Bond trader Ed Bradford recently linked to a Reuters article about how a recession or extended bear market might impact the stock market’s valuations. The article quotes the opinion of Mr. Robert Kessler, who thinks stock multiples will compress mightily:
I cannot speak to how Mr. Kessler arrived at his conclusion, but I think it would require some extreme scenarios to play out, namely a huge surge in inflation, a deflation-induced credit crunch and resulting financial crisis, or recession so deep that earnings would fall off a cliff.
That being said, I thought it would be helpful to examine the various post World War II recessions to see how the variables of inflation and GDP contraction impacted market valuations. Borrowing some data I found from Jeremy Robson, I, like he, made the decision to exclude the 2007-2009 “Great Recession,” because the data were fairly skewed, and I do not think such an anomalous event is relevant to the current environment.
Here were the results using trailing-twelve month (TTM) P/E:
*It should be noted that I used data from multpl.com, and the “beginning” data in all simulations was approximated to the best of my ability due to not having exact data as of the pertinent start dates, etc.
The results using Shiller/CAPE P/E were not all that different:
Now, the fun part:
I wanted to see how how market valuations were impacted during recession when inflation was above 3.5% (the postwar average – see here) and below, as well as when the “Rule of 20” was at 20 or below, or above (for more on this, see here).
What is surprising to me is that multiples actually contracted more when inflation was less than 3.5% than when inflation was greater than 3.5%. However, I think this is largely due to a higher average starting multiple (15.58 vs 13.45), and is also somewhat skewed by the tech bubble of the late 1990s.
Therefore, I think the better way to look at the data is through the lens of the Rule of 20, which somewhat smooths out the relationship between inflation and P/E, which are theoretically supposed to equal about 20. Then you can see that when inflation + P/E were less than or equal to 20, stocks contracted much less during recession than when inflation + P/E exceeded 20.
What does that say about our current environment? Inflation is running about .75%, though you can say different measuring tools say it’s possibly higher or lower. The current trailing 12 month P/E on the S&P 500 was around 18 at the end of last year. The most recent CAPE/Shiller reading I found was about 24. So using those numbers, we can see that at least using the trailing 12 month P/E of 18 + inflation of .75%, the resulting 18.75 is less than 20. If the post WW2 results are indicative of what may happen if we are currently in a recession, you might expect the S&P 500 multiple to contract about 13%, or to about 1827, which, coincidentally, is slightly above the 1812 level reached at the most recent cycle low intraday on January 20th. Using the CAPE/Shiller data, a 15% retracement takes us to about 1785 on the S&P, thus implying a little more downside from here.
Whatever the case may be, it is difficult to see stocks dropping to the depressed levels asserted by Mr. Kessler simply because any recession would likely be shallow (more on this later), and valuations, when put in the context of inflation and historical precedent, are not so stretched as to imply a massive contraction if recession should occur.
It should be noted that stock prices can decline more than actual valuations do during a recession (see Doug Short’s post on here). It is also true that valuations can contract more than stocks actually fall during a recession. Finally, it can also be the case that stock declines can be very steep even without a recession (see Ben Carlson here). The point I am trying to make is that anyone’s assertion about where the market must bottom to reach “fair value” should certainly be taken with a grain of salt.