Treatment Effect Heterogeneity of Farmers Participating in Cooperatives
利用亚美尼亚大规模农户调查数据，采用因果森林方法分析合作社参与对农户绩效的异质性影响，发现对市场信息有限、金融素养低、经验少和年长农户的益处更大。
ABSTRACT Despite the many studies providing evidence of the benefits of cooperatives, most small farms and farm households, particularly in transition economies, do not participate in them. This study analyses the effect of cooperative activity on farm performance using a large‐scale farm survey conducted in Armenia. In addition, it addresses the issue of heterogeneity, which is often overlooked in the literature. It looks at heterogeneity in terms of observed effect modifiers, which are measured covariates that can modify the causal effect of cooperation on a specific farm performance indicator. This is the first study that pays special attention to heterogeneity and uses the causal forest approach in the context of cooperative membership to characterise the drivers of treatment heterogeneity (i.e., variation in the causal impact of cooperative membership across farmers with different observable characteristics). Participation in cooperatives is associated with greater perceived farm performance among farmers with limited access to market information and low financial literacy. Similarly, higher benefits are observed for less experienced and older farmers, suggesting that cooperatives play a particularly important role in supporting more vulnerable subgroups.