in fisheries and livestock activity have similar and significant effects on both types of leasing. The
proportion of students has a negative effect because it leads to less disposable family labor if the schooling
choice is assumed exogenous to the leasing decisions. Interestingly, the engagement of the household in
fisheries and livestock activities has a positive impact on land leasing. This may be because such
households are more enterprising, or because the presence of interests outside of cultivation in the village
makes it profitable for these households to remain as farmers and not seek outside employment. The
positive education effect may reflect lower transactions costs (lower information and negotiation costs)
they face in the leasing market.
Multinomial Logit Sample Selection Model
The multinomial logit selection model is reported in Table 4. The first stage is a contract choice
equation with the three choices, 0) self-cultivation, 1) share-rent and 2) fixed-rent. The second stage has
leasing equations for share and fixed-rent tenants corrected for sample selection. The last sub-column for
each type of tenancy report the marginal contribution of each variable to the odds ratio between that type
and owner-farming. These odds ratios help us interpret the logit coefficients in a meaningful way. For
example, an increase in paddy land ownership makes the farmer 0.109 times as likely to be a share-tenant
and not a self-cultivator. Similarly, a unit increase in skills increases the likelihood of share-tenancy by
4.2%, but this effect is not significant. But an increase in skills has a positive and significant effect of 25%
on the probability of fixed-rent tenancy.
Since it is more useful to directly compare the relative probabilities of share and fixed-rent
tenancies, the last column computes the marginal effects on the probability of fixed-rents relative to share
contracts. These estimates show that an increase in skills makes it 19.9% more likely that a tenant
chooses a fixed-rent contract. Other variables that make the likelihood of fixed-rent tenancy greater are
paddy land ownership (408%), proportion of male adults (84%), and ownership of tractors (52%).
37