About 12 years ago, I began writing a weekly online investment column for MSN Money. My colleague, Jim Jubak, focused on market trends and specific stocks. I couldn't do that. I was afraid. I suggested my column be called "The Cautious Investor." I would write about safe, namby-pamby, diversified investments and asset allocation for novices.

I considered myself a value investor at the time. But I soon got smug. The correspondents on the MSN Web site continually advised others: "Dump your advisor. Dump asset allocation. Dump value. Growth is where the action is." I listened to the people I was supposed to be teaching, sold value funds like Oakmark Small Cap and the Dodge & Cox Stock Fund and bought growth stocks like Qualcomm and Cisco and TransSwitch. Within a year, my $39,200 portfolio grew to $235,000. I wrote a column called "The Unmaking of a Value Investor," for Bloomberg Wealth Manager. Plenty of money managers agreed with me. Remember Robert Markman? He published a book in early 2000 called Hazardous to Your Wealth: Extraordinary Popular Delusions and the Madness of Mutual Fund Experts. Markman claimed asset allocation was dead. The only fund category worth buying was large-cap growth, he said.

Had there been a structural change in the market? Was it the end of the Cold War?

I hadn't heard about black swans. But some people had: Those who had read the book The Black Swan by Nassim Nicholas Taleb. The book examines the influence of highly improbable and unpredictable events that have massive impact, an important and timely topic since we're now facing yet another crisis in people's faith about asset allocation.

Taleb told Joe Nocera, who wrote a story on risk for The New York Times Magazine, that as a trader he made money only three times in his life-in the crash of 1987, in the dot-com bust and now. Not only did he make money-he made a killing. His fund was up 65% to 115%.

But Taleb also told Nocera that people cannot imagine a future different from their own past experience. That's why most of them thought of the 20% drop in the Dow Jones Industrial Average in October 1987 as "a worst-case scenario." But the prize goes to those who can imagine a future worse than that.

Like Taleb, Horace "Woody" Brock, believes that using historical returns to predict the future is folly. Brock is the founder and president of Strategic Economic Decisions (a company that provides research and macroeconomic analysis for high-net-worth individuals and institutions) and he has five academic degrees-a B.A., an MBA and an M.S. from Harvard and an M.A. and Ph.D. from Princeton.

I talked with Brock about modern portfolio theory and the future. It's not that the theory is broken, Brock says. "That's like saying Galileo's theory is broken." In other words, a feather and a lead ball still fall at the same speed in a bell jar with no air, Galileo found. And Harry Markowitz's portfolio theory works the same way. "MPT gave us a theory that works with no wind," Brock says.  

But in physics, in economics, in science, we have to go from an extremely simplistic explanation of a phenomenon under controlled circumstances in a laboratory to a more complicated one in the real world, often with many variables that are unknown. We start with Galileo, but then move through Isaac Newton's theory and then go on to Albert Einstein, who took it all away. Knowledge builds on past knowledge.

And so Markowitz's modern portfolio theory, the idea of creating a portfolio that balances risk and return and then locating the portfolio on the efficient frontier, is the best an advisor can do for a client. Markowitz told us how to optimally diversify based on your own risk tolerance, and that, Brock says, was a major accomplishment.

Yet Markowitz made two faulty assumptions. First, he ignored serial correlation. He assumed that the joint distribution of returns on stocks and bonds is just like flipping a coin. Every time you flip, the probability of heads and tails is the same. "It's like a coin flip in that the probability of returns tomorrow is the same every day," Brock says. The Markowitz theory assumes that the probability of stocks' and bonds' return tomorrow is the same every day. The solution is to keep rebalancing your portfolio to keep it 60% in stocks, 40% in bonds.

To show the folly of that assumption, Brock uses a weather analogy. For people who live in Tahiti, he says, the probability of rain or sun is the same every day. A coin flip. But what if you live in Vermont? The mean temperature for the year is 43 degrees. But the mean temperature in winter is 18; in summer, 73. Likewise, the mean average for the stock market price is 14 times earnings. But in 1981 it was 8. In 2000, it was 30.

The probability of the joint distribution of returns of stocks and bonds is not the same on a day in 1981 as a day in 2000. "The good news is that Paul Samuelson in 1969 said there is no optimal portfolio," Brock says. "It depends on the state of the world." Markowitz would be right if the conditions in the world were always the same, like the weather in Tahiti. But the return on stocks and bonds depends on many things, including whether it is a bear market or a bull market.

The second mistake is that all modern finance assumes stationarity, Brock says. "It assumes that the probability doesn't change." But the problem is that the process is random rather than deterministic. Stationarity assumes that the way you go from summer to winter and back again is always the same. "All modern finance assumes stationarity," Brock says. "But finance is all about structural changes and how we interpret them: What changes will an Obama administration make? What will China do? What about global warming? We don't live in a world that is stationary."

Instead, everyone is wrong, he says. Markets don't price things correctly. One person will not ever know the entire truth, he might just be less wrong. "If you want to make money, you have to be ahead of others in predicting structural changes," Brock says. "If your theory of connecting the dots is less wrong than mine, you make more money. No one knows how the stimulus package will work or how global warming will play out."

The true source of market volatility is mistakes, Brock says. Just as Galileo's physics was wonderful in a simplified world with no friction and no air, modern portfolio theory would be fine in a world with no mistakes. MPT did not assume that people knew the weather but that they knew the probability. Mistakes are what generate risks.
"The world is not knowable just by crunching the data," Brock says. "I am forced to have my theory. I connected the dots better than you did. The main point is simple: Portfolio theory created a fantasy world of perfect hedging.

The crucial mistake in portfolio theory that Samuelson pointed out was that you don't live in Tahiti. There is nothing to rebalance to."

MPT assumes that there are no structural changes. "Nothing changes so you just punch in the data and you optimize A robot can compute Markowitz. That's what consultants have been doing for 40 years, optimizing for the correct portfolio," he says.

But that's wrong, Brock says. Instead, you need to constantly change the portfolio to keep it in line with the structural changes going on in the world.

Efficient market theory assumed people were rational. That's wrong, as behavioral finance continues to show. Aristotle was wrong because he thought the sun went around the earth. That doesn't make him worthless. It just makes him dated. "Biases, mistakes. It doesn't matter why, but it only matters that we're wrong."

Brock has little time for what he considers small-minded economists. "It's the coin-flip problem," he says. "It's only when you consider that the future will not be like the past that you're getting somewhere."

Optimizers are shuffling around in a whole mix of historic returns. That's what I take away from Brock. Historical returns don't recur. Why should they? Advisors offer "expected return," such as "you can expect a 6% return from emerging markets over time and you should put 5% of your portfolio there. It sounds so precise and well documented. But those are really historical returns. The historical return becomes the expected return." Another truism that's being tested: If the time period grows longer, risk goes down. But now some economists say that the shorter your holding period, the less likely you are to encounter a black swan. As Brock says, in order to invest, you must have a notion about the future.

Mary Rowland can be reached at r[email protected]. She has been a business and personal finance journalist for 30 years and has written two books for financial advisors: Best Practices and In Search of the Perfect Model.