February 22, 2012
Volatility & Asset Allocation
Earlier this month I wrote about the upcoming launch of two foreign bond index funds from Vanguard—the firm’s first step into the market for international fixed-income products. Foreign bond funds are nothing new these days, although the proposed Vanguard funds are a bit out of the ordinary because the portfolios will hedge foreign exchange risk. Vanguard’s reasoning is that forex hedging dampens volatility, which is true. But as I noted, volatility per se isn’t a problem in the context of a broadly diversified multi-asset class strategy. In fact, volatility can be quite helpful in that case. Although I mentioned this point briefly, the issue deserves more attention vis-a-vis rebalancing.
The basic message is that in order to capture the rebalancing bonus, a certain amount of volatility is necessary. As I noted in my book Dynamic Asset Allocation, a 1988 paper by Andre Perold and Bill Sharpe ("Dynamic Strategies for Asset Allocation," Financial Analysts Journal) outlined the standard framework for thinking about the relationship between volatility and rebalancing. Although you can forecast any outcome you prefer under specific assumptions, market history and the Perold and Sharpe paper imply that we should see volatility as productive within the environment of broad diversification across the major asset classes. Granted, there's no guarantee that rebalancing will produce higher returns vs. the same portfolio that's left untended, but the argument for not rebalancing is hardly a no-brainer either. In my view, a fair reading of the literature, along with a careful review of market history, suggests that some degree of rebalancing is useful if not essential.
A number of researchers have come to similar conclusions over the years, including a monograph (“The Rebalancing Bonus”) by the financial planner William Bernstein, who advised:
The actual return of a rebalanced portfolio usually exceeds the expected return calculated from the weighted sum of the component expected returns. A formula for estimating this excess return is derived and tested. It is demonstrated that assets with high volatility and low correlation with the remainder of the portfolio provide considerable excess return, or "rebalancing bonus."
As a simple test, consider how my proprietary index of all the major asset classes fared in recent history in three forms. One variation is the passive, unmanaged mix that’s initially set to market-value weights as of December 31, 1997 and left to drift with the market’s ebb and flow—the Global Market Index (GMI). The rebalanced version of GMI simply returns the portfolio mix to its 1997 market weights at the close of each year. Finally, the equally weighted GMI starts out with equal weights and resets to that mix at the end of every year. The chart below tracks the results of an initial $100 investment in each strategy from the end of 1997 through December 31, 2011.
It’s clear that the rebalanced and equal-weighted versions of GMI earned substantial premiums over the unmanaged GMI. In both cases, the higher return is due to rebalancing. In this simple example, the rebalancing is annual at the close of each calendar year, although in practice there’s a case for embracing a more opportunistic strategy. Meanwhile, let’s also recognize that the rebalancing bonus is dependent to a degree on volatility. Mindlessly adding assets to capture higher volatility per se isn’t wise, but if there’s an economic rationale for a given set of asset classes, and individually they exhibit relatively high volatility, intelligently managing the mix through time can pay off handsomely.
That brings us back to hedging forex risk. Yes, it’s clear that hedging reduces volatility. That may be beneficial if we’re looking at the assets in isolation, but as part of a broadly diversified portfolio it’s not obvious that keeping the volatility of the individual asset classes to a minimum is beneficial as a general rule. Indeed, depending on how you analyze the possibilities, hedging might end up taking a bite out of the rebalancing premium.
Markowitz long ago told investors that the primary focus should be on designing and managing the portfolio. That’s a message that still resonates and it reminds us not to obsess over the parts if—if—we’re embracing broad, multi-asset class diversification.
Posted by jp at February 22, 2012 10:25 AM
Yes, all true. There's always room for debate. I have a feeling that no matter how much history I show you, you'll be skeptical. Investing inevitably requires a leap of faith, but some leaps are more persuasive than others. By the way, did you notice that you use history to indict history? In any case, this post is just a tip of the iceberg. If you'd like to see more, I have a couple hundred pages of additional research in my book. But here's the real question: What do you recommend, and why?
Posted by: JA at February 23, 2012 10:41 AM
Sure since Dec 31 1997 rebalancing shows superior results, but what about other time periods??
Also, correlations amongst investment alternatives are typically short lived. Assets will correlate (or not) until they just do not (or begin to).
Correlations can be developed that persist, literally forever, and probabilities calculated to as many decimal places desired for coin flips or die tosses.
However, once you toss into the fray the irrational, unpredictable, emotional, and yes, panicked behaviour of people, all bets are off.
Benoit Mandelbrot has shown conclusively that financial (and commodity) markets are non-linear, non-deterministic , non-predictable and chaotic.
Despite all the fancy math and "laws" used by financial gurus it all depends, literally, upon the future unfolding as in the past.
History has shown this is simply a false assumption.
Posted by: JA at February 23, 2012 9:20 AM
Who cares about Peter Lynch in this context? Rebalancing controls risk, tends to increase returns, is consistent with any number of strategic investment policies and can be implemented systematically.
Posted by: Capital Markets guy at February 23, 2012 6:42 AM
True, although Warren Buffett has even more fame and a longer record of action (and success). Let's just all invest like Warren Buffett, right? Hmmm, maybe some statistics and strategic-minded/forecast-free history lessons are appropriate after all for those of us with a shade less talent than Buffett and Lynch.
Posted by: JP at February 23, 2012 6:40 AM
Peter Lynch is much more famous than Markowitz.
Formulas and statistics are fine, but actually doing it is another thing.
Posted by: Jack Reacher at February 22, 2012 3:52 PM