We clearly see that the drawdown risk of stock markets increases as the CAPE increases. Thus, in direct contradiction to the received wisdom of modern portfolio theory, higher valuations clearly lead to lower returns and higher risks in the future. Again, we could have created similar charts for all the countries in our sample. In Exhibit 7 we show the average drawdowns for developed markets and emerging markets. In order not to rely on a few big countries in each category, we have used equally weighted averages of both developed and emerging markets. It is interesting to note that emerging markets show only slightly higher drawdowns for a given valuation level than do developed markets.

Looking at the current valuation levels in different markets, it is interesting to observe the trade-off between expected returns and possible drawdown risks. In Exhibit 8 we show the expected real returns together with the average drawdown. The expected returns are from our model in Exhibit 3, and the average drawdown reflects the experience of the past five years, following CAPE valuations similar to today’s current CAPE of ±10%.

The relationship is clear. Investors in markets with the lowest expected returns face the highest drawdown risk. In other words, in markets like those in the US or Ireland it is highly likely that a buy-and-hold investor will experience severe losses from current index levels. Or, as John Hussman put it, the returns of the next five years are already on the table now. 

In fact, in the United States, when the CAPE has been at levels comparable to today’s, the average drawdown over the next five years has been an eye-watering 26%. We note, however, that this is a historical average. The most extreme drawdown came after the 1929 stock market crash, when the market fell by 80%. Since 1900 there have been only two occasions when CAPE levels like today’s were not followed by a drawdown of 15% or more: 1995 and 2003. In both instances the market continued climbing for five more years, reaching even more exaggerated valuation levels, before crashing. We would add here that the 2008 financial crisis has already wiped out any profits made since 2003 and more.

Eoin Treacy's view

Here is a link to the full report. 

The authors highlight that elevated valuations can be justified today because of low inflation and rock bottom interest rates but also point out that surges in Cyclically Adjusted P/Es (CAPE) to well above their mean generally results in mean reversion at some point over the subsequent five years. In other words, mean reversion occurs following overextension. Alternatively, the bigger the overextension, the larger the correction when it eventually loses momentum.  

The reliance of returns on perpetually low interest rates and below par inflation suggests that when a meaningful change occurs to these factors fundamentally oriented investors will be robbed of an important justification for paying ever higher prices. The additional effect such developments would have on corporate bond spreads would result in an extra headwind as funding for balance sheet optimisation programs would become more expensive. 

What is also clear is that the number of reports which have crossed my desk in the last couple of weeks pointing out these risks is higher than normal. This is not so surprising with the S&P500 testing the 2000 level but it does suggest that sentiment is cautious rather than ambitious which is interesting at these levels. Considering the consistency of the advance, the integrity of the progression of higher reaction lows remains the most important factor from an analytical perspective. (Also see David’s comment yesterday.)