Space matters: Reconstructing a Local-scale Okun's Law for Italy

Abstract

The present study develops a local-scale analysis of Okun's law for short-term changes in district product and unemployment rate in 686 labour market areas in Italy (2004-2005) based on a Geographical Weighted Regression (GWR) model. The relationship between changes in district product and unemployment rate was poorly significant using both parametric and non-parametric (a-spatial) techniques being more intense when space is incorporated in the model. The GWR results pointed out the spatial patterns characterizing the Okun's law at the local scale. The elasticity of district income to unemployment rate showed important spatial variations being higher in dynamic rural districts around metropolitan areas. The highest model performance was found in specific areas of both northern and southern Italy. However, the classical Okun negative relationship between district product and unemployment rate was mainly observed in northern Italy while the reverse pattern was identified primarily in southern Italian districts. While providing a novel approach shedding light on the spatially-varied income-employment dynamics, GWR models allow evaluating the complex geography of economic growth in Italy. The use of local-scale data and spatial approaches may inform policies regulating job markets on the base of their territorial specificity.The present study develops a local-scale analysis of Okun's law for short-term changes in district product and unemployment rate in 686 labour market areas in Italy (2004-2005) based on a Geographical Weighted Regression (GWR) model. The relationship between changes in district product and unemployment rate was poorly significant using both parametric and non-parametric (a-spatial) techniques being more intense when space is incorporated in the model. The GWR results pointed out the spatial patterns characterizing the Okun's law at the local scale. The elasticity of district income to unemployment rate showed important spatial variations being higher in dynamic rural districts around metropolitan areas. The highest model performance was found in specific areas of both northern and southern Italy. However, the classical Okun negative relationship between district product and unemployment rate was mainly observed in northern Italy while the reverse pattern was identified primarily in southern Italian districts. While providing a novel approach shedding light on the spatially-varied income-employment dynamics, GWR models allow evaluating the complex geography of economic growth in Italy. The use of local-scale data and spatial approaches may inform policies regulating job markets on the base of their territorial specificity.