Skills, Partnerships and Tenancy in Sri Lankan Rice Farms

where Zi is a vector of all other characteristics of the farmer. Then, the notional demand for land (Y^*i) in
the land leasing market is;

Yi * = C₁ - H_i = f (F_l, X₁, Z₁ ) - H₁

[38]

The difference between the actual and notional demands for land depends on the transaction costs in the

land leasing market. Suppose the actual net amount of land leased-in is Yi. Then

Y = h(Y_i*) = h(f (F_i, X_i, Zi)- Hi)

[39]

where h(.) is the transformation function in the leasing market.

The Choice of Contract

The main hypothesis of this paper is that relatively skilled tenants obtain fixed-rent contracts while

those endowed with family labor get share contracts. The theoretical analysis showed that the level of

managerial skill required to enter in to a sharecropping contract is less than that for a fixed-rent contract

because the landlord typically provides some of the skill-intensive managerial inputs in sharecropped

farms. The empirical model presented in this section identifies some key differences between the two

contracts as resource adjustment mechanisms.

We focus on the leasing behavior of tenants because data constraints prevent us from estimating a

more general model that includes the landlords. The simplest method of estimating the determinants of

sharecropping and fixed rent leasing is to carry out separate leasing or contract choice estimates for each

type of contract. The starting point of our analysis is probit and tobit models for the following underlying

equations based on the model outlined in [37]-[39]:

Si = βo + β1 Fi+ β2 Hi ₊ β3 X_ι + β4 Zi+ V

Fi = ^δo ^{+ δ} Fi ^{+ δ}> H_i ^{+ δ}3 ^χ_i ⁺ δ4 Zi ⁺ ψi

[40]

where ε_i and φ_i are distributed N (0, σ_ε²) and N (0, σ_φ²) and Si* and Fi* are the underlying

variable for sharecropping and fixed-rent leasing, Fi, is the endowment of family labor, Hi is the amount of

20