Join our community of smart investors
OPINION

CAPE storms

CAPE storms
June 18, 2014
CAPE storms

For the uninitiated, the CAPE is like any other price-earnings ratio, giving the multiple of share prices over earnings per share (EPS). But the EPS figure is not a reported figure for the last year, or a broker forecast of the kind we include in our company results analysis, but a 10-year average. This approach - usually applied to markets as a whole, rather than individual stocks - is designed to smooth out cyclical variations in corporate performance to give a kind of all-weather PE ratio.

The CAPE has become controversial for two reasons. First - to take an anthropological view - its current elevated level in the US pits the habitually bullish stock-broking community against academic economists, with investors nervously between them. Second, it has a better track record of predicting stock market returns than any other metric. The CAPE predicted 43 per cent of future 10-year stock market real returns in the US between 1926 and 2011, according to research by fund group Vanguard - far more than any other indicator bar the trailing single-year PE ratio, which predicted 38 per cent of returns.

Most critics of CAPE simply argue that its current high level is misleading. Reported earnings plummeted after the financial crisis, partly as a result of harsher accounting rules introduced after the Enron scandal in 2001. US companies also pay out less in dividends than they used to, and the growth funded by the retained earnings should leave stock prices commensurately higher relative to past earnings. Adjusting for these two factors, which skew historical comparisons, consultancy Capital Economics finds the US market "only" 30 per cent overvalued relative to the long-run average.

A subtler criticism of CAPE is that its usefulness as a predictive tool may be exaggerated. In last year's Credit Suisse report on Global Investment Returns, London Business School academics Elroy Dimson, Paul Marsh and Mike Staunton point out that the CAPE - or rather a similar metric based on dividends, for which data is more widely available - has been a much better predictor of returns in the US and UK than elsewhere. In Austria, the relationship between the 10-year market dividend yield and future returns has even been negative over the past century.

One explanation for this is that London and New York have had a capacity to bounce back after crises that has not been matched in other financial centres. Indeed, the impact of just a few collapse and recovery cycles is decisive. If the professors exclude the bust-boom years of 1929-1933 and 2008-09 in the US, and 1920-22 and 1973-75 in the UK, the relationship between the 10-year market dividend yield and subsequent returns becomes statistically insignificant.

Another problem is that hindsight is built into the predictive models. When the professors repeated their regression analyses using only data available at any given time, they found that the 10-year dividend yield had no predictive power at all. For every country in which a higher yield was associated with higher returns, in another it was associated with lower returns.

But most instructive of all are their simulations of a market-timing strategy based on their CAPE equivalent. Investors who sold stocks and bought treasury bills whenever the 10-year dividend yield was so low that it forecast negative returns would, in every country, have been worse off than investors who simply stayed in stocks. Gains in those years when the strategy did work were invariably more than offset by foregone returns in those years when it didn't.

Fortunately, the UK CAPE is nothing like as stretched as its US equivalent, suggesting stocks are, if anything, undervalued (see chart). Even if this were not the case, however, the best response would probably be to do nothing. "Time in the market, not timing the market" is, it seems, more than just a fund manager's marketing line.

CAPE: a comparison