Regional income convergence in India: A Bayesian Spatial Durbin Model approach

The discussion on regional disparity is essential for addressing politically sensitive policy issues in any federal polity. The research outcome of regional disparity analysis is, however, often ambiguous and is not robust to choice of strategies, namely β and σ convergence analysis. The regression based theoretically appealing β convergence approach have not given adequate attention to spatial effects. Spatial interactions would make the outcomes of this approach less reliable. This study, on reviewing various growth models found that Spatial Durbin Model of Fingleton and Lopez-Bazo(2006) is theoretically useful and empirically appropriate in β convergence analysis. This study estimated parameters of Bayesian Spatial Durbin Model using statewise real per capita GSDP data computed from Central Statistical Organisation (CSO) during the period 1980 – 2010. The study concludes that the later reform period has witnessed beta convergence due to feedback effect. The inclusion of spatial effects, the study contends, helped to explain the contemporary debate in β convergence analysis in India