The right time to buy: Earnings

But don't be fooled by these figures – these are snapshots in time and they refer to actual earnings and the data contained within it is historical. They tell you next to nothing about the likely direction of earnings and dividends, which are both falling fast.

If you want to understand why these stated, current earnings are viewed by some as next to useless it's worth looking at the detailed quantitative analysis produced by most of the major investment banks. SocGen's Andrew Lapthorne's weekly Global Equity Market Arithmetic is a must read for anyone trying to determine whether the markets represent 'good value' – the table below contains summary data from SocGen's analysis. At the moment the world's stock markets are valued not in single figures, but at a rather high 13.9 times current estimates of earnings, only falling to 11.2 times earnings estimates for 2010. But that fall in the PE ratio next year is only because analysts expect earnings to grow by 24.6 per cent. A great many economists – including those at the IMF and OECD – think those growth estimates are fairly ambitious. And looking at individual markets, a PE ratio in 2009 of 14.7 earnings for the US isn't terrifically cheap and a yield of 2.6 per cent seems almost miserly. The Japanese market is valued at a princely 39 times 2009 earnings (although that's at the lower end of the range for the last decade) and only European countries such as the UK and France look decent value at just under 10 times 2009 earnings and a yield of between 4.5 and 5 per cent. But many really bad bear markets have seen PE ratios shrink to less than nine or even eight before the bottom is reached – so even if you accept these estimates for earnings, markets are only good value at best.

The sad reality, according to many analysts and quantatitive analysts, is that no one really has any idea of what the earnings bit of the PE ratio for 2009 will be. The table below spells out this confusion in horrific detail – it's from US commentator John Mauldin and it looks at the combined estimates for the EPS produced by the 500 companies in the S&P 500 – this is usually expressed as a single bundle of dollar profits. At the beginning of 2007, analysts estimated that the S&P 500 would produce $92 worth of earnings yet by April of this year the figure was down to a frankly jaw dropping $14 (the market is valued at around $850). For 2009, analysts thought earnings would hit $81 in March 2008 yet by this April that had dropped to $28, and it's still falling.

The problem is that profits move in a cycle and earnings are only now coming down from peak earnings seen over the past few years – most analysts’ estimates are proving to be woefully inadequate. If one cuts through the data and looks at actual operating profits or EPS for the S&P 500, one comes to a truly horrifying figure – at the end of 2008 the S&P 500 was valued at 60 times actual reported earnings.

Many analysts have completely given up using the current PE ratio – they prefer to use an alternative measure called the 'cyclically adjusted price-earnings ratio' (CAPE). You might also see it termed the 'Graham and Dodd cyclically adjusted earnings measure' after the investor Ben Graham who first developed the tool. In recent years a variant on the same approach has been publicised by Yale economist Robert Shiller – his version of CAPE compares the current share price with average earnings during the past decade, rather than to the most recent year's earnings. The idea behind this alternative look at the all-important PE ratio is to even out the ups and downs in the ratio over a long-term (10-year) profit cycle. Because profit margins are mean-reverting, in boom times, companies will probably boast high margins and big earnings. In busts, profit margins collapse and companies have small earnings. According to one enthusiast of this approach: "Taking a single-year PE ratio can therefore provide a misleading picture of value – in booms, with high profit margins, stocks look cheaper than they really are. In busts, with low margins, stocks look more expensive than they are".

The CAPE measure now stands at around 14 times earnings according to Professor Shiller, well below the average since 1870 of 16.34, and the lowest level since 1986. Most analysts reckon that this is a positive sign although Professor Shiller is rather more bearish – he warns that these cyclical PE ratios have tended to bottom out at around six, implying falls of at least another 50 per cent from current levels.

In Mr Lapthorne's view, the Dodd and Graham version of CAPE shows that the US is only fair value, although the UK and Europe are good value – ie, cheap. Mr Lapthorne's colleague, SocGen Strategist James Montier believes that the UK and Europe markets are now a buy on most measures.

The big question is whether or not this measure has any predictive power – according to the Financial Time's investment editor, John Authors, the CAPE measure has proved to be a great market timing tool as "highs and lows for this metric have overlapped almost perfectly with highs and lows for the market".

■ Prognosis: The US is probably at worst fair value and good value at best. Most analysts reckon the UK and Europe are already cheap.

Falling earnings estimates for the S&P 500 in 2008 and 2009

Société Générale data on global stock markets