Estimating regional input-output multipliers in Australia: a methodological study.

Abstract

The objectives of this study are to review the existing methodologies for estimating regional input-output multipliers, test the regional input-output model, and recommend changes to multiplier estimation techniques if appropriate. The study reveals that the input-output model, after a number of adjustments are made to its structure, can produce accurate estimates of regional impacts. This conclusion provides support for a model that has achieved widespread use, but little rigorous empirical testing. Secondly, the study concludes that there is no one preferred method for generating regional input-output tables in order to generate multipliers. A GRIT-type model is considered the preferred alternative, but the final specifications of the model depend on the amount of data that needs to be acquired by survey. It is suggested that the amount of survey work necessary depends on a number of factors, including: . the amount of available published data; . the type of region being examined; . the number and type of industries being analysed. It is considered that use of survey-based sum of intermediate coefficients is a cost-efficient data gathering technique. The third major finding of the thesis is that the accuracy of impact forecasts are improved if Service sector impacts are estimated using a simple Service employment-population model, rather than the traditional input-output model. In order to achieve consistency in impact forecasts, it is necessary to nest this Service employment-population model within a demo-economic model.