| |
 November 2010 Asset Class Mythology Rethink All Those Weird Sectors and Odd Corners of Market Allocations
Now in 2010, only a few investors question the value of foreign stocks in their portfolios. In fact, even if they don't directly buy foreign shares or mutual funds, they probably still hold them in their domestic equity funds.
But something has changed: Over the past twenty years stock markets have become increasingly global. Not only do foreign shares have greater distribution in he U.S., more overseas investors also own U.S. shares. True multinational companies headquartered both here and abroad also increase globalization. As a result, the world's stock markets have become much more correlated, diminishing the benefit of international diversification. Advisors and investment managers realizing that investors still seek to diversification -- perhaps even more today than twenty years ago -- are offering products that allow the average investor to participate in arcane corners of the financial markets, from small cap developed market funds and emerging market stocks, to commodities and specific niches in real estate. The big selling point isn't the fact that these alternatives are new or unique, but rather that they, like foreign equities in the 1990s, can diversify portfolios. Sales suggest this is a claim that resonates with investors.
Is More Really Better?
Nevertheless, from the proliferation of alternative investment strategies and niche mutual funds, advisors, investment managers, and most importantly investors, all seem to believe not only that they can help diversify portfolios, the more the better. Advisors rarely recommend simple portfolios of stocks, bonds, and cash but rather embellish them with capitalization, sector, credit quality, and duration alternatives. Often they'll also add real estate, precious metals, commodities, and/or specific emerging country allocations resulting in a portfolio "pie" with lots and lots of little slices. But if individual alternatives offer little diversification potential, will a greater number provide more? Conceivably they could, but there's no reason why they'd have to. Rather than speculate, we compared the results of actual asset allocation models with three different levels of risk. The models and the asset classes used for each are shown on Chart 1.
The risk levels are defined as Conservative, Moderate, and Aggressive. This is fairly standard procedure for most investment managers who first profile their clients and then assign them to a particular strategic asset allocation based on their risk tolerance. We thought it would be informative to consider more than simply a balanced allocation because the addition of more esoteric asset classes could potentially be more beneficial for more aggressive allocations -- or vice-versa. By looking at allocations with different levels of risk, it's possible to see if this is truly the case. To test the benefit of adding additional asset classes, we used three levels of strategic allocation models. The first is composed of only two different classes, stocks and cash. This extremely basic mix was suggested by our previous analysis which suggested cash was perhaps the best diversifier for equities. If true, additional arcane classes would not be necessary. In essence, this level is the baseline or control level. Its models appear in the first row of Chart 1. The second level uses allocations offered by many advisors to individual clients and 401k participants. It uses seven different equity and fixed income asset classes along with cash. Its models and asset classes are the second set of charts and classes in Chart 1. The third set appears in the bottom row of Chart 1. This has the largest number of potential asset classes classes (16) and the widest range of allocation (14-15 asset classes) at each level of risk. For comparison, the second set, on average, uses 6-7 of its 7 potential classes and, as you would expect, the first set uses both of its two potential asset classes in each of its allocations.
The Rules of the Game Although long-term investing is often encouraged, strategic asset allocation models are typically revisited every five years or so in order to make adjustments for significant market changes. To capture this, we reoptimized the strategic asset allocation models every five years based on the prior five years' historical performance. In other words, the first inputs for optimization came from January 1, 1990 through December 31, 1994. Based on them, the first asset allocation models became effective on January 1, 1995 and remained in place until December 31, 1999. Throughout their duration, each model was rebalanced on a quarterly basis as investment managers often do. The models themselves are all index-based. Not only were index series used to create the models, index returns, risk, and correlations were the basis of comparison. While it's not possible for an investor to directly invest in an index, this approach was used for consistency across the various asset classes. Finally, because the test was index based, the effects of transaction costs were ignored. If included, they would pose a bigger drag for the more complex models given that costs are higher to invest in newer, less main-stream asset classes. Using the methodology outlined above, we measured each allocation's returns and risk over three measurement periods: 1995-1999, 2000-2004, and 2005-2009. For completeness sake, we also reoptimized the models based on data from 2005-2009 and include results through August 2010 on Charts 2. The risk/return summary for the entire period is reflected for all models on Chart 3.
Simply the Best Chart 3 goes beyond simple return to include risk as well. The simple Stock/Cash blend dominated in all three levels of risk. Despite the additional nine asset classes, the most complex blend ended up incurring the greatest risk in each instance. Inasmuch as diversification is supposed to help control risk, the inclusion of the additional asset classes can only be seen as a failure.
Of course this is only one fifteen year-plus period which covered the major domestic large cap run-up of the 1990s, the bear market which followed in 2001-2002, the subsequent recovery, and finally the fallout from the 2008 credit crisis. These are unique events and aren't likely to be repeated -- at least in such a sequence -- again. This is the problem of relying on strictly historical data which can be heavily influenced by one of a kind events. Nevertheless, this is how many money managers and advisors construct and manage their clients' asset allocations. Over the past fifteen years, it' likely many investors got results quite similar to those illustrated for level two allocations if not level three. It's also evident they would have been better served in level one's stock/cash blends. Are you surprised by this? You shouldn't be. Mean-variance optimization has greatly improved countless investors' strategies, but like most tools, its success is heavily dependent on the skill and experience of the one using it. The fact that a wide variety of asset classes that have proven to be good diversifiers in the past are now more readily available available does not guarantee their mere presence in a portfolio will improve its future results. You or your money manager may not take as simple an approach as we did in this test. Future results can be improved through the use of appropriate constraints, credible forward projections rather than historical data, or a more precise collection of asset classes. Maybe you can do that. But the one thing this exercise should underscore is the fact that the simple addition of more asset classes is no guarantee of better results -- particularly risk-adjusted results. More may be better, but only if it's more experience and expertise, not just more inputs. Search this site! Just enter you key word or words:
Get current quotes or follow your own custom portfolio,
courtesy of E-Line Financials:
|
|||||||||||||||||||||||||
WENTY YEARS AGO, INVESTORS fretted about adding foreign stocks to their domestic holdings. Although the great U.S. large cap runup was poised to begin, stocks of developed foreign countries were outperforming many domestic alternatives. In hindsight, adding foreign investments was not only the right decision from a return standpoint, it also helped diversify portfolios against some of the sharp gyrations of the 1990s.
