On regional growth convergence in Great Britain

This paper examines recent sub-regional output data for Great Britain to identify possible economic convergence. It concludes that sub-regional data are subject to substantial spatial autocorrelation. Conventional estimates of ‘beta’ convergence are subject to misspecification bias if spatial autocorrelation is not taken into account. Unconditional models fail to find any evidence for economic convergence – indeed, the most recent data point to significant economic divergence. Conditional models, controlling for region-specific steady-states and the influence of human capital accumulation, provide estimates closer to the ‘stylized fact’ of 2% per annum convergence. A further conclusion is that the use of regional price deflators may affect rates of convergence estimates. Spatial autocorrelation suggests that growth ‘hot spots’ can influence surrounding areas positively, but that poor economic performance in lagging areas may also have wider regional impacts