Agricultural Development and Land Tenancy in a Peasant Economy: Reply

Abstract

It is by now well known, particularly due to Bell and Braverman, that a model with constant returns to scale, sharecropping, and the wage-contract choice leads to a tendency for corner solution. In Bardhan (1979) I introduced a two-season production process interlinking land and credit markets and landlord monopoly power, not to avoid-as Datta and Nugent seem to imply-the problem of corner solutions, but only to introduce a dosage of realism in the model. The main problem for interior solutions in the context of the model was caused by the special assumption of constant returns to scale, which I used here, unlike in Bardhan and Srinivasan and Bardhan (1977), primarily to avoid cluttering up the equations. In any case, in peasant agricultural production functions with only land and labor as arguments and other factors suppressed (particularly managerial and animal labor input), the assumption of constant returns to scale is not very realistic. If my 1979 model is taken with diminishing returns to scale, the interior solution and most of the comparative-static results are more robust. It is not inconsistent for the landlord to be a monopolist in the land-lease market and at the same time be a competitive buyer in the labor market. In fact, I think this may be the more realistic case. As indicated in footnote 3 of Bardhan (1979), the demand for labor by others has been suppressed for convenience in equation (9); but, if one likes, one may introduce on the left-hand side of that equation a term D(W2) to represent the net demand for labor by others (say, by pure owner-cultivators) with usual properties of demand functions. None of the subsequent results in Bardhan (1979) changes. As far as I can see the empirical result reported in table 1 by Datta and Nugent on the basis of a normalized dependent variable is illegitimate. They combine the proportion of cultivated area under total tenancy given in Bardhan (1979) and the area under sharecropping as percentage of area under sharecropping and fixed-rental tenancy in Bardhan (1977) to get their normalized dependent variable, which is the percentage of land under share and wage contracts only. This they cannot do unless they use independent information on land under leasing arrangements other than share and fixed-rental tenancy, which in some states are significant.