A Sensible Mutual Fund Selection Model

We present a rigorous framework for asset allocation and selecting mutual funds that takes into account the unique preferences and constraints of an individual investor. The framework is based on the analytic hierarchy process (AHP), and the model generates reasonable asset-allocation recommendations and identifies the most suitable funds within an asset class. We include sample mutual fund selection for a hypothetical investor. A mutual fund selection model that uses the AHP framework is flexible, is user friendly, and ensures consistency throughout the portfolio decision process. Modern portfolio theory states that the investment decision process can be separated into two independent steps. In one, investment professionals, such as mutual fund managers, specialize in constructing risky portfolios. In a second process, individual investors choose complete portfolios by combining the optimal risky portfolio, based on their risk tolerances, and the risk-free asset. Finance research has traditionally focused on the first part of the investment process, especially on identifying some risky portfolio that would be universally optimal for all investors.In the real world, however, the investment process is more complicated; in addition to risk and return, factors such as taxes and human capital must be considered. Because different investors face different constraints in their investment decisions, a risky portfolio that is universally optimal may be impossible to find. Investors are left to their own judgments when determining a proper risk level and constructing their own complete portfolios. But many individual investors may not be competent at these tasks. We propose a model that assists investors in determining their asset allocations and selecting mutual funds for their complete portfolios.Our model divides the mutual fund selection problem into two hierarchies: The first hierarchy addresses alternatives at the asset-class level (asset allocation), and the second hierarchy addresses alternatives at the fund level (fund selection). Each hierarchy consists of three levels—the mission of the hierarchy, the selection criteria, and the available alternatives. In the first hierarchy, the mission is to obtain a suitable asset-allocation recommendation for an investor. The selection criteria are the investor's investment objectives and constraints, and the alternatives are the asset classes. In the second hierarchy, the mission is to obtain the most suitable mutual fund within each asset class. The selection criteria are based on structural and operational characteristics of the funds, and the alternatives are the funds available in each asset class.We provide an example to illustrate application of the model to the mutual fund selection problem for a hypothetical investor. We start with the investor's investment objectives and constraints as identified by choices the investor makes through pairwise comparisons. We empirically estimate the suitability of the asset classes according to each investment objective by using a sample from the Morningstar database covering the 1976–98 period. Then, we combine the relative importance of the investment objectives and the strength of the asset classes according to each objective to obtain the asset-allocation weights suitable for the investor. Once the weights of the asset classes are determined, the next step involves evaluating individual mutual funds in each asset class. We use a specific asset class to illustrate the ranking of individual mutual funds based on the selection criteria.The method we describe is an objective procedure for choosing mutual funds and thus a valuable addition to a financial advisor's toolkit. This method has the following distinct advantages.First, it incorporates the unique preferences and constraints of an individual investor into the mutual fund selection problem by allowing the importance of investment objectives and fund attributes to vary by investor.Second, it prevents the investor from making inconsistent preference assignments. Because the fund selection decision involves more than one attribute (e.g., return objective, tax efficiency, risk), enforcing consistency is essential.Finally, the framework minimizes the amount of technical input required from investors because it allows the integration of the financial advisor's recommendations, the findings of academic research, and historical mutual fund data. The result is a user-friendly yet rigorous model.