The Next Step in the Evolution of Decision-Based Attribution : Micro or Rules-based Attribution (and the Atoms of Attribution)

There has been a lot of focus on macro decision-based attribution (DPA), both in terms of decisions (e.g., asset allocation, manager selection, manager allocation and security selection) and who makes the decisions (decision-maker attribution). However, there has been little focus on the rules underlying the decisions – what we term micro DPA. For example, asset allocation value added may be generated by a hierarchy of rules that drive the decision at a very high level (e.g., a sub-model based on seasonality and a sub-model based on economic data) and then drilling further, by further rules that indicate the conditions under which the investor should go overweight or underweight. We argue that without a clear understanding of micro DPA, there is a danger that focusing solely on macro DPA may allow inefficient decisions to creep into a portfolio. We provide a simple case study of how rules-based micro DPA can be applied to extract a better understanding of decision-making in an investment organization and apply it to the asset allocation decision, but it is obvious that this same procedure can be applied at all levels of the portfolio and for all decisions. While this technique is clearly most easily applied to a quantitative process which has clearly specified rules, we also discuss how it can be applied to a qualitative investment process