PT - JOURNAL ARTICLE AU - Steve Q. Xia AU - Joseph Simonian TI - Measuring Investment Skill in Multi-Asset Strategies: An Empirical Study of the Information Coefficient as Weighted Rank Correlation AID - 10.3905/jpm.2021.1.208 DP - 2021 Feb 28 TA - The Journal of Portfolio Management PG - 135--144 VI - 47 IP - 4 4099 - https://pm-research.com/content/47/4/135.short 4100 - https://pm-research.com/content/47/4/135.full AB - Over the past decade, the investment industry has come to appreciate the importance of asset allocation and its role in achieving clients’ financial objectives. The result has been a profound change in the landscape for the industry in terms of both product design and distribution. Increasingly, the asset flow to investment managers has been heavily driven by multi-asset funds and balanced funds rather than individual stock or bond funds. With the growing prominence of multi-asset investing, the role of active asset allocation has also become more important. For investors who are evaluating multi-asset funds and for asset allocators trying to set expected asset allocation alpha, the measurement of asset allocation skill is a crucial task. To the latter end, this article provides an intuitive methodology to measure asset allocation skill within the formal context of the fundamental law of active management, a well-known characterization of portfolio managers’ alpha generation process. Specifically, the authors show how weighted rank correlation provides an intuitive and transparent version of the information coefficient. Their study is framed within a novel simulation-based framework that they use to analyze the impact of asset allocation skill and its implications for estimating reasonable expectations of asset allocation alpha.TOPICS: Portfolio construction, exchange-traded funds and applications, performance measurement, statistical methodsKey Findings▪ Weighted rank correlation provides an intuitive and nuanced way to express the information coefficient—the relationship between ex ante and realized returns.▪ The bespoke simulation framework described in the article can provide insight into the relationship between different skill levels and expected asset allocation alpha.▪ We find that there is a near-linear relation between asset allocation skill and alpha and asset allocation skill and tracking error, but a less-than-linear relationship between asset allocation skill and information ratio and asset allocation skill and bet size. The relationship between asset allocation skill and trading frequency is found to be more or less uniform.