Estimating Interregional Economic Impacts: An Evaluation of Nonsurvey, Semisurvey, and Full-Survey Methods

Abstract

Literature shows that nonsurvey input–output tables tend to produce regional multipliers with systematic upward biases. This paper explores the related, relatively uncharted territory of nonsurvey versus survey impact studies by means of a series of simulations. The base case is provided by a very detailed five region survey of both the forward and the backward impacts of the energy-distribution sector in the four northern provinces of the Netherlands. To deal adequately with the two-sided dependence between a firm or sector and a region, as opposed to using the traditional (gross) multipliers, we argue in favor of using a new multiplier concept: the net multiplier. Next, from examining alternative impact study methods ranging from quick and dirty, via semiextensive, to full-survey methods, we conclude that using even aggregate, first-order impact information is more important than using a very detailed survey-based input–output model.