The last four years in the market have led some to question whether diversification failed investors in 2008, says Tom Idzorek, the chief investment officer of Morningstar's investment division.
But he took issue with that in his presentation, called "A Non-Normal World: Rethinking Diversification, Risk, Asset Allocation and Modern Portfolio Theory" at the IMCA 2012 New York Consultants Conference, held in New York at the Times Square Marriott Marquis on Monday and Tuesday.
He said that about 25% of U.S. listed stocks lost at least 75% in 2008. However, only four of more than 6,600 open-ended mutual funds tracked by Morningstar lost more than 75% that year. That alone is an advertising pitch for diversification, he said.
In the future, he continued, investors will have to take a new view of risk, a holistic approach that separates a client's risk preferences from his risk capacity. Advisors often do questionnaires asking clients their risk preferences, but that's not necessarily the characteristics they ought to be using, since capacity is different.
When clients are younger, he says, most of their wealth stems from human capital, their ability to earn and save and the skills they bring to the table. When they get older, that human capital diminishes in terms of present value of dollars. As they get closer to retirement, what they have converted into financial capital becomes more important, but they lose the inflation hedge of human capital.
The human capital they have when they are younger, he said, is itself is one undiversified investment, which requires much more diversification and risk be taken in their investment portfolios. And different investors have different risk associated with their human capital. A university professor with tenure, Idzorek said, has human capital on his hands that's more bond-like since it's likely he won't get fired, while an investment banker whose job is often on the line has something more like an equity.
Idzorek also went into an extended presentation explaining how normal distributions are coming more into question as an explanation of returns, and thus how standard deviation is becoming a more questionable way of determining risk and risk capacity. Expounding on the work of his Morningstar colleague Paul Kaplan, he says that the bell-shaped curve for measuring the risk of catastrophic events is flawed.
When you overlay a normally distributed bell curve over the S&P 500 annual distributions from 1926 to 2009, he said, they appear to follow a normal distribution. But it's the details that matter, and years like 2008, he says, "really highlight the flaws in assuming that returns follow the normal distribution"-even as the vast majority of tools used assume that there is a bell shaped curve. When Idzorek and colleagues looked at monthly returns and put them in histogram buckets, then zoomed in on the left-hand tail representing market cataclysms, the histograms suggesting turmoil stacked up are higher than a simple bell shaped curve allow.
"Our risk models are generally understating how often we should expect 2008 type events," he said. "They are occurring ten times more often than our standard tools would suggest."
A log-stable distribution, on the other hand, shows more accurately often how the bad events occur, he said.