Serious Money:  Reversion to the mean is a powerful and all-pervasive idea in modern finance. From price-earnings ratios (p/e ratios) to gross domestic product (GDP) growth, analysts of virtually all persuasions believe that there is an average value that acts as a powerful magnet for almost all relevant metrics, writes Chris Johns

Robert Shiller is the Yale professor who coined the term "irrational exuberance". He has been almost deified by the markets for writing the book of that name that came out just before the tech bubble finally burst in 2000. Actually, his timing wasn't that good - he had argued that the market was expensive in 1996.

Anyone who followed his advice would have missed the crash, but would also have failed to participate in the fabulous second half of the 1990s. And I think he has remained bearish throughout the post-2003 market bounce. But his book is a marvellous primer for anyone interested in the branch of theory called "behavioural finance". It is also a book about mean reversion.

About the only chart in the book is one that illustrates the S&P 500 p/e ratio since 1870. Shiller calculates the average of this ratio and declares that statistical analysis more or less proves that the market always reverts to this mean. It may drift off for long periods, but it always comes back. Sooner or later, the market will end up at or close to its long-term average.

I have updated Shiller's calculations to the present day (and reworked them slightly) and reckon that the average trailing (backward looking) p/e ratio is about 16. Shiller concluded back in 2000 that because we were so far above that average, some kind of market correction was inevitable. The only other way for the p/e ratio to fall is if earnings, the denominator, grow very quickly - back in 2000, "very quickly" was an implausibly fast rate.

Today's level of the trailing S&P 500 ratio is about 17. The market p/e ratio has been falling steadily for six years. Even though the market itself began to recover about three years ago, the p/e ratio has continued to fall as earnings have grown faster than share prices. So, in Shiller's terms, the correction in the p/e ratio multiple still has a small way to go if we are to revert fully to the mean.

If analysts' forecasts are right we will, on unchanged share prices, be on a ratio of just below 15 times earnings by this time next year, thanks to robust profit growth. The market, if the earnings forecasts are right, will finally have strayed into "cheap" territory. Cheap, in this context, implies a p/e ratio below the long-term average - remember, mean reversion works both ways.

Why a long-term average of 16? This is justifiable in terms of trends in dividend growth and plausible assumptions about risk, but it would be equally easy to justify a ratio of 10 or 20. The number 16 simply has the support of history.

How valid is drawing a straight line through 136 years of very volatile history? In my opinion, not very. That average of 16 is pulled down by prolonged periods when the p/e ratio was very low - during world wars and 1930s-style depressions, for example.

If we make the bold claim that such events are things of the past, we might conclude that the "sustainable" p/e ratio is a little higher than its simple past average. But those would be the arguments of a natural bull.

Another bear of equities who buttresses his arguments with loads of detailed statistical analysis is Andrew Smithers. He focuses on something called Q theory (don't ask) but his argument is formally equivalent to Shiller's: Q reverts, eventually, to its long-term average.

Smithers has recently produced a new twist on the mean reversion story by pointing out that profit margins in most major economies are at all-time highs. Margins, like p/e ratios and Q, are mean reverting variables and, therefore, must come down. If profit margins decline at a time of rapidly slowing growth, overall profits will collapse. Smithers was reported earlier this week as saying that such a scenario could easily see stock prices halve in markets such as the US.

A forecast of a 50 per cent correction in equities is a big and brave call. Even if it is only 50 per cent right, it means that we would be well advised not to have very much invested in the markets at all.

Could Smithers be right? In principle, it is tough to argue with the observation that profit margins (and profits themselves) are cyclically very high. It is easy to agree with the notion that margins are either close to or at their peak for this cycle. If they are to go any higher, all of those income disparities so noticeable in many economies will only get worse as real wages stagnate and profits grow further.

But just because it is reasonable to argue that margins may not go up much further, it is a bit of a stretch to suggest that they are in imminent danger of collapse. Here we get into the argument about the impact of rising raw material costs and the future behaviour of the labour market. I find it tough to see how workers are suddenly going to start squeezing employers, given all that has happened, particularly globalisation, in the past decade.

A value trap is a market or stock that looks cheap but isn't really. Such a trap is sprung when it becomes clear that the earnings upon which "cheap" valuations are based are illusory. Such is the fear of the value trap right now that I suspect that markets will only respect actual earnings growth and will be supremely indifferent to forecasts, no matter how soundly based (Irish banks are examples of apparently cheap companies that the market thinks could be value traps). And earnings disappointments will be punished in the extreme.

The market's motto for the foreseeable future looks like being "show me the money". Fortunately, many - but by no means all - firms look well placed to do just that.

Chris Johns is an investment strategist with Collins Stewart. All opinions are personal.