(MONEY Magazine) – For three or four years now, poor performance and bad publicity have hung like storm clouds over Fidelity Investments, the nation's biggest fund manager. But I'm just back from a visit to Fidelity's Boston headquarters, and I can report that there's good news for the firm's 12 million customers and their $639 billion in fund investments: The clouds are lifting.

Just look at the turnaround in performance. So far this year, nine of Fidelity's 27 diversified U.S. stock funds, including $74.6 billion Magellan and $35.4 billion Contrafund, are beating Standard & Poor's 500-stock index. And all five of Fidelity's diversified international stock funds are outperforming their benchmark as well. That's in contrast to the past three miserable years. In 1995 and in 1996, only five of Fidelity's diversified U.S. stock funds beat the S&P; last year, not one of those 27 funds beat the market.

To give you a better idea of how dramatic the recent changes at Fidelity are, let's review the ailments the firm was suffering from:

Elephantiasis. When funds get huge in a hurry, their returns suffer. It's much harder to run a fund with $10 billion than one with $10 million. Contrafund, Growth & Income and Magellan all saw their returns dwindle as their assets ballooned after 1993. Amazingly enough, not only did Fidelity insist that size was not a problem, but in 1995 it even went back on its promise to close New Millennium to new investors when the fund passed $500 million in assets. Instead, Fidelity kept New Millennium open until it hit $1 billion in 1996; since then the fund, which used to clobber the market, has badly lagged.

Aggressiveness. Instead of avoiding risk, some Fidelity managers chased it. By Halloween 1995, Magellan's Jeffrey Vinik had sunk 43% of the fund's then $53 billion into technology stocks, essentially turning a diversified portfolio into the world's largest sector fund. Then, by year-end, Vinik dumped his tech stocks and put a third of the fund into bonds and cash. In 1996, Magellan trailed the market by an unheard-of 11 percentage points.

In late 1996, Emerging Markets manager Richard Hazlewood plunked 38% of his fund's assets into one country, Malaysia, whose stock market promptly fell 68% in 1997. The fund lost 40.8% last year, compared with a loss of just 2.4% at the average emerging markets fund; Hazlewood left Fidelity in January.

Disorientation. Ever since 1995, with giant growth companies like Coca-Cola, General Electric, Intel, Microsoft and Pfizer leading the market, Fidelity's traditional focus on reasonably priced mid-size stocks had turned its funds into market misfits. In early 1996, Fidelity had only 7.47% of its total equity holdings in the S&P 500's 10 biggest stocks--even though those companies accounted for 17.37% of the total value of the market index. While those giant stocks streaked ahead, Fidelity's smaller companies fell behind.

Brain drain. As performance declined, top managers and analysts fled Fidelity: Bob Beckwitt, who had run Asset Manager; Michael Gordon (Blue Chip Growth); Marc Kaufman (Select Electronics); Brian Posner (Equity Income II); Jeff Ubben (Value); and Vinik himself. The departures set off rounds of musical chairs as Fidelity shuffled managers to fill the gaps. That not only confused investors but slammed them with capital-gains taxes when the new managers sold off their predecessors' holdings. After Erin Sullivan replaced John Hurley in 1997, Emerging Growth paid out a monstrous $6 a share in taxable gains.

Schizophrenia. By 1995, Fidelity Retirement Growth owned Sungard Data Systems, Nintendo and Sierra On-Line among its top 10 holdings. Meanwhile, Blue Chip Growth had 39% of its assets in such racy tech stocks as 3Com, Informix and PeopleSoft. Those were hardly the big, safe stocks--or cautious diversification--that investors expected in funds with such stodgy names.

As you can see, Fidelity had plenty to fix. Over the past few years, Fidelity's headquarters, scattered across several old buildings in the narrow streets of downtown Boston, has had the feel of a city under siege. So I was both surprised and impressed to find that its executives now frankly admit their stumbles. This new candor is not only refreshing, but it makes the talk of a turnaround more credible. After all, you can fix problems only if you admit they exist.

Listen to this comment from Bill Ebsworth, Fidelity's head of equity research: "We're really facing up to the problems of size. Last year we realized that we had gotten to be the biggest very fast, and we asked, 'Are there growing pains? Did we make mistakes?'" Ebsworth pauses, then adds: "I feel no shame in telling you our answer was yes."

LIVIN' LARGE

So what has Fidelity done? The firm's biggest change is huge news, not just for its investors but for the whole U.S. stock market: Fidelity is buying tens of billions of dollars of America's biggest growth stocks in a concerted effort to improve its performance.

Under Robert Pozen, who became the head of all its funds early last year, Fidelity has exhaustively studied why it had been faring so poorly. Teams of eight to 10 employees analyze every fund's portfolio, preparing quarterly fund reviews (QFRs). These reports, which can run up to a dozen pages, compare the performance of the stocks each fund owned with those it didn't. They also analyze how the fund would have fared if it had owned the same stocks in different proportions, or traded faster or more slowly, or made bigger or smaller bets on a given industry, or kept higher or lower cash reserves.

The biggest revelation: Performance was helped by managers' "positive bets"--the stocks they owned. But it was hurt by their "negative bets"--the stocks they didn't own or owned little of. Put simply, Fidelity's funds were suffering from a shortage of superstar growth stocks like Microsoft and Merck. "What we realized," says Ebsworth, "is that we were copping out if we kept saying, 'Fidelity doesn't do well in large-cap markets.' We couldn't just throw up our hands and say, 'We'll do better when the market broadens out.' No--Fidelity will do well now."

IT'S OKAY NOT TO OWN MICROSOFT, BUT...

Thus Fidelity, which built its reputation as a masterful investor in mid-size value stocks, is no longer sneering at big growth stocks. Steve Kaye, who runs the $46 billion Growth & Income fund, puts it this way: "I'm not going to buy a stock just because it's in the S&P, but I do try to make some adjustments." In mid-1997, Microsoft was only Kaye's 21st largest holding--even though it was the fourth biggest stock in the S&P; now it's his fifth biggest position.

I asked Beth Terrana, manager of the $8 billion Fidelity Fund, if she has been tanking up on stocks like Microsoft and Merck over the past year because she likes them--or because she's afraid that her returns would suffer without them. "Both," she shot back. "I grew up in a tradition where cheap is always good and expensive is unthinkable. But there are only so many times the market has to beat me over the head with a frying pan before I get the point: These are great companies, and the market will continue to value them highly."

"It's okay for a manager to own no Microsoft," says Bob Pozen, "but I don't want someone doing that inadvertently or unconsciously. It should be a conscious choice." Translation: If a manager doesn't own a big stock that's going up, he or she had better have a darn good reason. Each morning, in fact, Pozen gets a report on how Fidelity's overall ownership of every stock compares with its weighting in the market as a whole. "We're not here to yell at people or give them a hard time," says Pozen, "but we want to make people much more aware of the impact of what they don't own."

Pozen and the half-dozen analysts and fund managers I spoke with all claim that the QFRs were not imposed from the top down but were developed mutually by the investment staff and administrators like Pozen. "This is a group of people who want to win more than anything," says senior vice president Rick Spillane. "If they're not winning, they want to know why."

And they take action: From March 1996 to March of this year (the latest data available as Money went to press), Fidelity's stake in Microsoft shot from 0.22% of its total stockholdings to 1.51%; its position in General Electric jumped from 1.19% to 2.43%; its holdings in Merck swelled from 0.37% to 1.32%. As of March, Fidelity had 10.3% of its total equity assets in the 10 largest stocks in the S&P 500 (Coca-Cola, GE, Intel and the rest). That's up from 7.5% two years earlier. And at last count, Fidelity had 33.9% of its total stock assets in the S&P's 50 biggest stocks, up from 27% in mid-1996 (see the graph on page 80).

If you invest in individual stocks, pay attention: Fidelity's burgeoning bet on big growth companies suggests that these stocks will keep outperforming smaller companies, at least until the market turns bearish as a whole.

LET'S DO LUNCH

Now let's look at the second big change: Fidelity's research buildup. Since the end of 1995, it's added 26 stock analysts, bringing the worldwide total to 181. That's enabled Fidelity to raise the average time that each analyst is assigned to cover an industry from less than 12 months ("That was embarrassing," winces Bill Ebsworth) to two full years, allowing them to deepen their base of knowledge. And it's helped Fidelity's 52 senior analysts reduce the number of stocks they cover from an average of 68 to just 35. As Fidelity's top utilities analyst Nick Thakore says, "Performing well on the big companies is more important now that we're bigger." So the senior analysts are applying their experience to big stocks, where it's harder to get an edge; their younger colleagues cut their teeth on smaller stocks.

Fidelity is also restructuring its foreign research. Instead of covering stocks by geographic region, its analysts in London, Tokyo and Hong Kong will cover one of the same seven industries as do the U.S. analysts and traders. (Last year, Fidelity created a family of sector funds for those seven groups--consumer products, cyclical industries, finance, health care, natural resources, technology and utilities--and tied part of the analysts' bonus pay to the returns of these portfolios.) As international funds leader Rick Mace puts it: "It's almost impossible to look at U.S. companies without analyzing their foreign competitors. If you think you can analyze Ford without studying Honda, you're nuts." That kind of information now flows much more freely.

In fact, Fidelity has made better communication between analysts and portfolio managers a top priority. Over the past few years, the two camps had become less collaborative. Says Beth Terrana of the Fidelity Fund: "As we grew up, one of the things we stopped doing was talking to each other." Now almost every day there's a lunch at Fidelity's headquarters, open to all and devoted to a different theme: mid-size companies, utilities, high-yield stocks. This informal idea sharing has boosted morale--and probably returns too. "We've raised lunch to a high art," says Pozen.

Finally, when Fidelity assigns a new manager to a fund, it mandates a transition period in which the new skipper serves under the old. This more gradual handoff should reduce the tax shocks that sudden manager shifts have caused in the past.

RENDEZVOUS WITH REALITY

Along with rethinking its portfolios and bolstering its research, Fidelity has made a series of smaller fixes.

Two years ago, its stock funds kept an average of 7.3% of their assets in cash; today that's down to 4.7%. By keeping less money idle, Fidelity's funds can more easily match, if not beat, the stock market. (Of course, they've raised their risk of losing money if stocks crash.)

Fidelity has also traded in its old way of trading stocks. Each fund used to have its own trader who might buy and sell hundreds of different stocks over the course of a year. Now Fidelity's 34 traders are each assigned to one of the firm's seven industry sectors. That way, each trader always deals in the same stocks, tapping into a steady stream of information that he can feed back to analysts and portfolio managers. And Fidelity is cutting costs by trading less. In 1996, its biggest funds turned over their typical stock after only 12 months; today they're hanging in for 18 months. Bob Pozen guesses that better trading could be adding as much as half a percentage point to Fidelity's returns.

And in a long overdue rendezvous with reality, Fidelity is finally closing funds to new investors. Last year, Magellan (now with $75 billion in assets) shut its doors to nearly all but pre-existing 401(k) accounts; in 1998, Contrafund ($35 billion), Growth & Income ($45 billion) and Low-Priced Stock ($11 billion) followed suit. These jumbo funds should have closed long ago. "It's almost like success is your enemy," admits Pozen. That suggests to me that Fidelity may now be more willing to close funds before they get too big. Fidelity still has a long way to go in aligning its interests with those of its shareholders, as smaller fund companies like Longleaf Partners have done; closing more funds (such as Equity-Income and Dividend Growth) before they get bloated would be a step in the right direction.

THE BOTTOM LINE

Add up all these things, and it becomes clear that Fidelity's comeback is not just luck. The company deserves credit for recognizing its problems and for tackling them head on. Not only has performance begun to shine, but no big domestic stock manager has left the firm since early 1997, and funds such as Blue Chip Growth and Retirement Growth now invest the way their names suggest they should.

Let's not forget, however, that these changes, like any major transformation, come at a price. For one thing, at some point the stock market will again favor Fidelity's traditional strength: reasonably priced, mid-size companies. Could Fidelity see such a shift coming and reorient its funds in time? History shows it's all but impossible for fund managers to anticipate that kind of change.

What's more, Fidelity is homogenizing its funds. Just two years ago, only 59% of the performance of its diversified U.S. stock funds could be explained by the movements of the S&P 500; today, however, 75% of the return at those funds is driven by what the market is doing overall. At least to my eye, Fidelity's stock picking looks less adventuresome than it used to. In 1996, Dividend Growth had such oddities as Earthgrains Co., Stanley Works and Tupperware among its top holdings. Today that fund's top 10 includes stocks like Citicorp, GE and Microsoft.

All of this suggests to me that Fidelity is creeping closer to the kind of "closet indexing" or timid mimicry of the market that most big fund managers practice. Not surprisingly, Bob Pozen disagrees. "I reject that argument emphatically," he says. "Each manager makes his or her own decisions--and they disagree all the time about the same stocks."

The bottom line is that Fidelity has created greater predictability of performance and lower risk of poor returns. For the vast majority of its investors, especially those who turn to Fidelity for 401(k) funds, that's very good news. On the other hand, Fidelity's new strategy means that most of its big funds are unlikely ever again to beat the market by the huge margins they once did.

I don't think Fidelity will regain the glory days of 1993, when an amazing 19 of its 20 diversified U.S. stock funds beat the market, many by margins of more than two to one. At two-thirds of a trillion dollars, Fidelity is just too huge to load up on smaller stocks with bigger potential. And with the firm's recent stumbles still aching in their memories, most of its managers are far more cautious today. I think pigs will fly before Magellan manager Bob Stansky puts more than 5% of his assets in a single stock, as the legendary Peter Lynch did with Chrysler back in 1982. And only a few Fidelity funds (among them Capital Appreciation, Emerging Growth, Export and Trend) still have an open mandate to "swing for the fences," as Pozen puts it.

All in all, Fidelity strikes me as a smarter, and safer, fund manager than ever before. Surprises--good and bad--now look a lot less likely. Overall, that's a good thing.