Testing the Role of Input-Output Linkages in Industry Agglomeration with the Spatial Econometric and Input-Output Model

This paper examines the relationship between input-output linkages and industry agglomeration by combining the strength of spatial econometrics and the input-output model. This paper uses the location quotient as a measure of industrial agglomeration and regresses it on the intermediate demand variable, which is an important variable in the integrated econometric and input-output models. The nature of the dataset and the construction of key variables are consistent with the use of a spatial Durbin model, estimated by “borrowing” spatial panel model techniques. With the model labeled as the spatial pseudo-panel industry fixed-effects Durbin model, this paper estimates it with the maximum likelihood method using manufacturing industries with 3-digit NAICS codes at the county level in the U.S. The results confirm the importance of input-output linkages in promoting the formation of industry agglomeration in a region.