Does democracy influence social spending differently from other regimes? This study tackles the challenge of estimating regime effects—rarely changing variables—in dynamic panel data models. Conventional methods often fail due to potential correlations between these unit-level effects and explanatory factors, creating an endogeneity problem. We propose a Bayesian approach specifically designed to address this issue in time-series cross-sectional analysis.
To demonstrate our method's effectiveness, we apply it to 18 Latin American countries over several decades of data. Our findings show that democracy has a statistically significant positive impact on social spending both short-term and long-term compared to non-democratic regimes.
📊 Data & Methods
* Analyzed time-series cross-sectional (TSCS) data from 18 Latin American nations
* Addressed endogeneity by correlating unit fixed effects with variables in model specification
* Employed Bayesian estimation techniques for accurate parameter recovery and inference
🔍 Key Findings
* Democracy consistently shows positive correlation with social spending levels
* Both immediate short-term and sustained long-term impacts are statistically significant
📝 Why It Matters
This research offers a robust methodological solution to estimating persistent political variable effects, providing clearer insights into how governance systems shape economic policies across different regions.
| A Bayesian Approach to Dynamic Panel Models with Endogenous Rarely Changing Variables was authored by Tsung-han Tsai. It was published by Cambridge in PSR&M in 2016. |
