Updating Different Types of IO Tables

What is it about?

The iterative scaling and rescaling of the rows and columns of an old IOT until they equal the new row and column totals, known as RAS, outperforms alternative techniques for updating old IOTs. Still, the errors of this information gain minimizing technique remain large, which is why adding survey data for the target year is required. In case of updating bi-regional IOTs, adding the known values of new national IO cells, while using the multi-proportional scaling of MR-RAS, considerably improves the accuracy of the updates. When negative cells are present, RAS and MR-RAS needs to be replaced by Generalized RAS and MR-GRAS. Finally, it is shown how time series of old IOTs may be used in CRAS to improve the estimates of the cells in RAS or GRAS.

Featured Image