In the 1990s, fund manager James O'Shaughnessy caused a stir with his bestseller, What Works On Wall Street (McGraw-Hill, 1998). The book's conclusion: The best strategy for superior returns was to invest in small-cap value stocks. Among other things, O'Shaughnessy found that the most consistent metric to define value was with a low price-to-sales ratio. That's not exactly what the investing minds of the time wanted to hear. But O'Shaughnessy — a quant who cares less about what a company does than how its equity scores — says he had the proof in over 70 years of back-testing analysis using the CompuStat and, more recently, FactSet databases. The advice in What Works seemed obvious to many, but the timing was poor. The book was released in an era when QQQQ was the name of the game. O'Shaughnessy's mutual funds, which were based on his research, languished. He sold the funds and closed his business. (But if investors had followed his advice in the 1990s they would have avoided much of the bloodletting of 2000 to 2002, he says.) Now, O'Shaughnessy is a managing director of Bear Stearns, where he runs $9 billion in Systematic Managed Accounts offerings (he still runs a value and a growth fund under his own name for the Royal Bank of Canada).

In March, O'Shaughnessy released a new book, Predicting the Markets of Tomorrow, published by Portfolio, an imprint of the Penguin Group USA. He stands by his earlier research, but he tells investors to ignore 70-year average returns; the market tends to move in 20-year waves. Small stocks stay in favor for about 20 years and then big stocks take over. The payoff: The book argues that the worst 20-year period for small stocks ended in 1999. That means small stocks should lead for at least another decade (but, he cautions, your clients should have some exposure to growth strategies). Put another way, if you think you can protect clients by putting them in “safe” large-cap stocks or in, say, an S&P 500 index fund, you'll be disappointed. Jim O'Shaughnessy recently chatted with Registered Rep.

Registered Rep.: So, growth is dead. Is that the point of your book in one sentence?

Jim O'Shaughnessy: I am not saying investors shouldn't have exposure to growth strategies. I think the growth strategy that will work best over the coming years is a “growth at a reasonable price”-type model. I think the traditional “we-don't-care-what-the-price-is-but-it's-got-a-sexy-story” growth style that was popular in the late 1990s is going to suffer. I don't think that style is going to have a return or resurgence any time soon.

RR: Tell me about your methodology style.

JO: Sure. We used the Ibbotson data that starts in 1926. And so our first 20-year cycle comes out in 1947. So we looked at an investment made as of 1926, and then began the 20-year rolling compound [returns] 20 years later.

RR: And why a 20-year rolling period?

JO: Well, we chose 20 years for two reasons. Number one, when you look at the average age that people actually seriously start to invest for their retirement, it's about 45. And so I said, if people are going to start at 45 and still maintain the idea that they might really want to retire when they're 65, that's a 20-year time horizon. Also, it happens to be the time period that ERISA requires those who are governed by its regulations to make forecasts for pension funds. At the 20-year mark, basically all the market noise disappears and it's all truly signal. Take a look at a one-year stock chart. It looks like a constant heart attack.

RR: Looking at the Ibbotson charts you can see clear trends over long periods?

JO: In the short term, returns are all over the map. And then you try to say, “Let's look at three-month intervals.” Quarters. That's how we actually make investment decisions on very short periods of time. And when you look at short-term returns on a graph, the graph is all over the place. You see that it's just all noise. And at one year, it isn't much better on the noise factor. Once you start saying, “Well, let's take a longer frame, let's look at five years,” then you start to see a real clear picture. When you get to 20, it is the ideal picture for seeing long-term trends in the marketplace.

RR: Let's say you're right, over that period the market moved in 20-year periods. But you know the unprecedented happens all the time. So, people saved and moved money around in 20-year increments. But how can you be so sure going forward?

JO: The argument I make here is that in financial markets, everything reverts to the mean. And it's not just one type of stock or another type of stock. It's all styles, all stocks, everything. When you look at the average 20-year real rate of return for any of the indexes you're interested in, when you look one standard deviation above and one standard deviation below — unusual times — what you see is, at least in all of the data that I have access to, 20 years later, over the next 20 years, the index, whichever one you want to talk about, spends the next 20 years reverting.

RR: Reverting either way, right?

JO: Yes, over long periods of time, the real rate of return to stocks is about 7 percent. Other researchers, like Jeremy Siegel of Wharton, found similar return patterns. So if you're well below your long-term average, you spend the next 20 going up. If you're well above your 20, you spend the next 20 going down. Today, we're coming off a market environment where the S&P 500 had its highest 20-year real rate of return in history, the 20 years ending March 2000, with average returns of 13.85 percent.

On the other hand, we have large-cap value and small-cap stocks, in general, one standard deviation below their long-term 20-year mean returns. As recently as March 2003, small-cap stocks were still one standard deviation below their average long-term real rate of return. So essentially, this marries very nicely into considerations of our own valuation.

RR: Why? People go momentarily — for years, I mean — insane?

JO: Yes, over long periods of time valuation matters. But in the short run, what we've found is there are market environments where investors are not willing to pay a lot for every dollar of earnings they get or what they are willing to receive in dividend income. Those are periods were low P/Es and high dividends are popular.

RR: And we're still in a hangover.

JO: So essentially, the data are telling us to be overweight large-cap value on the large-cap side, small- and mid-cap names, too. It's been a great five-year rally for value. My contention is, yeah, there'll be bumps along the way, but you ain't seen nothing yet.

RR: And, if you're a financial advisor, what should you do? Of course, every client's situation is different, has different risk/return profiles.

JO: The one thing that I would say is — and I say it in the book — I once was the ultimate do-it-yourselfer, but I have come full circle on that. I think anyone who can work with an advisor or with a broker should, because they're going to lack the emotional objectivity that their advisor is going to be able to bring to the table.

RR: Our readers will love to show their clients that line.

JO: What I think the advisor should do is understand these underlying themes of reversion to the mean, demographic patterns and valuation patterns, and adjust their clients' portfolio allocations to take full advantage of it. You know, the stakes could not be higher. Everyone is forecasting the S&P 500, after inflation, to return the long-term average of 7 percent over the next 20 years. What if it were half that? If it were half that, all of those Monte Carlo simulations would be out the window. All of the plans that you've made for your client are going to be wrong.

So what you've got to do is take a look at your underlying assumption set, and then say, “What's my target for this client?” If the target is, “I need to maximize the total return to the portfolio,” then you're going to have to use a lot more small- and mid-cap names, and you're going to have to make certain that on the large-cap side, you're value, or on the growth side, you're in the growth at a reasonable price. You're clients should have almost no exposure to traditional growth. By traditional growth, I mean when you're paying more than 30 times forecasted earnings, that's a traditional growth stock.

RR: It's funny, I was talking to a professional value investor today who says he found a stock for 35 times earnings and he bought it anyway.

JO: We actually recommend looking at the price to sales rather than the price to earnings. And we say anything above 10 is in toxic territory. When you look at price to sales in terms of consistency and its persistence in working well, it works the best of all of the investments.

RR: Explain.

JO: So the more you're willing to pay for a dollar of sales, the worse your returns. That is very counterintuitive for many growth guys, and we say simply by adding that and getting yourself back into the growth-at-a-reasonable-price category, you'll be able to maximize that side of large growth. But then we also want you to be in large value.

Again, think about the underlying factors. What I recommend in the book is you take a hard look at high dividend-paying stocks. Why is that? Well traditionally, at least over the 54 years of data that I have, they provided the best risk-adjusted rates of return. But going forward, what types of stocks do you think those 78 million aging boomers are going to like? I think that they're going to like stocks where they're getting good dividends.