The bank has laid off or reassigned the dozen or so employees at its Hudson Street Services unit, which offered data and independent research to investors. Goldman also sold its minority stakes in most firms that were producing the research, generating an overall profit in the process.

Hudson Street's failure is the latest sign of how difficult it is for smaller research houses to thrive in a market where everyone from the big Wall Street banks to major mutual fund firms are seeking to cut costs.

It is also a sign that major investors may no longer be prepared to pay for a diversity of opinion about the markets.

It wasn't always that way. Independent research was all the rage on Wall Street after a $1.4 billion settlement in 2003 between Wall Street banks and regulators led by then-New York Attorney General Eliot Spitzer concerning allegedly tainted research.

"It was like a big gold rush when everybody wanted to be an independent research provider," said Sanford Bragg, Chief Executive of Integrity Research, which tracks the independent research industry. "But all the hoopla around independent research in 2003-2004 has died down."

The settlement followed accusations by Spitzer and other regulators that securities analysts at some major banks were glorified shills for companies' shares, instead of providing the objective advice they claimed to offer. Analysts often got bigger bonuses if their positive ratings, or help on sales calls, allowed a bank to win investment banking business.

As part of the agreement, a dozen banks had to spend $460 million to furnish clients with independent research for a five-year period that ended in 2009 for most banks.

As that requirement fell away, independent firms had less revenue to rely on. Spending on such research will be down 24 percent this year compared with 2008, according to Integrity Research.

Hudson Street was not created to comply with the settlement's independent-research requirement, but because of industry trends that were sparked by the settlement.

In the end, customers did not want to buy what Goldman was selling. Research firms said they gained just a handful of clients from the platform. In a statement for this article, Goldman said, "Hudson Street no longer exists for the simple reason that weak demand from our clients did not warrant continuing the effort."

Spitzer compared the independent research industry's current dilemma with that of media companies battling against free news articles on the web. "People want the content, they don't want to pay for it," he said in a recent interview with Reuters.

A NEW WORLD FOR RESEARCH

The settlement forced Wall Street banks to restructure their research departments. Banks could no longer pay analysts more for winning investment banking revenue, so analysts' pay dropped.

Many responded by leaving banks and setting up their own independent research firms. Investors, after learning more tawdry details about how banks' investment research was made, grew increasingly willing to work with independent analysts.

Goldman Sachs began looking for ways to profit from the new world of research, and Hudson Street was a central pillar of that strategy.

The unit was the brainchild of two then-Goldman executives: Thomas Conigliaro, who was hired from Merrill Lynch in 2003 to develop Goldman's independent research business, and J. Michael Sanders, a Goldman veteran who was a managing director and co-head of its U.S. institutional sales operation.

The two men decided the platform should sell research from independent analysts to Goldman's vast array of institutional clients, including mutual funds, hedge funds and pensions. For every sale Hudson Street made, it received a commission. Goldman also made minority investments in the firms whose research it was selling.

Conigliaro and Sanders spent a year working with Goldman's principal strategic investments group - the bank's internal private equity arm - and an outside consultant to figure out which research firms to distribute. They sifted through more than 300 firms, and ended up with 11 that met their standards. They named the business "Hudson Street" for a street near Goldman's headquarters in Lower Manhattan.

Some of the bigger names were TrimTabs, which keeps track of supply and demand dynamics for stocks and exchange-traded funds, and Medley Global Advisors, which speaks to central bank officials and government employees globally to do macroeconomic research. Asset4, which was later acquired by Thomson Reuters Corp, Investars and iSuppli also joined.

Once Goldman invested in the firms and set up commission agreements, its sales force went to work introducing Hudson Street to important institutional clients, first in the United States and then globally. The bank set up 500 client meetings related to Hudson Street in 2007 alone, a source said.

The head of one research firm said Goldman told him he could double or even quadruple revenue, because the bank knows nearly every major investor globally and has a legendary sales force.

That growth never materialized.

"The distribution arrangement was a disappointment, in the sense that it didn't lead to the jump in sales that we expected," said Paul Warme, the chief executive of Lusight Ltd, an emerging-markets research firm that had been part of Hudson Street until 2009.

A Goldman source echoed that sentiment.

"We realized - and maybe this is the fundamental flaw in the business model - that in order for clients to pay for a product, it couldn't be a 'nice to have,' it had to be a 'have to have' - the kind of product that clients can't live without," said the source.

STAKE SALES

Hudson Street's failure is a modest setback for Goldman as it tries to reinvent itself in a post-financial crisis world. Investors have some faith that whatever happens, Goldman will find a way to build a profitable business. Its shares are still trading at among the highest valuations for banks with big investment banking businesses.

Even though the business itself performed poorly, Goldman managed to make a profit on Hudson Street by selling off the stakes it bought.

Goldman would not comment on specific transactions, but sources familiar with the business said that revenue from just a couple of asset sales more than offset losses on other money-losing deals, as well as the money Goldman put into building the overall business.

One company that was part of Hudson Street, Epocrates Inc, provided the most details about its Goldman partnership through public securities filings. Its story highlights the kind of troubles that independent research firms can run into by partnering with big Wall Street banks.

Epocrates, which sells software and technology to physicians, sold Goldman 3 million shares for $40 million in 2007. As part of the Hudson Street pipeline, the company collected data from its vast network of doctors and distributed the information to institutional investors who analyzed medical-industry stocks.

When the company went public in February 2011, Goldman sold some shares into the IPO. The stock soared on its first day of trading, and Goldman's remaining stake was worth more than $60 million. Since then, the company's shares have plunged 62 percent, and Goldman has overall lost more than $15 million on its investment.

Epocrates has had other problems. Three weeks after its IPO, the U.S. Securities and Exchange Commission subpoenaed the company for information and documents related to Hudson Street as part of its broader look at expert networks, which had become one focus of an insider trading investigation. In January 2012, the SEC told Epocrates that it had closed its inquiry without recommending enforcement actions.

A former executive said that the inquiry cost the company a significant amount of time and money, and made some regret having entered the investor-data business because it never became a substantial source of revenue.

Many of Goldman's other Hudson Street stakes were sold at a loss, either back to the companies or to third parties that acquired the firms.

People at independent research firms said that the inherent flaw in Hudson Street was not the product, but the partnership with Goldman.

"If you're going to be independent, you want to really be independent," said Barbara Steiner, founding partner and head of institutional sales at the independent research firm Portales Partners. Portales was not distributed on Hudson Street.

"As an adjunct to a bank, or another very large broker, whether you say you're independent or not, you're not perceived to be."

(Reporting by Lauren Tara LaCapra, Editing by Dan Wilchins, Martin Howell and Leslie Gevirtz)

• A Permanent Economic Emergency