Recent attacks on modern portfolio theory have been overstated and unhelpful, said David Blanchett, Morningstar's head of retirement research, at the 2012 Morningstar Investment Conference. Blanchett suggested enhancing the standard framework for asset allocation that was initially introduced by economist Harry Markowitz, rather than scrapping it for a new approach.
Blanchett had five primary points.
1) Yes, the basic principles of diversification worked in 2008.
It is true that stocks globally struggled, across regions and styles, but many other assets performed well. Also, the simple act of spreading investments across many stocks via a mutual fund rather than owning a single security greatly reduced risk. A full 25% of U.S. stocks plunged 75% or more in value during 2008, whereas less than 0.1% of U.S. mutual funds lost that much money.
2) Consider human capital in addition to financial capital.
Investment decisions should not be considered in isolation. A young person can be expected to generate a large amount of wealth through work. This means that all things being equal, she can assume more risk with her financial assets than can an older investor. Also, the nature of one's occupation affects financial decisions. A teacher who has a stable, predictable income base has human capital that resembles a bond, and therefore should own more equities with financial assets. Conversely, an investment banker who has unstable and variable income should put more into bonds.
3) Market returns are not normally distributed.
The standard Markowitz model assumes that risk can be measured by standard deviation, which in turn assumes that asset returns conform to a normal, bell-curve distribution. But financial markets do not obey this assumption. For example, since 1926 the U.S. stock market should have dropped by 15% or more only in one month out of the roughly 1,000 observations, according to the predictions of normal distribution. However, such a decline has occurred on 10 different occasions. Thus, standard deviation is an insufficient tool for measuring the risk of financial assets--and for forming asset allocations.
4) Investors need better allocation tools.
Not only do financial assets lose more than is assumed by the current breed of allocation software, but they also lose more often. These two tendencies are called, respectively, negative kurtosis and negative skewness. Incorporating kurtosis and skewness into allocations can dramatically affect the results. Blanchett demonstrated two retirement situations, with an investor withdrawing an aggressive amount of income against a fluctuating asset base that was invested in the financial markets. The portfolio created by a risk measure that incorporated kurtosis and skewness had a sharply higher success rate in simulations than did the one that used only standard deviation.
5) Bonds are unlikely to outperform stocks during the next 10 years.
Many people have radically altered their asset mixes in recent years on the assumption that bonds will continue to outgain stocks, as they have done during the most recent decade. Some have argued that a leveraged portfolio of bonds is clearly superior to owning stocks. Blanchett demonstrated that this claim has been true at some times in history--for example, during the 1930s and early 1940s, and again during the Noughties--and very much incorrect at other times. Blanchett also stated that current bond yields imply a low return on bonds. Stocks might fail to meet that very modest return hurdle, but probably will not. At the least, it doesn't make sense to expect such behavior.
John Rekenthaler is the Vice President of Research for Morningstar.