Illiquid Asset Allocation and Policy Weights

Abstract

The author develops a simple approach for computing the probability that an initial asset allocation will breach a pre-specified policy weight over a given time horizon. The model is consistent with assumptions made in most Asset Liability Management (ALM) studies and the closed-form analytic expression “buys” the user a variety of robust insights. After calibrating this model to broadly defined alternative investment asset class data, the author concludes that a conservative 5% commitment to an illiquid asset class has a 1/3 chance of doubling (i.e., to 10% of the fund) within 5 years, and tripling (i.e., to 15% of the fund) within 15 years. Paradoxically, the lower the effective correlation between the performance of a given asset class and the remainder of the portfolio—which is normally something to be coveted in strategic asset allocation—the greater the chances of breaching a given policy weight. The results suggest that initial target allocations to illiquid asset classes should be reduced relative to their liquid counterparts, when a conventional mean-variance analysis was used to obtain these policy weights.