Social networks, adoption of improved variety and household welfare: evidence from Ghana

In this study, we examine the effects of own and peer adoption of improved soybean variety on household yields and food and nutrient consumption, using observational data from Ghana. We employ the marginal treatment effect approach to account for treatment effect heterogeneity across households and a number of identification strategies to capture social network effects. Our empirical results show that households with higher unobserved gains are more likely to adopt because of their worse outcomes when not adopting. We also find strong peer adoption effect on own yield, only when the household is also adopting, and on food and nutrient consumption when not adopting. However, the peer adoption effect on consumption attenuates when the household adopts the improved variety. Furthermore, our findings reveal that adoption tends to equalise households in terms of observed and unobserved gains on consumption and can thus serve as a mechanism for promoting food security and nutrition in this area.