(FORTUNE Magazine) – In the 1980s, colorful, cigar-chomping Lewis Ranieri made himself Wall Street's megastar of mortgage-backed securities. In the 1990s he sold a batch of fixed-income funds to investors on the proposition that mortgage-backeds would be a better deal than boring Treasuries. Those investors have since relearned the oldest of stories. There is no free lunch--not even when "Lewie" Ranieri is serving it.

Given that shareholders of investment companies seldom raise a ruckus about anything, the annual meeting of the Hyperion 1997 Term Trust in Manhattan this October was downright remarkable. In angry attendance was a collection of shareholders, many along in years, who were genuinely bewildered as to why they had lost money in what they'd believed a completely safe investment. Not in attendance was the creator of the fund and one of its directors, Lewis S. Ranieri, 49, a man who gained Wall Street fame in the 1980s by building mortgage-backed securities from a speck on the scene to a powerhouse business. Too bad. Had Ranieri come, he could have explained that his own terrible judgments about mortgage-backeds had done the shareholders in.

Until this sourness developed, the Ranieri myth, given technicolor life by the book Liar's Poker, had been doing nicely, thank you. Born into a Brooklyn family of chefs--in beam and beard, he actually looks like a chef--Ranieri left college at age 21 to work in Salomon Brothers' mailroom. From that 1968 start, he made a Horatio Alger climb into trading and then to head man in the firm's mortgage-backed-securities department. That business was nothing when he took it over in 1979. But the savings and loans were soon frantically unloading mortgages and Ranieri was there to catch them, next converting them into newfangled securities that he sold with missionary zeal and great talent. By the mid-1980s, he'd built his MBS department into an enormously profitable operation that was the envy of Salomon's competitors.

But in 1987, at the height of his success, Ranieri was suddenly fired by Salomon's then chairman, John Gutfreund. Ranieri later told some reporters that Gutfreund couldn't stomach Ranieri's celebrity. Gutfreund suggested to the press that the skills Ranieri displayed in building a business didn't extend to managing it. To this day, the full reasons for the firing remain obscure.

Although he initially slumped into dejection, Ranieri quickly rebounded. Forming a private partnership called Hyperion Partners--Hyperion was a Titan god--he raised more than $400 million from such bigtime investors as insurers Prudential and Equitable and Leslie Wexner, chairman of the Limited, with most of the funds going into real estate, money management, and financial services. On that money Ranieri made money, scoring particularly with $90 million that bought a reeling Houston S&L, United Savings Association of Texas, in 1988. Now prosperously revamped as Bank United, the company went public this year, and the world got a view of how well the Ranieri-led group has done. It got back its $90 million, plus a bit more, before the IPO; reaped $182 million on the offering; and still holds stock valued at about $520 million. Not bad.

In his post-Salomon years, however, Ranieri also gathered money from small investors, selling four "term trusts," all carrying the Hyperion name. Fanned by Ranieri's reputation, the four lifted an imposing $1.82 billion from investors.

These term trusts, as well as counterparts sponsored by other Wall Street firms, were sold in the early 1990s when interest rates had dropped and the whole world was scavenging for extra yield. The trusts were fixed-income, closed-end investment companies, which meant they had a set number of shares, traded like ordinary common stocks, and were obliged to pay out virtually all the income they earned. Typically, the trusts were offered to the public at $10 per share. "Term" meant the number of years the investment was to run--say, five or seven or ten.

Here was the heart of the selling proposition advanced by the creators of the funds: "We, your investment advisers, will conservatively but cleverly invest your money in mortgage-backed securities etc. and for the life of this trust attempt to deliver you a return that exceeds what you could get on a Treasury security. You'll get your dividends monthly. And at the end of the trust's life, we will hope to return at least $10 to you. No promises, you understand. But we think we can do it."

In that spiel, Wall Street had a story that it could both sell aggressively and amplify grandly. In an unknowable number of cases--but certainly some--retail brokers led their customers to think the $10 was not just an objective, but a solid gold commitment. Yes, the fund prospectuses said differently, but they were often not read or not understood. When the sales hubbub was over, all manner of new shareholders in the term trusts thought they'd happened upon an excellent free lunch: high returns coupled with no risk.

Fact is, though, the trusts were behind the eight ball even before the game started. The brokers selling the offerings took a slice of each $10--usually 50 cents to 60 cents, or even more. That meant an investor had no more than $9.50 per share working for him from day one. Furthermore, the investment advisers--Hyperion Capital Management, in Ranieri's case--were owed advisory and administrative fees that annually run to around 0.6% of a fund's net assets, so these too reduced share values.

Beyond all that, the funds had to reach their $10 mountaintop by a preset deadline while dealing with an up-and-down bond market and while investing "conservatively" in mortgage-backed securities, which can be tricky little devils. All in all, the term trusts had bitten off a large, tough-to-chew mouthful.

Many, in fact, have choked on that wad. Of the 22 mortgage-bond term trusts tracked by Lipper Analytical Services, only one--a fund called BlackRock Advantage that will terminate in 2005--had a net asset value in mid-November that exceeded its offering price, and in that case the margin was tiny.

Clumped miserably at the bottom of Lipper's list are Hyperion's funds, all four of them sold originally at the standard $10 per share. Hyperion's best performer, if you will pardon the expression, is a fund that terminates in 2005 and has a net asset value of $8.90. Hyperion's cellar dweller is its 1999 fund, at $7.19. All told, the four Hyperion funds have at this point lost a stunning $410 million out of the $1.82 billion originally invested.

But wait. Maybe the dividends paid to the shareholders helped soften the blow. Alas, no. Consider the record of the Hyperion trust that terminates soonest, in November 1997. Though its dividends were originally a very respectable 7% on the $10 offering price, the fund proceeded to earn so little interest and swallow so many capital losses that it was forced to drop the payout five times, to a current level of 4.76%. When the dividends are averaged, the fund turns out so far to have delivered annual income of 5.84% on the $10. That rate is almost precisely what an investor could have earned had he chosen, back in October 1992 when the fund was offered, to buy a five-year U.S. Treasury note instead. That security, of course, would have been risk-free, certain to return 100 cents on the dollar at maturity.

The 1997 Hyperion fund, in contrast, had a net asset value in mid-November of $7.32 per share, or around 73 cents on the dollar. This fund has also owned up to defeat, having announced, in colossal understatement, that it is "unlikely" to be paying out $10 per share to investors next November. For the fund even to strive for that amount, Hyperion said, would require its taking risks in its short remaining life that would be unacceptable. So the fund has retreated to owning mainly--you will grasp the irony--U.S. Treasuries.

Okay, how did this debacle happen? How did Ranieri, the founding father of mortgage-backed securities, sire these investment runts? The answer is, first, that he started out with an absolutely wrongheaded and risky investment strategy for his funds. And second, having dug himself into a hole, he made subsequent bets that only managed to deepen it. To sum up, the king of merchandising mortgage-backed securities has been a klutz at managing them.

To understand Ranieri's investment strategy, you need to realize that by their nature, fixed-income funds hold portfolios of things--like bonds--that go up when interest rates go down. But rate increases can kill a fixed-income fund, and that was a threat for the term trust managers, who had a $10 appointment to meet some years out. So on occasion the term trusts simply hedged, trying to insulate themselves against market swings of any kind by offsetting their long positions in bonds with short positions of some kind. That wasn't going to get them the superior returns they were after, though. So on many other occasions, they tried to anticipate interest-rate swings, tilting their portfolios in one direction or another in hopes of riding a trend. Tilts are risky, of course.

At the start Ranieri tilted like crazy, positioning his funds to benefit from a rise in interest rates. He did that by larding his portfolio with "interest-only collateralized mortgage obligations," better known as IOs. These highly exotic derivative securities, which Ranieri helped popularize in the 1980s, do the opposite of bonds, climbing in value when interest rates go up--and absolutely evanescing in value when rates go down. In a parallel bet, Ranieri hedged with futures contracts in 1993, thereby shorting the market.

And bang! Interest rates didn't go up in 1993; they went persistently down. During that year, a period in which pedestrian, unhedged bond funds did splendidly, the Hyperion 1997 trust, to take one example, collapsed to a net asset value that was more than 15% below the fund's $10 offering price. In short, Ranieri took a great bond market and turned it into a bad one.

Then, in the disastrous bond year of 1994--a time when betting heavily on rising interest rates would have been genius behavior--Ranieri got weak knees about his short positions and failed to capitalize on the trend. The final descent for the 1997 fund came in 1995 and 1996, when Ranieri tilted one way and then another, only to get whipsawed because he had it wrong each time.

Now, to a fundamental question: Is it fair to assign blame for the Hyperion funds' woes to Ranieri personally rather than to the investment managers working for him? FORTUNE couldn't explore this question directly with Ranieri because he wouldn't speak to us about anything having to do with the funds. He can't talk, his spokesmen say, because there's litigation out there: Three different lawsuits against Ranieri et al. are in the courts, each charging in some way that the funds' prospectuses misrepresented the risks. Ranieri's camp says that just isn't so--the prospectuses were in fact loaded with caveats--and in the case furthest along and now on appeal, a federal judge has agreed that these documents carefully explained the risks.

In any case, those same prospectuses help to clarify Ranieri's role in this mess. They say Ranieri was to have "general oversight" of investment strategies and approve any "significant changes" in strategic course. Those sailing orders fit Ranieri's forceful, opinionated, cigar-punctuated ways as well. Says a Wall Streeter who knows Ranieri's style: "I doubt that Lewie was hands-on in running the funds. But he probably had views as to which way to take exposure. And when Lewie has views, he can be very persuasive."

What's more, when the funds were being offered in the early 1990s, Ranieri was personally active in telling retail brokers what a neat thing they'd be for the customer. He even did some "hoot 'n' holler"--that's what brokerage firms call their in-house broadcasting systems--to whip up enthusiasm for the funds. Considering everything, we'll put this rat in Ranieri's pocket.

Since Ranieri was not hooting, much less showing up, at the October annual meeting of the 1997 trust, the job of facing its shareholders fell to Hyperion Capital Management's president, Kenneth Weiss, 45. He was nervous and defensive, except in dwelling on one piece of placating news: HCM's voluntary decision to take a 60% cut in the annual advisory fee owed it by the fund. (Weiss maintained, however, that the cut simply reflected the fund's strategic swerve into U.S. Treasuries, which don't need much management.)

Weiss was also forced by a questioner to deal with a subject guaranteed to infuriate rather than placate: legal expenses. Turns out that under the original agreement between the 1997 Trust and its organizer/promoter/adviser, HCM (a.k.a. Ranieri), the trust itself is required to pick up most litigation expenses. The exception would be legal costs arising from fraud or malfeasance. In other words, as long as Ranieri can prove that he just managed dumbly, not venally, he can duck the tab.

So far, the 1997 trust's legal duty has cost its shareholders nearly $600,000, and the meter's still running. Though no figures have been disclosed, two other Hyperion trusts have been stuck with bills as well. Some days there is no justice.