A new approach to determining optimum portfolio mix
Shahriar Khaksari, Ravindra R. Kamath, Robin Grileves · The Journal of Portfolio Management · 1991
v) 2 K $: $ tion of the AHP Model for pension fund optimum asset mix decisions was to demonstrate use of the model. We did not intend to provide exhaustive discussion of the underlying theory or derive specific conclusions with respect to our hypothetical example hierarchy. Figure 1B shows an alternate type of hierarchy, which, at first glance, looks like Figure lA, but actually not every element in Level 2 affects each element in Level 3. The hierarchy in Figure 1B is not a complete results. Gordon McMeekin states that the primary purpose of his comment on our paper is to correct some of the errors he perceives to exist in our computational procedures. He also suggests an extension to the model and advocates the use of a software package he developed for this application. We appreciate Dr. McMeekin’s interest in our work, but we believe that most of his comments stem from a misunderstanding of our methodology. Dr. McMeekin refers to ”Saaty’s Analytical Network Process.” This is the first time we have seen this term, and without a full reference we are unable to verify whether this term was actually used by Saaty or if it is Dr. McMeekin’s capsulation of the literature. AHP in its general form is modeled so that each level in the hierarchy dominates the level below it. The asset allocation problem, for example, was modeled hierarchically from higher to lower levels to develop a framework for measuring the priority of elements in different levels of the hierarchy, with respect to elements in higher levels, and with respect to the overall purpose of the hierarchy. fgures 1A and 1B show two different types of hierarchies. Figure 1A shows a model where each element in the higher level influences all the elements in the level immediately below it. Because of this complete set of impact of higher elements on the lower one.