Tuesday, June 24, 2014

Andrew Smithers on using CAPE in international markets

Link to Part 1: The problem of Cape and non-US markets (part 1)
The cyclically adjusted price-earnings ratio (Cape) has become well known as a way of valuing the US equity market. Its moderate success in this role has led to the assumption that the same approach will be valid for other markets. Unfortunately this seems doubtful, as I will try to explain. I should warn readers that, despite trying to make my explanation as simple as possible, I have been unable to avoid raising some quite technical points.

Link to Part 2: Why the cyclically adjusted PE usually fails for non-US stock markets (part 2)
Cyclically adjusted price-earnings ratio (Cape) appears to be a valid way to measure the value of the US stock market, but this does not mean that it can sensibly be used for other indices. As I explained in a previous blogpost, Cape is only valid if it can pass two tests: first, that the real return on equities has been mean reverting; and second, that profit margins have also been mean reverting and have rotated quickly around their average. 
Real returns on equities has been less strongly mean reverting in other markets than they had been in the US. This weakens the case for Cape in other major stock markets, but does not, I think, necessarily rule it out. Even if returns would otherwise have been mean reverting, they will not have been if countries had suffered unexpected and catastrophic losses, such as occurred in world wars. I had already explained in Growth and returns, another blogpost, that these losses were the probable explanation for the exceptionally low returns on equity investment in the first half of the 20th Century in countries such as Germany and Japan. 
If this proviso is accepted, it means that the calculation of Cape should exclude pre-war years. As earnings per share (EPS) data are generally not available before second world war this is almost a help rather than a hindrance. Even using postwar data, however, it is clear that the failure of real returns to exhibit mean reversion in Japan means that Cape cannot sensibly be used to value Japanese equities. 
This leaves the second test to be considered and, unfortunately, this suggests that, with the possible exception of the UK, Cape should not be used for non-US markets.