How to lose $9 trillion in a bull market

Jason Zweig is the author of the new book 'Your Money and Your Brain.' Email him at investor@moneymail.com.

It looked good on paper Investors plow more money into stocks at the top and sell near the bottom, so their returns are much lower than the market's. NASDAQ annual return, 1973-2002: 9.6% Return NASDAQ investors earned: 4.3% Source:Ilia Dichev, University of Michigan.

More from Money Magazine

The safest bets in bank stocks Is Wal-Mart recovery-proof? Making up your employer's match

Best Places to Live Current Issue Subscribe to Money

NEW YORK (Money Magazine) -- The day-to-day fluctuations of the stock market are hard to fathom, but fortunately they mean very little in the long run.

There is, however, a bigger market mystery well worth pondering: In 1982 the total value of the U.S. stock market, as measured by the Wilshire 5000 index, was $1.2 trillion.

The index has since returned an average annual rate of 13.3 percent - enough to turn that $1.2 trillion into $28.2 trillion. Yet the value of Wilshire stocks, as of Sept. 30, was $18.7 trillion, meaning investors earned far less than 13.3 percent a year.

Did $9.5 trillion disappear? And how can investors earn less than their investments?

An accounting professor at the University of Michigan named Ilia Dichev has cracked the case, and his findings, published recently in the prestigious American Economic Review, have huge implications for how you should invest.

I've written before about the gap between the returns reported by mutual funds and the money earned by their investors. Reported returns almost always look better than investor returns because people pile in after a fund gets hot and then sell or freeze after it's gone cold.

What Dichev's research shows is that the same thing holds true for the stock market as a whole.

By looking at how much money was flowing into publicly traded companies through initial and secondary stock sales and how much was flowing out via dividends, buybacks and buyouts, Dichev was able to measure the return on the typical invested dollar.

So what about that $9.5 trillion? "The money did not disappear," says Dichev. "It was never there in the first place." In other words, reported long-term returns aren't historical, they're hypothetical.

The U.S. stock market was never worth $28 trillion. That 13.3 percent "average" return was only for a strict buy-and-hold investor, a description that hardly applies to the big institutional players that move the stock market.

Consider that between 1973 and 2002, Nasdaq stocks gained an annual average of 9.6 percent. But that assumes money was invested on Jan. 2, 1973 and stayed put until Dec. 31, 2002 (with no taxes paid on the gains).

In reality, because investors pumped $1.1 trillion into Nasdaq stock offerings between 1998 and 2000 - just before the worst crash in modern history - the typical dollar invested in the Nasdaq earned only 4.3 percent a year, less than half the historical return.

So what can you learn from Dichev's research?

First, the more an investment jumps around in price, the more likely you are to underperform its average return - big swings will give you more opportunities to buy high and sell low.

Technology stocks in the 1990s, and energy and real estate stocks in this decade, earned much higher returns than the people who invested in them. The same thing will happen with today's hotties, like emerging markets funds.

You aren't mistaken to own them, but you're wrong to overdose on them in pursuit of high reported returns that very few people ever got to enjoy.

Second, for you to match the market's historical return, three things have to happen.

For starters, history has to repeat itself, with stocks continuing to produce high returns well into the future. (That, by the way, is no sure thing.)

Next you must invest in the entire stock market at rock-bottom cost, preferably through a total stock market index fund.

Then you must refrain from trading along the way.

If you take any other course of action, your results will differ from those of the market - most likely for the worse.

Based on decades of data from 19 countries, Dichev thinks that the average investor incurs a "timing penalty" of 1.5 percentage points a year by buying high and selling low. Impatience will cost you dearly.