Managing a 401(k) Account: An Experiment on Asset Allocation

The study reports the results of an asset allocation experiment in which subjects managed an endowment of money over a 20 "year" time period. While grounded in theory, the study takes an applied look at the ability of subjects to efficiently and effectively make asset allocation decisions similar to those found in 401(k) accounts. The main conclusions are as follows. First, efficient portfolios are more easily created when the set of assets to choose from is carefully constructed. Thus, financial engineers should be given the responsibility for choosing the assets available to plan participants and ensuring that combinations of these assets will fall on the efficient frontier. If followed, this advice would likely significantly reduce the amount of individual company stock offered in Defined Contribution (DC) plans in place of well-constructed low cost index funds from multiple asset classes. Second, if the assets selected for inclusion in DC plans allow the investor to easily create portfolios on the efficient frontier, then the challenge for the investor is not how to get onto the frontier but where to locate on it. The simplistic surveys that are commonly used by DC plan providers to determine risk tolerance and to recommend asset allocations are woefully inadequate for this task. More sophisticated and theoretically driven instruments must be created to educate investors on the risks and the benefits available at different points along the efficient frontier.