Should you use earnings forecasts?

The price/earnings ratio (p/e) is one of the simplest and most popular valuation measures in investment. To calculate it, you take a company’s share price and divide it by the earnings per share. In effect, it shows how much investors are willing to pay for a given £1 or $1 of earnings from that company. A high p/e indicates that the share is highly valued – usually because investors expect earnings to grow quickly – while a low p/e suggests the stock is cheap.

But which earnings figure should you use? One option is the “trailing” earnings – the most recent full-year profit figure. On the one hand, this has the benefit of being a known figure. On the other, there’s no guarantee it will reflect future earnings. So forward p/e ratios – where the “e” is based on analysts’ forecasts for the year ahead (or even further forward) – are often used instead of trailing earnings. These give you a view on how cheap or expensive a firm is compared to what analysts expect its earnings to be – but clearly, these forecasts can be wrong.

So which p/e is best? Last year, Joachim Klement of the CFA Institute produced what he reckons is a definitive answer. He looked at four major stockmarket indices: the S&P 500 in the US, the
FTSE 350 in the UK, the Euro Stoxx 300 in the eurozone, and the Nikkei 225 in Japan. For each market he took the 20% of stocks with the lowest p/es, and the 20% with the highest, then compared average returns going back monthly over 20 years. He did this for both trailing and forward p/es.

To cut a long story short, Klement found that investing with the help of trailing p/es was far more effective than using forward p/es across all four major developed markets. Using trailing p/es, the least expensive stocks outperformed the most expensive by anything from 1.2% a year (in the US) to 10.1% (in the UK). Using forward p/es, the least expensive stocks still outperformed the dearest in the UK, Japan and the eurozone, but by a much lower margin than using trailing p/es. And in the US, the cheapest stocks by forward p/e actually did worse than the most expensive ones – “in other words, relying on forward p/e ratios destroyed performance”, as Klement puts it.

The fundamental problem, says Klement, is that analysts’ earnings forecasts are overoptimistic – on average by about 10% (although the figure is not consistent enough to be useful in “discounting” this optimism). As a result, even though they’re not perfect, you are far better off using trailing earnings to find cheap stocks. In short, says Klement,  “Never use forward p/e ratios. Ever.”