This article can be read by selected Monevator members. Please see our membership plans and consider joining! Already a member? Sign in here.
Why you can’t trust the CAPE ratio
![What to do about extreme US market valuations [Members]](https://monevator.com/wp-content/uploads/2025/10/CAPE-ratio-99th-percentile-1024x485.png "What to do about extreme US market valuations [Members]")
Next post: End in sight for renewable infrastructure trusts?
Previous post: Weekend reading: Break glass in case of emergency
Comments on this entry are closed.
Great article.
What’s the antonym phrase for ‘catching a falling knife’. I.e. Holding on to rising assets for as long as possible, as opposed to buying back in at depressed prices?
Typically I’m not very active, but have been thinking of these CAPE valuations as a threshold for rebalancing.
More food for thought, thanks @TA! I wonder how things look if you use a different averaging period (for CAPE)? The cyclically adjusted bit is meant to cover the economic cycle including boom and bust.
I suppose the last 10 years does cover a recession, in that we had Covid but it was far from typical as there was massive stimulus.
On the flip side by averaging I also wonder to what extent the historical earnings are being “left behind”. If there has been strong earnings growth relative to GDP then historic earnings may not be telling you much about future earnings prospects. So may also be worth looking at price earnings or price to forward earnings measures not just CAPE.
An excellent piece (as always) @TA.
Returns come from the future, not the past (i.e. from future EPS and multiples changes).
Starting prices impact that return but over long periods (20 or 30 years plus) those variables drown them out.
TTM and smoothed out PE measures only tell us what happened looking backwards, especially for CAPE 10. That may, or may not, be useful.
Mean reversion is not a law of nature. Sometimes it really is different. Mostly it’s not, but not always. The past rhymes, rather than repeats. And there is no ‘natural’ level of stock prices to earnings.
Moreover, the evolution of the CAPE across it’s whole history in US markets (1871 to date) suggests great sensitivity to the interest rate regime. Higher long term rates, lower CAPE. The Price and the Earnings need to be seen in light of the ‘risk free’ rate.
Does that mean that CAPE shouldn’t be used? No. Context is everything.
Valuations *eventually* matter at the *extremes*. Eventually can end up being a very long time; and extreme can be very much over or under valued indeed.
When (as with Japan in 1989) CAPE hits 90 then it’s perfectly rational to say that the risks outweigh the benefits, even if that turns out not to be the case with hindsight.
But that’s less clear when the CAPE is at 40 (as it now, and as it was once before back in 1999/2000) for the US, as against, say, 25.
CAPE is very useful to help calculate a starting SWR for an equity SIPP. For deciding when to buy and sell shares, not so much.
But again, it all depends.
CAPE peak ‘predictivity’ (R squared) is at 18 years out (where, in the US, it can explain nearly 70% of returns over such a period). At one year out it has nearly no predictivity. Over all periods from one to 30 years out it explained about 40%, and, as @TA notes, at 10 years out then it’s only 38%. Better than nothing but hardly enough to rely upon.
And we also have to bear in mind the possibility (I put it no higher here) that passive flows might have ‘buggered up’ the efficiency of the market in setting prices (the constant bid effect of passive fund flows into regular investment plans – like 401Ks in the US – which buy at the same rate regardless of prices in the market).
NB/PS: when I refer to “1871”, one needs an initial 10 years’ data (from that date) in order to get a first CAPE 10 reading for 1881, when the US data for this metric begins.
Rhino wanted a term for investing during a period of rising CAPE values. Having just come back from a trip to Pompeii and Vesuvius, how about “riding an active volcano”?
@NLC – yes not bad, another thing that popped into my head was a quote from Danny the Dealer in Withnail & I – ‘If you’re hanging onto a rising balloon, you’re presented with a difficult decision. Let go before it’s too late or hang on and keep getting higher, posing the question: how long can you keep a grip on the rope?’
@DH – Thank you! I’m also essentially convinced that mean reversion doesn’t exist. I must write a piece on that sometime. I’m further nodding along to your comment that CAPE seems better suited to modifying the sustainable withdrawal rate. That fits with your observation that CAPE’s significance rises over longer-time horizons (as per @Prospector’s intuition.)
@No longer civil – heh, that’s a good metaphor for investing in general.
@Rhino – if you rebalanced now, could you avoid beating yourself up if the US kept rising for another ten years? I would guess you could but I can’t remember how sanguine you are about these kinds of decisions. I’m sure I could. I essentially never look back at my counterfactual portfolio choices – unless forced to by a Monevator article 🙂
I’m probably more uncomfortable with the concentration of US large caps in my portfolio than valuations per se but solved that with my allocation to global small value.
@Prospector – there are tons of arguments about the relevance of past earnings figures – largely due to changes in accounting standards. I’ve read a batch of studies that adjust CAPE in various ways but they don’t seem to make much difference to the pattern overall. The top line message is generally that adjusted-CAPE brings the market temperature down a few notches but it still looks historically high. I think DH’s point about interest rates likely covers whether investors would be prepared to pay more for earnings at Time A relative to Time B. I agree with you that paying attention to a suite of metrics would make me more comfortable about making a valuation based decision.
Interesting, food for thought.
In my understanding, the S&P500 is expensive, or extremely expensive, by pretty much all metrics. Or are we back to using “eyeballs” or “burn rate”?
An analysis I read recently said that ~20% of US stocks outperformance can be attributed to growth (US economy outperforming the rest) and ~80% to valuation multiple (investors paying more and more for the same cash flow).
Always during bubbles, popular opinion finds plenty of justifications and derides time-tested metrics. The practical problem is that extreme valuations can go on for a long time, and how to predict the turning point. Most people can’t, even the professionals.
Anyway, we should have an upper limit to any individual country in the portfolio. This is for diversification and basic risk management. And this will cap US exposure.
@TA outstanding bit of analysis — worth the price of admission for the scatter plot alone. I was looking at your blue vs red thinking to myself I’d like to make the time independent series and boom you had it. Lovely stuff.
If you were to draw a vertical line on that graph at the CAPE = 40 then we could all go home. The point about the poor prediction and the 25% is that it covers all the mid points where we’re wandering nicely. That broad — one hesitates to say bubble — of spreads in the middle of your scatter plot. If you look at the extremes, it would be a brave assumption that … if the past is connected to the future (pace @deltahedge your point is well made) … the it would be a brave assumption that 10 year returns are strongly positive. That point would be a very long way from home.
In plotting your scatter graph, you presented the data in an insightful and revealing way. No need to torture it to prove the alternative thesis!
@Rhino – perhaps we should all use our hair as our aerials?
@No Longer Civil – tremendous trip: was there recently myself. Naples seems to be a city that has adjusted to imminent destruction by not bothering to tidy its bedroom for two thousand years.