IOSCO Annual Conference
Istanbul, 21 May 2002

CONTRIBUTION TO THE PANEL ON SECURITIES ANALYSTS

Luigi Spaventa
Chairman

Commissione Nazionale per le Società e la Borsa

1. The incentive structure of the analysts' industry has become inherently flawed. The output of that industry affects the prices and volumes of traded securities. On the other hand, the analysts' activity is not economically independent, being horizontally integrated within the financial services industry It is a department of global (or universal) banking, where analysts are employed, and paid, by banks that act as intermediaries, underwriters, global coordinators, advisers, lenders. The success and profitability of most of these operations depends to no little extent on the prices and volumes of the securities traded. It is therefore affected by the reports and studies issued by analysts, who are in turn employees of entities that stand to lose or gain by the effects of their research on the market.

This inherent flaw leads to a large number of situations of potential or actual conflicts of interests. Before the pricking of the bubble and the Enron case caught everybody's attention, the voices pointing to such situations were (surprisingly) few and (less surprisingly) unheeded. Now, at least for the moment, shooting at the analysts has become the game in town.

A preliminary IOSCO paper on the subject provides a (probably non exhaustive) catalogue of the conflicts that may arise:

- When an analyst's firm has underwritten or may underwrite a company's offering, the analyst may be under pressure to praise, or at least not to be negative on, the company's prospects.

- The analyst's firm may own shares, trade securities of, or provide finance to the companies covered by the analyst, who may be under pressure to provide favourable reports.

- The firm's brokerage activity, and the resulting commissions may be boosted by the advice provided in the reports.

- There may be relevant financial and corporate links between the analyst's firm and the companies covered by the analysis.

These conflicts are exacerbated when the analyst, as an employee, receives compensation (in the form of bonuses) the amount of which depends on the firm's overall profitability and when he has to report to the sell-side departments of the firm. I neglect the pathological possibility (which may occur independently of the situations of conflicts of interest to which I am referring) of the analyst trading in the shares of the covered company and front-running the outcome of their studies.

2. There often are other, and perhaps subtler, flaws in the analysts' research.

Economists are taught and have learned that we live in a world of uncertainty, where micro- and macroeconomic forecasts and projections are not only interdependent, but contingent upon the realization of events to which a 100 percent. probability can never be assigned. Not so the analysts, whose reports seem to ignore any "if " or "but". Macroeconomic forecasters nowadays use fan charts, assign probabilities to their valuations and assess their sensitivity to a change in the assumptions regarding exogenous and uncontrollable variables provide alternative scenarios with respect to the base line. Not so the analysts, who usually produce one unqualified number (the "target price") without disclosing the events and trends upon which that number is contingent. When considering sectoral trends analysts are prone to entertain backward looking expectations: what happened in the past is assumed to continue in the future, as was the case with telecommunications and the e-economy. The herd behaviour of the studies produced by analysts has been documented: it is safer to go along with the crowd than being an outlier. Finally analysts are often prone to base their assessment on all sorts of pro formameasures of performance and profitability that heavily depend on elusive and unreliable accounting conventions: as a prominent international banker once told me, after ebitda's and the like, we shall soon be offered valuations based on ebe's - earnings before expenses.

Of course, these technical flaws - an example of fuzzy economics - are in a way connected to the wrong incentive structure of the industry: when it pays to believe, little room is left for doubts.

3. We know by now the bitter fruits of this situation. What matters here is not whether the analysts' views on individual securities are correct in the very short period: they may well be, insofar as the very short-term outcome may itself be the result of those views. Chairman Greenspan recently reported that "three- to five-year earning forecasts for each of the S&P 500 corporations averaged almost 12 percent per year between 1985 and 2001", while actual earning growth over the period averaged about 7 percent. "The persistence of the bias year after year - he adds - results, at least in part, from the proclivity of firms that sell securities to retain and promote analysts with an optimistic inclination", the bias beeing especially large when the analyst's firm also serves as an underwriter for the company's securities. Almost three years ago the former chairman of SEC Arthur Levitt stated that a review conducted by SEC "indicated that analysts, all too often, are falling off [the] tightrope on the side of protecting the business relationship" and do "their bit to market their own firm's underwriting talents and to sell a company's prospects". I refrain from referring to more lurid cases that have recently come to light.

What is true for the US is also true elsewhere. I myself drew attention to the inherent bias of the analysts' research more than two years ago and Consob, as I shall say presently, has sought to introduce some remedies. As in Italy the market fell by 25 percent in 200,. the share of the studies with a "buy" recommendation only decreased from 58 to 48 percent, and that of the studies with a "sell" recommendation only rose from 6 to 9 percent. The analysts' industry in Italy is highly concentrated. Of the 10 entities that produce over 60 percent of the studies, eight are groups active in merchant and investment banking. As was found by Consob, the analyst's firm has operated in a direction opposite to that suggested by their published studies in 40 per cent of the cases; never was a negative view expressed when the company concerned was, or was about to, become a client of the bank.

4. It is far from easy to find convincing and cost effective solutions to these issues, as always happens when we are faced with wrong incentives. As often happens, we are confronted with three broad - and not necessarily mutually exclusive - choices: let the market take care of the problem; rely on self-regulation; introduce more statutory regulation.

The first is an attractive possibility. Chairman Greenspan feels confident that "it is just a matter of time before the ex post results of analysts' recommendations are compiled and published on a regular basis" and trusts "that with such transparency, the current upward bias of analysts' earnings projections would diminish rather rapidly". But who will engage in such costly compilations and publications, the outcome of which is in a sense a public good? It is in the nature of a public good that everybody benefits from it, but nobody is ready to bear its cost, which, for this very reason, must be covered by general taxation. It is difficult to imagine the banks and intermediaries as sponsors of the initiative, for only too obvious reasons. It is perhaps easier to hope for a joint initiative of institutional investors, which have a direct interest in less biased and more robust analyses, the alternative being to entrust this task to some public agency

Another consideration may cast some doubt on the effectiveness of a purely market remedy. The market is not only myopic, but tends to have a short memory. The present furore against analysts is partly explained by the disillusion caused by a gloomy post-bubble environment, marred by Enronitis. Let another bubble start and everybody will soon forget the past and be happy to follow the most extravagant recommendations on the beauties of Dutch tulips. Thus perhaps, while endeavouring to implement chairman Greenspan's suggestion, something more may be required.

5. Self-regulation has fallen out of fashion nowadays. Remember the profession's oversight board for accountants? It never ever passed a negative judgement on the accounting firms' practices - as we say in Italy a dog never eats another dog - and is now being dismantled to be replaced by an independent body under the supervision of the SEC. Nearer to our subject today, I have never met a banker who would not boast about the impregnable Chinese walls in place in his firm between research and other departments: the public at large, judges and regulators have now discovered that they were paper walls - Japanese walls perhaps - that were only too easy to penetrate. Those stricter rules that are now being introduced by some firms have been forced upon them by pervasive judicial initiatives.

Rules, to be effective, must be enforceable and sanctionable. There can be no certainty that internal rules or codes of conduct will be enforced and their violation be sanctioned. The Enron case provides again sobering lessons regarding the ineffectiveness of internal codes of conduct. Nor can one rely in these cases on reputational sanctions: what happens behind the scenes is not known by market participants until it is too late.

More effective rules can instead be enforced by self-regulatory bodies such as the exchanges, that have the power to set the conditions required for admission to listing and to sanction intermediaries. They could play a greater role of supervision in the matters we are discussing by monitoring the effects of the analysts' research and their links with the companies covered by that research, as is now happening in the US.

6. So far the analysts' activity has been outside the field of regulation, with a few exceptions. The trend now is to extend the scope of regulation.

In Italy studies have to be filed with the regulator and with the exchange and the latter is required to make them available to the public. The issuer has to make immediately available to the public all the information provided in meetings with analysts. As from last year, the published studies must report "in bold letters" the nature and extent of the business and financial relationships between the analysts' firm and the company covered by the research.

The IOSCO study which I cited before provides a rich menu of possible regulatory interventions: all kinds of detailed disclosure; limits to the analysts' compensation structure, up to the point of forbidding that it be linked to banking activity; prohibition of, or limits to, the trading of shares covered by the analysis; accountability and responsibility of compliance with the rules.

Only two weeks ago the SEC approved the new rules set by the NASD and the NYSE, that cover some of the points raised above: analysts should not be supervised by the investment banking department nor can they discuss their reports with investment banking personnel; analysts' compensation cannot be connected to specific investment banking transactions; existing interests of the firm in the company covered by research must be disclosed; restrictions are introduced on personal trading by analysts and their financial interests in the covered companies must be disclosed; the past record of research in a company's assessment must be published. The terms of the Merill Lynch settlement overlap, at least in part, with these new rules.

Many wonder if these rules, though representing a great progress, are robust enough and if they go far enough. Some of them seem to replicate the disreputed Chinese walls though strengthening them with some form of external sanction. More poignantly, it is objected that they address the symptoms but not the deep cause of the conflicts of interest. Thus Henry Kaufman - the well-known former head of research at Salomon - asks for more. At a minimum he requests that the head of research be a member of the senior management of the firm, noting that otherwise it will be impossible to defend the independence of research. But in his view "the logical solution to [the] conflict [of interests] is for sell-side institutions to provide no research reports to clients". This, I am aware, is a bit extreme, but he may be right, also because Kaufman's solution would reduce the amount of detailed regulation.

As an economist I am surprised that the none of the proposed regulatory innovations seem to deal with the technical pitfalls of the analysts' research which I mentioned before. The problem here is not with independently and identically distributed shocks with zero mean which are and should be disregarded. The problem is that of assessing the probabilities of alternative events that may affect the forecast and of disclosing the assumptions used and the sensitivity of the forecasts to a change in those assumptions.

7. One very last point. Regulatory intervention in the field of research is a case in which international coordination is particularly necessary. We have experienced that in Italy, which, until now, was one of the very few countries to have introduced some rules. The Italian industry has complained bitterly of the damages caused by regulatory arbitrage. Firms established in Italy are subject to Consob rules. Not so the big London based or New York based firms, which are not obliged to file their studies with Consob and the Milan Stock Exchange for prompt dissemination nor to disclose with any precision their conflicts of interest.. This is an instance in which regulatory competition encourages a race to the bottom.

* * *

Regulation is costly and burdensome, and if possible the same objectives should be reached by other more flexible means. But those to which the currently regulatory reaction is directed have little to complain: there are now paying the retribution for their past excesses.

*************************