This study examines the effect of cooperative participation on household income among livestock-raising households in Inner Mongolia, China, with particular attention to selection bias and income heterogeneity between cooperative members and non-members.
The analysis employs a two-stage Endogenous Switching Regression Model (ESRM) to correct for non-random selection into cooperatives. Household-level survey data are used to estimate income outcomes under actual and counterfactual participation scenarios. Results from the ESRM are compared with pooled ordinary least squares (OLS) estimates to highlight the consequences of ignoring selection bias.
The results show that cooperative participation generates a modest but positive income gain for member households. In contrast, non-member households would experience a significant income loss if they were to join cooperatives, indicating potential adverse selection. Education and household size positively influence income, while cooperative membership remains a key determinant. Training participation is unexpectedly associated with lower income. Pooled OLS estimates produce weaker explanatory power and inconsistent coefficients, confirming the importance of correcting for selection bias.
This study is based on cross-sectional survey data from livestock households in Inner Mongolia, which limits causal inference and generalizability. Future research should use longitudinal data and larger samples to better evaluate the long-term effects of cooperative participation on household income.
This study provides micro-level evidence on the income effects of cooperative participation among livestock households in Inner Mongolia using an endogenous switching regression model. By accounting for self-selection bias and estimating heterogeneous treatment effects, it offers new insights into the role of cooperatives in improving rural household income.
