Agribusiness Management · Agri-sector specialisation · also AEC-512
Why institutions and transaction costs matter for how agricultural resources are used, and how agrarian institutions such as tenancy work.
In the exams: agricultural economics PhD entrance and fellowship examinations (light)
Institutional economics asks why the rules, customs, contracts and organisations around a transaction matter as much as prices and technology. A Bihar farmer leasing land on a half share, a Kenyan smallholder growing beans under an export contract and an Iowa corn grower with a marketing agreement are all choosing among institutional arrangements. Each arrangement shapes who bears risk, who invests, who cheats and how much output is lost along the way. Neoclassical theory, which assumes complete information and costless contracting, cannot explain why these arrangements exist or why they differ across places; institutional economics can.
The subject has two traditions. The old institutional economics of Veblen, Commons, Mitchell and Ayres treated institutions as habits of thought and collective action, and was sceptical of equilibrium theory. The new institutional economics (NIE) of Coase, Williamson, North and Ostrom keeps the tools of mainstream economics — scarcity, choice, comparative statics, game theory — but adds positive transaction costs, bounded rationality and opportunism. In agriculture the NIE has been unusually productive: the theories of share tenancy, interlinked contracts, common-property management, group lending and contract farming all come from it.
This manual follows the ICAR M.Sc. Agricultural Economics minor course AEC-512 but is written for a global reader, since the same models are taught in Europe, China, Africa and the Americas. In Indian entrance and fellowship papers the subject is examined lightly and mostly as author-idea matching (who wrote what, which concept belongs to whom) with occasional reasoning questions on sharecropping and the Coase theorem. Postgraduate programmes elsewhere examine it through short models: share tenancy, a principal–agent incentive constraint, hold-up, a repeated game. This manual works those models numerically.
Douglass North defined institutions as 'the rules of the game in a society or, more formally, the humanly devised constraints that shape human interaction'. They have three parts: formal rules (constitutions, statutes, land records, contract law, grading standards), informal constraints (customs, norms, taboos, codes of conduct, reputations) and the enforcement characteristics of both. North separated institutions from organisations. Organisations — firms, cooperatives, panchayats, trade unions, regulated market committees — are the players: groups of individuals bound by a common purpose who act within the rules and who also try to change them. A government that registers thousands of producer companies has created organisations; whether it has changed institutions depends on whether the rules on leasing, contracts and market entry let them do anything new.
Institutions reduce uncertainty by making other people's behaviour predictable. They do this at a cost: they must be designed, learned, monitored and enforced, and they can be captured by those with power. An institution is therefore neither good nor bad in itself; the institutional economist asks what it does to the costs of exchange and to the incentives to produce, invest and innovate. Informal institutions often fill gaps the state leaves: kinship networks enforce repayment without courts, and traders' reputations substitute for written contracts. Formal institutions that ignore such informal ones frequently fail, which is a recurring theme of this subject.
Thorstein Veblen (The Theory of the Leisure Class, 1899) attacked the hedonistic, calculating individual of marginalist economics and argued that economic behaviour is governed by evolving habits of thought. His followers Clarence Ayres and others developed the Veblen–Ayres dichotomy between 'ceremonial' behaviour, which protects status and existing power, and 'instrumental' behaviour, which solves problems with technology. Wesley Mitchell built empirical business-cycle research. John R. Commons (Institutional Economics, 1934) made the transaction the unit of economic analysis and defined an institution as 'collective action in control, liberation and expansion of individual action'. He classified transactions as bargaining transactions (between legal equals in a market), managerial transactions (between a superior and a subordinate inside an organisation) and rationing transactions (allocation of burdens and benefits by a collective authority such as a legislature).
The old school was holistic, evolutionary and inductive. Coase's criticism was that it produced 'a mass of descriptive material waiting for a theory'. Its attention to power, habit, law and history has returned through evolutionary economics and Geoffrey Hodgson. Examiners often ask students to match these names to ideas and to contrast the old school's methodological collectivism with the new school's methodological individualism.
Oliver Williamson coined the term 'new institutional economics' in 1975. The NIE accepts that individuals are rational but boundedly so (Herbert Simon's 'intendedly rational, but only limitedly so'), that some of them are opportunistic ('self-interest seeking with guile'), and that transacting is costly. From these premises it explains why firms exist, why contracts take the forms they do and why some societies grow and others stagnate. Its main branches are transaction cost economics (Coase, Williamson), property-rights economics (Alchian, Demsetz, Grossman, Hart, Moore), agency and contract theory (Jensen and Meckling, Holmström, Stiglitz), the economics of collective action and the commons (Olson, Ostrom) and the economic history of institutions (North, Greif, Acemoglu and Robinson).
Williamson (2000) organised the field into four levels. Level 1, embeddedness, holds informal institutions, customs, traditions, norms and religion; it changes over centuries or millennia and is largely taken as given by economists. Level 2, the institutional environment, holds the formal rules of the game — property law, the judiciary, the polity, the bureaucracy — and changes over decades to a century. Level 3, governance, is the play of the game: the choice of contracts and organisational forms (spot market, contract, hybrid, firm) to align with transactions; it changes over one to ten years. Level 4 is resource allocation and employment — prices, quantities and incentive alignment — adjusted continuously, the domain of neoclassical economics and agency theory. A sharecropping contract (Level 3) is shaped by tenancy law (Level 2) and caste and gender norms (Level 1), and it determines how much labour and fertiliser are applied to the plot (Level 4).
Ronald Coase asked in 'The Nature of the Firm' (1937) a question neoclassical theory could not answer: if the price mechanism allocates resources efficiently, why do firms exist, inside which resources are directed by command rather than by price? His answer was that 'there is a cost of using the price mechanism' — the cost of discovering relevant prices, negotiating and concluding a separate contract for each exchange, and writing contracts for the future when the future is uncertain. A firm replaces many market contracts with one employment contract under which the employee agrees to obey directions within limits. The firm expands until the cost of organising one more transaction inside it equals the cost of carrying it out through the market or in another firm. Internal organisation also has costs, which is why all production is not carried out by one large firm.
Carl Dahlman (1979) gave the standard classification of transaction costs: search and information costs (finding a buyer, learning the quality of seed or the price in a distant market), bargaining and decision costs (negotiating terms, drafting the contract) and policing and enforcement costs (checking that the other side performs, and acting if it does not). Kenneth Arrow called transaction costs 'the costs of running the economic system'. A useful distinction is between ex ante costs, incurred before the deal (search, drafting, safeguards), and ex post costs, incurred afterwards (monitoring, haggling over adaptation, dispute settlement, the cost of a breakdown). In agriculture, perishability, weather and quality variation make ex post costs large.
In 'The Problem of Social Cost' (1960) Coase used the example of a rancher whose straying cattle damage a neighbouring farmer's crop. The Pigovian tradition said the rancher should be taxed or made liable. Coase showed that the problem is reciprocal: preventing harm to the farmer harms the rancher. If property rights are well defined and transaction costs are zero, the two parties will bargain to the outcome that maximises their joint value whichever of them holds the legal right; the assignment of the right changes only who pays whom. Suppose a fence costs 200 a season and the crop damage without a fence is 300. If the rancher is liable he builds the fence, because 200 is less than 300. If the rancher is not liable, the farmer pays for the fence (or pays the rancher up to 300 to build it), and the fence is still built. The allocation is the same; the distribution differs.
The lesson Coase drew is the opposite of the one often attributed to him. Because transaction costs are never zero, the legal assignment of rights and the design of institutions matter for efficiency. When the farmer faces fifty ranchers, or when the damage is hard to measure, bargaining breaks down and the law must place the right where the parties would have placed it had they been able to bargain, or choose the cheapest institution for the job — a liability rule, a regulation, a tax, a merger. The theorem is a benchmark, not a prediction. It also assumes no wealth effects and symmetric information.
Transaction costs are hard to measure because many are opportunity costs or are avoided by choosing another arrangement. Applied work measures observable costs directly (travel time, commissions, grading, chasing a defaulted payment), infers them from behaviour, or tests whether observed governance choices match transaction attributes. De Janvry, Fafchamps and Sadoulet (1991) showed that transaction costs open a band between the price at which a household could sell and the higher price at which it could buy; households whose shadow price falls inside the band stay out of the market. Fixed transaction costs, which do not rise with quantity, penalise small sellers per unit. This explains why smallholders often sell to a village trader rather than a distant wholesale market, and why producer organisations that bulk produce can raise members' net prices without any change in the market price.
Williamson built transaction cost economics (TCE) on two behavioural assumptions. Bounded rationality means that all complex contracts are unavoidably incomplete: the parties cannot foresee and write down every contingency. Opportunism means that some parties will exploit gaps in a contract, misrepresent information or renege when it pays. If people were fully rational, comprehensive contracts could be written; if people were never opportunistic, incomplete contracts could be completed by promises. It is the combination that creates a contracting problem.
Transactions differ along three dimensions. Asset specificity is the degree to which an asset can be redeployed to alternative uses and users without loss of productive value; it is the most important dimension. Uncertainty concerns disturbances that require the parties to adapt the contract. Frequency matters because recurrent transactions justify the fixed cost of a specialised governance structure. Williamson identified six types of asset specificity: site specificity (a sugar mill and the cane fields around it, since cane loses sucrose quickly after cutting), physical asset specificity (a dryer designed for one processor's specification), human asset specificity (knowledge learned by doing a particular job), dedicated assets (capacity added to serve one buyer that would be idle if the buyer left), brand-name capital, and temporal specificity (where timing is critical, as with perishable milk or fruit that must be processed within hours). The last was emphasised by Masten and is especially important in agriculture.
Before an investment is made, a buyer may choose among many suppliers; this is large-numbers bidding. Once one party has made a specific investment, the relationship becomes bilateral: each side is locked in. Williamson called this the fundamental transformation. Klein, Crawford and Alchian (1978) defined the appropriable quasi-rent as the excess of an asset's value in its current use over its value in its next-best use. That quasi-rent is what the other party can try to appropriate by threatening to walk away. The resulting hold-up problem is that a party which anticipates being held up after investing will under-invest beforehand.
A worked example makes this concrete. A grower can build a cold room costing 160 that is useful only for supplying one exporter. If the grower supplies the exporter, the joint surplus is 300; without the relationship the cold room is worth nothing. After the cold room is built the cost is sunk, so if the parties split the 300 equally by bargaining, the grower receives 150. Since 150 is less than 160, the grower will not build the cold room, even though the investment creates 300 of value for a cost of 160. If the cold room had cost 120, the grower would build it, because 150 exceeds 120. Hold-up therefore destroys efficient investments whenever the investor's bargaining share of the surplus falls short of the cost. Remedies include long-term contracts with price formulas, the other party co-financing or owning the specific asset, posting a hostage or bond, reputation in a repeated relationship, and vertical integration. The famous case of General Motors and Fisher Body was presented as an example of hold-up resolved by integration; later historical work by Coase and others disputed the facts, so the case is contested and best cited with care.
TCE's central hypothesis is discriminating alignment: transactions, which differ in their attributes, are aligned with governance structures, which differ in their costs and competences, so as to economise on transaction costs. The three generic governance structures are the market (classical contracting, strong price incentives, autonomous adaptation, court ordering), the hybrid (long-term contracts, franchises, alliances, cooperatives, contract farming; intermediate incentives and administrative controls) and the hierarchy (the firm; weak incentives but strong coordinated adaptation by fiat and internal dispute resolution). Williamson (1991) showed that if governance cost is drawn against asset specificity, the market is cheapest at low specificity, the hybrid at intermediate specificity, and the hierarchy at high specificity.
A numerical version: suppose the governance cost of a transaction with specificity k is 20 + 6k under the market, 30 + 3k under a hybrid and 45 + k under hierarchy. The market and hybrid cost the same at k = 10/3, about 3.3; the hybrid and hierarchy cost the same at k = 7.5. For k below about 3.3 the market is cheapest, between 3.3 and 7.5 the hybrid, and above 7.5 the hierarchy. At k = 5, the costs are 50, 45 and 50, so the hybrid is chosen. Greater uncertainty steepens the cost lines of the market and hybrid more than that of the hierarchy and so shifts the switch points to lower k. In agri-food chains, fresh milk, cane and processing vegetables (site and temporal specificity) are rarely traded spot, while storable grains usually are. Ménard extended the analysis of hybrids, showing that they need mechanisms such as joint planning, pooled resources and private dispute resolution to work.
A property right is a socially enforced right to select uses of a resource. Economists treat ownership as a bundle of separable rights: the right to use an asset, the right to the income from it, and the right to transfer or alienate it. Schlager and Ostrom (1992) refined the bundle for natural resources into five rights: access (entering a defined area), withdrawal (taking resource units such as fish or water), management (regulating use and improving the resource), exclusion (deciding who may have access) and alienation (selling or leasing the management and exclusion rights). Holders range from authorised users (access and withdrawal only), claimants (plus management), proprietors (plus exclusion) to owners (all five). A tenant cultivator typically holds use and part of the income rights but not alienation; a member of a forest user group may hold withdrawal and management rights without exclusion.
Four property regimes are distinguished. Under private property an individual or firm holds the bundle. Under state property a public agency holds it. Under common property a defined group holds rights and duties and excludes outsiders. Open access is the absence of property: nobody can exclude anybody. Ciriacy-Wantrup and Bishop (1975) and later Ostrom insisted that common property and open access must not be confused. Garrett Hardin's 'tragedy of the commons' (1968), in which each herder adds animals to a pasture because he gets the full gain and bears only a fraction of the overgrazing cost, is a tragedy of open access. Many real commons — Swiss alpine meadows, Japanese village forests, irrigation systems in Nepal and Spain, the shamlat and gochar lands of India — have been governed sustainably for long periods by their users.
Harold Demsetz (1967) argued that 'property rights develop to internalise externalities when the gains of internalisation become larger than the cost of internalisation'. His example, drawn from the anthropologist Eleanor Leacock, was the Montagnais (Innu) of the Labrador peninsula: as the commercial fur trade raised the value of beaver pelts, hunting territories became private family territories, because over-trapping a common stock had become costly enough to justify the cost of marking and policing boundaries. The same logic predicts that land rights become more individualised as population pressure and market access raise land values.
The theory is a demand-side story and has been criticised as 'naive' because it ignores who supplies property rights and how. Rights are supplied through politics, and those with power may create rights that benefit themselves rather than rights that maximise total value; enclosure movements show both effects. Technology such as barbed wire or digitised land records lowers the cost of exclusion. Barzel stressed that rights are never complete: attributes too costly to measure are left in the 'public domain'.
Secure land rights are expected to raise agricultural investment through three channels: assurance (a farmer who expects to keep a plot invests in trees, terraces and soil fertility), collateral (a titled plot can be mortgaged, easing credit) and gains from trade (secure rights permit renting and selling land to more productive users). Hernando de Soto (The Mystery of Capital, 2000) popularised the collateral argument, calling untitled assets 'dead capital'. The empirical evidence is mixed. Besley (1995) found that stronger transfer rights raised tree planting in parts of Ghana, while noting that causation can run the other way because planting trees can itself strengthen claims. Goldstein and Udry (2008) found that in Ghana farmers with less political power in the local lineage fallowed their plots less, for fear of losing them, and so had lower yields. Title raises credit little where foreclosure on land is socially or legally difficult.
In the theory of the firm, Grossman and Hart (1986) and Hart and Moore (1990) redefined ownership as the holding of residual rights of control: the right to decide how an asset is used in circumstances not specified in a contract. Because contracts are incomplete, the owner's position in renegotiation is stronger, so ownership encourages the owner's relationship-specific investment and discourages the non-owner's. Ownership should go to the party whose investment matters more for joint surplus. Hart and Holmström shared the 2016 Nobel prize for contract theory.
An agency relationship exists whenever one party, the principal, engages another, the agent, to act on its behalf and the agent's actions or information are not fully observable. Landlord and tenant, processor and grower, and cooperative members and their manager are examples. Jensen and Meckling (1976) defined agency costs as the sum of the principal's monitoring expenditures, the agent's bonding expenditures and the residual loss that remains. Alchian and Demsetz (1972) explained the classical firm as a solution to team production: when individual contributions to joint output cannot be separated, members shirk, so a monitor is needed, and the monitor is given the residual claim so that he has an incentive to monitor.
Two information problems are distinguished. Moral hazard (hidden action) arises after a contract is signed, when the agent's effort or care cannot be observed: an insured farmer may neglect his crop, a sharecropper may apply less labour than agreed. Adverse selection (hidden information) arises before a contract, when one side knows its type: Akerlof's (1970) market for lemons showed how a market for used cars, or for uncertified seed, can unravel when buyers cannot distinguish quality and pay only for average quality. Responses are signalling by the informed party (Spence: certification, warranties) and screening by the uninformed party (Rothschild and Stiglitz: a menu of contracts that makes types self-select). Akerlof, Spence and Stiglitz shared the 2001 Nobel prize.
The standard moral-hazard model has the principal choose a payment schedule to maximise expected profit subject to two constraints. The participation (individual rationality) constraint requires that the agent's expected utility from the contract is at least his reservation utility. The incentive-compatibility constraint requires that the agent prefers the action the principal wants over any other. When the agent is risk-neutral and has no wealth constraint, the first-best can be reached by 'selling the firm to the agent' — a fixed rent — because the agent then becomes residual claimant and bears all consequences of his effort. When the agent is risk-averse, a fixed rent loads all risk on him; a fixed wage gives him no incentive. The optimal contract trades off risk-sharing against incentives, which is the basic logic of the sharecropping literature. Holmström's (1979) informativeness principle states that any signal that carries information about the agent's effort should be used in the contract — for instance, a contract grower's pay can be based on his yield relative to the average of other growers in the same area, which filters out common weather shocks.
A numerical example with limited liability: a project succeeds with probability 0.8 if the agent works hard and 0.4 if he shirks; success yields 1,000 and failure 200. Hard work costs the agent 60. The agent is risk-neutral, his reservation payoff is zero and he cannot be paid less than zero. With a bonus b paid only on success, the incentive constraint is 0.8b − 60 ≥ 0.4b, so b ≥ 60/0.4 = 150. Paying exactly 150, the agent's expected payoff is 0.8 × 150 − 60 = 60, more than his reservation payoff of zero: this 60 is a limited-liability rent that the principal cannot extract. The principal's expected profit is 0.8 × 1,000 + 0.2 × 200 − 0.8 × 150 = 720, higher than the 520 (0.4 × 1,000 + 0.6 × 200) earned by paying nothing and accepting shirking, so the principal induces effort.
Complete contract theory assumes that everything relevant can be written down and enforced by a court. Incomplete contract theory accepts that many variables — the quality of fruit, the timeliness of collection, the effort of a field officer — are observable to both parties but not verifiable to a third party. Contracts then rely on renegotiation, on ownership of assets (which determines bargaining positions) and on self-enforcement through relational contracts: agreements sustained by the value of future dealings. A relational contract holds if, for each party, the one-time gain from reneging is smaller than the present value of the future surplus that would be lost.
Suppose a farmer under contract to a processor is offered 24 per kg on the spot market against a contract price of 20, on a harvest of 100 kg; side-selling would gain 400 this season. If the contract yields the farmer 150 per season more than his alternative and he discounts each future season by a factor of 0.8, the present value of continuing the relationship from next season is 150 × 0.8 / (1 − 0.8) = 600. Since 600 exceeds 400, a farmer who expects to be dropped after side-selling will not side-sell. If the spot price rose to 27, the gain (700) would exceed 600 and the relationship would break down.
Land that is not cultivated by its owner can be worked under three broad contracts. Under a fixed-rent (lease) contract the tenant pays a fixed sum in cash or kind and keeps the rest of the output, bearing all risk. Under a wage contract the owner hires labour at a fixed wage, bears all risk and keeps all output. Under a share contract (sharecropping, métayage, batai, mezzadria) output is divided in a fixed proportion — half-and-half is common, though shares vary by region, crop, irrigation and who supplies inputs. Sharecropping remains widespread in South Asia and parts of Africa and Latin America.
Alfred Marshall pointed out that a sharecropper who keeps a share s of output will apply a variable input such as labour only until s times its marginal product equals its cost, while an owner-cultivator or fixed-rent tenant applies it until the full marginal product equals its cost. Output and input use are therefore lower under sharecropping: the share acts like a tax on output. A worked example: output is Q = 40√L, the output price is 1 and the opportunity cost of a unit of labour is 2. The marginal product of labour is 20/√L. A fixed-rent tenant sets 20/√L = 2, so L = 100 and Q = 400; the surplus over labour cost is 400 − 200 = 200. A half-share tenant sets 0.5 × 20/√L = 2, so √L = 5, L = 25 and Q = 200; the surplus is 200 − 50 = 150. The Marshallian inefficiency, the lost surplus, is 50, and output halves. The same logic applies to fertiliser if the tenant bears its full cost but receives only part of the output; Heady (1947) showed that sharing input costs in the same proportion as output removes the distortion for that input.
If sharecropping is inefficient, why does it survive? Steven Cheung (1969) argued that the landlord need not accept low effort: he can specify the tenant's labour input and plot size in the contract and enforce it, in which case share tenancy is as efficient as fixed rent or wage contracts. D. Gale Johnson (1950) earlier noted that short leases and the threat of eviction can discipline tenants. These monitoring explanations make the inefficiency a matter of how costly enforcement is.
Joseph Stiglitz (1974) gave the risk-sharing explanation. A fixed-rent contract gives full incentives but places all output risk on the tenant; a wage contract places all risk on the landlord but gives the labourer no incentive to work hard. When the tenant is risk-averse and effort is unobservable, the optimal contract combines a fixed payment with a share of output, trading incentives for insurance. Under a mean-variance illustration in which output has mean 100 and variance 400, the tenant's risk-aversion coefficient is 0.05, and certainty equivalent is mean minus 0.025 times variance, a fixed rent leaving the tenant an expected 50 gives him a certainty equivalent of 50 − 0.025 × 400 = 40, while a half share with the same expected 50 exposes him to only a quarter of the variance (100) and gives 50 − 0.025 × 100 = 47.5. Newbery showed that a mix of fixed-rent and wage contracts could replicate the same risk-sharing, so risk alone does not explain sharecropping.
Other explanations are screening (landlords offer a menu of contracts; more able tenants self-select into fixed rent, as in Hallagan 1978), limited liability (poor tenants cannot pay a large fixed rent in a bad year, so the landlord cannot sell them the farm and must use a share), and the sharing of non-labour inputs such as bullocks and management skill held by the landlord or tenant (Bliss and Stern's study of Palanpur, Bell's work, and the 'double moral hazard' models in which both parties provide unobservable inputs).
Radwan Shaban (1987), using ICRISAT village data from India, compared plots owned and plots sharecropped by the same households, which controls for household ability. Input use and output per hectare were significantly lower on sharecropped plots, supporting the Marshallian prediction where monitoring is imperfect. Other studies find smaller or no differences where monitoring is close; results vary by region and crop.
Tenancy reform tries to improve tenants' security and bargaining position. West Bengal's Operation Barga, launched in 1978, registered sharecroppers (bargadars), gave them heritable protection from eviction and regulated the crop share in the tenant's favour. Banerjee, Gertler and Ghatak (2002) estimated that it raised agricultural productivity significantly. Their model separates the channels: a higher share raises effort (the Marshallian channel); security of tenure raises the tenant's incentive to invest but also removes the landlord's threat of eviction as an incentive device, so its net effect is ambiguous in theory and had to be settled by the data. Elsewhere in India, many state laws ban or restrict leasing, cap rents or confer ownership on long-standing tenants. These rules often drive tenancy underground as oral, insecure 'concealed tenancy', which shuts tenants out of institutional credit, insurance and support schemes. The Model Agricultural Land Leasing Act, 2016, circulated by NITI Aayog, proposed legalising and recording leases while guaranteeing the owner's return of land at the end of the lease, so as to encourage leasing without loss of ownership. Leasing law differs widely across countries and Indian states and must be checked locally.
An interlinked (interlocked) transaction is one in which two or more exchanges between the same pair of parties are made jointly, with the terms of one conditional on the other. Common forms are land-credit (a landlord lends to his tenant), labour-credit (an employer advances credit against peak-season labour at a stated wage), credit-output (a trader lends on condition that the crop is sold to him) and input-output (a dealer supplies fertiliser on credit against the harvest).
Interlinkage complicates the measurement of prices. A trader-lender who charges no explicit interest may still be earning a high return through the output price. Suppose a farmer borrows 10,000 for six months at 'zero interest' and must sell 50 quintals to the lender at 5 per cent below the market price of 2,000 per quintal. The farmer gives up 0.05 × 50 × 2,000 = 5,000, which is 50 per cent of the loan for six months, or 100 per cent a year in simple terms. Analysing one market alone can badly mislead.
Amit Bhaduri (1973) offered the 'semi-feudalism' view: a landlord who is also the tenant's creditor may resist technical change, because higher productivity would let the tenant escape debt and cut the landlord's interest income, which can exceed the gain from his share of the extra output. Interlinkage, on this view, is an instrument of exploitation that keeps agriculture backward. Later theorists, including Braverman and Stiglitz (1982), showed that a landlord who can adjust the share and interest rate together would usually prefer to capture the gains of innovation rather than block them.
The efficiency view treats interlinkage as an institutional response to missing and imperfect markets. Lending to a tenant or labourer is cheaper for a landlord or employer who already observes the borrower's effort and output, and the threat of losing access to land or employment acts as a substitute for collateral. A trader-lender collects repayment directly out of the crop he buys, which saves enforcement costs. Braverman and Stiglitz showed that a landlord who lends to his sharecropper can use the loan terms to influence the tenant's effort and choice of technique, partly offsetting the Marshallian inefficiency. Most economists now accept that interlinkage can both reduce transaction costs and concentrate market power, and that which effect dominates is an empirical question.
Informal rural credit markets show high and widely dispersed interest rates alongside rationing. Irfan Aleem's (1990) study of moneylenders in Chambar, Pakistan, found that lenders' screening and collection costs were high, so high interest rates reflected costs and default risk more than monopoly profit; Hoff and Stiglitz (1990) explained rural credit markets with screening, incentive and enforcement problems under imperfect information. Markets are fragmented because each lender knows only a few borrowers well. Stiglitz and Weiss (1981) showed that lenders may ration credit rather than raise the interest rate, because a higher rate attracts riskier borrowers (adverse selection) and induces riskier projects (moral hazard).
Policy has tried to break exploitative interlinkages through institutional credit, warehouse receipt finance and regulated markets. Subsidised formal credit was often captured by larger farmers. Modern contract farming, input-supplier credit and digital platforms re-create interlinkage in new form — a processor that supplies seed and fertiliser on credit and deducts it from payment for the crop is running an interlinked contract. The task for a policy-maker is not to ban interlinkage but to ensure competition among lenders and buyers, transparent price and quality terms, and accessible dispute resolution.
Mancur Olson (The Logic of Collective Action, 1965) challenged the assumption that groups with common interests will act to further them. Because the benefits of a collective good — a higher support price won by a farmers' lobby, a repaired canal, a common brand — go to all group members whether or not they contributed, each rational member has an incentive to free ride. Olson argued that large groups are less likely to provide collective goods than small groups, because each member's share of the benefit is smaller, contributions are less noticeable, and the costs of organising are higher. He classified groups as privileged (at least one member gains enough from the good to provide it alone), intermediate (no one will provide it alone, but the group is small enough that members notice each other's behaviour, so coordination may succeed) and latent (large groups that will not provide the good without coercion or selective incentives). Selective incentives are private benefits or penalties given only to contributors, such as a cooperative's input supply or insurance available only to members, or social sanctions.
This explains why small, concentrated producer groups often win more policy favours than millions of dispersed farmers or consumers. Critics note that heterogeneity, communication, leadership and trust matter as much as group size.
Collective-action problems are often modelled as games. In a one-shot prisoner's dilemma with payoffs temptation T = 5, reward for mutual cooperation R = 3, punishment for mutual defection P = 1 and sucker's payoff S = 0, defection is a dominant strategy and both players end with 1 though both would prefer 3. If the game is repeated indefinitely and each player discounts the future by a factor δ, a grim-trigger strategy (cooperate until the other defects, then defect forever) sustains cooperation when the one-period gain from defecting, T − R, is no more than the discounted loss, δ(R − P)/(1 − δ). Rearranged, cooperation is sustainable if δ ≥ (T − R)/(T − P) = (5 − 3)/(5 − 1) = 0.5. Repetition, patience and the ability to observe and punish defection therefore favour cooperation; this is the logic of the folk theorem and of community enforcement in villages.
A linear public-goods game gives a group version. Four members each have 20 units; whatever they put in a common fund is multiplied by 1.6 and shared equally. The marginal per-capita return of a contribution is 1.6/4 = 0.4, less than 1, so each individual gains by keeping his money: if the other three contribute everything, a free rider earns 20 + 1.6 × 60/4 = 44, while a full contributor earns 1.6 × 80/4 = 32 when all contribute, and only 8 if he alone contributes. Yet all contributing (32 each) beats all free riding (20 each). Some collective-action problems are instead assurance (stag-hunt) games, in which cooperating is the best reply if others cooperate, so the task is building trust.
Elinor Ostrom (Governing the Commons, 1990) showed that communities often manage common-pool resources — resources from which exclusion is costly and whose use by one subtracts from others' — without either privatisation or state control. She distilled eight design principles of long-enduring institutions: (1) clearly defined boundaries of the resource and of who may use it; (2) congruence between appropriation and provision rules and local conditions; (3) collective-choice arrangements allowing most users to take part in changing the rules; (4) monitoring by people accountable to the users; (5) graduated sanctions for rule-breakers; (6) quick, low-cost conflict-resolution mechanisms; (7) minimal recognition by government of the community's right to organise; and (8) for larger systems, nested enterprises in which appropriation, monitoring, enforcement and governance are organised in multiple layers. Robert Wade's study of irrigation and grazing in South Indian villages (Village Republics, 1988) and N. S. Jodha's work on the decline of common property resources in India's dry regions are classic South Asian contributions.
Ostrom's Institutional Analysis and Development (IAD) framework organises the study of any 'action situation' by the biophysical conditions, the attributes of the community and the rules in use. Experiments found that communication and sanctioning raise cooperation well above selfish predictions. A simple fishery model shows what is at stake: if total catch is Y = 40E − E² for fishing effort E, the fish price is 1 and the cost of effort is 8 per unit, open access drives effort to the point where average revenue equals cost (40 − E = 8, so E = 32, catch 256, and rent 0), whereas a sole owner or a well-governed community sets marginal revenue equal to cost (40 − 2E = 8, so E = 16, catch 384 and rent 384 − 128 = 256). Open access dissipates the whole rent; good common-property rules can capture it.
The 'group approach' channels services through organised groups of farmers or households rather than individuals. Its rationale is institutional: groups lower the transaction cost per member of delivering credit, extension or inputs; they permit bulking for sale; and they use members' local information and social ties to screen, monitor and enforce. Examples include bank-linked self-help groups, joint-liability groups, water user associations, joint forest management committees, cooperatives and farmer producer organisations.
Joint liability lending illustrates the mechanisms. Stiglitz (1990) emphasised peer monitoring: because members are liable for each other, they watch each other's use of loans. Ghatak (1999) emphasised peer selection: safe borrowers group with safe borrowers, which reduces adverse selection. Besley and Coate (1995) emphasised social collateral: members pressure each other to repay. A simple calculation shows the insurance effect: if each of two members' projects succeeds independently with probability 0.8 and either successful member can repay for both, the loan is repaid with probability 1 − 0.2 × 0.2 = 0.96, against 0.8 for an individual loan. Members can also collude against the lender, and many lenders have moved toward individual liability. Groups work best when Ostrom's conditions are met: clear membership, rules made by members, monitoring and graduated sanctions.
Why do institutions change? Vernon Ruttan and Yujiro Hayami extended their theory of induced technical innovation (technology evolves to economise on the scarce factor, as land-saving biological technology in Japan and labour-saving mechanical technology in the United States) to institutions. In 'Toward a theory of induced institutional innovation' (1984) they argued that changes in relative factor and product prices, and in technology, create a demand for new institutions, and that the supply of institutional change depends on the cost of collective action, on cultural endowments and on the political power of those who would gain or lose. Their Philippine example was the shift in rice harvesting from the hunusan system, open to any villager for a customary share, to the gama system, in which only workers who had weeded the field without pay could harvest. As population growth lowered real wages, gama brought the harvesters' effective share down to the market wage without openly breaking custom.
Lance Davis and Douglass North (1971) had earlier modelled institutional change as occurring when the expected net gains from a new arrangement exceed the costs of making the change, with 'primary action groups' as the entrepreneurs who perceive and organise to capture those gains.
North's later work (Institutions, Institutional Change and Economic Performance, 1990) rejected the efficiency view that institutions evolve smoothly toward the best arrangement. Institutions are created by those with the bargaining power to devise rules in their own interest, and inefficient institutions can persist for centuries. Path dependence arises because institutions display increasing returns: large set-up costs, learning effects, coordination effects (people invest in skills and relationships that fit the existing rules) and adaptive expectations. The idea draws on David's QWERTY and Arthur's lock-in models. Organisations that have grown up under a set of rules have a stake in keeping them. Change is mostly incremental; revolutions alter formal rules faster than informal constraints, so outcomes often differ from reformers' intentions.
North also stressed mental models and ideology: people interpret the world through shared beliefs, and change in beliefs is part of institutional change. With Barry Weingast (1989) he argued that England's Glorious Revolution of 1688 created credible commitment by the Crown not to expropriate, through parliamentary control of taxation and an independent judiciary, which lowered the cost of public borrowing. Avner Greif's work on the medieval Maghribi traders showed how a coalition with multilateral punishment of cheating agents enforced long-distance trade without courts — an example of informal institutions supporting markets.
Agricultural history provides many cases of institutional change. Land reforms after the Second World War in Japan, Korea and Taiwan transferred land to tenants and are widely credited with supporting rapid agricultural growth, while reforms in many other countries stalled. China's household responsibility system, which spread from about 1978 and returned farming decisions and residual output to households while land remained collectively owned, was followed by a large jump in output; Justin Yifu Lin's analysis attributed a large part of the growth in the early 1980s to the change in incentives. In India, the shift from administered and regulated marketing toward private markets, electronic trading platforms and contract farming has been gradual and contested, and the 2020 farm laws were repealed in 2021 after farmer protests — a reminder that the political supply of institutional change can block reforms for which economists see demand.
The lessons for a practitioner are consistent across countries. Changes in prices and technology create pressure for change, but the outcome is shaped by who has power, by existing informal norms and by whether those who lose can be compensated. Formal rules copied without supporting informal institutions and enforcement often remain on paper.
Why are some countries rich and others poor? Geography, culture and institutions are the three classes of answer. The institutions hypothesis holds that societies with secure property rights, constraints on rulers and broad access to economic opportunity invest and innovate, while those whose institutions allow elites to expropriate do not. The empirical challenge is that institutions and income cause each other: rich countries can afford good courts.
Daron Acemoglu, Simon Johnson and James Robinson (2001) addressed this with an instrumental variable. In former colonies where European settlers faced high mortality, colonisers set up extractive institutions designed to transfer resources to the metropole; where settler mortality was low, settlers came in large numbers and built institutions that protected their own property. Those early institutions persisted. Using settler mortality as an instrument for current institutions (measured by protection against expropriation), they found a large effect of institutions on income per head today. Their 'reversal of fortune' paper (2002) showed that among former colonies, those that were relatively prosperous in 1500 are relatively poor today, which they attributed to the imposition of extractive institutions in densely populated, prosperous areas. Acemoglu and Robinson's Why Nations Fail (2012) generalised the argument into a contrast between inclusive and extractive economic and political institutions. Acemoglu, Johnson and Robinson received the 2024 Nobel prize for studies of how institutions are formed and affect prosperity.
The institutional view has critics. Glaeser, La Porta, Lopez-de-Silanes and Shleifer (2004) argued that the usual measures of institutions capture outcomes rather than durable constraints, and that human capital may be the more basic cause of growth, with institutions improving as countries grow. Rodrik, Subramanian and Trebbi (2004) found that institutions 'rule' over geography and trade, while noting that this says little about which specific reforms to make; many different arrangements can deliver security and incentives.
Agricultural evidence is strong. Abhijit Banerjee and Lakshmi Iyer (2005) compared Indian districts where the British vested land-revenue liability in landlords (zamindari areas) with those where it was settled directly with cultivators (raiyatwari or mahalwari areas). Decades after independence, former landlord districts had lower agricultural investment, lower productivity and lower public investment in health and education, which they linked to the class conflict and weaker collective action the landlord system left behind. Dell (2010) found persistent lower consumption in Peruvian areas once subject to the colonial mita forced-labour system. Such studies support the view that land and labour institutions have long-lasting effects on rural development.
Williamson's fourth level, resource allocation, is where institutions show up in the numbers a farm economist measures. Insecure tenure lowers investment in land; sharecropping lowers labour and fertiliser use; missing credit and insurance markets push households into low-risk, low-return crop mixes; high transaction costs keep them out of markets. Each of these is a misallocation relative to the neoclassical benchmark, and each has an institutional cause.
This gives a practical diagnostic sequence. First, describe the transaction and its attributes (specificity, uncertainty, frequency, measurability). Second, describe the institutional environment: property rights, contract law and its enforcement, market regulation, informal norms. Third, identify the governance structure actually used and the incentives it creates for each party. Fourth, compare the resulting allocation with feasible alternatives, remembering that the relevant benchmark is another real institution with its own costs, not a frictionless ideal — what Demsetz called avoiding the 'nirvana fallacy'.
Vertical coordination is the way successive stages of a value chain — input supply, farm production, assembly, processing, distribution — are harmonised. Mighell and Jones (1963) introduced the term in agricultural economics. The forms range along a continuum from spot markets (open-market prices coordinate), through specification and marketing contracts, production contracts, strategic alliances and cooperatives, to full vertical integration (ownership of successive stages). The TCE prediction is that coordination moves toward the integrated end as asset specificity, uncertainty, quality requirements and the need for timing rise.
Contract farming is agricultural production carried out under an agreement between a buyer and farmers that sets conditions for production and marketing. Marketing contracts specify price or a price formula, quantity, quality and delivery time, leaving production decisions to the farmer. Production contracts (resource-providing contracts) go further: the buyer supplies inputs such as chicks, feed, seed or technical advice and often owns the product, and the grower is paid for services, frequently through a tournament or relative-performance scheme. Broiler production is a well-known example of production contracts; fruit and vegetable export schemes often use marketing contracts with some input provision.
For farmers, contracts can provide an assured market, reduced price risk, access to inputs, credit and technology, and higher incomes; for buyers they provide a reliable supply of specified quality and traceability for food-safety standards. The costs are also institutional. Farmers face hold-up after specific investments; buyers may reject produce on quality grounds when market prices fall; contract terms may be unclear or written in an unfamiliar language; and dependence on one buyer creates monopsony power. Buyers face side-selling when spot prices exceed the contract price, and diversion of supplied inputs.
Whether contract farming excludes smallholders is debated. Fixed transaction costs per supplier favour larger growers, but firms do contract with small farmers when land is held in small units, when the crop is labour-intensive and needs careful handling (where family labour has a supervision advantage), or when producer organisations lower the cost of dealing with many suppliers. The welfare effects on participants are generally found to be positive in reviews of the literature, but selection bias — better-off farmers choose to participate — makes many estimates fragile, and results vary by crop, country and contract design.
Good contract design follows from the theory. Price formulas that share risk (a base price plus a share of the spot price premium) reduce side-selling incentives when spot prices rise and protect farmers when they fall. Transparent, verifiable quality grading, ideally by a third party, reduces opportunistic rejection. Input credit should be recovered at a stated rate. Contracting through producer organisations lowers transaction costs and adds peer monitoring, and dispute resolution should be quick and cheap.
India's Model Agricultural Produce and Livestock Contract Farming and Services (Promotion and Facilitation) Act, 2018 was circulated to states as a model law providing for registration of agreements, pre-agreed prices and dispute settlement at the local level, and protecting farmers' land from transfer under such agreements; adoption is left to states. A central law on farm agreements enacted in 2020 was repealed in 2021. FAO, UNIDROIT and IFAD published a Legal Guide on Contract Farming (2015). The United States regulates poultry and livestock production contracts under the Packers and Stockyards Act. Practitioners must check the current law of the jurisdiction concerned.
A student must be able to match each founding author with the idea and the decade, state the Coase theorem with its conditions, list the dimensions of transactions and types of asset specificity, distinguish common property from open access, recite Ostrom's design principles, and explain why a sharecropper applies less labour than an owner. Above all a student should be able to work the short models — the Marshallian share-tenancy example, a principal-agent bonus under limited liability, a repeated-game threshold and a hold-up calculation — because these turn up in fellowship papers and in every postgraduate course on agrarian institutions.
A manager five years into an agribusiness or development job uses this subject when designing a contract-farming scheme, a procurement arrangement or a farmer group. The decisions are concrete: whether to buy on the spot market, contract or integrate; how to set a price formula that limits side-selling; how to grade quality so that rejections are not seen as opportunistic; how to recover input credit; how large a group should be and what rules it needs; and how local tenancy and leasing practice affects who actually farms the land and who can sign a contract or receive credit.
A senior professional with 25 years of work experience — heading procurement for a processor, running a cooperative federation, or leading a rural finance or land programme — uses institutional economics to judge reforms and organisational strategy. Examples are deciding whether to own a cold chain or contract it out, negotiating with a state government over leasing and contract-farming rules, restructuring a cooperative whose members side-sell, and assessing whether a land-records or titling programme will change investment or merely shift power. The habit is to compare real alternatives, each with its own costs, rather than an ideal.
A doctoral researcher with long field experience works at the frontier: identifying causal effects of tenure security, tenancy reform, contract farming or producer organisations with credible designs; testing between screening, risk-sharing and moral-hazard explanations of contract choice; measuring transaction costs; modelling relational contracts and collective action with game theory and experiments; and studying institutional persistence and change with historical data. Such a researcher also advises governments and multilateral agencies, where the task is translating theory on incomplete contracts and commons governance into enforceable, context-specific rules.
This manual does not yet cover in depth: formal mechanism design beyond the two-outcome moral-hazard model; the econometrics of contract choice and of institutional persistence; evolutionary and behavioural institutional economics after Veblen; the details of tenancy and land-leasing law in individual Indian states and other countries (which change and must be checked locally); cooperative governance theory (covered in the cooperatives subject); and quantitative estimates from individual empirical studies, which are described qualitatively because exact effect sizes vary across samples and were not all re-verified.
Original practice questions written to the syllabus and pattern of the exams named in each question's examTags. They are not past papers and do not reproduce any board's questions.
These notes are free to read here, with no account.
What the Laisha app adds:
Free for ten days, then a subscription. We would rather write that here than have you find it later.