The investing world is full of rules of thumb, and while some are helpful simplifications of otherwise complex concepts, others are in fact misguided or even flat out wrong. One of the roles we play as a trusted advisor is to steer clear of these types of misconceptions, and to that end, an investing "axiom" that falls into the misguided category is the belief among investment practitioners that asset allocation explains 90% of a portfolio’s returns. This statement is in fact a gross misrepresentation of a 1986 article in the Financial Analysts Journal1, and yet many investors use the supposed 90% rule as the justification for their emphasis on making tactical, asset class-based investment decisions. Subsequent literature has helped correct the misunderstandings of that article but hasn’t addressed how investors can improve their investment processes.
The 90% Rule
Brinson, Hood and Beebower’s 1986 article “Determinants of Portfolio Performance” unfortunately had a confusing title, as their work actually had nothing to do with the absolute level of portfolio performance, but was rather focused on variability of performance. To conduct their analysis, the study used a sample of 91 large U.S. pension plans and studied their associated policy portfolios and actual returns. A policy portfolio is a set of asset class weights that a plan sponsor believes gives it the best chance of meeting its long-term return goals relative to the risks. The weightings are relatively static, and any changes should represent a shift in long-term thinking. A plan’s actual allocations, though, fluctuate around these policy targets depending on a range of factors such as the performance of different asset classes. When the authors compared the hypothetical return of each plan’s policy portfolio with its actual return, they found that the policy portfolio’s return, on average, explained 93.6% of the variation in the actual return. The residual 6.4% they assumed was explained by market timing and security selection.
Many readers of the article then went on to claim that 90% of a portfolio’s total return was explained by asset allocation. While even the authors of the study have pointed out that this assertion is false2, the article’s real findings were also relatively uninteresting. As mentioned above, the authors define three variables – asset allocation (setting of a policy portfolio), market timing and security selection – that explain all of the variability of returns in a portfolio. However, when analyzing the entire universe of investors, market timing and security selection are a zero sum game. That is, when you add up the returns of all investors, by definition you get the market return. If your portfolio has a higher return than the market, someone else must have a lower return, and thus relative performance amongst all investors will always add up to zero. Therefore, on average, asset allocation will always wind up explaining 100% of the variability of performance.
The study has spawned a long running controversy in the investment community on the importance of asset allocation, as well as a slew of articles that attempt to fill in the gaps in our understanding of how asset allocation impacts returns. With regard to performance differences amongst a peer group, a study by Roger Ibbotson and Paul Kaplan suggests that asset allocation explains roughly 40% of the difference in return among a broad sample of balanced U.S. mutual funds. Most of this debate we do not find overly engaging though, because the focus centers on short-term relative performance. Knowing how much your performance deviates from the market over anything but an extremely long period of time is not a useful exercise in our view, as we are more concerned about how to generate attractive long-term absolute returns. As the adage goes, “You can’t eat relative performance.”
Nevertheless, the same Ibbotson and Kaplan study does reveal one useful finding on generating performance. The return differential between the balanced funds in their sample was actually a function of the degree of active management, or “active share,” that funds exhibit versus one another3. The more active the funds in the sample are, the less asset allocation determines performance. Interestingly, a Yale study showed that funds with higher active share tend to outperform over long periods of time. Furthermore, when the high active share results from stock selection as opposed to factor bets, performance was even higher, suggesting that bottom-up stock picking is a valuable exercise4.
It is difficult to overstate the influence that the original Brinson article has had on the investment industry. Its mischaracterized findings played a significant role in creating the top-down approach to asset allocation driven by macroeconomic analysis that is prevalent among the investment advisor community. It also played a role in the growth of index investing, as it led the investment community to think that returns from active investing were not meaningful. It stands to reason that if asset allocation were responsible for 90% of portfolio returns, then advisors could more easily justify spending the majority of their time allocating between various index funds. Evidence of this is not difficult to find: There are numerous examples of firms citing the Brinson study in their marketing materials. This top-down way of thinking pervades even the way the industry broadly communicates with clients, as conversations often times begin and end with a discussion of asset class (as opposed to manager) performance.
While it is valuable to test the logic of common investment rules of thumb, demonstrating that this dogma itself is not based entirely in fact still does not enlighten us as to the correct approach. For insight on this, we have found it helpful to study the investment processes of investors we respect who have had success generating exceptional long-term returns.
Alternative Approaches to Multi-Asset Class Investing
One school of thought on allocating capital is exemplified by Warren Buffett’s management of Berkshire Hathaway’s assets, which are generated through float from its insurance businesses, as well as retained earnings from its other operations. Warren Buffett enjoys the unique and envious position of having virtually no constraints on where he can invest his capital. While Berkshire has a substantial portfolio of short-term investments used to hedge against future insurance losses and to run the day-to-day operations of the business, we consider those assets to be separate from the long-term return seeking investments that Buffett is well known for managing.
If Buffett chose a traditional asset allocation approach, he would likely devote his time to forecasting asset class returns and allocating money according to the results of those findings. Instead he has eschewed macro forecasts and has pursued a bottom-up approach to allocating capital. This approach most notably consists of the ownership of several businesses outright and long-term common stock holdings, but has also included ownership of fixed income securities, preferred stocks, cash equivalents and more opportunistic investments, such as merger arbitrage. With regard to asset allocation, in Buffett’s own words he says, “We have no particular bias when it comes to choosing from these categories. We just continuously search among them for the highest after-tax returns … limiting ourselves to investment alternatives we think we understand5.” Buffett and his partner Charlie Munger have bought equities within their circle of competence for many years, but recently, as their assets have grown and succession planning has become a concern, Buffett has hired additional managers to buy stocks. Buffett’s main focus has always been on adding value to his enterprise through long-term business ownership, but his opportunistic investments are also worthy of note. Through his conservative management of Berkshire, Buffett has been able to provide capital during periods of market dislocation on extremely attractive terms, and this has been an additional source of return to his shareholders.