results show clearly that landlords play a significant role in providing these tasks in sharecropped lands and
a minimal role in the fixed-rented farms. Roumasset [1995] reports further evidence of this division of
managerial tasks by type of contract. Under Roumasset’s classification, the landlord carries out land and
asset management as well as most production decisions, whereas the tenant provides the control of labor
and some of the more routine production decisions in sharecropped lands. This mostly anecdotal evidence
tells us that sharecropping allows for a skill-abundant landlord and a time-abundant tenant the right
incentive schemes to specialize in providing the labor tasks in which they possess a comparative
advantage.
Eswaran and Kotwal [1985a], in their seminal theoretical work, focus on two types of skilled labor
inputs, management and supervision. They argue that tenancy contracts are designed to offer self-
monitoring incentives to the owners of these inputs by tying their incomes to farm output. In this setting,
owner-cultivation is chosen if both inputs are provided by the landlord, and a fixed-rent contract is chosen
if they are provided entirely by the tenant. Sharecropping provides a solution to the double-incentive
problem that arises when both parties provide unenforceable inputs. This is of course a second best
solution because of the usual Marshallian inefficiency costs associated with sharing rents. 8
The Eswaran-Kotwal formulation takes the absolute advantage of landlords in management and
tenants in supervision as given. The assumption that tenants are somehow endowed with supervisory
skills is questionable [Hayami and Otsuka 1993]. To begin with, tenants in share arrangements appear to
provide a much broader range of tasks than just supervision. Even if supervision of casual workers is
important, it is likely to require very little skill and a large amount of time. Hayami and Otsuka [1993]
8 Shaban [1987] supports this argument with empirical evidence that yields (controlling for other observed factors)
are lower in sharecropped lands. We do not contest this view that sharecropping is less efficient than fixed-rent
farming. Our aim is to determine the conditions under which share-tenancy arises in spite of the apparent
inefficiencies.