(FORTUNE Magazine) – Eliot Spitzer was at the apex of his campaign of scaring the daylights out of Wall Street securities analysts when he was invited to address a whole conference of them. Gazing out over the dinner crowd, he smiled brightly and said, "Good evening! I'm delighted to meet you all. It's nice to put faces with the e-mails."

There was a burst of laughter followed almost immediately by haunted silence. The line was funny--e-mails had embarrassed more than a few analysts--but also deadly serious. That much the group could sense. What they couldn't know was how much the man at the podium would alter their world.

For cynics who ask whether Spitzer has really changed anything on Wall Street, the answer three years later is clear. Has he ever. Life is different for the companies the analysts cover, for the firms that pay the analysts, for the investors who read their reports, and most of all for the analysts themselves, who now work harder for less money and worry about whether they can brush their teeth without a compliance officer's permission.

One other thing: On matters of substance, they almost never use e-mail.

As for Wall Street's major firms, they're lighter by some $15 billion in fines, penalties, and civil lawsuit settlements. Increased compliance costs are unreported but certainly total millions a year. Of course, Spitzer wasn't just out to upend the system. His declared objective was "to protect the small investor and restore integrity to the marketplace." So the big, bottom-line question on Spitzer is this: Has he done it? Is the small investor better off?

Spitzer revolutionized the financial-services industry in three major ways, with varying effects on individual investors. They barely noticed his crackdown on the insurance brokerage industry. Mutual funds are closer to investors' hip pockets, but the slimy practices he stopped, known as late trading and market timing, amounted to small change in a multitrillion-dollar business. His biggest scalp was that of Strong Capital Management founder Dick Strong, who had to pay $60 million personally and was barred from the industry for life. Spitzer did get fund firms to reduce fees by nearly $1 billion over five years, though even that amount isn't huge when spread across 54 million fund-owning households.

The individual investor's verdict on Spitzer thus comes down to the so-called global settlement of December 2002. It was negotiated on the enforcement side not just by Spitzer but also by the SEC, the NASD, the New York Stock Exchange, and other state regulators, but it was modeled closely on a deal Spitzer had cut with Merrill Lynch several months earlier, back when those other enforcers were apparently still snoozing.

The goal of the global settlement was stopping fraudulent research. The one fact everyone remembers from those days is that Merrill Lynch analyst Henry Blodget was recommending a stock (24/7 Media) in his public reports while privately calling it "a pos"("piece of shit") in an e-mail discovered by Spitzer's staff. There were plenty of other such e-mails from other analysts at other firms.

The reason for the duplicity, of course, was investment banking. Brokerages with I-banking operations collected mammoth fees by arranging financing for companies, and those companies liked giving their business to firms whose analysts were friendly. From the individual investor's perspective, has severing that connection been beneficial? Let's take a look.

You can't wipe out venality, but the global settlement has dramatically changed the substance of Wall Street research. At the ten major firms that settled, research must be separated from investment banking in every way--financially, operationally, even physically. Analysts' pay cannot be based on I-banking results, which is why $20-million-a-year analysts like Salomon Smith Barney's Jack Grubman are long gone.

Plenty of evidence says these rules have made a difference. Most tellingly, analysts are actually recommending that investors sell certain stocks, even stocks of client companies. Back in 2001 only 1% of all buy-hold-sell recommendations issued by the ten major firms were sells; 65% were buys. In retrospect, of course, we know that practically every stock in the market should have been a sell. Today, though prices are lower, 11% are sells and just 38% are buys. Merrill Lynch even had to give up some lucrative investment banking business--helping on Warner Music's IPO earlier this year--after its famous media analyst, Jessica Reif Cohen, insisted the projected offering price was way too high. (Events proved her right.) Such things were simply unimaginable pre-settlement.

So research is more honest. Is it any more reliable? It makes sense that it would be, but if what made sense really happened, then the Yankees would be World Series champs. Specifically, it seems intuitively good that analysts are issuing more sells and fewer buys--yet are they the right sells and buys?

StarMine, which tracks analysts' calls, figures that in aggregate analysts slightly underperformed the market in 2002 (pre-settlement), then beat it moderately in 2003 and by a slimmer margin in 2004. That's encouraging, but evidence going back decades shows that analysts have been consistently terrible, even in the days before investment banking became the driving force. So we'll have to wait a few more years before concluding that the settlement has had a lasting impact.

In any case, we still haven't answered the most important question: Considering everything, are individual investors better off? Here's why they might be. The I-banking conflict was a big, fat obstacle to efficient functioning of the markets. It's what economists call an externality. Remove it, and you get more analysts trying to determine stocks' real value rather than trying to drum up banking business. Even if collectively they can't beat the market, they are jointly creating a market that more properly reflects stock values, and that's good for investors.

But here's why investors ultimately may not be better off. The big message of the settlement is that it's safe to go back into the markets, that research is trustworthy again, that a deeply conflicted Jack Grubman won't tell you to buy WorldCom at the top. And it all may be true. But it reinforces a view that has been about as thoroughly discredited as it could be: that ordinary investors should be picking individual stocks at all. Some 50 years of research shows that you can't beat the market that way, except by sheer luck. If people plunge more lustily into stock picking as a result of the settlement, most of them will regret it.

So the final verdict on Spitzer's Wall Street legacy is a paradox. He eradicated a practice that deserved to die and that we're certainly better off without. Yet it's tough to say exactly how--or even whether--individual investors have benefited. Princeton professor Burton Malkiel, author of the classic A Random Walk Down Wall Street, sums it up: "It's good that analysts are no longer doing marketing for the Wall Street firms. But even today, if you're going to get into and out of stocks based on analysts' recommendations, it can still be very hazardous to your wealth."