If this sounds like a reason to sell, it shouldn't. There’s a big problem here. The Cape has been above its long-term average since late 2009. But the S&P 500 has doubled since then. Anyone using the Cape as a guide would therefore have missed out on big gains.
One reason for this is that the Cape's long-term average is depressed by very low share prices in the 1920s and 1970s. However, the fact that shares were cheap 90 years ago is not a reason to sell them now.
Another reason is that the Cape hasn't been an especially powerful predictor of annual changes in prices. In the last 20 years the correlation between it and subsequent annual changes in the S&P 500 has been just minus 0.3, implying that it explains just nine per cent of the variance in annual returns.
However, if we add just two variables to the Cape, we can beef up this explanatory power a lot - to over 46 per cent.
One of these variables is foreign buying of US equities. Big foreign buying has often been a sign of excessive optimism, which has led to prices falling. The other is the shape of the yield curve, as measured by the gap between 10- and two-year yields. When this has been inverted (with two-year yields above 10-year ones, it has led to higher returns, for given levels of the Cape and foreign buying.
Given these three things, several other factors lack predictive power. These include the conventional price-earnings ratio, the fed funds rate and yields on corporate bonds.
This augmented Cape has done a better job than the plain Cape of predicting returns on several occasions. For example, it predicted high returns in the late 90s when the Cape forecast only ordinary ones. And it correctly told us to sell in 2002 and in 2007 when the Cape did not. However, the plain Cape has had its successes too. It got us out of the tech bubble earlier than the augmented Cape. And it told us to buy in 2010 when the augmented Cape didn't.
Right now, however, we don't much need to choose between the two indicators. Based on post-1995 relationships, the Cape is predicting a rise in the S&P 500 of just under 8 per cent in the next 12 months. This is because it is only slightly above its post-1995 average, and an average valuation should imply an average return. The augmented Cape, however, is pointing to better returns because foreigners have recently been net sellers of US shares.
Granted, even this augmented Cape implies that there’s around a 30 per cent chance of the S&P falling in the next 12 months. But this is no higher a probability than normal. If you're looking for reasons to be pessimistic, you should look elsewhere.
