Abstract
Current effect-decomposition methods have three major weaknesses: (1) the conceptual framework used, although appropriate for recursive models, is too narrow for use in nonrecursive models; (2) the discussion of nonrecursive models has been limited to single-loop simultaneous models; and (3) the computation procedures are complicated and laborious or they require additional computer works. This article proposes the use of a flowgraph method that eliminates these difficulties. It is shown that with a limited number of flowgraph analysis rules, the effects in either recursive or nonrecursive models can be easily decomposed.

This publication has 7 references indexed in Scilit: