Yield effects of agricultural cooperative membership in developing countries: A meta-analysis

Abstract

This study uses a meta-analysis to synthesize the effects of agricultural cooperative membership on the yield of crops and livestock. It collects 158 estimated yield effects from 42 studies, covering 19 developing countries. Our analysis finds evidence that there exists positive publication bias in the empirical literature, confirming that researchers and journals have a preference to publish articles that report positive and significant results. After correcting for publication bias, we find that cooperative membership has a small-sized and insignificant effect on the yield. The meta-regression analysis reveals that variation in the reported yield effects can be largely explained by the study attributes such as the sample type (full sample vs. subsample), membership ratio, econometric approaches (instrumental-variable based parametric approach, non-parametric approach or ordinary least square regression), effect size types (average treatment effects on the treated, average treatment effects, or coefficient), agro-product type (grain or others), and climate zones (tropical or non-tropical).