Comparison of Two Methods of Identifying Input-Output Coefficients for Exogenous Estimation

Input-output (IO) updating research indicates substantial improvements in the forecasts when some of the coefficients have been exogenously estimated and included in the updating process. Several methods for identifying the appropriate subsets have been proposed. The present paper attempts to assess the relative performances of two such approaches: 'the largest coefficients' and 'the most important parameters' criteria. Utilizing these criteria, a set of coefficients from the 1966 IO table of the former Soviet Union were selected and exogenously determined. The remaining coefficients were updated to 1972 by means of naive, RAS, and Lagrangian techniques. Comparison of the results with the 1972 benchmark table provided the desired answers.