A sorry tale of fortune tellers and forecasters

Let's imagine for a moment that you're a salesman. You sell photocopiers and enjoy a reasonable bonus scheme, paid quarterly according to the number and value of the photocopiers you sell. Now, if you were to predict your income next year, to the last dollar, what is the probability of your estimate being within, say 5%, of the eventual figure?

We suspect it would be pretty low. You might snare a large account, or you may lose one. You may be retrenched from your position, or you may be promoted. Or a more powerful, cheaper machine may supersede the brand that you sell. It makes the point: the future, to a large extent, is unknowable.

A tough call

So, if a salesman finds it difficult to predict next year's income, imagine the complexities of doing likewise for a company selling numerous products, in a variety of markets, with thousands of employees. It has to be even harder doesn't it?

Why, then, do brokers, fund managers and, to a lesser extent, people like ourselves, devote so much time to predicting what is unknowable, only to be proved wrong a matter of months down the track?

That's what this article, and those that follow it, is all about. We think it will help you to develop a healthy scepticism to so-called expert opinion and, in so doing, realign your thinking, if you haven't already done so, towards a more profitable mindset.

Whilst we do usually make earnings forecasts for the next 12 months, we don't refer to them very often. But take a look at a broker's report and you'll see forecasts for the next two, three or even five years. And yet the regularity of profit upgrades and downgrades from the companies themselves is testament to the futility of these figures. So, why do they bother?

There are a couple of reasons. The first is explained by a contemporary phenomenon. Numbers are precise, even when they are wrong. Investors, being schooled in the disciplines of science from a very early age, are inherently more disposed towards a figure than they are a generality.

-Playing the numbers game with the big boys-

A figure suggests a rigorous scientific process while a generality suggests an opinion. In essence, numbers are facts.

Analysts play on this to produce research that sounds more believable, especially when cloaked in an impenetrable language that has more in common with the science lab than it does with the art of investing. Of course, the fact that their forecasts are not held accountable a year or so down the track only encourages them. Forecasts can be wrong and believed at the same time.

The second reason is a little more pragmatic. Have you ever marvelled at how a company announces an increase in profit and yet the share price falls (Toll Holdings and CSL are recent examples)? Or how a company announces a record loss and the share price rises (News Corp performs this stunt regularly)?

The answer lies in what is known as the 'consensus forecast'. This is the average figure once all the analysts have crunched their numbers and pinned an exact value on the next reported profit. It's simply an average of the range of forecasts.

Playing safe

So there's an expectation of a company's result before it is announced. And that expectation might be, for example, a rise of 25% on last year. Then, if the company delivers a seemingly respectable 20% increase in profit, those with higher hopes and shorter time horizons are likely to sell. The reverse is true for companies reporting smaller losses than the consensus forecast; a share price rise is likely to follow.

In other words, the share price responds to the expectations of others in the market rather than the performance of the company itself. For investors, the problem is that these expectations are filtered through a range of commercial imperatives that don't just reduce the reliability of the forecast but also reward those making them for being overly optimistic.

Most investment banks have a stockbroking arm and a corporate advisory division. The corporate advisory fees (structuring tax deals, advising on takeovers and the like) are usually quite large and extremely lucrative while the broking operation often isn't. It is sometimes used as a loss leader.

Imagine for a moment that you're in charge of such an outfit. Reclining in your leather chair, you're pondering the beauty of these advisory fees when a report on AMP from your stockbroking division crosses your desk. The insurance analyst has derided the company's management - much as we did in issue 98/Mar 02  (Sell/Switch to Suncorp - $19.12) - and paints a pessimistic picture for the immediate future. How do you react?

You know very well that a client like AMP is a wonderful honey pot of never-ending fees. It requires a whole host of specialist advice from firms just like yours, especially given its poor management depth.

Protecting interests

What would be the chances of AMP awarding you a slice of their corporate work after reading this unfavourable report? Pretty slim we suspect. How, then, will you react? Upgrade the recommendation from Sell to Hold? Slide the report into the bottom draw to gather dust? Or maybe go the whole hog, call it a Buy and upgrade the forecasts?

This rather colourful and imaginary example explains why negative broker reports are as rare as rocking horse poo. You may also have guessed that analysts also tend to overstate their forecasts for fear of offending management.

So forecasts aren't usually accurate just because the very nature of forecasting is difficult. There are other forces at play. So how should you approach earnings forecasts?

Instead of looking at the science, we'd rather take a step back a few hundred years and look at forecasts, not from the position of the compromised expert but the indifferent generalist. As a layman, do you, for example, expect one of the big four banks' earnings to be much higher in a decade's time? We certainly do. And if our forecasts are out by 10 or 20% for a few of those years, does it matter? Not really.

The accuracy of forecasts is only critical to those trading the stocks to which they relate immediately before and after it reports. Why complicate something that is essentially quite straightforward?

Having a general idea of whether the companies you own shares in will increase their profits over time is crucial. But pulling your hair out over next year's profit forecast is unproductive and painful.

As the British economist and investor Lord Keynes once said 'I'd rather be vaguely right than precisely wrong'. Amen to that.

Next time we'll talk about our own forecasts. How do we come up with them? How much credence should you give them and do they matter anyway?