Decision maker based attribution
Decision Based attribution is getting more and more attention lately. The technique that allows for capturing all decisions that are made in the investment decision process from the high level strategic allocation decisions down to the lowest operational decisions. In this year’s Dietz award winning article, Arun Muralidhar proposes to extend the analyses by the ‘who’ dimension. In this way you can show per responsible person the added value that he or she has.
According to the article, a good performance report should provide insights in the following 5 questions:
- Who made the decision?
- What did they do?
- How much was their impact?
- How good were those decisions?
- Were they skillful?
The decision maker
For the first question, ‘Who made the decision’, it is important to have a clear picture of the investment decision process and the Organizational Network. The mapping of the one into the other provides insight into who made the decision. It can be that multiple people are responsible for a decision, for example the board and an external consultant together determine the Strategic asset allocation and the rebalancing policy. Then the added value of this step should be split between them according to an pre-determined weight.
It is not only important to look at the added value of each decision, but also at who is responsible for the decision.
Impact and actions
Decision based Performance Attribution can be used to answer the next two questions, e.g. to determine what the decision maker did and the level of his impact. Within such an analyses, every hierarchical decision is evaluated against its predecessor. For example a tactical deviation in the asset class allocation is evaluated against the strategic allocation measuring if the decisions added value to the total fund. Next to an allocation decision, also a tactical deviations in the rebalancing frequency could be implemented. The allocation decision and the deviation in rebalancing frequency decision could be taken by different people in the organization. Therefore they should also be evaluated and shown separately.
Quality of the decision
To determine whether the decisions were good decisions, Arun Muralidhar argues that you should look at the risk adjusted return, best measured with the M-cube risk measure. This measure measures the risk adjusted return in basis points of performance, similar to the M-square (Modigliani-Modigliani) measure. In contrast to the M-square, the M-cube measure is normalized for differences in correlation between the portfolio and the benchmark. This measure is helpful to compare the returns of portfolio managers. And it would be interesting to see how this framework can be extended to cover all decisions in the investment decision process.
Determine how much confidence you have that a portfolio manager is skillful
Skills or luck
To determine whether a portfolio manager is skillful, often insufficient history is available. Again, it would be interesting to see how this can be extended to cover all decisions, especially the decisions which are not taken to maximize return, but to minimize risk (like implementing an overlay). Altogether, Arun proposes a very insightful way to report for pension funds where not only the decisions are captured but also the ‘Who’ dimension is taken into account. He proposes workable solutions for challenging questions like, how good are the decisions and was any skill shown, or was it just luck. Furthermore it was a very readable article, so a well-deserved winner of the Dietz award and a must read for anyone in the field of performance measurement.