Uncertainty Treatment in Input-Output Analysis

Abstract

This work provides an extensive overview of the input-output (IO) literature, both theoretical and empirical, dealing with the inherent IO data uncertainty issues. The survey is carried out on the basis of a specific uncertainty technique used, rather than taking a chronological overview approach, which also allows for easier comparisons and linking of the outcomes of the individual contributions. Thus, we discuss the literature within seven methodological blocks (sections), which include deterministic error analysis, econometric and other (non-Bayesian) statistical approaches, random error analysis and probabilistic approach, full probability density distribution approach, Monte Carlo analysis, Bayesian approach, and other techniques. Within each section, the literature on a certain topic is reviewed in its historical context, which helps to clarify the state of the art. Our main findings from this survey, related discussions, final remarks and observations are given in the concluding section.

Keywords: input-output uncertainty, deterministic and random error analysis, stochastic input-output analysis, Monte Carlo simulations, Bayesian approach

JEL Classification: C67, D57, R15