Regional input-output analysis, data and uncertainty

This paper addresses two major issues. The first is the introduction of uncertainty into the framework of classical deterministic regional input-output (I-O) analysis. Secondly, the role of the use of measurable flow data on the fundamental structure of the models is explored. The opening section uncovers an anomaly in the classical approach when flows between regions of final demand products, external exports and imports are neglected in comparison to the flows of internal intermediate inputs. Next, we introduce uncertainty into the analysis by extending an entropy formulation developed by Wilson, which itself evolved from the formalism established by Leontief-Strout (L-S). The main enhancements include (i) the introduction of regional output capacities to capture spillovers from regions operating close to capacity, (ii) the inclusion of flows of external imports and exports and (iii) the use of total measurable flows as input which are readily available from surveys, yielding as output not only the total flows of each sector between each pair of regions, but (optionally) the further disaggregation of these flows to include their final destination sector or their use as final demand. In fact, a key objective is to structurally account for the joint influence of technology, output capacities and transportation costs on the pattern of intermediate and final demand flows between regions. Further extensions of the approach lead to the generation of probabilistic supply functions as tools within a potential CGE analysis. This option requires the introduction of prices, permitting a profit constraint to replace the simple transport cost constraint of the earlier models. Copyright Springer-Verlag 2004