It has been five years since veteran investment consultant and celebrated author Peter Bernstein invoked the word "obsolete" to describe the policy portfolio, which encourages fixed weights for multi-asset class investing strategies. By contrast, dynamic asset allocation is the superior alternative, he argued in a widely discussed 2003 article in his newsletter Economics and Portfolio Strategy.
The idea wasn't new in 2003, although--due in part to the timing--Bernstein's counsel was provocative. The U.S. stock market had just come through one of its deepest and longest corrections in history. Meanwhile, here was a high-profile, widely respected financial analyst and historian challenging what some--perhaps many--considered conventional wisdom.
In fact, the choice of dynamic or static asset allocation probably has not made a huge difference one way or the other in recent years. Midway through last year, all the broad asset classes--equities, bonds, commodities, REITs and most of their subdivisions--were in the fifth year of an extraordinary bull run that left almost nothing behind. Fixed and active asset allocation strategies alike had ample chance to shine. So it goes when almost everything is virtually flying, year after year.
No one will confuse the bull run between 2002 and 2007 with what has been unfolding over the last 12 months or so. Higher volatility and red ink are again harassing investment strategies, elevating the relevance of asset allocation along the way. Although there's always likely to be a bull market somewhere, the expectation of tidy gains in almost everything is probably history. That raises the stakes for the gritty work of picking asset class weights and deciding when and how to adjust those weights.
Bernstein's five-year-old counsel may face the acid test in the months and years ahead--arguably for the first time since his 2003 essay appeared. Why? The case for a relatively active approach to asset allocation looks timely for 2008 and beyond. The reasoning boils down to a belief that actively managed asset allocation is well suited in a world of divergent and evolving expectations for risk and return.
In fact, the conviction is supported by a growing body of academic research that has been piling up empirical evidence on the side of dynamic strategies. A crucial finding: Securities markets appear to be at least partially predictable after all.
Professors Robert Shiller (Yale), Ken French (University of Chicago) and scores of others have documented what many investors have known (or suspected) all along: Markets go to extremes from time to time. That's another way of saying that expected returns vary through time, which inspires active asset allocation.
Yes, the S&P 500's annualized return is 10 percent over the past 80 years, but even if that holds for the future, no one should expect 10 percent year in and year out, as the chart at the bottom of this page reminds us. Expected return rises and falls in connection with changing prices and valuations. And as academic studies strongly suggest, when expected returns are relatively high, the weight of the asset class should also be relatively lofty, and vice versa.
Sound familiar? The concept is at the heart of Graham and Dodd's Security Analysis, the 1934 classic that formalized value investing, or buying securities at a discount to their estimated intrinsic value. The book focuses on individual securities rather than markets. But embedded in the strategy is the belief that the relationship between market price and intrinsic value is forever in flux, which means that expected return fluctuates, too. For those who agree, the intellectual leap to active asset allocation is only natural.
Academics are inclined to agree in the 21st century, but not without rethinking certain aspects of modern portfolio theory as some have used it in the past. Notably, the intellectual evolution that now favors active asset allocation conflicts with the random walk theory (RWT), a particular version of the efficient market hypothesis (EMH).
RWT, which helped spawn the indexing revolution, asserts that returns are 1) independent--meaning that yesterday's return has no effect on today's or tomorrow's; and 2) returns are "normally" distributed over time, as per a gently sloping bell curve. Assuming the two assertions accurately describe market behavior provides statistical aid and comfort for fixed-asset allocation strategies premised on the idea that expected return is fairly stable.
In fact, returns don't strictly follow a random walk. So-called "fat-tail" distributions prevail, meaning that large price changes occur in the real world with far more frequency than a normal distribution predicts. That has been clear since at least Benoit Mandelbrot's research in the 1960s, and over time the literature has only confirmed the point. The message is periodically repeated, often to deaf ears. Eugene Fama, who coined the term "efficient markets," recognized the case for non-random distributions in his landmark 1965 paper that helped launch EMH. More recently, financial scold Bill Jahnke argued that a proper reading of finance literature leads to the conclusion that "the policy portfolio deserves to be buried" (The Investment Think Tank, Bloomberg, 2004).
Indeed, the academic bibliography is now flush with 20-plus years of empirical studies showing that fundamental data (dividend yields, interest rates, etc.) offers a richer source for predicting returns than what was thought possible via the early conceptions of EMH that focused on price alone. For example, one line of research shows that returns are mean reverting in the medium to longer term, which implies predictability, and so asset allocation weights should change. Such notions clash with the random walk account of EMH.
A modest degree of return predictability may be respectable in academic circles, but debate still rages about the underlying cause. One camp says irrational investors drive valuations to extremes. A competing view sees markets through the prism of an updated EMH: Expected returns vary in order to compensate for risk associated with the business cycle--a recession beta, if you will.
Either way, the lesson is that some degree of dynamic asset allocation is warranted in a world where expected return cycles.
How can active asset allocation coexist with notions of an efficient market? Economist Paul Samuelson, one of the intellectual fathers of EMH, has bridged the chasm by observing that markets can be microefficient and still be macroinefficient.
In fact, the seeds of active asset allocation in the context of modern portfolio theory were planted long ago. After all, MPT's founding document (Markowitz's 1952 paper on optimal portfolios) allows for asset classes--"aggregates," as he calls them--in portfolio construction. Meanwhile, the Capital Asset Pricing Model--the theoretical foundation for capitalization-weighted indexing--assumes shifting weights for assets, as per Mr. Market's incessant repricing. And in the 1970s, some of the pioneers of the original index funds--Bill Fouse, for one--took the original but overlooked MPT ideas to heart and developed so-called tactical asset allocation, which eschews the principle of the policy portfolio.
The chief inspiration for accepting some form of dynamic asset allocation is the market, suggests Bill Reichenstein, CFA and professor of investments at Baylor University. In a recent interview, he discusses an example drawn from the then-current market condition of the S&P 500--off roughly 14 percent from its high of last October. "From this point forward," Reichenstein explains, "the risk premium that's embedded in stocks is higher than the risk premium of a few months ago. There's nothing inefficient about that. That doesn't mean that the next three months are going to deliver well above average returns. What it means is that over the next five or 10 years, stocks will do a lot better than they would have if you would have started out three months ago."
Market excess becomes conspicuous from time to time, says William Bernstein, author of The Four Pillars of Investing. That was true in 1990, for example, when "a lot of people saw all the BS in the Japanese market," which was trading at astonishing nosebleed valuations, he recalls.
The past, of course, never fails to offer trustworthy guidance about what you should have done. Handicapping the future in real time is the challenge. Yes, the academics now counsel that returns are somewhat predictable, but at best it's still only a partial solution because forecasting still entails risk.
"The level of predictability [in equity returns] tends to be 25 percent, 35 percent," says Reichenstein. That suggests that asset allocation strategies should be only partially dynamic--such as allowing for shifts in equity weights within a modest range without betting the farm on predictions, he advises. Even then, adding value to the portfolio assumes the strategist has the talent to correctly read the market's signals and make portfolio adjustments at opportune moments. Nonetheless, while the academic literature shows that there is opportunity for generating alpha with dynamic asset allocation relative to a fixed policy, there are no guarantees. Expected returns vary, but that doesn't mean everyone will profit from the trend.
Perhaps, then, it's no surprise to learn that even advocates of dynamic asset allocation recognize that the policy portfolio is still useful--even if it's theoretically inferior. A fixed asset allocation can capture "a big part" of the results generated by a more active strategy, says Jim King, president and chief investment officer of National Penn Investors Trust Co., which manages money for high-net-worth clients.
Bill Bernstein, too, says static asset allocation "isn't a bad idea," although he favors a dynamic approach for his client portfolios. Active asset allocation demands "industrial amounts of discipline," he reminds. Taking advantage of higher expected returns and then pulling back when the outlook is less alluring is inherently a contrarian philosophy. "You have to buy when everyone else is selling."
Yes, mustering the discipline to buck the crowd can reap big rewards. One has only to consider Warren Buffett, Ben Graham and George Soros for inspiration. But such names tend to be the exceptions. Mediocrity or worse is still the likely result in money management generally. And that leads us back to the bedrock principle that was quantified all those years ago by Markowitz and his intellectual heirs: Risk and return are joined at the hip.
Some things remain the same no matter what the academics say.
James Picerno (jpicerno@highlinemedia.com) is senior writer at Wealth Manager.
