There exists a significant body of research that attempts to determine the relative importance of different investment decisions with respect to return outcomes. The onset of the investigation into this topic can be traced back to a paper originally published almost 40 years ago: Determinants of Portfolio Performance, written by Brinson, Hood, and Beebower in 1986. The authors’ analysis in that seminal paper demonstrated that long-term asset allocation drives 90 percent of a portfolio’s performance, leading to a generalized – and quite prevalent – hypothesis: investment policy asset mix decisions are far more important than investment fund implementation decisions.
In the ensuing decades, this hypothesis has been the subject of much discussion among investment professionals. Some have fully accepted it, while others have severely critiqued it; criticisms have ranged from arguing that the original analysis missed important nuances, to stating that it was outright flawed. The ongoing debate even induced one of the authors of the original paper to wade back into the discussion decades later, stating, “We would not have guessed that a six-page article would be the focal point of a 20-year discussion.”1
There is no doubt that the subject is both contentious, and one that is fundamentally important to institutional investors. At the same time, the complexity and variety of the investment decisions faced by institutions have expanded significantly since much of the original research was done. In light of these two considerations, this article aims to take a fresh stab at the topic. Starting with a review of the main points and findings that have been put forth on the subject over the decades, we conclude that the biggest point of disagreement is the way in which asset mix decisions are isolated and measured on a relative basis. Despite multiple proposals on how to define an appropriate baseline, we argue that a crucial consideration has been overlooked: the uniqueness inherent to institutional investors’ individual circumstances. To address this, we devise a framework that introduces the concept of a reference portfolio that can serve as a customized baseline to measure and guide asset allocations decisions, allowing us to reinvestigate the original question. Then, within our proposed framework, we use historical benchmark and fund manager returns to represent a broad selection of fixed income, equity, and alternative investments for a modern-day investor, test different objective-oriented cases, and compare our results to those previously found.
Our primary finding is that the relative weight of one type of decision’s impact on risk/return outcomes can vary significantly depending on the following factors:
The investor’s return objectives and/or risk tolerance
The breadth of the chosen asset class opportunity set
The alpha potential and tracking error profile of active management (which varies by asset class)
Consequently, despite extensive past efforts and our own more recent attempt to conclusively establish which investment decision matters more – asset mix or fund implementation – we find there is no universal result that can be relied upon given the large amount of variables at play. Therefore, we would argue that the default mindset for institutional investors should be to ascribe equal weight to both when making ex-ante investment decisions, even if one will likely end up being more consequential than the other on ex-post outcomes.
To read about our findings in more detail, find the full piece here.