How Many Active Funds Should You Hold?

Traditional quantitative approaches to portfolio construction have drawbacks for investors or advisors who combine multiple active managers in typically producing large numbers of disperse positions. We develop a new methodology for sequentially allocating to active funds that results in parsimonious numbers of funds and test our results on US active equity mutual funds. To initiate the algorithm, we choose the fund with the highest information ratio (IR). Then, we select the next fund that produces the highest increase in the portfolio’s IR, noting that the IRs of the remaining funds depend on the current active portfolio. This procedure is repeated as long as the active risk of the portfolio is above a minimum threshold. Under certain conditions, the problem nests the well-known Knapsack Problem. The algorithm generates approximately the same IRs with significantly smaller numbers of funds than traditional mean-variance optimizations