Ranking Investment Funds Based on Limited Information

Abstract

A new approach for the ranking of investment funds by the fund of funds manager is presented. Often the manager has to make a decision regarding ranking of funds and allocate capital between them for the next few months (for example, for the next quarter), but the available information/data is limited to daily returns for a relatively short time period (for example, the ending quarter). No other information regarding investment funds may be available, or it may be too costly to obtain and process. Or the manager may choose to use limited information/data for a very short time period because after unexpected events or disasters the best strategy for the investment funds is completely different from one that was best a few months or years before. The method of ranking funds presented here is simple and follows a natural meaning of daily and quarterly returns. As a computational example typical data for ten funds for a period of 106 days are used and processed. One can see that the obtained ranking performs better than statistical one based on values of sample means and sample variances of returns of these funds. This new simple method finds its niche in analysis of investment funds and fills the gap between more sophisticated approaches (that need more information/data available and more powerful cluster of the analytic computing procedures) and non-numerical Laplacian approach (where capital is allocated more or less equally between all funds that did not fail in a previous quarter).