CHICAGO -- The Dow at 120,368 in 2025? It could be, if Roger G. Ibbotson's projections are correct.

Don't scoff. Twenty-five years ago he correctly projected the Dow at 10,000 in 1999. Investors didn't believe him then, either.

Mr. Ibbotson, the academic half of the team that did the seminal Ibbotson and Sinquefield study of investment returns between 1926 and 1974, projects large-capitalization stocks could return 11.6% compounded annually in the next 25 years.

That sort of growth would result in a Dow of 100,000 in early 2024.

Reaction from economists and money managers to Mr. Ibbotson's predictions ranged from "ridiculous" to potentially right on the money.

It might be wise to hold any bets against his projections. Mr. Ibbotson, a finance professor at Yale University who updates the historic study yearly with Ibbotson Associates Inc., Chicago, made similar projections in 1975 out to 2000, and he was remarkably close. Mr. Ibbotson is the founder and chairman of Ibbotson Associates.

In 1975, Mr. Ibbotson and his partner at the time, Rex Sinquefield, said the Dow Jones industrial average would hit 10,000 in November 1999 -- just eight months later than it did.

And, the pair came within 100 points of correctly predicting the Dow's year-end 1998 return of 9181.

Mr. Ibbotson's asset class return projections for 1999 to 2025 are lower than for the previous 26-year period. He said large-cap stocks are expected to return a compound annualized 11.6% and small-cap stocks, 10.7%. He predicts long-term government bonds will return an annualized 4.7%; 90-day Treasury bills, 4.5%. Inflation will be 3.1% on a compound annualized measure, he predicts. Each forecasted return is a median return.

Others have similar projections.

Jonathan Francis, senior vice president and head of global strategy at Putnam Investments, Boston, said Putnam is roughly in agreement with the Ibbotson forecast of an 11.6% average annualized return for large-cap stocks for the 25 years. Such a return could reasonably push the Dow firmly toward Mr. Ibbotson's 100,000-plus range by 2024, Mr. Francis said. He would, however, lower the return on 90-day T-bills to 3% and inflation to 2%.

Some dissent

Then there's Jeremy Siegel, finance professor at the Wharton School, University of Pennsylvania in Philadelphia. Said Mr. Siegel: "I don't want to bad-mouth Roger Ibbotson, because he's done some wonderful work, but I'm very familiar with his (earlier) predictions and he was completely wrong on many things, especially about inflation.

"His stock-return predictions came close, but because he was so wrong about inflation, he was nowhere near predicting what those dollars made in those stocks would be worth in 2000, the real return."

As for the future, Mr. Siegel would place large-cap stock returns in the 5% to 8% range over the next 25 years, closer to average historical returns.

Through 1998 at least, the Ibbotson and Sinquefield projections for asset class returns between 1974 and 2000 were close to the actual returns:

* Large-cap stocks were projected to return an annualized 13%; the actual return through the end of 1998 was 14.9%.

* Long-term government bonds were projected to return an annualized 8.4% between 1974 and 1998; the actual return through year-end 1998 was 10%.

* 90-day T-bills were expected to return an annualized 7.1%; the actual return as of Dec. 31 was 7%.

* Inflation was projected at 5.9%; it was 5.2% at year-end 1998.

For his forecasts through the first quarter of the new millennium, Mr. Ibbotson used the same methodology he and Mr. Sinquefield developed in the early 1970s.

Mr. Ibbotson described the method as taking the Treasury yield curve at year-end 1998, which has forward interest rates built into it, and adding historic risk premiums from 1925 to 1998.

"Back in 1975 and today, people thought we were bullish on the stock market. But what our research reflects is the historic pay-off for assuming investment risk. We're trying to be neutral and we assume that prices are fair and fluctuate randomly around this fair payoff for risk. We assume that the market is efficient," Mr. Ibbotson said in an interview.

Range of probability

Mr. Ibbotson reaffirmed the explanation he gave to Pensions & Investments in 1975 regarding the predictive uses of his forecasts: "This is not an attempt to arrive at a precise prediction of market behavior. Because we use historic rates of return, we know what the forecast error is and we know what the probability distribution curve is. We are providing a range of probability of returns."

In fact, Mr. Ibbotson said his critics in the 1970s who said his estimates of stock returns were too high ended up being badly wrong, as the actual returns were higher than he predicted.

"In fact, the market beat us, but nobody was closer," he said.