* PEs set to return to relevance
* Ratio more useful at certain points in cycle
* Difficult to determine at times of great uncertainty
By Brian Gorman
LONDON, June 17 (Reuters) - Price to earnings (P/E) ratios, consigned to the doghouse over the past year, are creeping back into favour with investors as visibility on company profits improves. The ratio, traditionally a key measure of a stock's value, was jettisoned by many analysts in the midst of a credit market crisis that tipped economies into recession and made company sales and earnings difficult to predict with any certainty.
But improving macroeconomic indicators suggest the worst may be behind us, making P/E relevant again.
"If you have greater confidence in predicting earnings then suddenly P/E ratios become a lot more relevant and useful," said Nick Wilson, cyclicals companies analyst at ICAP.
Wilson said that the market was entering a stage of the cycle where P/E ratios would be widely used again, adding that aerospace and defence, with a P/E ratio of 8.5 for 2009, was the most predictable sub-sector within cyclicals.
Earnings at banks were the most difficult to forecast through the downturn, as losses and writeoffs piled up from exposure, direct and indirect, to a collapse in risky U.S. mortgages.
While analysts turned to price to book ratios for the sector during the past year or two, there are indications that many are now willing to consider P/E again in addition. The return to P/E for banks is part of a cross-sector move back to the ratio.
"The problem before was that if you used trailing numbers, you didn't know what was telling you, with all the writeoffs, and forward numbers had been distorted because analysts hadn't adjusted their expectations. But now they have," said Georgina Taylor, strategist at Legal & General Investment Management.
"Once we're past asking how strong the recovery is going to be -- then we ask 'Where are the bargains? Where is the value?' Valuation metrics like P/Es become more helpful."
LOGIC BREAKS DOWN
Some of the limitations of P/E ratios became more apparent than ever in the early part of 2009.
P/Es tell investors how expensive a share is relative to its earnings. The lower the ratio, the "cheaper" the stock. However, the elementary mathematical logic breaks down once the market goes into negative territory.
For example, a share price of 10 pence, divided by earnings of -20 pence per share gives a P/E ratio of -0.5, a meaningless number to gauge investment value.
In February, Royal Bank of Scotland reported a loss of 24.1 billion pounds ($39.6 billion) for 2008, the biggest loss in UK corporate history. Many other companies also reported losses, albeit not on quite the same scale.
But the market may be heading for a period of more stability, as more companies swing back into the black.
"The rapid pace of earnings decline is probably behind us," said ICAP's Wilson, adding that the P/E ratios for 2010 were likely to be more relevant than those for 2009.
AND DIFFICULTIES REMAIN....
But some analysts would always question P/Es on the grounds that earnings volatility in their sector is a way of life. Oil prices have fluctuated wildly in recent times, going from more than $147 a barrel to less than $35 in a few months.
It follows that oil companies' profits will fluctuate at an even greater rate. If total costs are 80 percent of revenue one year, it only takes a 20 percent drop in oil price -- well within the bounds of possibility -- to wipe out profit completely, assuming costs remain the same.
Some of the sector's analysts have long since regarded P/E ratios as, at best, of limited relevance.
"To cut a long story short, we don't worry about P/Es. It's something you calculate and put in the research, because people want to see the number," said Richard Griffith, oil analyst at Evolution Securities.
He also pointed to other reasons for volatility in earnings, such as some high-risk exploration companies that may report little revenue, far less profit, if they fail to find oil.
But other analysts said P/E had its place, at the right time, even if it needed some support from other metrics.
"At the turning point in the cycle -- either peak or trough -- it becomes very tricky, and very much need to be supported by, at minimum, two or three other valuation metrics," said ICAP's Wilson, adding that he tended to balance "snapshot" P/Es with other ratios like enterprise value to sales.
As with any valuation measure in an efficient market, P/E is unlikely to be an easy way of spotting bargains for very long.
"The problem is that if investors truly believe in the earnings streams of the companies, they get more and more expensive," said L&G's Taylor. (Editing by Sitaraman Shankar)