The S&P 500 is currently experiencing a huge bull market, but it’s now clearly expensive on long-term valuation metrics like Robert Shiller’s CAPE. When I last looked the S&P 500 was up to 1,980, and at that level it has a CAPE ratio of 26.
That may not sound very interesting, but the S&P 500’s CAPE ratio has averaged just 16.5 over the last hundred years which, if I were to be sensationalist about it, would mean that the S&P 500 is more than 57% overvalued.
A statement like that is enough to strike fear into the heart of many a value investor, or cause much rolling of eyes from sceptics, bulls and growth investors.
The problem is it sounds so definite: 57% overvalued. It makes it seem like the S&P 500 is about to drop back to fair value at any minute, which would mean the index dropping all the way back down to 1,253. That’s a decline of almost 40%!
I don’t think statements like “57% overvalued” are helpful, but on the other hand I’m not saying that the S&P 500 isn’t expensive, because it is. However, there are better ways of getting the message across, such as:
The S&P 500 is expensive but that doesn’t mean it can’t keep going up
The S&P 500 is in the midst of a massive low volatility bull market. It’s quite spectacular as you can see:
Since clearing the old “resistance” around 1,500 in January 2013 the index has shot up by almost 30% in just 18 months or so.
Volatility has also been low and getting lower, producing high returns with little “risk”; what more could investors want?
The thing about bull markets is that they produce positive feedback. As shares go up investors think they’ll keep going up, so they buy more, which creates demand which drives up prices, and so on and so forth until there’s nobody left to buy.
So what’s gone up keeps going up, and that could easily be the case here. The peak CAPE value for the admittedly bonkers dot com boom was an eye watering 44, so perhaps there’s room for another 70% of upside yet, though I doubt it.
The S&P 500 is expensive which means that valuations are probably likely to contract in the longer-term
That’s not the sexiest sub-title in the world but it does represent a more reasonable and possibly more useful statement than “the S&P 500 is 57% overvalued”.
57% overvalued makes it sound like the market is destined to return to its long-term fair value, but in reality it isn’t.
We may never see the S&P 500’s CAPE at 16.5 again. Investing isn’t physics and there are no valuation “laws”. But we do have a lot of data going back over a long period of time and it would be foolish to just toss that in the bin.
So let’s have a look at what sort of values CAPE has had for the US large-cap index over the past century:
I’ve highlighted where it’s at today in black. Hopefully this makes the situation somewhat clearer. Rather than simply saying that the current valuation is 57% above the mean we can see what sort of range of CAPE values the S&P 500 has had over a 100 year period.
Amazingly enough the range goes all the way from less than 5 to more than 40; that’s quite a spread.
But within that spread it’s pretty obvious that the market spends most of its time with CAPE values somewhere between 5 and 30, and within that it spends most of its time in the 10-20 range.
I could also mention that while the mean CAPE value is 16.5 the median is lower at 15.4, which means the index has spent about 50 years out of the last 100 below 15.4; compare that to today’s level of 26. Or to belabour the point, the S&P 500 has spent 88% of its time with its CAPE value lower than it is today.
The S&P 500 is expensive, so future returns are likely to be below average
Of course what investors really care about are returns; valuations are just one way of uncovering companies and markets that may offer good risk adjusted returns.
So what does a CAPE valuation of 26 imply for future returns?
Let’s start with a few simple assumptions. Real growth of earnings has been close to 2% for the S&P 500 over the past century, so I’ll assume that will be the real rate of growth over the next decade. Inflation has averaged 3% during that time, so I’ll assume that going forward too.
Putting those together means that a reasonable ballpark assumption for the growth of average S&P 500 earnings over the next decade is 5% a year.
That would take the 10 year cyclically adjusted average earnings from 76 index points today to 123.7 index points in 2024 (i.e. if the S&P 500 is at 1,980 index points today and its CAPE is 26, then its cyclically adjusted earnings are 1,980 / 26 = 76).
If cyclically adjusted earnings are 123.7 index points in 2024 then what sort of values could the S&P 500 have at that time?
Again, it’s down to the assumptions that you make.
I prefer to make market projections over 10 years because over 10 years the relationship, or correlation, between today’s price and the price in 10 years is almost non-existent.
Because the two are not linked, and because we have no way of knowing what level the market will be at in a decade, a reasonable course of action is to assume that the likelihood of CAPE being at any particular level is the same as it was in the past.
For example we can just assume that the odds of the S&P 500’s CAPE being 17 in 2024 are the same as they have been in the last century. Of course the odds won’t be the same because history doesn’t repeat, but it does rhyme and so making forecasts and projections based on past data is a reasonable thing to do.
Taking all of those assumptions into account, the estimated odds of the S&P 500 having any particular value look like this:
So as you can see there is a tiny chance that the market will be below 618 in 2024, and some tiny chance that it will be above 4,948. The most likely of these various ranges is 1,237 to 1,855, where there is a 29% probability of the market being between those two values.
That’s not exactly encouraging and the chart suggests that there is a near 50% chance that the market could be lower in 2024 than it is today. That would be especially unpleasant given that the S&P 500 spent most of the previous decade failing to make new highs as well.
I’ve highlighted the range that contains today’s value of 1,980 and clearly there’s a good chance that the market could still be at that level 10 years from now.
None of this bodes particularly well for long-term investors of large-cap US stocks.
Of course if you’re optimistic then you could always argue that we’re in a long-term bull market and that there’s a 50% chance the market could be higher in 2024 than it is today.
You could even be really optimistic and say that there’s a chance the S&P 500 could reach 5,000 by 2024. But it’s a pretty small chance.
I don’t like when metrics are used in isolation. Myself I don’t pay too much attention to PE anyway. In 1980 Coca Cola was trading at a 21x PE and that did not stop the company shares to grow at 13% compound.
E from the PE ratio is determined by some strange accountancy rules, some of them very different from 20 years ago. Some expenses are deducted immediately, some capitalised and amortised over the years, however from time to time some write-offs occur. IAS 32 and IAS 39 have brought more confusion than help in my opinion. As a result I tend to avoid this metric.
I prefer to use the EV metric and cash-flow metrics. For me it is not the same if the cash-flow is used to buy-back shares (I hate dividends as they are a nuisance) or used in the business to keep the business afloat / competitive.
To give an example I bought Facebook at $31 shares at a PE ratio of 85 and I still own some shares now worth $73 at a PE ratio of 120. However when I bought the shares I paid a 40x multiplier of the free cash-flow as the free cash-flow was higher than the profits (accountancy issues related to the cost of share options exercised by employees on the IPO). We had normalised earnings in the year ended in 2013, the earnings have doubled in the year ending in 2014 and probably will increase again by 60-70%. However the free cash-flow has disappeared because of the acquisition of What’s app. My target is around $100 per share.
Your metric falls foul in IPO cases, because you don’t have an historical P. So you can’t realise that Royal Mail is sold cheap by UK Government.
There is also the effect of interest rates and QE that inflate the real asset prices. Unfortunately QE is like drugs, the US may stop it, Japan , Europe and China will widen the taps. Obviously this could end up in a financial meltdown, but turning off the taps will mean a financial meltdown as well.
listening recently to warren buffet when asked if the US market was expensive ? he replied The US market was ‘reasonable’. so not expensive . and he said ‘one thing for sure stocks will be higher in 10 years time.’
Hi Dawn, it’s interesting to hear him say that, but I’d still disagree – although of course as you can see I don’t like to have a fixed view and prefer to think of things in terms of probabilities. I would say that there is no better than a 50% chance (approximately!) that the US market will be higher in 10 years.
Hi John,
Great article as ever – few people are willing to take such a longitudinal view of the markets.
One question. How do you reconcile your projections for low future returns for the US market with your more optimistic projections for the UK (in your June FTSE 100 update you summarised thus: “overall though I think the odds are likely that we’ll see better than average returns over the next decade”).
Hi Sean, it’s all in the valuation.
US stocks are high relative to their historically “normal” valuations (CAPE specifically) while UK stocks are on the cheaper side of normal. So assuming both markets have similar future mean valuations as they have in the past, then the US is more likely to face valuation ratios fall, which will of course reduce returns, while we’re more likely to see them rise. Plus we get a higher dividend yield which is a significant contributor to total returns in the long-run.
In the short-term though it’s anybody’s guess, but investors shouldn’t be focusing on the short-term anyway.
But doesn’t the FTSE tend to follow the US market – i.e. high correlation – or do you think it can rise irrespective?
There is some correlation , but just look at the last year: FTSE 100 up 5%, S&P 500 up 20%. It’s a pretty loose sort of correlation and so each market is more or less free to do as it chooses.