afsane sherkat; Mohammad Jelodari Mamaghani; Ali Asghar Banouei; Ashkan Mokhtary Asl Shouti; Sonia Sabzalizad Honarvar
Volume 15, Issue 56 , April 2015, , Pages 135-160
Abstract
In this paper, we have used four conventional, Adjusted, Generalized, and Adjusted Generalized RAS methods to update input-output Coefficients (IOC). Conventional and Adjusted RAS methods can only update positive and zero cells and are not sensitive to the existing negative cells like net exports and/or ...
Read More
In this paper, we have used four conventional, Adjusted, Generalized, and Adjusted Generalized RAS methods to update input-output Coefficients (IOC). Conventional and Adjusted RAS methods can only update positive and zero cells and are not sensitive to the existing negative cells like net exports and/or net taxes in Input-Output Tables (IOTs). To solve this drawback, analysts have proposed Generalized RAS (GRAS). This method can update positive, zero and negative cells. From the application view point, this method has two limitations: First, it is more focused on numerical examples rather than real IOTs. Second, extending the GRAS to AGRAS has not been attempted yet. The above limitations will lead us to pose the following questions: is it possible to extend Generalized RAS to Adjusted Generalized RAS? And which method has more statistical errors? For this purpose, we have used the two aggregated survey based IOTs of Iran for the years 1996 and 2001. Our findings show that it is possible to extend GRAS to AGRAS. Whit respect to the measurement of statistical errors, the following results have been obtained: first, the statistical errors of GRAS is lower than Conventional and ARAS, and the second statistical errors of AGRAS are much lower than corresponding figures in Conventional, ARAS and GRAS.
