New gods found wanting

SELL-SIDE ANALYSTS have been busy reforming the way they work as the backlash to the Internet-era glut of compromised research intensifies.

Several months ago HSBC set the ball rolling when it said it would balance the number of 'buy' recommendations it issued with an equal number of 'sells', a commitment it later appeared to backtrack on.

Morgan Stanley has reformed its research practices and, most recently and memorably, Merrill Lynch agreed to pay a US$100-million fine and introduce changes to settle high-profile charges brought by New York Attorney-General Eliot Spitzer.

The central charge against sell-side analysts is they issue overly positive reports on companies from which their employers hope to win lucrative corporate finance deals.

Then they got to share in the spoils.

It remains to be seen whether the sound and fury emanating from the United States changes that fundamental dynamic. The conflicts inherent in sell-side research are widely recognised and long-standing, although they have been exacerbated by the decline in the importance of brokerage commission revenues and the huge amounts of money to be made by arranging pre-2001 Internet initial public offerings.

The Merrill settlement has been criticised as tame, while suspicions remain that the voluntary reforms of other houses will prove to be largely window-dressing.

If the wind of change is blowing through the US brokerage industry, it certainly does not appear to have reached Hong Kong's shores.

'Sell' recommendations account for just 4.03 per cent of the consensus recommendations on Hong Kong and China stocks tracked by data provider Thomson First Call. In the US, they reportedly average about 3 per cent of the total, compared to about 15 per cent in London.

'Strong buy' and 'buy' recommendations account for 58.06 per cent of the consensus recommendations tracked by Thomson First Call.

Of the 2,532 recommendations, a more respectable 369 or 14.5 per cent are 'sell', although there are nearly three times as many 'strong buy', at 1,039, or 41 per cent.

Research studies have shown that analysts systematically overestimate firms' earnings growth. It feeds through to over-rosy assessments of the stocks' investment prospects.

In these cynical post-bust times, the question of why analysts err on the positive side is easily answered: Conflict of interest. Mr Spitzer's release of internal e-mails, showing Merrill analysts privately disparaging stocks they were touting to the public as good investments, confirmed what many had suspected: Analysts keep their true opinions to themselves, and perhaps their best clients.

Such cynicism is understandable after debacles such as the collapse as Enron - when many Wall Street analysts maintained their 'buy' recommendations almost till it declared bankruptcy. The case for prominent and clearly labelled 'health warnings' on sell-side research is unanswerable.

Yet the idea that sell-side analysts blithely sell their souls for the investment-banking lucre and abandon their responsibilities to investors without a second thought is fanciful. For a start, it assumes that analysts have no regard for their professional integrity and ethics. There will always be bad apples, but it is unlikely this applies to an entire body of industry professionals. The truth, as always in such cases, is probably more complex.

An experiment from the field of behavioural finance offers a more convincing explanation of the systemic upward bias.

Martingale Asset Management president and chief executive Arnold Wood outlines the experiment in Investing Worldwide.

Two games are played with two decks of cards, one blue and one red. For US$1, players receive a blue card. The dealer pulls it from the pack, shows it to the player and puts it back.

The game is played the same way with the red deck - with one crucial difference - the player touches the card before it is put back in the deck.

The players are told that if their card is pulled from the deck, they will win US$100.

Before proceeding, the dealer asks whether the players are willing to sell their cards to the dealer.

'Here is where the game gets interesting,' Mr Wood said.

'In the test, of those using the blue deck, 19 per cent said they were unwilling to sell their cards to the dealer. By contrast, of those using the red deck, 37 per cent were unwilling to sell their cards.'

The dealer then asked how much the players would want for their cards. The blues asked an average $2; the reds asked for about $9.

The behaviour of the players in this experiment is irrational, according to the edifice of thought on which modern financial theory is constructed.

There is no difference between the two games, except that the reds touch their cards. Does this give them a better chance of winning?

What has this got to do with analysts' earnings estimates?

The red-deck players in the card game appeared to believe - probably unconsciously - that touching the card gave them a better chance of winning. It gave them an illusion of control and increased their confidence.

As Mr Wood writes: 'Illusion of control is influential because most of us have been brought up to believe that decision-making is built on rationality, that we can and do analyse circumstances objectively . . . in reality, we are not as rational as we believe or have been taught to believe. We succumb to temptation and that age-old duo - fear and greed.'

Mr Wood reports a research study showed that the average analyst's earnings estimate was off by 45 per cent a year over a 17-year period.

Why should this be so, when analysts have such an array of tools and information to help them forecast accurately?

'Analysts presumably have skill. They meet the company managers; they 'touch' them. They are familiar with the game; they have confidence that they can forecast. But how confident should investment analysts be in light of their performance?'

Experiments such as the card game suggest analysts are subject to 'cognitive errors'. They are overconfident about their ability to forecast.

We may offer some common-sense observations of why this might be so. Analysts are often relatively young and usually well-educated. They are paid large amounts of money and put into close contact with powerful and often charismatic business executives who have a strong motive to court their favour. It would be a robust individual who could maintain complete objectivity in such circumstances, particularly when the potential rewards for looking on the bright side are so much greater.

It does not help that standard finance theory in which most analysts are schooled does not permit such cognitive errors. In this artificial world, perfectly rational investors make perfectly informed choices based on the expected mean-variance trade-off. The standard model provides a useful framework for thinking about market behaviour, but as anyone who has watched a speculative mania knows, it far from captures the full reality of how investors behave in all situations.

This assumption of perfect decision-making leads to virtual deification of the market (reflected, for example, in the mystical overtones of the popular phrase 'the market is the market').

If a stock (or a market) is trading at a certain price, it is possible to work backwards and infer certain assumptions about growth and risk. Since the market is always right, if eHubris.com is trading at 500 times earnings, that must be because it is going to grow at 50 per cent a year forever, right?

The analysts were the gods of this new-economy theology. Puffed up with importance, they probably believed they were right. Unfortunately, they found out they were only human - as were the poor saps who followed their recommendations.

Graphic: hk04gbz