Thursday, November 12, 2020

The Case Against Using the CAPE Ratio for Relative Valuation Across Markets

Bloomberg has an article You May Regret Staying Parked in U.S. Stocks which made the case that there’s "widespread agreement" and "the answer isn’t in dispute" that foreign stocks will outperform going forward. 

Simplified version of my view of that statement.... c'mon now. Extended version of my view of that statement is in line with what Jamie Powell outlines here:

So really, it’s hard to argue that US equities are any more expensive than their European counterparts. Sure, the States might have far more stocks at trading at rich prices, but that’s by virtue of its sectoral composition, rather than it being a uniquely American phenomenon. 

But the point of this post isn't to get into the relative fundamentals of US vs foreign stocks, but rather to outline how the author made a basic error with his choice of valuation metric.


Cyclically Adjusted Price to Earnings Can Not Be Used to Compare Two Very Different Markets

The Bloomberg article states clearly that U.S. fundamentals have been superior over the last decade, which in their view is unlikely to repeat.

The jaw-dropping earnings growth generated by U.S. companies in recent years — and which propelled U.S. stocks past foreign markets during the last decade — is unlikely to return anytime soon.

Fair... but then the author uses the cyclically adjusted price-to-earnings ratio ("CAPE"), which averages the real EPS from the last decade, for the denominator in the price multiple used to compare markets.



So what's the problem with the CAPE across markets? 

The problem is the U.S. has experienced "jaw dropping" growth in EPS, while foreign stocks have not.

To the numbers... the chart below shows what the CAPE ratio would be for any given market assuming a constant trailing P/E over time (i.e. if a market traded at 10x, 15x, or 20x and it continued to trade at that exact multiple for the entire time frame). The only thing that changes in the below is the EPS growth from year 1 to year 10. Note I ignore inflation in the below, so assume growth rates are real instead of nominal.



For a company with 0% EPS growth each year, it's easy... the starting and ending multiple is the CAPE ratio. But for a company with positive or negative EPS growth, the answer may be less intuitive. EPS growth makes the CAPE higher all else equal and EPS contraction makes the CAPE lower all else equal.

So... if we assume:
  • Stocks in a market with favorable sectors and growth prospects "should" trade at 20x (and do throughout) and have grown EPS at 10% / year, the CAPE is now ~30x.
  • Stocks in a market with less favorable sectors and growth prospects "should" trade at 15x (and do throughout) and have EPS that have contracted at 2.5% / year, the CAPE is now ~13x.
Let me rephrase this to be clear... in this example the trailing P/E for the favorable market was at a 33% constant premium to the less favorable market (20/15 -1), but the end CAPE ratio of the favorable market appears as a premium of more than 100% to the less favorable market.


Conclusion

CAPE normalizes earnings when those earnings are cyclical. If your view is technology stock earnings are at a cyclical peak and industrial / banking earnings are at a cyclical trough then foreign AND US value stocks look cheap. If not, they very well may not be. 

That's your bet.

Your bet is not that valuations are remotely as extreme the CAPE implies... and there's no disputing that. 


No comments:

Post a Comment