Proposed Objectives of the Project:
Given recent trends in shrinking agricultural labor pools and rising wage bills, we aim to develop a theoretical framework for the adoption of mechanized harvesting. We then wish to use this model to further explore the general equilibrium effects of mechanized harvesting on labor markets as well as environmental impacts of the same. This model can be used to predict the regional disparities in the spread of mechanized harvesting and can be used as a useful tool for weighing the benefits and costs of mechanized harvesting to local economies.

Summary of Results:
We implemented a randomized controlled trial providing 40-50 kg of new machine harvestable chickpea (MHC) seed variety to 310 farmers in 60 villages free of cost.

We find that the proportion of treatment farmers harvesting MHC using combined harvesters is 10 percentage points higher than the proportion of control farmers using combined harvesters on their traditional chickpea crop, and this difference is statistically significant. Since the quantity of MHC provided was just sufficient for cultivating one acre, most treatment farmers also grew the traditional variety of chickpea on their remaining land. Interestingly, the proportion of treatment farmers using combined harvesters for harvesting their traditional chickpea crop was similar to the proportion of treatment farmers using combined harvesters for harvesting MHC (no statistically significant difference). This points to the fact that the introduction of the MHC variety actually led to an increase in mechanization, not only for the MHC itself but for other varieties grown by the treatment farmers.

Now, we wanted to find what other factors play a significant role in mechanization. Interestingly, we find that the only factor that has any significant effect on the likelihood of using combined harvester for harvesting chickpea crop is the cultivated area. This seems to be a pointer to the fact that mechanization is limited to large landowners. Given that the rental market for combined harvesters in this context is fairly well developed, it is interesting that mechanization has actually not spread beyond farmers with large landholdings. This makes us believe that there are barriers inherent in the networks amongst farmers which prevents the diffusion of mechanization more widely.

The role of social networks in shaping existing levels of mechanization is a first-order issue for both academic and policy reasons. Important among these include the specific role of initial adopters or agronomic field trial study participants (“seed” farmers) who could shape local demand and price-expectations for new technologies. Therefore, in a follow-up study, we propose to examine the wedge introduced by narrow network ties of initial adopters in the demand for new technology. In addition, we examine how return heterogeneity among initial adopters affects subsequent adoption via networks. Specifically, we propose a referral system that incentivizes seed farmers via bonuses for adoption via referrals. To estimate the network wedge, we compare the adoption outcomes when we limit the referral system to specific target groups set ex-ante with a referral system without any such limitation (both within and across villages). We anticipate that appropriate incentives along with target setting should expand the reach of the technology much beyond the narrow social network of the seed farmers and subsequently shape market demand.

We hope that this follow-up study will help us construct a fuller picture of the drivers of mechanization which in turn will enable us to develop a coherent theoretical framework for it. We hope to extend the model to predict the spread of mechanized harvesting across space by incorporating a model of agricultural technology diffusion -- as this appears to be one of key mechanisms needed for a study of the general equilibrium effects of mechanization.