Publications

A Decision-based Approach to Risk-adjusted Performance Attribution [PDF]
Winter 2024

Investment decisions change the risk profile of a portfolio. This is usually on purpose: a manager taking an active investment decision foresees that exposing the portfolio to a different level of risk will eventually pay off, given his outlook on the market. It is therefore important to evaluate the performance of a portfolio not only by measuring the return contribution of each decision, but also by making explicit how the risk decision impacted this contribution. In this article, we propose an attribution method which combines the concepts of Brinson-Fachler attribution and risk-adjusted returns. For each investment decision and the corresponding attribution effect, we identify a component due to risk-adjusted outperformance and a second component due to the incremental risk resulting from the active investment decision. We demonstrate the importance of applying this decomposition method to the levels on which the investment decisions take place. Finally, we provide guidance on how to apply this generic method to various investment strategies and the corresponding attribution methods.

This article won the Peter Dietz Award 2024 for excellence in performance measurement writing presented in The Journal of Performance Measurement.

Attribution-driven Investment Decision Processes [PDF]
Spring 2020

In the investment management industry, performance attribution is commonly used to explain the performance of an investment portfolio by quantifying its components. In practice, the process of analyzing the performance of a portfolio by applying an attribution method takes place separately from the portfolio’s investment process. Even when the attribution effects are linked to specific investment management decisions, usually the execution of such a decision by the responsible manager is not directly targeted at impacting the resulting value of the attribution effect. In this article we introduce the concept of an attribution-driven investment process. In this setup, each involved manager agrees in advance with the definition of the attribution effect specific to his decision, and strives to contribute positively to the portfolio’s overall result by directly managing his attribution effect via his decision. However, to enable a manager to do so, the definition of the effect must satisfy the so-called manageability criteria. We show that some of the traditional multi-period attribution methods lead to unmanageable attribution effects. As a solution to this problem, we propose a generic method to derive multi-period attribution effects from single-period effects such that the manageability criteria are preserved.

Geometric Attribution and the Interaction Effect [PDF]
Summer 2018

It is often assumed by practitioners in the field of investment performance attribution that the geometric attribution method developed by Burnie, Knowles and Teder (BKT) is interaction-free. That is, the method explains excess return by allocation and selection effects, without the need for an interaction effect. In this paper, we show that, despite this belief, an interaction effect does arise when the BKT method is directly derived from the arithmetic Brinson and Fachler (BF) method. We present this derivation, offering a variant of the BKT method which features an interaction effect and, in line with BF, a selection effect which is independent of the allocation decisions (and vice versa). Finally, we show that in the newly introduced BKT variant, an interaction effect does not only occur within each individual market segment, but also between segments. This results in an additional cross-segment interaction effect, which is not present in the arithmetic BF method.

Annual Risk Measures and Related Statistics [PDF]
Spring 2017

A common method to annualize monthly risk measures is to multiply the outcome by 12 or the square root of 12. Paul D. Kaplan has shown that for one particular measure, the standard deviation, this approach is incorrect, given that returns are compounded over time rather than summed. Kaplan has also presented a method to calculate the annual standard deviation correctly. This article expands on Kaplan’s work. It shows for a wide palette of risk measures and related statistics how the annual variants can be calculated correctly, under the assumption of compounded returns.