فصلنامه پژوهشهای اقتصادی ایران (Mar 2014)
Evaluation of Conventional and Adjusted RAS Methods in Updating Input-Output Coefficients of the Iranian Economy: Emphasizing Various Scenarios of Exogenous Information
Abstract
There are several methods of updating input-output coefficient matrix in last six decades, but there are still issues about RAS and adjusted RAS methods which have been focus of input-output analysts in recent years. One challenging issue is the relationship between more exogenous, superior or additional information of target year in adjusted RAS method and its statistical reduction error relative to conventional RAS method in updating the input-output coefficient matrix. Some analysts observe the positive relationship, whereas others by focusing on the nature and criteria of exogenous information opine that using more exogenous information in the adjusted RAS will not necessarily reduce the statistical errors compared to conventional RAS method. The existing evidence in Iran is around the findings of the positive relationship which has in fact lead to the common belief between compilers and also users of table in Iran. In this article we attempt to examine this issue by means of two symmetric input-output tables of the years 1996 & 2001 by posing two main questions. The first question: Is there any relationship between more exogenous information in adjusted RAS method compared to conventional RAS method in reducing statistical errors? Second question: Do the nature and criteria of more exogenous information, irrespective of more or less cells; have an influence on increase or reduction of statistical errors in updating coefficients? Our findings do not support the existing common belief among the compilers and users of Input-Output Table in Iran and reveal the followings: 1- The adjusted RAS method, in some of exogenous information, is not preferable to conventional RAS method. 2: Measurement of credibility of updated coefficients depends on choice of the nature and criteria of exogenous information, and 3: Using more exogenous information of the target year would not necessarily lead to decrease of statistical errors in the updated coefficients.
