Subnational MRIO building: the added value of household consumption and aggregate value-added data

Abstract

Data scarcity makes subnational input-output accounts inaccurate. Builders of such accounts must resort to allocating output, value added, imports, and exports to regions using readily available industry-wise data, like shares of national jobs by industry. Meanwhile, population shares are typically used to allocate other final demand components. Lately, however, some statistical agencies have been releasing more subnational data. Surprisingly, builders of subnational input-output accounts do not appear to use them. This is probably due to uncertain trade-offs between the costs and benefits of deploying such data. We, therefore, explore the degree to which using some value-added and household consumption data can improve subnational multiregional input-output accounts. We find that integrating either household-consumption or some value-added data improves account accuracy little. Using both datasets in combination, however, does improve estimated accounts somewhat. Plus, together they portray rather accurate estimates of interregional income multipliers and consumption-based greenhouse gas emissions