I just completed a post discussing the pros and cons of the Shiller PE. After reading it I thought, “Geez, as much as I like discussing cyclically adjusted earnings, this is a boring read.” While some of you would like it, I think it would put too many readers to sleep. Therefore, I decided a more interesting way to discuss the topic was to publish an email exchange I had with a friend and fellow absolute return investor.
Discussion of the Shiller PE with fellow absolute return investor…
Me: Many articles and studies on stock market valuation state today’s market is the most expensive in history, except for the 2000 peak. After looking at the long-term Shiller PE chart this appears true. However, is it an apples to apples comparison? Specifically, during 1991-2000, central banks were not artificially suppressing real interest rates through asset purchases. What do you think about calculating a Shiller PE using a historical average real rate of around 300 bps, or similar to what we experienced in 1991-2000? In effect, making the periods more comparable.
Absolute Return Investor: To be clear, your point is if real interest rates were near average, borrowing costs would eat into EPS, hence a higher Shiller PE? If so, I’m in favor of any ratio that shows valuations more on an operating basis. I’m curious to know how you would calculate. Just take interest expense/debt for every company and add 200 bps?
Me: I was thinking 200-300 basis points and using the S&P 500’s aggregate debt to determine the higher interest expense. The yield on high grade corporate bonds during 1991-2000 (period used to calculate the 2000 Shiller PE) averaged approximately 7.6% versus 4.4% during 2008-2017 (period used in today’s Shiller PE), or a 300 basis point difference. The 10-year Treasury’s yield averaged approximately 6.5% in 1991-2000 versus 2.5% in 2008-2017, or a 400 basis point difference. Inflation during these periods was approximately 2.7% in 1991-2000 and 1.6% in 2008-2017. Therefore, in real terms, rates used in the Shiller PE calculation in 2000 were approximately 200 to 300 basis points higher versus 2017 (depending on if you use corporates or 10yr Treasury). For what it’s worth, real interest rates in 2000 were near long-term averages, while in 2017 they are significantly below (goes without saying!).
Absolute Return Investor: Before you go through the hassle of calculating, you might want to run P/EBIT or EV/EBIT to see the output. I know you’re not stripping out taxes but those stats are probably easy to find.
Me: Right. A Shiller EV/EBIT would be nice!
Absolute Return Investor: As you already know, you will have to preempt the argument that today’s low rates make a higher PE more acceptable.
Me: True. And I’m also aware real rates and profits will never normalize again! 🙂 But my point is since real rates are significantly below average and are artificially suppressed, the Shiller PE may do a good job of normalizing earnings over ten years, but not normalizing real interest rates (there’s nothing normal about 0% real yields!). It just seems this market is more expensive than 1999-2000 (especially broadly speaking), but based on the Shiller PE, it is not. However, based on my calculations, most of this discrepancy appears to be due to lower real interest rates.
Absolute Return Investor: It’s a good idea. We made some calculations on median PE. I occasionally see references to median PE being higher than the tech bubble, but wanted to run the numbers myself. Verified. But your idea takes it a step further. The only issue with doing EV is it gets messed up with all the financial companies in the index. And that if you increase rates for borrowing costs you should also inflate bank profitability. I’m not saying I agree, but I often omit financials from my analysis like these.
Me: Good points and makes sense. My calculation also wouldn’t include the impact higher real rates would have on operating profits of non-financials. Maybe P/S remains the best aggregate valuation metric, which weeds out a lot of this and of course verifies what we already know – stocks are expensive! I suppose debating valuations and what year stocks were most expensive doesn’t really matter at this point, especially given positioning (almost all cash). It’s more out of curiosity. I’m confident my opportunity set is the most expensive it’s ever been…just trying to get a better understanding why some valuation measures are more expensive than others.
Absolute Return Investor: Plus, the Shiller PE is going to lose its utility in a couple years when the last recession’s earnings fall off. Have you thought about a modified Shiller PE where the trailing period fluctuates to include two full business cycles? Or at least one? An arbitrary 10 year time period doesn’t hold up in extreme environments.
Me: That’s a great idea – including one or two complete cycles instead of a fixed 10 year period. I remember in 2015-2016 thinking the Shiller PE was including two periods of earnings booms and only one earnings bust. It wasn’t balanced or complete cycles.
Absolute Return Investor: Exactly. So what’s your adjusted Shiller PE look like?
Me: After normalizing real rates and adding 200-300bps to the S&P 500’s interest expense, I’m getting a Shiller PE of 39-46x, which is quite a bit higher than today’s 30x and similar to the 1999-2000 bubble peak. The calculation is an approximation as I don’t have a database on aggregate S&P 500 debt and I’m still without a Bloomberg. I backed into the S&P 500’s debt by using published debt to equity and price to book ratios. Again, not that any of these aggregate valuation measurements influence decision making…just helps support what we already know, but I think it’s interesting nonetheless.
Absolute Return Investor: Ha! You’re like an analyst operating using smoke signals without a Bloomberg. Very impressive. I believe it.
Me: Life without Bloomberg…it’s like that Seinfeld episode when Jerry shares his “moves” with David Puddy but withholds them from George. When George discovers, he gets upset with Jerry and yells, “I’m out there rubbing two sticks together while you’re walking around with a Zippo!”
Thanks for your feedback. Very helpful!