The reason to believe the US stock market is overvalued can be summed up in one four-letter word: CAPE. The cyclically adjusted price/earnings multiple used to do a great job of signalling when the US stock market was excessively cheap or expensive. That signal was valuable; it did not help predict the timing of a peak or trough, but was very powerful in predicting returns 10 years into the future; the cheaper the market, according to CAPE, the stronger the 10-year returns.
At present, it suggests very clearly that US stocks are too expensive, and at its latest reading, last week, passed a very ominous landmark. The benchmark calculation for the indicator, kept by the Yale University Nobel laureate economist Robert Shiller, compares share prices to the average of real earnings over the previous 10 years, and now shows that the US stock market is even higher than it was on the eve of the Great Crash of 1929. Only the last two years of the dotcom bubble have seen comparable overvaluations.
On its face this is terrifying. But the market does not feel anything like as overblown as it was, by all historical accounts, in 1929. And CAPE has been signalling an expensive market almost uninterruptedly since 2010. Consequently, there is a thriving industry of academics and investment management research departments in either knocking down CAPE, or adjusting it in various ways to show that markets are not that overvalued after all.
First, one important adjustment, suggested in a recent paper called CAPE Fear by Rob Arnott and others of Research Affiliates, is to note that the CAPE does have an upward trend over time. If we want a “fair value” for CAPE, we should get it by comparing the current number to an upwardly sloping line over time, rather than to the historic average. As Mr Arnott points out, the US has journeyed in the course of more than a century from being effectively an emerging market to being the world’s largest economy. Of course the earnings multiple will have risen overtime.
Using Mr Arnott’s suggested trend line, the market shows up as significantly expensive, but not as expensive as in 1929 (where the market in turn now looks roughly as overblown as it did in the later dotcom bubble).
The most popular point made against CAPE is that it still includes the crash in earnings that followed the credit crisis. It was a brief and drastic fall, of more than 90 per cent, that was driven largely by writedowns taken by banks (meaning that it reflected a correction for prior years’ earnings that had always been overstated), and was over in barely two years. The rap against CAPE is that once the dreadful earnings slump drops out of the 10-year total, then the earnings in the denominator will rise sharply, and the CAPE will fall. Stocks will suddenly not look so expensive after all.
This is an appealing argument, but testing it with some arithmetic shows that is not as clear-cut as it might at first appear. The decline in earnings was drastic, but the recovery was equally swift. If we base a CAPE on earnings starting in June 2010, by which point earnings had completed their recovery, then the multiple is now only 27.3. This is far below the current official Shiller CAPE of 33.2, but it is still comfortably above Rob Arnott’s trend line. It is also only enough to take the CAPE ratio back to where it reached, according to the Shiller calculations, at the end of 2016. It is almost exactly where the Shiller CAPE peaked in late 2007 before the collapse of the credit crisis.
Most startling of all, if we change the denominator to the average of seven years and six months of earnings (a period that excludes the post-crisis earnings slump completely), then the CAPE now is 29.76, while its peak ahead of the Great Crash was 29.16. The terrifying finding that the market is as overpriced as it was before history’s most damaging crash remains intact, even if we ignore the 2008-10 earnings slump altogether.
And in any case, earnings appear to have been rising unsustainably fast, and are due for a decline. It is after all precisely this effect that led Shiller and others to use CAPE as a metric in the first place.
None of this means that CAPE is telling us that the market will crash tomorrow. To quote Mr Arnott: “Most of us who use CAPE as a valuation tool readily accept that CAPE is only one measure of valuation, and that — as American investor Stan Druckenmiller likes to point out — capital flows drive short-term market behaviour far more powerfully than valuation.”
A final complaint is that CAPE does not include interest rates. If rates are low, discounted future earnings are higher, and it is reasonable to pay more for stocks. Rates are of course very low. Just as CAPE has signalled expensive stocks for years, so bond yields have for years seemed poised to break above a downward trend that has lasted since the early 1980s, but this has never happened.
However, nobody is advocating using CAPE completely in isolation. Mr Shiller’s own presentation of the metric is on a chart that also includes long-term bond yields. And yields have started 2018 by rising menacingly. If 10-year yields, pass last year’s high of 2.63 per cent, speculation that the trend. has at last been broken will intensify.
Many have convinced themselves that CAPE does not matter and that the bond bull market will continue. They should at least consider the possibility that these judgments are wrong.