Significance of EG Index with Spatial Clustering Testing for Indian Manufacturing

EG index, the most widely adopted agglomeration index has two major limitations. First the index does not yield a statistical significance value and hence prone to Type-I and Type-II errors. The paper uses the method proposed by Cassey and Smith (2014) to overcome this limitation. Second, the index treats space symmetrically and as a result does not capture agglomeration when the clustering range extends beyond the boundary of a single spatial unit. Moran’s I, a spatial correlation measure, that measures the correlation between the spatial units and the units contiguous to them is used along with the EG index to overcome this limitation. The paper also uses LISA, a spatial exploratory technique to break down the Moran’s I vale for each spatial unit, to locate the industry clusters using data for 144 Indian manufacturing industries for the year 2005-06.