An holistic approach to investment assessment

A paper I co-authored is finally published (in early view, .pdf) in the journal of Managerial and Decision Economics. Along with Konstantinos Kirytopoulos, Athanassios Rentizelas and my professor Ilias Tatsiopoulos, we proposed an holistic approach to investment assessment, a process of value that may be highlighted under the current economic situation.

To make a long story short, the typical approach to investment assessment includes the following steps, as these are described in the figure.

However, one may argue that the process remains fundamentally suboptimal, with the discrete nature of its steps serving as one of its most important flaws (given that the optimization of each step does not necessarily result into the optimization of the process as a whole). We tried to address this concern by introducing an integrated holistic approach, as in the figure that follows.

The input variables to the selected criterion’s function (typically Net Present Value) is broken into two categories, the ones directly defined by the investor being the first (like investment’s height or operational specifications), and those that cannot be modified  being the second (interest rates etc).

The approach suggests initially assigning to the latter ones their most probable values, so as to be able to optimize (using genetic algorithms, because of the tough nature of the problem) the values of investor defined variables, then moving on to an extended risk analysis by Monte Carlo simulation to accurately (probabilistically that is) compute the implied risk of the investment. We also introduced NPV Expected Shortfall and NPV Risk Preference Index, providing a customizable metric of high informational value picturing the whole probability density function, next to demonstrating the whole process by an extended case study regarding the comparison of two renewable investment scenarios.

As a closing note, I should mention that, to me, the concept of studying the problem at hand by separating its variables to the ones you can influence and the ones you cannot, while then attempting to optimize the first set and track the total risk of the second, remains an approach of great practical -if not philosophical- value, next to its academic one. And I do hope that is of interest to you, too.