Portfolio Diversification - Is Lunch Still Free?
Portfolio diversification is often referred to as the only free lunch in finance. With brutal declines across almost all asset classes in 2008 and into 2009, some of the benefits of diversification were clearly undermined. Instead of enjoying a meal on the house, many investors are feeling more like a miserable third-grader whose lunch money was pinched from the locker room.
What went wrong? Didn't Investing 101 promise that if portfolios were constructed properly, certain asset classes should "zig" when others "zag"? Evidently - with the notable exception of gold bullion and government bonds - that was not the case. In the US exchangetraded fund space, not one long-only equity ETF ended 2008 in positive territory.
First, it will be helpful to review the conventional claims of portfolio diversification. From there, we can test those claims and determine if lunch is still free after all. The case is straightforward. If diversified investments with suitably low correlations are used to construct a portfolio, poor performance of one investment can be ameliorated by other investments in the portfolio that perform better. Therefore, fluctuations from single investments are minimized ... and portfolio risk should be reduced. The other key reason for diversification is that it expands the opportunities for returns beyond those that are available from domestic investments. Gains can be amplified by taking advantage of international and, in some cases, alternative-type investments.
It's an elegant concept ... but, as is often the case with theory, does not always function as it should in the real world. As such, a revised and more complete approach to portfolio diversification and risk management is required. Let's examine specific aspects of these concepts, and determine some of the lessons investors should be learning.
Observation #1: Diversification will often fail to yield the promised benefits just when they are needed most. Research has repeatedly shown that cross-asset correlations increase during financial crises (as compared to their historical averages). This is most evident in equity markets. The figure below confirms this conclusion in examining the most recent broad market decline. The chart separates correlation figures between the bull market period (March 2003 - October 2007) and the bear market period to date (November 2007 - January 2009). Most notably in stock markets, a broad index of emerging markets showed a near perfect correlation to the S&P 500 during the last period of declining markets. More encouraging, correlation figures actually improved in some asset classes. Note the drastic correlation reductions in international fixed income markets and gold bullion (an investment class with historically low correlations to broad equity markets).
Lesson Learned #1: Active asset mix is crucial. If diversification breaks down when it counts most, what strategies should be employed to limit drawdowns? Our view is that a static asset allocation is simply not a prudent investment strategy. Rather, actively shifting asset mix is a crucial component and can limit downside by shifting away from overvalued assets toward those that present better value. Make no mistake. We are not advocating for a hyperactive trading strategy ... one that dodges in and out of asset classes to chase returns. Rather, material shifts should be implemented when risk levels or valuations reach extremes. And, a strong risk management framework should drive asset allocation decisions. That may mean reducing overall equities when risk premiums are hitting record lows (such as in the period leading up to 2007). Or, it may mean re-deploying capital to investment classes that have exhibited defensive qualities when probabilities of heightened market volatility are high.
Many dismiss dynamic or tactical asset allocation as "market timing". Actually, it's quite the opposite. Critically, dynamic asset allocation focuses less on short-term events and more on intrinsic values and risk measures. As the ever-insightful Dr. John Hussman has commented, "our activity as investors is not to try to identify tops and bottoms - it is to constantly align our exposure to risk in proportion to the return that we can expect from that risk."
Contrast dynamic asset mix strategies with the typical emphasis on short-term forecasting in the investment industry and the picture becomes even clearer. A recent report from Société Générale studied the accuracy of analyst forecasts over the period 2001-2008. Not surprisingly, abysmal results were found. The average annual forecast error of predicted stock returns was 25% (in absolute terms, not relative!). Clearly, short-term forecasts are not adding value to the investment process. Rather than attempting to predict next quarter's earnings or the level of the Dow at the end of 2009, a more prudent investment approach should focus on determining fair values and positioning portfolios in areas that provide the best risk-return dynamics. That removes the bare forecasting element from the investing process.
Observation #2: While correlations may increase during inclement market environments, returns over longer periods vary significantly. Consider the case of developing markets versus US stock markets over the last decade. In terms of rising correlation figures, there were almost no diversification benefits over certain measured periods. But longer running return figures paint a different picture altogether. Average annualized returns for the iShares Emerging Markets ETF (NYSE: EEM) and the iShares S&P 500 (NYSE: IVV) over the last five years were 5.78% and -4.29%, respectively (as at January 31, 2009). Looking at the underlying indices (prior to the launch of these ETFs), the difference is even more apparent. The emerging markets index produced average returns of 8.46% while the S&P 500 returned just - 2.65% annually.
Lesson Learned #2: De-emphasize shorter term volatility and employ all available investment opportunities. Given that asset correlations may increase over some measured periods, many have presented that as evidence to invalidate the case for diversifying into international markets. To quote Mark Twain, "rumors of my death have been greatly exaggerated". Similarly, the benefits of global diversification are not dead and the case remains strongly intact.
Two points are noteworthy. The first is that investors should recognize that short-term price swings are the norm with risky assets ... even in balanced portfolios. Effectively, investors are purchasing varying levels of risk when constructing diversified portfolios. That means investors should be able to tolerate short-term volatility ... and ignore much of the market noise. A view to longer term return figures is critical. Many developing markets are poised for much higher economic growth rates and concomitant investment returns. Higher short-term price volatility does not justify eliminating these exposures from portfolios.
Secondly, excluding exposures to certain investment classes is an imprudent limitation. Rather, investors should take the generalist approach and utilize the full set of investment opportunities ... of course, providing those assets present attractive risk/reward dynamics. International investment opportunities and other special situations will always present themselves (emerging markets, inflation-linked bonds, precious metals, and so forth). Confining exposures to purely domestic investments is a sure way to increase volatility and lower returns. (Canadian investors take note: we are particularly prone to a strong homeland bias).
Observation #3: Traditional quantitative risk models deliver an incomplete measure of risk. Modern finance theory has distilled risk into a single number -- standard deviation (a simple measure of the dispersion of a data set). This figure has become industrystandard and is widely utilized in portfolio construction techniques.
Lesson Learned #3: Fundamentals are also a key measure of risk. Limiting risk measures to a strictly quantitative science is perilous. Risk is more than just volatility, as history has repeatedly shown. Structural factors and irrational human expectations also determine risk. Volatility, after all, is only determined after the fact. What's more, mathematical models can breed a false sense of confidence when constructing diversified portfolios. That does not mean they do not have their place in an investment process. They certainly do. But a wider analysis of risk is necessary when constructing diversified portfolios.
Fundamental risks are perhaps the most common missing element of the typical risk analysis approach. Purchasing assets with a "margin of safety" is at the heart of classic value investing -- the price we pay for an asset can limit losses. Most quantitative risk models ignore this as a key input ... in fact, normally, no fundamental data is used at all.
Looking back at the example from observation #2, emerging markets vastly outperformed the S&P 500. Why? A large part was due to valuation differences -- emerging markets were priced much cheaper than US markets. In 1999 (the beginning of the measured period), emerging markets were hitting very low valuations while the S&P 500 was at record high valuation levels. Starting investment valuations are the best predictor of longer term returns.
Where to from here? Turning our attention to today's investment climate, where to next? And can we expect certain virtues of portfolio diversification to return? It is instructive to reflect on the situation that financial markets found themselves in 2007. Almost all asset classes were in definitive bubble territory. Risk premiums across the board were hitting historic lows. In that environment, diversification was of little value in limiting losses. Rather, that was a time to reduce overall portfolio risk.
Today, we have a very different situation. Yes, the global economic news is bleak. But that is already discounted in many asset prices. In fact, some risks are now being well-compensated and longer term expected returns should be ratcheted higher. In client portfolios, we are now taking steps to move back up the risk curve ... gradually increasing exposures to select emerging market ETFs and other value-priced assets. Indeed, the full benefits of diversification are likely to shine again.