Current Uses of the RAS Technique: A Critical Review

Improved Estimation of Regional Input–Output Tables Using Cross-regional Methods

“…Lynch, 1986;Polenske, 1997). For the dataset we study here, most WAPEs would decline sharply to 0.1-0.2 when the 5% most important cells would be replaced by the true, survey-based values (see Jiang et al, 2009 Third, turning our attention to the cross-regional methods, we can conclude that the four methods perform very close to each other on average, but that there are some marked differences at the level of individual regions.…”

“…RAS can be seen as a method that tries to reconcile the old intermedite input structure as well as possible with the new column and row totals. Despite regular complaints about the poor performance of RAS, reviews of empirical results such as Polenske (1997) and Jackson and Murray (2004) tend to conclude that RAS results are seldom outperformed by alternatives using the same type of information.…”

“…The 'RAS' technique developed by Stone and Brown (1962) has been acknowledged as one of the most widely-used ways to update tables, based on the input-output structure of an older survey-based Morrison and Smith, 1974;Sawyer and Miller, 1983;Polenske, 1997;Jalili, 2000;Jackson and Murray, 2004). RAS can be seen as a method that tries to reconcile the old intermedite input structure as well as possible with the new column and row totals.…”

“…the so-called RAS technique, is still a very popular method. In terms of estimation performance, it is hard to beat if no supplementary information is available (Oosterhaven et al, 1986;Polenske, 1997;Jackson and Murray, 2004). Alternatively, regionalization using location quotients is an often used method if a survey-based national table for the object year is available (see, e.g.…”

The Solution of Updating or Regionalizing a Matrix with both Positive and Negative Entries

“…The RAS-algorithm works, i.e. minimizes information loss, as long as the start matrix A contains only non-negative entries (see Bacharach, 1970;Lecomber, 1975;and Polenske, 1997, for overviews of RAS). Budavári (1981) discusses a generalization of RAS based on a general set of convex goal functions and on a general set of linear constraints.…”