While so-called "Chinese walls" were set up after regulators and legal cases shed light on the role analysts and bankers played in inflating the 1990s technology bubble, research teams still appear conflicted between their conviction and their bank's client list.

A mere 9 percent of analyst recommendations by investment banks and brokerage firms globally are a "sell" right now, based on 120,029 recommendations issued on nearly 17,000 companies, according to a Reuters study of StarMine data.

Ten years ago, sell orders jumped to nearly 20 percent after a series of rules were put into place to wipe out banker-analyst conflicts. It's now back in the single digits, and in some cases headed to levels last seen in the 1990s.

"Research is associated, rightly or wrongly, with an organization and if somebody puts out a sell recommendation people don't like that," said David Baran, co-founder of Tokyo-based hedge fund, Symphony Financial Partners.

An analyst with a negative view on a stock will often prefer to keep that belief tight, fearing pressure from top bankers seeking business from the company or being shut out by its senior executives.

"In general it's difficult for a lot of the analysts who could be negative or negatively biased on a company and expect to see them welcome at the next investor meeting or get access to the management," said Baran.

<^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Recommendations by region: link.reuters.com/sax27s

Before and after US reforms: r.reuters.com/pab37s

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^>

The survey of data from StarMine, a Thomson Reuters product, showed just 6 percent of the recommendations in the United States were "sell" as compared with 10 percent in Asia and 14 percent in Europe. Some analysts, when convinced of a stock's imminent drop, have been confident enough over the years to issue "strong sells". Around the world those calls now account for just 2.5 percent of the recommendations.

That 6 percent figure in the United States is nearing the 3 percent mark that stayed in place before and during the tech bubble, according to a study published in the Journal of Accounting and Economics, before post-crash rules were put in place to encourage more honest research.

DWINDLING COMMISSION

A major pressure on analysts to award a "buy" rating has always been from their bank's "sell side" -- the equity sales staff who can pitch the recommendation to a fund client and earn a commission on a resulting order to buy a block of shares through the bank.

That remains the case, even though the analysts' role in directing revenues to the trading desk appears to be shrinking.

Greenwich Associates produced a recent study saying that most of the growth in the Asian commission pool between the third quarter of 2010 and the same period in 2011 occurred in the fourth quarter of 2010, while trading conditions were challenging for much of 2011.

Globally, the profit margins made on stock trades is shrinking rapidly at banks, with J.P. Morgan disclosing that the bank made 1.5 cents per share on average in North American cash equities last year. From a fund's standpoint, it is cheaper now to trade electronically.

Commission rates on self-directed e-trades of Asian equities averages 5 basis points, compared with rates as high as 25 basis points through traditional methods, Greenwich says.

"To be a full service house, you need to provide a number of things that include sales, research, trading, investment banking and usually prop trading," said Jesse Lentchner, Asia Pacific CEO of broker-dealer BTIG.

"The problem is that the incremental increase in business you get from research doesn't pay for itself, and investment banking revenues are highly cyclical. It's a somewhat unstable and anachronistic business model."

A number of regulations enacted in 2002/03 in the United States to demolish the conflicts of interest between analysts and investment banks resulted in a spike in "sell" recommendations and gave a fillip to independent research firms.

For example, NYSE Rule 472 and NASD Rule 2711, adopted in July 2002 by the New York Stock Exchange and Nasdaq, sought to break communication between investment bankers and analysts and banned analyst compensation being tied to investment banking business.

A Global Settlement reached in April 2003 between U.S. regulators and 10 major investment banks banned analysts from going on "road shows" at which investment bankers give presentations on new securities issues to potential buyers.

Regulators had alleged that the banks allowed their bankers to pressure equity research analysts in ways that could cause them to issue misleading research and made the banks to pay $1.4 billion in penalties. About $450 million of that was to support independent research.

But many analysts say the culture of bank research never changed.

"There's just no incentive for me to issue a 'sell' call," said the industry head of a research team at an investment bank, who did not want to be named. "I'd say only 5 percent of what clients pay us now is for my recommendations. What they're paying for is for me to sit down with them and explain an entire industry to them in fifteen minutes."

CONFLICTS OF INTEREST

After years of building an army of research analysts to cover companies, especially in Asia's fast-growing economy and corporate sector, banks are expected to slash the size of sell side desks as the industry pulls back on spending.

Aside from costs, financial industry sources inside and outside the banks say that the quality of research remains a lingering concern throughout the sector.

"It should be about writing independent research and a bank comes with a huge conflict of interest," said Gillem Tulloch, founder of Hong Kong-based independent research firm Forensic Asia Ltd and a former telecommunications analyst at Japanese brokerage Nomura.

"Essentially what the research side has become is simply the marketing arm for the companies that they follow," Tulloch said, adding that the main reason fund managers make time for sell side research is to hear what the corporate executives are telling them or connect them directly to the companies.

Tulloch left Nomura in 2002 after he issued a negative report on a South Korean company, which prompted an apology from Nomura at the time, according to media reports.

It is not hard to find examples of analysts or research firms who have taken a hit after publishing unfavorable reports against companies.

Oversea-Chinese Banking Corp (OCBC.SI) barred a Morgan Stanley analyst, Matthew Wilson, from meetings involving its executives as it was unhappy with the analyst's coverage, according to a report in the Business Times in 2007.

At CLSA, analysts gave Samsung Electronics a low corporate-governance rating of 44.1 on a scale of 100 in its inaugural survey on the topic in 2001. Next year, Samsung was not rated in the survey because the brokerage did not have enough information, according to a Wall Street Journal report.

In 2002, UBS analyst Jonathan Dutton downgraded his rating of Samsung Electronics (005930.KS). Soon afterwards the local securities regulator started a probe into trading conflicts at brokerage houses, according to a report in Financial Times.

"You have to realize that it's really difficult for an analyst to come out and actually have a very negative view. Their careers are on the line," said a Hong Kong-based hedge fund manager who previously worked as an analyst.

LIMITED USE

In Asia, where countries and regulations vary widely causing the region to be difficult to understand even for the biggest fund houses, sell side research does help filter ideas and assess where some of the investment flows might be heading, according to industry sources.

Hedge funds and others also use sell side research to test their ideas and measure how far off the street estimates are from what they believe to be the likely outcome. But most of them find little value in following sell side stock recommendations anymore.

"It allows you to understand where the herd may be heading," said Paul Cuthbert-Brown, a partner at Singapore-based hedge fund Vulpes Investment Management.

"They do build models, they do get information but you have to go to the basic information that they reveal to you and draw your own conclusions."

BTIG's Lentchner said that there has been a shift over the last 20 years in how analysts provide value.

Before regulations outlawed the practice they had a genuine information advantage as the first recipients of disclosures from the companies they covered.

Now, he said, there is still a large amount of routine work that needs to be done collecting and formatting data, but a smaller opportunity for revelatory research.

"How you can provide real value now is through synthesis and analysis. You don't need a 200-person research team for that."

(Additional reporting by Olivia Oran in NEW YORK; Editing by Michael Flaherty and Alex Richardson)

• A Permanent Economic Emergency