August 10, 2009
ASSET ALLOCATION: STILL RELEVANT AFTER ALL THESE YEARS
Some said it was dead. Others claimed it was misleading. Many simply ignored it, in good times and bad. But asset allocation is hardly dead. In fact, it couldn't be any more relevant.
The mistake that many investors make is comparing a multi-asset class portfolio to something riskier, such as any one asset class. In fact, there are countless ways to beat a multi-asset class portfolio over the short- as well as long-term horizons. Doing so after adjusting for risk, however, is far more difficult.
It's easy to find one asset class or one security with an expected return that's far higher than the market portfolio, which we're defining as all the world's major asset classes weighted by the market values. The problem is that there's something approximating an equal abundance of asset classes and individual securities with lesser prospects at any one time relative to the market portfolio. Distinguishing one from the other isn't impossible, but it's devilishly hard to do continually, year after year.
Does that mean we should simply buy and hold the market portfolio? Probably not, although it's worth pointing out that you could do a lot worse. Consider, for instance, one measure of the market portfolio, as defined by the Global Market Index (GMI), courtesy of The Beta Investment Report. For the 10 years through the end of July 2009, GMI posts a 3.8% annualized total return, which is considerably better than the 1.2% annualized loss for U.S. stocks, as per the S&P 500.
Still, the temptation is always there to do something else. There have been times in the past, and there will be times in the future, when U.S. stocks beat GMI, for instance. That's true for any of the various stock, bond, real estate and commodities components that collectively comprise GMI. Why not simply own the components with the highest expected returns and shun everything else?
The answer, of course, is that no one has 100% confidence in forecasting returns, in part because no one can see all the risks lurking in the distance. Then again, our confidence in peering into the future isn't zero either. Financial economists have turned up some useful insights over the years for projecting return and risk. In fact, in my upcoming book—Dynamic Asset Allocation: Modern Portfolio Theory Updated for the Smart Investor, to be published by Bloomberg Press in February 2010—I review, analyze and interpret the multi-decade history of research that should be the cornerstone of every strategic-minded investor's money management principles.
As for GMI, it should be no surprise that it beat the U.S. stock market over the past 10 years. After all, GMI holds a fair amount of bonds, both domestic and foreign, along with REITs and commodities, all of which post positive returns for the past decade. In fact, quite a few corners of GMI are doing well--quite well. Inflation-indexed Treasuries, for instance, post a 7%-plus annualized total return for the past 10 years through last month. Meanwhile, emerging market bonds are up by more than an annualized 11% over that stretch—the best performer among the major asset classes.
Ah, but who in July of 1999 recognized that emerging market bonds would beat everything over the next decade? Probably no one, or at least very few. The summer of 1999, you may recall, was a point when the equity market was soaring--U.S. stocks in particular. That convinced many that stocks were the only game in town.
Picking winners today isn't any easier today. Should we just abandon all hope and simply buy and hold the market portfolio? That's one solution, and it probably will do fairly well as a long-term strategy. But there are sound reasons for modifying the market portfolio, and that begins with adjusting it to fit each investor's particular circumstances, including time horizon and risk tolerance, which varies from person to person.
We can modify Mr. Market's allocation based on valuation and other factors that drop clues about the future path of individual asset classes. If stocks look attractive relative to bonds, for instance, we may want to change the passive mix as it currently exists. But we need to be careful here. To the extent that we alter the passive asset allocation based on forecasts (as opposed to risk tolerance and time horizon), we should do so cautiously and marginally, and in accordance with our confidence in the prediction.
A factor in an investor's confidence reflects his skills as an analyst. Suffice to say, some are better than others. Beyond individual talent, it's also important to recognize that expected return transparency ebbs and flows through time, along with the business cycle and other variables. At turning points in the cycle, which is to say moments of extremes, the general level of predictability is higher, if only slightly.
But make no mistake: we should practice dynamic asset allocation cautiously, and with eyes wide open. This is a difficult task and after adjusting for trading costs, taxes and the inevitability of error at times, few will come out ahead of the market portfolio in the long run. Alpha, as the saying goes, sums to zero, even when second guessing Mr. Market's mix. For those who are skilled at managing asset allocation, the market-beating results come exclusively from those who engage in this dark art and come up short of the unmanaged benchmark.
Asset allocation is still relevant, but no one ever said it was easy. The good news is that Mr. Market's asset allocation is pretty good template for considering the possibilities. But the further you move away from the passive mix, the more talent and confidence you'll need to make the trip worthwhile.
Posted by jp at August 10, 2009 10:14 AM
“To the extent that we alter the passive asset allocation based on forecasts (as opposed to risk tolerance and time horizon), we should do so cautiously and marginally, and in accordance with our confidence in the prediction.”
Good enough advice, but it misses the bullseye: all asset allocation -- all investment decisions! -- should reflect our considered beliefs about subsequent asset classes' returns, aware of our limitations and the fact that the market can stay “irrational” (i.e., contrary to our beliefs) longer than we can hold our beliefs. The "variance" part of Markowitz's equation can be replaced (perfectly!) by the square of our uncertainty about the subsequent returns that'll be experienced.
Every AA is the outcome of a joint forecast of expected returns and uncertainty about what those returns will be over the target time horizon. The consensus, as Sharpe was apparently the first to identify formally 50 years ago (!), is a pretty good forecast of the outlook, perhaps because it is just the consensus investor who will buy and sell assets up or down to the price that makes the subsequent risk-adjusted returns equal.
There's still plenty of benefit to getting ahead of the curve and forecasting what others will forecast. But to do so, we have to say that we are better at playing the Beauty Contest better than the dollar-weighted other participants. One who has only a handful of chips in front of him ought to be damned careful in claiming he has more insight than those with a mountain of them.
Posted by: Walt French at August 10, 2009 10:52 PM