Decision-based performance attribution (DPA) focuses on breaking down total performance into the various decisions made at investment funds. But, because duties are delegated at many levels, inefficiencies and poor risk management may ensue unless all individuals or institutions are held accountable for the decisions they make. Thus, decision makers should be held accountable for their attribution decisions. A report that tallies every dollar of added value—attributed to each relevant participant in the investment process—is needed. Creating a performance attribution process that focuses on who the decision maker is ensures accountability because only what is measured is monitored and managed, leading to improved governance and risk management.

In a typical investment process, individuals or institutions can make multiple decisions throughout the investment decision process. It would be useful to highlight how the participants did in aggregate and how they performed each task. The author calls this approach “attribution of returns using organizational networks” (ARUN).

DPA is important because it helps portfolio managers recognize sources of added value, both positive and negative, so they can emphasize the good and correct the bad. Although the DPA method accounts for every basis point of performance, it is silent on who in the organization made the decision, what the risk-adjusted return is, and how confident one can be about the skill content of the added value. Thus, compensation may very often not be in line with the contribution to the bottom-line added value. Because decisions are delegated and responsibilities can be diffused, a clear articulation of who made what decision becomes important. In addition to basic attribution, the reports should include risk-adjusted performance and the level of confidence in the skill of the decision maker. This approach ensures that big bonuses or fees are not paid for one quarter of outperformance, which could be due entirely to noise.

The author examines the organizational network of a typical public pension plan in the United States—how asset ownership is structured, who the decision-making entities are, what tasks they are responsible for, and to whom they report. Through this mapping, the author demonstrates the challenges of performance attribution, concluding that the reports must include risk-adjusted performance and the level of confidence in the skill of the decision maker in addition to basic attribution. He then reviews typical Brinson and DPA reports to show the challenges of using such reports and being able to determine whether the decision makers are adding or subtracting value and whether such value addition or subtraction is based on skill. The author goes on to illustrate techniques to calculate risk-adjusted performance when decisions are delegated and investors are worried about luck versus skill, and he provides a template for the ARUN approach to demonstrate its usefulness.

Calculation of risk-adjusted performance requires two statistics in addition to the returns and weights used for attribution: the actual relative risk taken by the agent (person or entity delegated to) in deviating from the benchmark set by the principal and the target relative risk set by the principal to control the agent. M-cube—created by the author as an extension of the M-square metric of Modigliani and Modigliani (MM, Journal of Portfolio Management 1997)—shows that the performance of asset managers can be normalized for differences in correlation with the benchmark while retaining the attractive MM feature of expressing the risk-adjusted statistic in return terms. The question about luck versus skill is addressed by useful equations derived by Ambarish and Siegel (Risk 1996) and further developed by Muralidhar and U (Journal of Pension Plan Investing 1997) to establish a measure of confidence in skill. The author offers a template for ARUN based on all these inputs. An ideal ARUN report would present the value added by decision makers, clearly showing who contributed how much to the total fund and by what decisions.

The author highlights a critical factor often ignored in performance attribution: the decision maker. Getting a clear fix on the “who” factor would ensure better accountability. Although much more work needs to be done to arrive at foolproof attribution, this paper is a good start. This research would be an extremely useful tool in managers’ hands to ensure that people are rewarded in accordance with their contribution rather than their contribution getting lost in a maze of layers of delegation.