Owning multiple asset classes is the foundation of sound investment strategy. As commonly analyzed and practiced, however, multi-asset class diversification may offer fewer risk management benefits than meet the eye.
Consider a simple portfolio comprised of domestic stocks and bonds, with the Russell 3000 and Lehman U.S. Aggregate Bond indices as proxies. We'll use a 60% equity/40% bond mix--a common asset allocation benchmark for institutional portfolios. If we define risk as volatility or standard deviation of total return, does this mix control risk relative to owning stocks alone?
The casual observer may be tempted to answer "yes." Moving a portion of stocks into bonds, after all, is widely touted as Risk Management 101. And for good reason since bonds generally exhibit low to slightly negative correlation with stocks. For the five years through this past August, for instance, Lehman Aggregate and Russell 3000 posted a -0.17 correlation, according to Morningstar Principia (1.0 is perfect positive correlation, 0.0 is no correlation and -1.0 is perfect negative correlation). The implication: Building a portfolio of the two spreads risk around, providing a smoother, more stable investment ride compared with owning just stocks.
On the surface, that certainly sounds accurate, as the numbers suggest. Russell 3000's annualized standard deviation was a brawny 14.3% for the 10 years through this past August--well above the 3.5% risk for Lehman Brothers Aggregate or the 8.7% risk for a 60/40 mix over that stretch.
End of story? Not quite. Looking at total portfolio volatility can be misleading, as it is in this case. Yes, shifting a 100% allocation in stocks to a 60/40 mix sharply lowers overall standard deviation, thanks to bonds. But when we look at the portfolio's components individually, the distribution of risk hasn't changed all that much relative to a stocks-only portfolio. Portfolio risk, for good or ill, is still almost fully reliant on stocks.
The lesson: Don't focus exclusively on the whole at the expense of the pieces in risk management. Each component contributes to overall risk, which leads to two critical questions: 1) How much of the total volatility does each component contribute? And 2) How do those individual contributions change, if at all, after shifting the asset mix?
These are basic questions, yet they're too often overlooked. That's partly because the popular analysis of asset allocation favors the dollar mix rather than the risk mix. Both are important, but focusing solely on the dollar mix can mislead us.
Consider again our 60/40 portfolio. The na?ve implication: Stocks contribute 60% of the total portfolio risk. In reality, that estimate is way off base. Virtually all of the volatility in the 60/40 portfolio comes from equities and only 1% from bonds, based on analysis of five-year standard deviations and correlations through the end of August 2008 (again per Morningstar Principia). The lesson: The dollar mix can be deceptive in terms of assessing the sources of volatility.
A 60/40 portfolio has a total volatility of 5.8, based on the five years through this past August. That cuts total portfolio risk nearly in half compared to equities alone. Why, then, is 99% of the volatility in a 60/40 portfolio still coming from equities?
The answer is that controlling overall portfolio volatility is more nuanced than generally recognized. Three factors determine volatility overall along with how volatility is allocated within a portfolio:
1) the standard deviation of each investment component;
2) the correlation of total return among the components; and
3) the portfolio weights of components.
The interaction of these variables often creates risk allocations in portfolios that differ sharply from what capital allocations imply.
What's a strategist to do? For starters, don't assume that a capital allocation is the last word in diversification. In other words, analyze risk from several perspectives.
Risk allocation deserves at least as much attention as capital allocation, but in practice it's usually ignored, says Kevin Kneafsey, global co-head of the strategic solutions group at Barclays Global Investors in San Francisco. "People aren't even looking at it." Overlooked or not, monitoring risk allocation is crucial, he asserts. That doesn't mean that capital allocation is irrelevant. "You want to look at capital allocation," he says. "Investing is in part a capital budgeting problem, and so you've got to see how you spend your money. But you also have to see how you spend your risk."
Ignoring risk allocation can nurture the illusion that a portfolio is diversified when it's not, warns Geoff Considine of Quantex Inc., a Boulder, Colo. portfolio analytics software firm. "The whole game of diversification is low correlation between asset classes--one zigs, the other zags," he reminds. "But if one of them has very low volatility, then if it goes in the other direction it doesn't really offset losses [elsewhere in the portfolio]."
Low correlation alone, in sum, doesn't always suffice. Take another look at our 60/40 portfolio. Yes, bonds posted a slightly negative correlation with stocks for the five years through August. Unfortunately, the benefit of bonds' low correlation with stocks is muted by fixed-income's low volatility: Lehman Aggregate Bond's standard deviation is roughly one-third of the Russell 3000's volatility. If the volatility of bonds was higher, the low correlation would be more potent as a diversification tool.
The strategic challenge is finding asset classes that exhibit low/negative correlation and sufficiently high volatility to make diversification worthwhile. The message may be clearer if we think of trying to diversify an all-equity portfolio with cash. On the surface, it looks like a winning idea since cash routinely posts a roughly zero correlation against stocks. By that standard, cash looks like a strong diversifier. But that's a hasty decision once we look closer and recognize that cash has almost no volatility. And so even a cash/stock portfolio weighted 90% in 3-month T-bills means that 85% of total volatility is still coming from equities, despite the fact that stocks still represent a mere 10% of the dollar-weighted asset allocation.
Even moving into a broader array of asset classes has its challenges in the pursuit of diversification. As an example, consider the 60/40 bond/stock portfolio once more and what happens after expanding its horizons with foreign stocks and bonds, REITs and commodities in this dollar-weighted mix:
U.S. stocks (Russell 3000): 25%
Foreign stocks (MSCI EAFE): 25%
U.S. Bonds (Lehman Bros. Aggregate): 15%
Foreign Bonds (Citigroup Non-$ World Gov't Bond Index, unhedged in $ terms): 15%
REITs (DJ Wilshire REIT): 10%
Commodities (DJ-AIG Commodity): 10%
The above profile boils down to a global capital allocation of 50% equity/30% bonds, with REITs and commodities at 10% each. It looks diversified, but nearly 70% of overall portfolio volatility still comes from equities, based on five-year correlations and volatility through August (see graph above).
That's progress in terms of diversifying the source of volatility relative to the 60/40 mix, which draws nearly all the portfolio's volatility from equities. Yet the broader mix above reminds us that lowering the reliance on equities for risk exposure may require building portfolios that look radically different from the usual asset allocation fare.
That explains why institutional investors have been moving into alternative assets like commodities, hedge funds, venture capital, private equity, etc. The basic allure is the prospect of finding low correlation married to high volatility.
A similar effect can be engineered with leverage, which can turn low-correlation/low-volatility assets like bonds into low-correlation/high-volatility assets. Ray Dalio, founder of Bridgewater Associates, a large institutional money manager, made this point three years ago in a widely read essay titled "Engineering Targeted Returns and Risks."
The article says that most investors face the dual challenge of diversifying equity risk without materially lowering prospective return. There are two basic strategies, Dalio advises: Diversify into asset classes with low/negative correlations and high volatility, or lever up assets with low correlation and low volatility. The problem with the former is that good choices are hard to find, which inspires Dalio to consider leverage as compensation for the shortcomings.
Nonetheless, it's hard to imagine leverage finding favor with individuals or their advisors. Meanwhile, Dalio and others argue that the best-managed portfolios, regardless of strategy, are looking at prospective returns and volatility at roughly one-to-one ratios. "If you look back on institutional research, everyone sort of agrees that the best you can plan on is for a 10% return for a 10% standard deviation on a forward-going basis for a fully diversified portfolio," says Considine. "That's sort of the benchmark." He notes that other celebrated strategists--including Yale's David Swensen and Pimco's Mohammad El-Erian--have made similar observations about the one-to-one ratio for risk and return for prudently managed portfolios.
The key to getting close to the ratio--if not besting it in the long run--surely demands a deft hand with asset allocation. And that starts with intelligent risk management, including keeping a close eye on risk budgeting. Managing outcomes is never easy, but it can get a whole lot tougher if you don't know where the risk is buried.
James Picerno (email@example.com) is senior writer at Wealth Manager.