Market tactics: CAPE crusader

The break to new highs on Wall Street has emboldened some bulls to argue that the long-term slump in stocks since 2000 has ended. Especially given the dearness of bonds, they argue, it is only natural that today's boom should continue throughout its fifth year and beyond. There are equally many sceptics who argue that the market is now running on fumes and that another major top is close.

By past standards, the present uptrend - at 50 months - is indeed long in the tooth. Since 1900, a typical bull market in US equities has lasted 33 months. Only six out of 20 have gone on longer than this, and most of those occurred during periods of economic and financial wellbeing, such as the 'Roaring Twenties' or the go-go 1980s.

Those arguing that the S&P is good to keep rising often cite the index's modest valuation using forecast earnings 12 months from now. The bears, by contrast, point out that US large-cap stocks are very dear in terms of their average earnings over the last decade.

It does seem more intuitive to use forecast earnings when weighing up the market's future prospects. What is yet to come should matter more than what has already been. But the S&P's forward price-to-earnings (PE) multiple has not been a particularly meaningful guide to the coming year's returns. In statistical terms, it has historically explained a mere tenth of the real return over the next 12 months.

After a forward PE valuation specifically like today's, the median gain over the next year has been 6.6 per cent after inflation, which is certainly not to be sniffed at. And the probability of the market going up over that period has been more than 70 per cent.

Unfortunately, these findings are from a very limited data set. Forecast earnings for the S&P only go back to 1985. What's more, a modest forecast PE is no safeguard against catastrophic losses. In 2008, the index sported almost exactly the same rating as today, but went on to plunge 40 per cent.

Using CAPE for long-term analysis

What about valuation based on the cyclically-adjusted price-earnings ratio (CAPE)? Thanks to the work of Professor Robert Shiller, we have more than 130 years of earnings data for the S&P. The US market currently trades on a CAPE of 23, which is roughly one standard deviation above the long-term average of 16.5.

Similarly high CAPEs have heralded some of history's most drawn-out bear markets. For example, the 'lost era' for US equities between 1968 and 1982 began from a CAPE of 22.2, while that of the early 20th century was ushered in by a CAPE of 19.2.

While this may seem scary, it is worth remembering that CAPE is not - and was never intended to be - a device for timing market tops. Overall, CAPE tells us next to nothing about likely returns over the next year. Today's reading of around 23 suggests a 50:50 chance of the market delivering a positive real return by next May - hardly head-for-the-hills stuff.

CAPE

In light of this, you might well ask if forecast PE ratio and CAPE have any use whatsoever. The answer is yes, but only if you are taking a much longer-term view. For determining real returns over the next decade or so, their explanatory power is decent enough. High readings have tended to lead to low returns, and vice versa.

So, what does the today's combination of a high CAPE and moderate one-year PE ratio suggest about the outlook for the period to 2023? By my calculations, the S&P's typical return over the next decade from here would be an annualised real return of minus 3.4 per cent, with only a one-in-five chance of a positive real return. This fits in with my view that lower profit margins and higher inflation are set to cause major difficulties.

As to the more immediate future, my stance on US stocks is bullish owing to the ongoing money-printing machinations of the Federal Reserve. If that stimulus is withdrawn or looks likely to be withdrawn, my instinct would be to sell up. In the meantime, I would welcome a 5 to 10 per cent dip in the S&P in order to buy anew.