Uncertainty in Integrated Regional Models

This paper examines the nature of uncertainty in integrated econometric+input-output (ECIO) regional models. We focus on three sources of uncertainty: (a) econometric model parameter uncertainty; (b) econometric disturbance term uncertainty; and (c) input-output coefficient uncertainty. Through a series of Monte Carlo simulations we analyse the relative importance of each component as well as the question of how their interaction may propagate through the integrated model to affect the distributions of the endogenous variables. Our results suggest that there is no simple answer to the question of which source of uncertainty is most important in an integrated model. Instead, that answer is conditioned upon the focus of the analysis and whether the industry specific or macro level variables are of central concerns.