Abstract
Data scarcity induces inaccuracy in regional (i.e., subnational) input-output (IO) accounts. Output, value added, imports, and exports are often allocated to regional industries using available, simple industrywise economic data—employment or jobs. When aggregate regional income data are lacking, population shares apportion final demand. Fortunately, governmental statistical agencies in certain areas of the world are releasing more subnational data. Oddly, literature on subnational MRIO accounting does not acknowledge their use. We conjecture that this is due to uncertain trade-offs between the expense of adopting such data into the accounts and the net accuracy gains their adoption yields. Herein, we explore the degree to which using some value-added and household consumption data might yield more accuracy in subnational MRIO accounts. When it comes to impact analysis, via regional input coefficients and multipliers, we find that integrating superior data does not produce major accuracy gains compared to MRIO accounts built using simpler data (regions' shares of national jobs and population). On the other hand, we do show that MRIO accounts built on both some industrywise regional value-added and household consumption data present fairly accurate interregional income multipliers and consumption-based greenhouse gas emissions.