Some of their colleagues on Wall Street call it crazy, but some Ivy League economists believe they have found an odds-on way to make money in the bond market.

The professors say investors should buy long-term bonds when their yield rises to a big premium over the rates available on short-term instruments.In late May, for example, investors could have locked in a yield of 9.35 percent on 30-year bonds at a time when three-month Treasury bills were yielding just 6.24 percent.

It won't work every time, but usually buying in the long end of the market at times of wide interest-rate differentials will put an investor well ahead of someone who sticks to the short end of the market, the professors say.

"These are not risk-free strategies, but on average, you do better going to the higher-yielding instruments," N. Gregory Mankiw, a Harvard University economist, said in an interview this week.

Wall Street economists say it's not that simple. They say when long-term rates are much higher it's because people are expecting an increase in inflation, which will erode the price of the bonds and wash out the gains from the higher yields.

In short, Wall Streeters say the attractive yields on long-term bonds are only there to make up for the risk of inflation. Let the buyer beware, they say: The spread could get even wider.

But people like Harvard's Mankiw, Yale University's Robert Shiller and Princeton University's John Campbell say the conventional wisdom has it backwards.

Most of the time, they say, a big spread between long- and short-term rates is a signal that long-term rates are going to fall - not rise, as the Wall Streeters say.

When long-term rates fall, prices of bonds go up. That means the smart investor who bought bonds when the yield spread was big is doubly blessed by both high yield and capital appreciation.

From 1961 through 1984, the yield spread between the 30-year bond and the three-month Treasury bill averaged about 1.2 percentage points, Mankiw has calculated. Spreads tend to narrow if they get much bigger than that.

According to Mankiw's calculations, if the spread widens to as much as 3 percentage points, an investor has a two-thirds chance of making a profit by taking money out of short-term bills and putting it in long-term bonds.

A two-thirds probability isn't a guarantee, but it's better than most other bets available on the sunny side of the securities laws.

Even now, Mankiw says, the spread is wide enough - about 2.3 percentage points - to justify shifting to long-term bonds.

Why would such a straightforward strategy have escaped the notice of sharp-eyed economists on Wall Street? The Wall Streeters say the strategy is wrong.

Robert Brusca of Nikko Securities Co. International Inc. said history is not a reliable guide to the likely "normal" yield spread between long- and short-term rates.

Investors are demanding bigger spreads on long-term bonds these days because they are fearful of inflation, and that fear is likely to persist, he said.

"During the '70s, bond investors were burned and burned again (by inflation). The steep yield curve tells me that people have learned something," he said.

Lacy Hunt, chief economist of CM&M Group investment firm, said anyone who followed the academic theory would have missed "the greatest buying opportunity of this decade, and in fact the entire 20th century."

That was in late 1981, he said, when long-term rates briefly fell below short-term rates. The long bond's yield of 15 percent seemed relatively unattractive at the time, but it proved astronomical in light of the recession that followed.

Of course, the academics do not claim the idea will work every time. Also, Yale's Shiller said it's possible that the "normal" spread has risen above its historical average, making it harder to know when the differential is advantageous.

Still, it seems as if the yield-spread notion may be an idea whose time has come. Princeton's Campbell said Wells Fargo Investment Advisers, for one, has an asset allocation model based on it.

If too many people start exploiting the spread, it will snap shut and the opportunity will be lost, the academics say. For now, they say, there is money to be made. They have even dabbled around the edges themselves.

"It's always attractive to think that people on Wall Street are smart and have figured everything out, but that's not always true," Shiller said.