Incorporating Time-Series Into an Interindustry Analysis to Model the Regional Economic Structure: A Case Study of the Illawarra

Abstract

Since the pioneering work of Glickman on embedding an input-output (IO) module into an econometric framework, 1 there has been a plethora of studies on combining an IO analysis with an econometric model in the literature. The popularity of the combined framework is due to its superior performance in economic forecasting and higher accuracy in impact analysis. There are a number of approaches through which an IO model is combined with an econometric model. This paper examines three approaches to combine IO analysis with econometric modelling, namely embedded, coupled, and linked. All three approaches are applied to the Illawarra economy in a series of ex-post forecasting experiments. Each approach is applied to a hypothetical scenario of sectoral reallocation of government expenditure, to investigate certain key sectors that provide more jobs per dollar of expenditure relative to the other sectors. The comparative forecasting performance and impact analysis accuracy of each approach is examined.