The price/earnings (p/e) ratio crops up everywhere, from newspapers to websites to analysts reports. But just how reliable is it as a valuation measure?

A refresher

The p/e ratio is deceptively simple. In its most basic form it compares a firms share price (for ordinary rather than preference shares – see page 48) to a years earnings (profits after tax but before dividends). Take a firm with a share price of £3.00. It made an after-tax profit of £4m last year and has eight million shares in issue. Earnings per share (EPS) are £4m/8m, or 50p. So the historic p/e ratio is £3/50p, or six times.

If next years profits are expected to come in at £6m, then forecast EPS, assuming no new share issues or buybacks, is £6m/ 8m, or 75p. So the forward p/e is just four (£3/75p). In other words, if profits stayed static each year, it would take you four years to earn back your original investment (not adjusting for inflation).

Whats the point?

A p/e ratio can suggest whether a share is cheap or pricey. As Jack Hough notes in Smart Money, US stocks have traditionally traded at about 15 times yearly earnings. The higher the ratio, the longer youll wait for a return and the more the chance of disappointment. For example, says Hough, a US stock trading on 21 times earnings will earn back its principal in about 15 years (assuming earnings compound up, like interest on a bank deposit). Trading on 15 times earnings, theyd take nearer ten years.

As for judging whether a p/e is high or low for a particular firm, benchmarks to use include past years, the firms peers and sector as a whole, and the wider market – the FTSE 100 for a UK blue chip.

A forward p/e should usually be more useful than a historic p/e – future earnings are more relevant to an investor. Unfortunately, analysts are dreadful at getting forecasts right. Besides which, both measures suffer other flaws.

When a p/e will let you down

Conventional p/es can let you down in at least four ways. First, a single-year p/e is a snapshot based on one years earnings. That makes it an unreliable guide to future performance for a cyclical firm with volatile profits. Next, earnings arent equally important in analysing all industries. In investment trust or property sectors, net assets under management tend to matter more. That means other ratios, such as the price/book ratio (the share price as a multiple of the book value of a firms assets per share), are more relevant.

Thirdly, the p/e tells you nothing about cash flow and hence dividends. Profitable firms that dont manage their cash flow properly can go bust – a p/e may give you little advanced warning. The ratio also reveals nothing about the proportion of annual profits paid out as a dividend, or whether current dividend policy can be maintained. Fourthly, it is a hopeless measure for a firm making losses, as you have no earnings figure available. Analysts try to correct for this by looking at other measures, such as the price/sales ratio (the share price as a multiple of one years turnover). But in many cases the best bet is to look at a firms earnings across an entire business cycle (see column, right).

The Shiller p/e

As the recession bit, many firms suffered steep falls in earnings. Some have since recovered, at least in part, thanks to record low interest rates and – for UK firms at least – a plunging currency that has made exports cheaper. The Shiller p/e (named after its creator, Yale University professor Robert Shiller) tries to adjust for these frequent boosts and declines. It takes the current share price and compares it to an average, inflation-adjusted earnings figure for the last ten years – roughly one full business cycle.

As Shiller is American, the measure is typically quoted for the US market (being based on the S&P 500). However, since other major markets, such as Britain and Japan, tend to follow the US lead, its a useful guide to value elsewhere too. And just now, its sending out a warning, says Steve Sjuggerud on Dailywealth.com.

Based on the Shiller p/e, the average American stock is now trading on 22 times earnings. Thats bad news. Why? Because Shillers data shows that, since 1900, its been a bad idea to buy stocks when theyre trading at 20 times earnings or higher. The higher above 20 times earnings you go, the less money you make over the next five years. Sjuggerud also notes that GMOs Jeremy Grantham (who has a formidable track record in calling peaks and troughs) forecasts an average return of just 0.4% a year for US stocks over the next seven years. Strip out inflation of, say, 2%, and an investor will lose money.

What to do

Sjuggerud recommends good defence rather than offence. Short term, that means tightening trailing stops (if you have an automatic stop-loss order triggered by, say, a 25% drop in a share price, reduce it); selling overpriced positions (stocks with above average p/es), and shrinking the size of your bets. To that wed add sticking to good-value, income-paying, defensive blue chips. In a toppy market, theyll deliver the best nights sleep.