crucial factor th a t restricts the peasant’s options in m arket dealings is the inflexibility in his survival needs. His subsistence consump tion cannot be postponed or reduced. It is this nonlinearity th a t the moneylender exploits, to ensure himself superprofits by reproducing the peasant’s need for a loan. From the peasant’s point of view, given his low level of production and consum ption, the insurance motive turns out to be overriding. Two observations on this pattern of transactions are in order. First, it is a trivial consequence th at the peasant remains indebted to the moneylender period after period. Secondly, while we have stressed the low productivity of the peas ant as the condition th a t ensures the continuation of the debt-trap, it is equally necessary th at he does not accumulate a stock of consumables or assets which can serve as an emergency consumption fund or as collateral to borrow against. 62 If these conditions are m et, however, the moneylender need not restrict the peasant’s consumption to the bare minimum, if he can ensure th at the peasant does not save between m arket periods (thereby accumulating enough resources to cover his consumption needs and hence dispense w ith the necessity of a loan). This can be ensured, for example, if the moneylender disburses loans in kind in the form of perishable goods, and if the peasant does not have access to a resale m arket. The peasant would then have to consume his entire current holding of consumption goods (”grain” ) within the current m arket period, and his need for credit in the following period would not be reduced. Allowing the peasant a higher consumption would of course correspondingly decrease the m oneylender’s profits. If usury profits were the latter’s only objec tive, then such action is irrational on his part. However, even w ith the higher consumption, the peasant cannot escape being recurrently indebted, and is thus tied to the moneylender. Suppose now th at the moneylender is also the peasant’s landlord. Further suppose th at he imposes on the peasant the condition th at, in order to obtain a loan from him, the latter also has to be employed in cultivating his land. If the peasant seeks alternative employment elsewhere, the loan is foreclosed. It is easy to see th a t this would force the peasant to continue employment w ith the same landlord-moneylender in every period, even if marginally more attractive employment opportunities become available occasionally. The patron, acting as moneylender and landlord, could thus ensure th at the land-tenure contract is renewed in perpetuity. If the employment contract is not interlocked w ith other contracts - especially credit - however, the peasant is not constrained to continue his association w ith the landlord. The moral hazard problem discussed in chapter II then cannot be circumvented, nor can the additional benefits mentioned earlier be reaped. 63 In sum , then, if certain initial conditions on the organization of production and agents’ asset endowments are satisfied, then a perm anent dependence relation would turn out to be pareto-superior compared to an arrangem ent where employ ment and credit contracts are m ade piecemeal and for lim ited periods. It should be noticed th at, while to the peasant it is the dependence per se th a t is im portant, for the landlord it is the permanence of the relation which is necessary to produce the desirable intertem poral consequences. 4.2 A Social Equilibrium (of Sorts) We can now turn to the second of the two questions raised in chapter 1: W hat are the mechanisms, economic or extra-economic, which support a system of perm anent dependence relations? Since it was argued in the last section th at such a system is pareto-superior to feasible alternatives, the present question may seem superfluous. This is not so. We claimed th a t under certain initial conditions patron-client arrangem ents would be optim al. However, the optim ality is conditional on the agents’ belief th a t the arrangem ent is perm anent, which leads them to adopt a certain behaviour. If an agent believes otherwise, then the same behaviour may not continue to be optim al for him. The situation is one which has been extensively analyzed in the game theory literature. Each agent, patron and client, can play one of two strategies, one of which is appropriate to a perm anent arrangem ent, the other optim al for a tem porary arrangem ent. The payoff in each period if both play the first strategy is pareto-superior to the payoffs if both play the second. However, the superior strategy-pair emerges as an equilibrium only if both agents believe th at the game will in fact be repeated a sufficient (theoretically, infinite) num ber of tim es.2 2This is essentially a repeated prisoner’s dilemma, in which the superior (cooperative) soluton can be enforced by a suitable choice of strate- gies if the game is repeated an infinite number 64 For each agent, then, the desirability of term inating the arrangem ent in some given future period would depend upon a comparison of two factors. One is the best alternative stream of payoffs he can subsequently attain. This m ust be greater (in present value terms) than his payoff stream under the current arrangem ent, and the gain from defection m ust more than balance the loss which he incurs as both agents revert to their respective inferior (tem porary) strategies for the intervening periods. We postulated earlier th at, if certain initial conditions obtained, neither party would find it rational to defect. By extension, a support mechanism would guar antee th a t these conditions continue to obtain, or at least not change sufficiently to make defection rational. We indicated th at one characteristic of agrarian trans actions (we claim) performs an im portant function in this respect-nam ely, peasant indebtedness. If the peasant’s income stream is risky, and if his income may frequently drop below the level necessary for survival, then he needs access to a credit source to ensure continued survival. A stock of consumption goods or m arketable assets may be substituted for a credit source. Such a stock may be drawn upon directly, or serve as collateral for a loan. Conversely, if he does not have such a stock, then his future labour is the only asset against which he can borrow.3 In an economy where labour is plentiful, however, a future comm itm ent to work is not necessarily an attractive investment for a lender, less so if the lender is not customarily an employer of labour. Especially in distress periods the supply of labour futures is likely to be plentiful, and hence would comm and a low price. This, however, is not true for a perm anent comm itm ent, since such a commitment has additional value. Thus if the peasant does not expect to build up stocks, of times. For finite repetitions the inferior equilibrium usually dominates. See Axelrod (1984), Rubinstein(1979), Kreps et.al.(1982). 3It should be noted that this does not require the peasant to customarily survive at some pre defined level of "bare subsistence” , it only requires that in some periods his income should fall below survival. 65 but does expect appropriate fluctuations in income, then it is rational for him to rem ain loyal to his perm anent commitment. The accumulation of stocks or wealth can affect the perm anence of his comm it m ent in other ways as well, two of which are especially im portant. Both function by changing the feasibility of alternative opportunities, one by affecting his ex pected income stream within the agricultural sector, and the other by providing entry into other sectors. First, the stock could function as capital. If the peasant has some land of his own, which previously was not sufficient to meet his needs, the application of capital may raise its productivity to the extent th at the peasant becomes self- sufficient on the average. This of course also implies an increase in the value of his land, which would then function as collateral, enabling him to borrow on the open m arket.4 The value of stock as capital depends on the technology which the peasant can acquire to increase productivity. A given level of savings is thus more likely to enable the peasant to become independent if more productive technology is available for him to use. M igration is another option. It is a well-accepted theoretical conclusion th at if the peasant is cognizant of opportunities to earn a higher income elsewhere (say in a nearby town), and if the expected utility of the alternative income stream outweighs the expected utility of his present state by an am ount which more than compensates for the costs of the move, then he is likely to make the move.5 In the absence of a credit-m arket, however, it is necessary for him to be able to meet the immediate costs of the move before the higher income stream is realized. These costs are not merely those of transportation; they also include the costs of 4In chapter 3 we showed that, if the peasant is not sufficiently productive on his own land, then a cred- it-agent can maintain him in a dependent position, and obtain a superprofit, purely by manipulating a credit arrangement. 5This is, of course, a result of the well- known Harris-Todaro model. 66 surviving in the new environment until employment is found. Since the waiting period is usually uncertain and the disutility of running out of funds high, the amount of funds which are needed to make the decision worthwhile is likely to be substantial, at least compared to the average income of the peasant. It has been observed in empirical studies on rural-urban m igration th at the poorest members of the rural population seldom tend to m igrate except in distress situations. More often it is the agents with some assets who go to towns in search of alternative opportunities.6 The peasant’s incentive to defect from a perm anent arrangem ent, then, is likely to be higher if he can accumulate some stock of commodities or assets, for at least three reasons: he may be able to meet his own distress needs, he may be able to increase productivity on his own land, and he may be able to m igrate in search of more rem unerative employment opportunities. The incentive to defect would, of course, also increase if these alternative opportunities became more attractive, but the opportunities cannot be availed of in the absence of some initial asset- ownership. Now consider the question of a support mechanism. If the landlord can in some way ensure th at the peasant is unable to accumulate assets, then he could be correspondingly certain th at the peasant, behaving rationally, would not plan to defect. Both agents would then play their perm anent strategy, and the pareto- superior outcome would be attained. A m om ent’s reflection will make it obvious th at the interlocking arrangem ent which is incorporated in a patron-client relationship serves also as such a support mechanism. Since virtually all of the peasant’s external transactions pass through the hands of his patron (or are known to him ), and the patron also necessarily 6See Skeldon (1985) p.56-57. W hile Skeldon notes that data on the wealth of rural- urban migrants is scanty, he argues that, ” (G)iven the higher general level of education of migrants vis-a-vis their communities of origin, it is likely that migrants tend to have above-average family incomes. The poorest cannot afford to move or can afford to move only season- ally.” p.57. 67 knows the peasant’s income, the latter has all the information necessary to evaluate the client’s asset position. This is only m ade easier by the close personal ties between the two agents. Further, the patron is also in a position to regulate the commodities which the client acquires. Typically, loans are not disbursed at pre-stipulated tim es, nor are they advanced in cash. R ather, the peasant is granted credit to m eet his excess needs as and when they arise, and these are advanced to him in the form of the specific goods he requires. The peasant’s share of agricultural output, similarly, is paid to him in the form of grain, p art of which he consumes, and p art he returns to the landlord as repaym ent for earlier loans. Grain prices are lowest after harvest, hence the peasant cannot profit from selling his share, and since he does not have holding power he cannot wait to profit from sales in the following pre-harvest season. Thus, as far as the patron is concerned, the patron- client relationship itself provides a support mechanism to ensure th at the dependence relationship remains perm anent. In order to implement this mechanism, however, it is necessary that the peasant’s transactions be interlocked - all or most of his trades being made w ith the landlord, and th at he come up against frequent shortfalls in consump tion. The landlord’s transactions with him would typically be conducted in kind (though accounts may be kept in cash, see references in previous footnote, espe cially Bhaduri), and closely geared to the peasant’s actual consumption pattern. This is done to ensure th at the peasant does not accumulate any assets, whether stocks of consumables or other liquid assets, which may make it worthwhile for him to default from the perm anent arrangem ent. Conversely, if a patron-client relationship is already in place between the two agents, w ith the expectation of a long-term association, then both would be aware of the constraints which are thereby imposed on the peasant’s options. As we suggested earlier (chapter 1), the peasant would not consider defection a feasible 68 option, since it can never be implemented w ithout the compliance of the landlord. Hence the relationship is self-perpetuating unless it is unilaterally canceled by the landlord. This completes our argum ent in support of the hypothesis put forward in the first chapter. The argum ent is summarized in the concluding chapter which fol lows, where some im portant consequences of the hypothesis are outlined. 69 C h a p t e r 5 C on e lu sio n s 5.1 A Relation of Dependence In chapter 1 we proposed the hypothesis that long term tenancy relations between landlord and tenant, which involve routine. extensions of credit by the former to the latter, may be .pareto- optimal in,certain situations, and may also be more profitable to the landlord. To achieve the pareto-superior solution, however, a mechanism to enforce the duration of the arrangement is necessary, and patron- client relationships incorporate a mechanism sufficient for this purpose. In chapter 2, a model was presented to illustrate the first part of this hypothesis. We found th at, if independent decision-making by the peasant is involved, a long term interlocked contract turns out to be optimal. It should be noted th at the conclusions there did indeed turn on the independence of decision making, and not on the possibility of shirking. The latter is a factor in many types of principal-agent relations, whereas in m odern production involving extensive division of labour very few decisions are usually m ade by subordinates. Thus our focus is on a problem which is somewhat specific to peasant agriculture. If the explanation is realistic, then we could claim th a t the corresponding production arrangement emerges as efficient given the particular technology and organization in use, rather than being caused by high supervision costs. 70 In chapters 3 and 4, we tried to illustrate the second part of our hypothesis— th at the constitution of patron- client relationships incorporate a sufficient mechanism to support the pareto-superior arrangem ent. The argum ent is as follows. If a share-cum-consumption loan contract can be guaranteed to continue for a long tim e, then the outcome is superior to one obtained by concatenating several short-term arrangem ents. The problem, then, is to ensure the renewal of the contract from period to period, perhaps indefinitely. Specifically, the peasant may on occasion find alternative employment opportunities which are more attractive than than the current arrangem ent. If he is then free to defect from the latter he would do so, and as a consequence neither peasant nor landlord would expect an association th a t is indefinitely long. They would thus play their short-term strategies, and the superior outcome would not obtain. In order to avail himself of these alternatives, however, the peasant has to pay a lumpsum cost which cannot be deferred. This cost can be m ade greater if he is also indebted to the landlord. He may pay this cost out of accumulated savings, or cover it by taking a loan. The second option is not viable since he has credit only w ith the landlord, who sustains a loss if the peasant moves to a new employment. The only remaining alternative is to accumulate savings in the form of money and durable goods. However, none of the commodities which the peasant actually consumes is durable, nor is any of the commodities he produces. Further, he does not have an initial stock of money or durables. In order to save, then, he m ust acquire these by exchanging for them commodities he does have access to, including labour-power. There are two preconditions under which he can do this in any given period. One is the availability of a surplus above his consum ption needs which he would exchange; the other is access to an agent (or a market) who would trade him savings goods for this surplus. If either of these conditions is not fulfilled, then the peasant is restricted from saving. 71 The patron-client nature of his relationship with the landlord enters crucially in actuating this restriction. If the peasant normally conducts all his transactions with the landlord, then the latter can certainly ensure th a t he does not acquire savings goods. Further, if the peasant never actually has in his possession any surplus th a t is m arketable, then he cannot trade with other agents. This need not restrict him from consum ption in excess of bare subsistence. The patron may well act as a trustee, and make consumption goods available as and when needed. W hether these goods are forwarded as normal earnings, a loan or a bonus does not really m atter, we noted th a t the payoffs to both agents are functions of the physical commodities distributed, not of the specific accounting term s th a t are used. Thus new clothes and festival meals may be regularly provided as traditional gifts, which the client fully counts on receiving. Funds to cover conspicuous consumption, which the client is required to undertake occasionally, may on the other hand be advanced as loans, which again he normally anticipates will be granted (see Rudra, 1984). Note th a t our context is form ulated so th at the only reason for the peasant to save is to make it possible for him to defect. As long as he is unable to leave the p atron’s employ, all his distress needs can be m et by drawing upon the p atron’s resources. 9 We need not conclude, however, th a t the patron and client are therefore en gaged actively in a game in which the former exercises constant vigilance while the latter is forever seeking surreptitious means by which to accumulate a stock of wealth. On the contrary, poor peasants in eastern Indian villages probably do not accumulate stocks simply as a m atter of course, and also as a m atter of course they rely on their patrons to provide support in case of sudden need. The security of patronage is viewed as a positive asset by the client, not as a restriction which prevents him from seizing an option. By the same token, the option of defection is not then an option at all, and would not be considered as such in the norm al 72 course of things. This is akin to the serf in medieval Europe, who may have idly wished for freeman status, but would not normally consider it a feasible option from which the lord and existing order were actively debarring him. This complex of interdependences and restrictions are, of course, sustained as much through social and political pressures as through economic ones. R eputation and traditional alliances play a significant role, as do a m yriad of institutions such as caste, gifts, conspicuous consumption, and community sanctions against ’deviant’ behaviour. Technology and organization of production form the basis upon which the entire complex rests, and are im portant constituents of the context within which the arrangem ents we have discussed axe meaningful. 5.2 The Possibilities of Change We may now ask two im portant questions. Given this structure of relationships, duties and restrictions (of which we have investigated only a limited aspect), in w hat ways if at all can we expect technological and organizational change to occur? Further, w hat socio- economic consequences would such change have in the given context? We will not pretend to answer these questions in precision and detail, but will rather propose in outline some answers th at are suggested by the preceding analysis. Three observations of a general nature are in place here. F irst, as a change (or innovation) benefits a specific agent or group, it may also render them inde pendent of other agents in some crucial way. The former group may then find it unnecessary to fulfill certain traditional obligations towards the latter, which may in turn severely depreciate the latter’s position. This process, however, does not take place overnight; there is usually a long period over which an innovation is first cautiously tested on a limited basis by more enterprising agents, is then in troduced selectively by others, and finally becomes commonplace or is rejected. In 73 the m eantime, agents who are experimenting with the innovation m ust continue to rely upon traditional ways to support themselves, and therefore upon traditional alliances with, and services from, other agents. Finally, innovations have a ten dency to become common property, especially if the resources to implement them are provided to all cultivators in an area through a development programme. Consider, for example, the hypothetical case of a backward village which is targeted for the promotion of high-yielding varieties of seeds and fertilizers, coupled w ith a credit scheme to make this technology available to all cultivators. Suppose further th a t many of the poorer peasants in the village initially own some land, hitherto insufficient to support themselves and their families. Also suppose, not unrealistically, th a t the large landowners wield m ost of the social, economic, and political power in the village. U nrestricted introduction of the proposed development scheme would then have the following consequences. First, of course, it would potentially raise the productivity of land all around. Secondly, since such technology requires more than the traditional lackadaisical effort, it would also increase the labour requirement per unit of land. (Thus it may not necessarily increase the productivity of labour (measured in efficiency units), but it may allow the employment of a larger number of such units.) Thirdly, the innovation may allow some of the poor landowning peasants to become self-sufficient on their own land. In order to do this, they would devote a greater amount of tim e on their own plots th an previously, thus withdrawing themselves to th at extent from the employ of their more affluent neighbours. This independence may render unnecessary the erstwhile allegiance owed to patrons, and lower the likelihood of long-term cultivation arrangem ents between these peasants and their former landlords. The larger landowners are then faced w ith a smaller supply of labour issuing from their traditional clients, and at the same tim e have a larger need for labour. Further, the labour now available is not bound, nor likely to be held in long-term 74 contracts, hence it needs to be more closely supervised. This is further true if the new technology requires more accurate and careful decisions to be m ade in the field. All this, of course, adds up to an externality which adds to the cost of the innovation, perhaps substantially, as far as the richer peasants and landlords are concerned. Like powerful agents in any economy, however, closely-knit clans in small villages may be well-equipped to internalise such externalities. Two distinct reaction patterns on the p art of the village elite may emerge in this situation. The first is to eschew the innovation altogether, and impose sanctions on any agent th a t adopts it. Exclusion from the benefits of collective services and security would normally be a sufficient deterrent for all but the wealthiest of peasants, in view of the subjective risks associated with non-traditional technology. Social excommunication would provide any further disincentive th at is needed. Such measures never need to be actively implemented, a consolidated opinion by an assembly of patrons decrying the innovation would normally be sufficient. This implicit threat may well be the most long-standing institution making for stability in Indian village structure. This reaction pattern is far from hypothetical. Indeed, this was the most general reaction to the first introduction of ‘green revolution technology’ in eastern India, where village after village failed to adopt new technology even though it was m ade available at low cost, accompanied by credit, and prom oted vigorously by government agencies. The alternative is more complicated. Essentially it consists of reallocating resources of the village between old and new technology and between agents so as to retain the relationships of dependence and allegiance. In other words, it consists of incorporating the innovation into the existing pareto-optim al equilibrium. The new conditions of productivity, coupled w ith the pre-existing distribution of asset-ownership and m arket powers, may however not be compatible w ith the 75 continuation of patron-client relationships as a viable institution. It was suggested in the first chapter th at as a condition for this kind of relationship to exist between two agents, one of them m ust wield considerable economic power, while the other is powerless both in relative and absolute term s. As a reason, we suggested th at if the client has no option but to be dependent on the patron, then he is likely to make the latter’s interests his own. Since the new technology changes the value of both assets and m arket powers, the abovementioned condition need not continue to hold. The peasant may himself become economically self-sufficient, if he previously owned some land or livestock or other capital. Further, small farmers who hitherto ran self-contained opera tions may enter the m arket for full-time attached workers, and independent credit i agencies, including traditional moneylenders, may find it worthwhile to cross over into m arkets which they previously considered inaccessible or unprofitable. More generally, each agent in the village economic structure may be expected to adopt a specific strategy in reaction to the new technology, depending upon his relative and absolute position and upon the changing pattern of his interests and opportunities. P art of this strategy may itself be to change the position, interests and opportunities of other agents. The actual pattern of change th a t emerges would depend upon these strategies as much as on the new technology. To comprehensively interpret actual patterns of change th at have been observed in villages in eastern India, and to causally relate them to initial conditions, ex ternal influences, and strategies, is a task larger than can be attem pted here. This would require a thorough analysis of economic, social and political realities, in other words a study of the political economy of the village as a complex unit, not merely the study of a specific relation existing within it. However, certain directions can be indicated at this point, even though the associated reasoning is susceptible to modification. 76 The m ost obvious predictions are twofold. F irst, as new technology becomes available and gains ground, we would expect growing inequality in the ownership of land and other assets, though smaller units may become more viable. Secondly, we would expect an increase in the incidence of indebtedness in peas ant-landlord relationships. The first observation is justified as follows. Since erstwhile clients who hold small plots of land now stand to become independent, it is to the interest of the landlords to annex these lands if possible. Besides, these lands are potentially more valuable, hence more attractive as investments. The landlord can enforce this acquisition in several ways, notably by exploiting existing debts. For the second observation, note th a t landlords now face a potentially tighter m arket for labour, and are aware th a t formerly faithful clients are confronted with a wider range of options since middle farmers may now enter the labour market as buyers. There may also be a tendency to attach casual labourers through indebtedness. In order to implement this, landlords may attem pt to curtail the influence of independent moneylenders. 5.3 A Look at the Bardhan and Rudra Surveys We do not have at our disposal data th at is collected specifically to test this hypothesis. Much relevant information is contained in several surveys conducted in recent times. For purposes of illustration, we can examine the results of an extensive and well known survey th at was conducted by B ardhan and R udra in the last decade, which included nearly 275 randomly chosen villages in Bengal, Bihar and eastern UP (see B ardhan and R udra, 1978, and also 1980 a, b, c ). The researchers broadly classified the villages as ‘highly advanced’, ‘m oderately advanced’, and ‘backward’, depending on the extent of the use of tubewells and pum ps and the use of chemical fertilizers and HYV seeds. The poorer peasants 77 (who may qualify as ‘clients’ in some or many cases) are classified as tenants, casual labourers, and attached labourers. It is a cross-section study and does not track individual villages over tim e, however, the differences between the three categories of villages may point to a tem poral process. If this is true, then some of our conclusions are indeed borne out. (This does not prove th a t our hypothesis is true, but th at it is at least substantiated by this particular study). The suspicion th at the incidence of peasant indebtedness would increase with the level of advancement is indeed vindicated. The survey quotes, in each category of villages and peasants, the proportion of peasants th at take loans from their landlords. Among tenants, the proportion th at take consum ption loans grows from 45% in backward villages to 56% in the interm ediate category, and falls to 50% in the advanced category. The proportion th a t take production loans grows steadily, from 32% in the backward villages to 43% and 44% in the higher categories. Among casual labourers, 44% in the advanced villages are found to take loans against future commitments of labour, as against 32% in the backward villages. We have not distinguished here between loans with and w ithout interest, as the authors do, since it seems th a t this distinction may not be significant. We argued (chapter 3) th at fluctuations in grain prices may impose a large own-rate of interest, w ithout a positive rate being explicitly charged. The authors do point out th a t the loans are ‘usually repaid in harvest tim e in grain and in labour’ (1978, p.382), which, we noticed, could facilitate a high implicit rate of interest. ( But see Gangopadhyay and Sengupta, 1986.) In the case of attached labourers, the proportion who take loans is higher in the interm ediate villages (82%) than in the backward villages (72%), but this proportion is lower in the advanced villages (60%) than in either of the other two categories. However, we find th at a much larger proportion of these peasants in the advanced villages live in hom esteads provided by the employer (24%, vs. 78 11% in the interm ediate and 4% in the backward villages). Providing homesteads is certainly an effective way to ensure long-term allegiance. Indeed, within our framework, the homestead functions as a loan which has to be repaid if the peasant leaves the employment of the landlord. It is indeed true th at a large proportion of attached labourers tend to remain attached for long periods of time. The authors found th at 30% of these labourers have in fact ‘ worked for the same employer for more than live years, while another 58% have had associations ranging between one and five years. This data, however, is not broken down by level of advancement. Proportionately fewer of the attached labourers receive allotments of land for cultivation in the advanced villages (27%) compared to the backward villages (45%). This is associated with the fact th at a much larger proportion of these labourers in the advanced areas work for employers whose m ain occupation is self-cultivation (73%, vs. 43% in backward areas). Presumably, more landowners turn to cultivation as the return to this activity grows. This tendency is also clear among the landlords of tenants who were surveyed (79% vs. 50%). The availability of new technology, as it increases the value of land, also changes the relative profitability of different activities, notably cultivation and money lend ing. The two are not independent— moneylending would normally produce a greater return when the borrower can use credit to improve his production possi bilities substantially (see Gangopadhyay and Sengupta, 1986). However, when a large proportion of borrowers are poor peasants whose economic conditions stand to change perm anently with the coincident availability of credit and technology, the possibilities of usurious profit may dwindle over time. Further, the balance of influence is also likely to change between landholders and moneylenders, the for mer becoming more powerful. Though this conclusion is somewhat ad hoc, it melds well w ith observation. The survey we have been quoting shows th at, compared to backward villages, a smaller proportion of advanced villages have professional moneylenders (32% vs. 51%), while a larger proportion have rich farmers who 79 also lend out money. Coupled w ith the earlier observation th a t the incidence of indebtedness is greater in the advanced villages, this may indicate th at, with advancement, either some moneylenders acquire land and become cultivators, or existing landlords are able to integrate the function of moneylending into their operations. From the d ata it is abundantly clear th a t there is a trend away from leasing out land to tenants, and towards greater control over cultivation by using casual or attached labour (1978, table 20). There is also a greater tendency for poorer peasants to lose their land. The proportion of casual labourers who have recently lost land, either through eviction or in settlem ent of a debt, is significantly higher in the advanced and m oderately advanced categories than in the backward category (1978, table 19). Ostensibly, then, the spread of new technology sets up a process of redistri bution, both in asset ownership and profitability of activities. A ttention turns more towards land and cultivation, and away from rent as the dom inant form of income from land. At the same time, increased attention is directed at enforcing attachm ents which breed allegiance towards the landlord, w hether through an en largement of debt relations, alienation of land from poorer peasants, or providing hom esteads and such. In order to do this effectively, landlords need to extend their auxiliary activities, such as trade and credit, which brings them into com petition w ith professional moneylenders and traders. However, the increasing im portance of landholding may well give the landlords an edge in this struggle. The result is in creasing horizontal integration of these activities with landholding and cultivation, and a decrease in credit (and perhaps trade) as independent activities. Ironically, this implies th a t the poorer peasant becomes more completely dependent on his landlord, and is therefore further insulated from the open m arket. 80 5.4 Reflections on Policy The following conclusions thus emerge regarding the p attern of development th at is induced by the introduction of new technology into the traditional village economy. As a result of the changing production possibilities, the concentration in landownership tends to increase, as does self-cultivation by rich farmers. The proportion of land under tenancy may decline. Along with self-cultivation, there is an increased emphasis on dependence relations between landlord and labourer. In order to implement this emphasis, there is a tendency towards horizontal in tegration of activities such as self-cultivation, credit extension and provisioning. This in turn may tend to reduce the importance of professional moneylenders who do not have landed interests. These movements condition the distribution of income and socio- economic power between groups in the village. The rich landowning farm er is certainly in ascendance, perhaps at the expense of the moneylender and middle farmer. The small peasant’s assets and independence are both on the decline, though his absolute income may not show much change. Generally, the small peasant is likely to find himself landless on a larger scale, and more dependent on the landlord. One aspect of this dependence is an increasing insulation from alternative trading partners, whether in the land, labour, credit or commodity m arkets. In some respects, this points to a ‘dualistic’ pattern of development. While the increasing production possibilities induce an increase in the level of contact between the village-as-unit and the outside world, this contact is m ediated en tirely by rich farmers; internally the village recedes further into a non-m arket, dependence-based organism. The overall effect of this movement on the internal distribution of income is one of polarization-the increased return from agriculture flows exclusively to richer farmers, while the poorer peasants barely retain their initial levels of income and consumption. It scarcely needs to be mentioned that 81 this polarization in income distribution, and consequent increase in inequality of income, is perhaps the m ost widely perceived trend in ‘developing’ agriculture. A final qualification is in order. The ‘thought experim ent’ in this concluding chapter was conducted on the premiss th at the new technology is m ade available to the village by an agency which remains passive w ith regard to the power-structure internal to the village. Thus the outside agency merely makes technology and credit available to the village agriculturists, perhaps on relatively easy term s, but it does not interfere in the (possibly changing) patterns of control and dependence between the members of the village society. This, for example, had been the strategy of the ‘Green Revolution’ throughout the seventies. Recently, however, the state government in West Bengal has been pursuing a much more vigorous, ‘interventionist’ strategy for rural and agrarian development - one which is aimed consciously at weakening the power which large landowners and moneylenders have over poorer peasants. This is achieved partly through strengthening village councils or ‘panchayats’ and partly by registering tenurial claims of small peasants in land. Such a strategy, even in this limited form, may have significantly different outcomes compared to a passive one. Indeed, the inherent claim of this last chapter is th at, in order to be effective in reducing poverty and inequality of income, agrarian policy in these areas m ust actively interfere w ith the existing socio-economic structure of power, dependence, and ownership. / ’ 82 R eferen ces Ahluwalia, Montek S.(1978): Rural Poverty in India: 1956/7 to 1973/4, World Bank S ta ff Working Paper no. 279, May. Alston, L., D atta, and Nugent (1986) : Tenancy Choice in a Com petitive Framework with Transactions Costs, Journal of Political Economy v.92 no.6. Alston, L. and Ferrie (1985): The Use of In-kind Benefits in Agricul ture: A Synthesis and Test, mimeo Axelrod,Robert(1984): The Evolution of Cooperation, Basic Books, New York. Bardhan, Pranab (1980): Interlocking Factor M arkets and Agrarian Development: a Review of Issues, Oxford Economic Papers, 32. ___________ (1983): Land, Labour and Rural Poverty, Oxford. and Ashok R udra (1978): Interlinkage of Land, Labour, and Credit Relations, Economic and Political Weekly Annual Number. ______ ___(1980a): Types of Labour A ttachm ent in Agriculture, Eco nomic and Political Weekly v.XV no.35. _________(1980b): Labour Employment and Wages in Agriculture, Economic and Political Weekly v.XV no. 45-46. _________(1980c): Terms and Conditions of Sharecropping Contracts, Journal of Development Studies, April. Bardhan and Rudra (1986): Labour Mobility and the Boundaries of the Village Moral Economy, Journal of Peasant Studies ,v.l3 no.3. Basu, Kaushik (1984): The Less-Developed Economy: A Critique of Contemporary Theory, Basil Blackwell, London. Bhaduri, Amit (1973): A Study in Agricultural Backwardness under Semi-Feudalism, Economic Journal no. 83. ________ (1983): The Economic Structure of Backward Agriculture Academic Press, London. Binswanger, H. and Rosenzweig, M.(1986): Credit M arkets, W ealth, and Endowments in Rural South India, Paper prepared for Inter national Econ. Association World Congress, New Delhi. Bloch, Marc (1961): Feudal Society (vol.I), Chicago, University Press. Brass, Tom (1986): Unfree Labour and Capitalist Restructuring in the Agrarian Sector: Peru and India, Journal of Peasant Studies v.14 no.l. Braudel, Fernand (1982): The Wheels of Commerce, Harper & ; Row, New York. Braverman, A. and Srinivasan, T.N. (1981a): Sharecropping and the Interlinking of Agrarian Markets, American Economic Review v.72. __________(1981b): Credit and Sharecropping in Agrarian Societies, Journal of Development Economics 9. Braverman, A. and Stiglitz (1982): Sharecropping and the Interlinking of Agrarian Markets, American Economic Review September. Brenner, Robert (1976): Agrarian Class-Structure and Economic De velopment in Pre-Industrial Europe, Past and Present no.70. Brock, William (1971): Sensitivity of Optimal Growth Paths w ith Re spect to a Change in Target Stocks, Zeitschrift fur Nationalokonomie, Suppl.l. Cheung, S.N.S. (1969): The Theory of Share Tenancy, Chicago, Uni versity Press. D atta, S. and J.B. Nugent (1985): B ahrain’s Pearling Industry: How it was, why it was th at way, and its implications, in Nugent and Thom as (ed). Desai, Rudolph and R udra (eds.)(1984): Agrarian Power and Agricul tural Productivity in South Asia, University of California Press. Eswaran, M. and A. Kotwal (1985a): A Theory of Two-Tier Labour Markets in Agrarian Economies, American Economic Review v.75 no.l. ------------------ (1985b): A Theory of Contractual Structure in Agricul ture, American Economic Review, v.75 no.3. 84 ____________ (1985c): Why are Capitalists the Bosses? (mimeo.) Foster-Carter, Aidan (1978): The Modes of Production Controversy, New Left Review no. 107. Frykenberg, R.E. (1969) (ed.): Land Control and Social Structure in Indian History, Wisconsin, University Press. Gangopadhyay, S. and Sengupta, K. (1986): Interlinkages in Rural Markets, Oxford Economic Papers no.38. Ghose, Ajit K.(ed.l983): Agrarian Reform in Contemporary Develop ing Countries, St. M artin Press. Ghosh, A. and Saith (1976): Indebtedness, Tenancy, and the Adoption of New Technology in Semi-feudal Agriculture, World Development v.4 no.4. H atcher (1981): English Serfdom and Villeinage: Towards a Reassess m ent, Past and Present no.90. Hilton, R. (1976)(ed.): The Transition from Feudalism to Capitalism, New Left Books, London. Klima, Arnost (1979): Agrarian Class-Structure and Economic Devel opment in Pre-Industrial Bohemia, Past and Present no 85. Kosinski, L.A. and Elahi, K.M. (eds.) (1985): Population Redistribu tion and Development in South Asia, D. Reidel, Amsterdam . Kotwal, Ashok (1985): The Role of Consumption Credit in Agricul tural Tenancy, Journal of Development Economics, 18. M cEachern, Doug (1976): The Mode of Production in India, Journal of Contemporary Asia, v.6 no.4. Mellor, J.W . and Desai, G.M.: Agricultural Change and Rural Poverty: Variations on a Theme by Dharm Narain, Johns Hopkins. Minchinton,W .E. (1968)(ed.): Essays in Agrarian History, vol.I, British Agricultural History Society, London. Mingay G.E. (1963): English Landed Society in the Eighteenth Cen tury. Routledge and Kegan Paul, London. 85 M itra, Tapan (1983): Sensitivity of O ptim al Programs with Respect to Changes in Target Stocks: The Case of Irreversible Investment, Journal of Economic Theory 29. Moore, Barrington, J r .(1966): Social Origins of Dictatorship and Democ racy : Lord and Peasant in the Making of the Modern World, Bea con Press, Boston. __________(1978): Injustice: The Social Bases of Disobedience and Re volt, M.E. Sharpe, New York. Newbery, D.M.G.(1975): Tenurial Obstacles to Innovation, Journal of Development Studies 11. (1977): Risk-sharing, Sharecropping, and Uncertain Labour M arkets, Review of Economic Studies,44. and Stiglitz (1978): Sharecropping, Risk Sharing, and the Im por tance of Imperfect Information, University of Cambridge Economic Theory Discussion Paper no.8. Nugent, J.B. and T.H. Thomas (ed.1985): Bahrain and the Gulf: Past Perspectives and Future Alternatives, Croom Helm, London. Postan, M.M. (1937): Chronology of Labour Services, in Minchinton (ed). Rubinstein,Ariel (1979): Equilibrium in Supergames with the Over taking Criterion, Journal of Economic Theory 21. R udra, Ashok (1981): Indian Agricultural Economics: M yths and Re alities, Allied Publishers, New Delhi. R udra, Ashok (1982): Extraeconomic Constraints on Agricultural Labour: Results of an Intensive Survey... ILO-ARTEP, Bangkok. _________(1984): Local Power and Farm-Level Decision Making, in Desai, Rudolph and R udra (ed.). Sarap, Kailash (1987): Transactions in Rural Credit M arkets in West ern Orissa, India. Journal of Peasant Studies, v.15 n o .l. Stokes, Eric (1978): Peasants and the Raj, Cambridge, University Press. Skeldon, Ronald (1985): M igration in South Asia : an Overview, in Kosinski and Elahi (ed.) Srinivasan, T.N. (1979): Agricultural Backwardness under Semi- Feud- alism -Com ment, Economic Journal. June.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Countertrade: A theoretical and empirical investigation

PDF

Irrigation institutions and technical change: A case study from South India

PDF

The political economy of the state and development strategies: A case study of Turkey, 1930-1980

PDF

The influence of industrial pressure groups within the People's Republic of China

PDF

Intensive growth vs. extensive growth and their historical implications

PDF

Weight variations among preschoolers: An analysis of evidence from rural Tamilnadu

PDF

An economic analysis of cooperative group behavior: Theory and evidence

PDF

Money, credit and economic growth in fragmented economies

PDF

Chinese and Japanese in California, 1900-1920: A case study of the impact of discrimination

PDF

Moving toward free trade in North America

PDF

Essays on some Lewisian themes of economic development

PDF

Essays on rent seeking

PDF

A case study of underutilization of canal irrigation in Bihar, India

PDF

A multidimensional analysis of regional integration and cooperation: The case of the Economic Community of West African States

PDF

China: Recent political and economic developments

PDF

Adaptive economizing in disequilibrium: Essays on economic dynamics

PDF

The role of structure, interdependence, and interventionary policy in the diffusion of innovations

PDF

The politico-economic determinants of ASEAN regional economic cooperation

PDF

Incomes policies in economic reform: A comparison of Hungary and Yugoslavia in the 1970s

PDF

The size distriubition of Turkish manufacturing establishments across sectors and over time and its responsiveness to economic policy changes

Asset Metadata

Creator Bose, Gautam (author)

Core Title Patronage, dependence and debt in peasant agriculture

Degree Doctor of Philosophy

Degree Program Economics

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Economics, Agricultural,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Nugent, Jeffrey B. (committee chair), Elliott, John E. (committee member), Hamilton, Nora Louise (committee member), Kuran, Timur (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c20-274208

Unique identifier UC11257958

Identifier DP23368.pdf (filename),usctheses-c20-274208 (legacy record id)

Legacy Identifier DP23368.pdf

Dmrecord 274208

Document Type Dissertation

Rights Bose, Gautam

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA