It’s time for another one of my semi-regular FTSE 100 valuation and forecast updates, using the latest data for both the index’s price and its earnings.
Valuing the FTSE 100 using CAPE
My tool of choice in the valuation game is Robert Shiller’s CAPE (Cyclically Adjusted PE). CAPE is basically the same as the PE ratio that most investors are familiar with, except that it uses 10-year inflation-adjusted earnings as the “E” part of the ratio.
It does that because the 10-year earnings average is a much more stable number from one year to the next than just a single year’s earnings. That’s important because we want something nice and stable to compare the FTSE 100’s price to.
In other words:
CAPE is like measuring the distance between two points on the ceiling while stood on a solid table, whereas the standard PE is like doing the same measurement stood at the top of a wobbly step-ladder.
You’ll get an accurate measurement with one and a dangerously inaccurate one with the other (dangerous if you fall off the ladder anyway).
Calculating “fair value” for the FTSE 100
Today, the FTSE 100’s 10-year inflation-adjusted earnings stand at 530 index points. I’ve put the value into index points because that’s how the index is measured.
So with a current price of 6,700 index points, the FTSE 100’s CAPE is 12.6.
That’s actually pretty low. The long-term average, across multiple international markets, is somewhere in the mid-teens, around 15 or 16.
If the market were at “fair value” today, i.e. with a CAPE ratio of 16, it would be at 8,500. At that level, it would have a dividend yield of 2.8%.
That sounds a little high (i.e. low yield) to me, but not significantly so as the market’s long-run average dividend yield is about 3% (we’re at 3.5% at the moment).
I could just stop there and say that the FTSE 100 is slightly below the typical long-run average of 16, and therefore it’s “slightly cheap”.
But that’s a bit dull, so here are a couple of visual tools to illustrate what the market’s valuation means for future returns.
CAPE is “mean reverting”, and that’s a good thing
First, I’ll take a little detour into the world of “mean reversion”, as that underpins everything else (skip this bit if you already know all this).
The FTSE 100’s CAPE is a mean reverting statistic, which simply means that the further it is from its long-term average, the more likely it is to move back towards that average.
This is a common feature of many financial statistics, such as the house price-to-earnings ratio.
Despite what some property pundits think, house prices are indelibly linked to the earnings of the people who either buy or rent those houses.
If a four-bedroom flat in London is put on sale for £1, there will be a mad rush of buyers who will bid up the price to something more sensible.
If that same flat is put on sale for £100 billion there will probably be silence from potential buyers (or perhaps even a little astonished laughter). In that case, the seller will have to lower the price (a long way) before any buyers become interested.
That’s why the UK house price-to-earnings ratio tends to hover around 4. Sometimes it falls to 3 or lower, as it did in the mid-1990s and sometimes it climbs above 5, as it did in 2008 and seems to be doing again today.
In the long run, assuming that a mean-reverting ratio will be at its long-term average in 10 years or so is probably as good a way to estimate future returns as any. And that applies to any supply and demand-driven asset class, including the stock market and property market.
FTSE 100 CAPE probability “fan chart”
So, having gotten that introduction out of the way, here’s my first visual too, the FTSE 100 CAPE probability “fan chart”:
A few notes:
- The black line – is the FTSE 100 from 1987 to today
- The left axis – is logarithmic, which just means that each step up is double the previous step, which you can pretty much ignore if you want to
- The green bands – represent various CAPE valuation levels from a CAPE of 8 up to a CAPE of 32 (the average, in the middle, is 16)
- Darker greens – are the valuation range where the FTSE 100 has historically spent most of its time. This is where it is more likely to found be in future.
- Lighter greens – are valuation levels that the market rarely reaches. This is where it is less likely to be found in future.
So for example, the central dark green band is where the FTSE 100 would be if it had a CAPE ratio between 14 and 20, which is what I call “normal”. It is also the valuation range where the market is most commonly found.
As the chart shows, in the dot-com boom in the late 1990s, the FTSE 100 was way above “normal”. Since the market is most likely to be found in the dark green “normal” band in, say, 5 years’ time, the most likely move for the market from 1999 onwards was downwards towards that “normal” range.
And that’s exactly what happened, somewhat abruptly, in the 2000-2003 crash.
Another example is 2009 when the market dropped to 3,500 and was way below the normal (dark green) valuation range. The implication was that the FTSE 100 would rise in value back towards that normal valuation range in the next few years.
Once again that’s exactly what happened, as the market leaped upwards in value during the latter part of 2009 and 2010.
So today we effectively have a small (probable) tailwind thanks to a low valuation and CAPE’s mean reverting characteristics.
The market is likely to revert back into that dark green band at some point over the next decade, and since the FTSE 100 is below that level at the moment that means a valuation expansion rather than a contraction.
That valuation expansion would of course be in addition to any earnings growth (typically about 4% a year) and dividends (3.5% a year with the FTSE 100’s current dividend yield).
That is, I think, the most likely outcome, but that’s very different from saying it’s the only possible outcome. The future’s not set and the FTSE 100 can do whatever it likes, within reason (as can the property market).
FTSE 100 CAPE-based future returns “heat map”
Because CAPE is mean-reverting, it influences the future course of the market:
- When CAPE is high – future returns are likely to be low
- When CAPE is low – future returns are likely to be high
- When CAPE is normal – future returns are likely to be normal
With this insight, it’s possible to re-draw the probability fan chart as a future returns heat map in order to highlight the relative “goodness” of low valuations and “badness” of high valuations.
A few more notes for the heat map:
- Red – implies low future returns
- Yellow – implies normal future returns
- Green – implies high future returns
Hopefully, this more clearly illustrates the implications of various CAPE valuations, and the probability fan chart should remind you that these are “likely” outcomes and not guaranteed in any way, shape or form.
We all know how badly the market performed after 1999, and this heat map shows why. Valuations were sky high, and obviously so.
Then again, in 2009 valuations were low and future returns likely to be outstanding (as they have been, as long as you could stomach the fear).
In summary:
The FTSE 100’s CAPE valuation is “slightly cheap” at 12.6 compared to a long-term average across many different markets of 16.
The implication is that in the medium-term (5-10 years) total returns are likely to be slightly above normal, where normal is around 5% total return after inflation (so about 7% if we assume inflation at the Bank of England’s 2% target).
“Fair value” for the FTSE 100 is currently 8,500 (i.e. the value that would give the index a “normal” CAPE valuation of 16), and I think it is more likely than not that the index will reach that level in the next 5 years.
Excellent and enlightening post as usual, John. Thanks for this.
It is rather interesting to see that by the metrics you provide that the FTSE 100 still remains somewhat undervalued. It feels that way through using the yield as a proxy, but to have it backed up by the CAPE is very interesting indeed.
Thanks again. Very interesting!
Hi DD, you’re welcome. Yes the dividend yield and CAPE generally play nicely together and to have a “reasonable value” signal coming from both is good.
Hi John
It really does seem to be a tale of 2 markets at the moment. On the one hand we have the US on a CAPE of over 27 and well above the 80 percentile which is forcing me to be underweight internationally. Then on the other we have the UK CAPE at 12.6 and “slightly cheap”. As they say – interesting times.
Cheers
RIT
Hi RIT, exactly. If I put my nerd hat on I might check the average ratio between the US and UK CAPE and see how that’s changed over time and what sort of percentile we’re at now.
Some investors have asked me how much I think a US market collapse would affect the UK market. It’s a hard question to answer. There would obviously be an impact, but from a valuation point of view I don’t see that the UK market can fall a great deal without rebounding strongly, largely because of its already quite high yield.
I think it would mostly depend on what economic factors were driving the bear market, if we get one.
And even then, bear markets are a good thing for long-term investors as they can buy assets more cheaply than before, so it’s not really something to “worry” about anyway.
Perhaps those who are about to become financially independent might worry, but they shouldn’t be heavily in equities anyway if they’re bothered by capital volatility or if they want to turn their capital into an annuity.
John
“…I might check the average ratio between the US and UK CAPE…” I have accurate data for both going back to 1993. It certainly looks interesting with 3 phases:
– first phase the run up to the dot-com bubble,
– second phase the period between dot-com and the GFC,
– third phase the period since the GFC
I’ll try and work it up into a brief post in the coming couple of weeks so I can get you a view of the chart.
Okay RIT, should be interesting.
I think this is excellent work but I wonder if the FTSE100 is now the most relevant index to use for UK investors when it is dominated by international shares & does not accurately reflect the UK economy. Clearly the world is changing and we are all subject to what happens in the international context BUT a long term investor in the UK has his long term savings/pension needs in sterling assets; Not Euro, $, or other currency assets. As a result I think the FTSE 100 is not the correct index to use for UK investors. Im not sure if there is a 100% UK domestic index but assume that the FTSE 250 is a closer proxy of UK Co Ltd.
This then begs the question as to the valuation of the REAL UK market. You are saying that the FTSE 100 is slightly cheap but it is depressed as a result of strong sterling & weak commodity prices on its international exposure.
Domestic UK shares, as measured by the FTSE250 have outperformed the FTSE 100 by a long way.
I therefore wonder if the UK market is really as cheap as you infer.
Bob P
Hi Bob, thank you, and that’s an excellent point about the FTSE 100 and the UK economy; the two do not necessarily have much in common.
I would say whether a UK investor uses the FTSE 100 as a relevant index depends on their portfolio. If their portfolio is a FTSE 100 tracker then of course the FTSE 100 is utterly relevant and the UK economy is less relevant.
If they invest in lots of smaller, more UK-focused companies then they should care about the FTSE 250 and small cap indices.
So it isn’t necessarily right to say that a UK investor is invested in sterling assets. If they are invested in the FTSE 100 then they are invested in whatever international currency exposure that index has. Although of course, their personal exposure to the UK economy is an entirely different matter.
Personally I use the FTSE All-Share as my benchmark of choice because all of my holdings comes from that index. My portfolio is split about 50/35/15 large/mid/small-cap, so I think the All-Share (which is split 85/12/3 I think) isn’t completely different and is a reasonable benchmark.
As for valuing and forecasting the FTSE 250, I may well have a stab at it soon. I have FTSE 250 data going back to 2008, which is 7 years now. It isn’t enough, but it is almost there and out of impatience I may do the analysis anyway.
The short version is that the FTSE 250, at 17,705, has a 7-year CAPE of 24.6 and a dividend yield of 2.46%. That’s a very different picture to the FTSE 100 and looks more like the S&P 500 valuations.
However, it’s a smaller-cap index, so I would expect it to be on higher valuations generally because of its higher growth characteristics. Also, the data does only cover 7 years, which is primarily through the period of the great recession, so the earnings are a bit depressed.
Having said that, I think the FTSE 100 is clearly the better value of the two indices at the moment.
As to your last point, on whether the FTSE 100 really is “slightly cheap”, it would be impossible to know for sure. CAPE is only an indicator of value after all, and the market could do any number of things going forward, as my probability fan chart hopefully shows.
But I do still think the FTSE 100 is one of the more attractively valued global indices out there, certainly compared to the FTSE 250 and S&P 500.
John
Good and profound writing and not boring. Another value metric, the dividend is high in the FTSE 100 relative to other developed countries (not mentioning the US).
Hi 123456london, thank you.
I agree that the dividend yield is a fairly good indicator of value for market indices. Combining yield and CAPE is a good idea, although I didn’t get round to it in this article.
Thanks for commenting
John
quite reassuring that ftse 100 cape is ‘slightly cheap’ its hard knowing where to invest in an already risen market. im only 15% in the US , europes bumpy with all this Greece thing going on and Japan ok its the yen that’s the problem , should you be currency edged or not???? I think my spare cash can head for the ftse 100. theres no cape for the FTSE250 Is there?
Hi Dawn, sorry no CAPE for the FTSE 250! I might publish an article with some limited 7-year data which shows that index as perhaps being slightly expensive.
As for currency hedging, I don’t bother personally. I just make sure my portfolio generates its revenues and profits from many different countries and at least 50% outside the UK.
I would certainly welcome your work on the FTSE250 but agree with you that 7 years is a bit short. Looking at the 10 year chart the index has been in a bull trend for the entire 7 years and it misses out the previous bear phase from 2007 when the index more or less halved. I would guess that makes it very difficult to make good decisions based on a 7 year time period – but I look forward to your work on the subject as it is very relevant.
Hi John, Interesting adoption of CAPE in the theory. On the housing market, the stats look even more bloated than 5. Given the average house price is about £186K and the average salary is £26K, we are sitting on a fantastical ratio of 7.15X.
This all seems pretty unrealistic to me and it’s crutches, housing benefit, help to cry (sorry I mean buy), funding for lending, alleviated stamp duty etc are all likely to get kicked away as the government piles on more debt. This and even a modest interest rate rise should do the job of mean reverting.
For the stock market, the trick is to find those stocks that don’t mean revert. Fundsmith seems to think they have that formula — time will tell I guess.
LR
Hi LR, I think I’m going to do a similar analysis for the UK housing market, using the same graphics as I think I’ve found a reasonable data set from Halifax. The key difference is that you don’t need to cyclically adjust earnings as household earnings are nowhere near as volatile from one year to the next as stock market earnings.
I shall try to limit my comments on the social cost of this demented housing market and the policies which are driving it!
I am always agnostic of valuations based on historical data like CAPE P/E etc. I have just read Mr Shiller book and he started to distance himself from these ideas.
I prefer to start from the idea that stock prices are a discount of future earnings, they do not price past earnings. One things I learned is that past earnings are not an indicator of future earnings, last example is Tesco.
Hi Eugen,
Obviously I’m going to disagree. I think there’s a pretty big difference between the earnings and dividends of a mature market and a single stock like Tesco.
A “dividend suspension” for the FTSE 100 is unthinkable in pretty much any situation other than nuclear war. So for large-cap, diverse(ish) mature market indices I think past earnings are a pretty good guide to future earnings, and therefore future stock market valuations (within reason of course, we cannot “know” what will come, e.g. if there is an all out nuclear war next year then future earnings will be zero!)
I believe that you did not understand the exact meaning of my first post and it is probably my fault.
I did not imply the “dividend suspension” or the fact that the future earnings of FTSE 100 companies will become Nil. For me in my investment process I do not care if a company pays or not a dividend, I will prefer only to repurchase shares when earning cannot be used for future investments.
What I was saying is that past earnings as measured by CAPE are not a good indication for future earnings of companies neither the future prices that investors will pay for a stock, a regional market (FTSE 100) or a global market (MSCI World).
Dimensional has looked at this extensively and they cannot find any coleration between CAPE and future investment returns. They done the whole exercise, sorted the stocks in five quintile after CAPE etc. This is why they defined the ‘value’ factor based on book value and not on CAPE. Even Mr Shiller in his latest book ‘Irrational exuberance’ (pages 204-205) says that the CAPE is not proven to have ‘statistical significance’.
I am also not sure that ‘smoothing the denominator of the ratio’ is of any help. So, if noise in the denominator (volatile accounting profit) is interfering with the signal about valuation, why not get rid of fundamental accounting inputs altogether?
Inflation is common to both numerator and denominator in the ratio so if we drop accounting profit we could focus on deflated equity price to find our signal. However, if we assume the signal results from some systematic equity-return process, as a function of an adaptive capitalist system, the numerator is better defined as deflated total return, with dividends reinvested. This for example could mean that over each 25 year period the investment return would be 5.5% annual real return.
This offers a noise-free estimate of valuation for a market index (itself a Darwinian construct). Together with assumptions about the strength and time dependency of reversion to trend, this could provide an objective estimate of horizon-specific, future, real returns and variances for that index. When fundamental denominators are distorted, this approach will provide better estimates and when they are not the estimates will be no worse.
Applying this approach to market value suggests that no major developed market is significantly different from normal except Japan which is still undervalued. The implication is that the coincidence of international accounting changes and the credit crisis is still distorting fundamental inputs.
This contrast with an above message from Retirement Investing Today based on fundamental accounting-based measures that the US market, among others, is overvalued.
I think CAPE is a useful I’ll tool, just not the only tool in the investors’ toolbox.
LR
Housing prices in many parts of the UK have started to decouple from earnings. Mainly because not too many people earning a living in the UK are buying those house any longer. Statistic shows that mortgage lending has not increased significantly as you would expect when house prices start to increase significantly.
In the South there is a sway of foreign money buying properties in London, who sends the sellers to buy properties around London who sends retiring people to buy properties as far as Dorset where I live. There is no lending involved in this, as my happy friends of mine’ estate agents told me. Obviously this could change one day and the foreign money could dry up.
There is a new wave now of buy to let investors coming, who are releasing pension funds using the pension flexibility and move into investing in property. This could create a bubble in the property market, expecialy that in this case leverage is used and these new property investors are not experienced. However if the UK economy stays buoyant and GDP is growing, it could go ahead for a few more years.
Properties more than stocks are bought behaviourally. If the economy looks good ahead, I will upgrade my house, try to buy a larger one, based on good future expectations.
Hi Eugen, I see your point, and I know you have to hang out your hat on something and make a decision. I’m always wary of the “it’s different this time” theory and only time will tell, but I’m not convinced that the old formula does not still apply. Still, I wish you luck with your venture.
I sold out of UK property, and a place in the US this year.
LR
LR
I am not saying it is different, I am only saying this bubble (if there is to be one) started differently.
The last one started with offerings of NINJA mortgages, this one started with foreign money, followed by pension money and certainly like every bubble it will suck others in, probably using high leverage if mortgage lending is going to increase.
I am not sure banks will stay cautious for ever. However it takes time for a bubble to form, in property there are three or four main issues that will trigger the bubble to pop: high increase in interest rates (unlikely as there is no inflation as such yet), deep decrease in the output as measured by GDP (possible, but unlikely to go back to recession this year), significant reduction in lending, high number of properties on offer (something like Spanish style).
This leads me to another issue, there are not enough properties on offer at this moment and we have a high shortage of skilled builders, mainly due to builders reducing costs from 2007 and not offering apprentenships to new entrants in the industry.
Let’s get back to CAPE now. I only do one deep value company at one time, and as I finished with Tesco in a profitable manner, I choose another target BHP Billinton.
I said before that I do not invest in mines, but this time I have just considered one. The forward PE is around 9.5, I heard the CAPE is around 12, not sure if this means any good.
It is the type of company everyone has given up on it and for this reason I like it. The management has now a very clear idea, keep only 14 mines which are very profitable and get rid of the rest either by selling them and using the money to pay off some debts or by demerging them and releasing shareholder value.
Both Dr Copper – the only commodity with an economic PhD and the oil price have found some bottom prices and these have been already priced in the share price of the stock.
As with Tesco I do not expect something miraculous to happen, I do not expect mean revertion in the earnings for the company, all I believe is that the earnings forecasted to be met by the company and as a result a decrease in the discounting rate used by the investors which in turn would offer me a 40% return.
As with Tesco, I am prepared for the shares to drop another 20-30% and I will triple my holding if this happens and if the main picture remains unchanged. I actually like events that generate a drop in share prices, I remember with Tesco that the gain I made was after the news that the accounts were cooked and people gave up on the stock completely and allowed me to buy shares at 165-170p to lower my average buying price.
Hello, thank you very much for your fantastic blog
Do you think it is good to invest in an ETF FTSE 100? I like the Vanguard ETF Vuke.
Hi Guastavo, I think that passive investing is probably the right approach for most people, and that for passive investors in the UK it’s entirely reasonable to have some assets invested in a FTSE 100 tracker, whether ETF, unit trust of investment trust.
However, I can’t really say whether it’s “good to invest” in a FTSE 100 tracker as that depends on the specific investment goals, i.e. what sort of risk and return the investor was after.
So whether a FTSE 100 tracker is “good” is something that the investor or their advisor would have to decide.
Gustavo
You are asking for financial advice, something is not within the remit of what John does on this site.
For questions like this you need to contact your independent financial advisor.
Hi John .Just wanted to say your articles and website are great . As inflation is currenly much lower than the average 2% could this affect the forecast above ? I worry that we may have a Japan type scenario ? Any thoughts?
Thanks
Raj
Hi Rajeev, yes inflation will have a big impact, the problem is with knowing which way it will go; up, down or staying where it is.
I pick a 2% inflation rate because the BoE has been able to maintain that, more or less, for quite a few years, and it’s their target so I’d say it has to be the most likely rate of inflation for the future. But yes, if we have Japan-style prolonged deflation the the forecast would have to be adjusted downward.
Unfortunately I don’t think this is something we can know in advance.
John
The Telegraph quoted a Barclays Research figure for CAPE of about 15 for the start of 2015. You’ve calculated it to be quite a bit lower. Where do you get your data and do you have any suggestions as to why their answer’s different?
I’ve found your idea of assuming the market takes about 7 years to correct itself extremely useful, Many thanks! Keep it simple and at all times bear in mind what one’s assumptions are. I think dividend yields are a red herring and a very unwelcome complication. Do you have any statistical evidence that they help? Evidence for CAPE needs quite a lot of data; evidence for a two factor model (CAPE & yield) would require huge amounts of data.
I’m a bit worried to see people commenting on CAPE for Russia or for comparing stocks. In tandem with some statistical evidence, I think it’s important to have some rationale for why CAPE works. I see no reason why it should work for a one sided and periodically lawless stock market like Russia’s. Probably the FTSE minus a few of the weirder mining companies is a more logical subject for CAPE than the whole FTSE, but I expect it doesn’t matter much.
Have you tried calculating CAPE for an equally weighted index of FTSE stocks? The whole FTSE is dominated by a few huge companies. There’s some evidence that a more equally weighted portfolio has lower variance and higher returns.
Hi Gilbert, I have noticed that my CAPE ratio tends to be somewhat lower compared to other sources; I’m not exactly sure why this is, although there are likely to be at least a couple of reasons:
1) I measure FTSE 100 or FTSE 250 once each year, at the end of the year. I assume that Barclays uses a higher frequency, i.e. daily, weekly or monthly. This will have some impact, but my assumption is that more granular data is not better. Knowing 10,000 things about someone rather than the most important 10 things is not – generally – going to give you any more insight into what they’ll be doing 5 years from now.
2) Barclays may use a different earnings figure. My earnings figures are reverse engineered from the official index level and PE quoted by FTSE. I have no idea what value Barclays uses.
3) My CAPE figure for the start of 2015 is 12.5, although of course the exact figure depends on the exact day of measurement. 12.5 and 15 are not a million miles apart, and as long as the same system is used consistently then both are just as likely to prove useful in the long-run.
As for the dividend, I think it’s a weaker signal than CAPE, but still much better than PE. As for statistical evidence, I don’t have any, but it can be built with a bit of work using professor Shiller’s data set:
http://www.econ.yale.edu/~shiller/data.htm
The best methodology I’ve seen is from Andrew Smithers’ book, Revaluing Wall Street, where he calculates annualised forward returns for periods of 1 to 30 years and averages them, and then compares those average forward returns to the market’s stats at the starting point. He settles on CAPE and Tobin’s Q, and throws dividend yield, “shareholder yield” (dividends plus buybacks) and PE out of the window as not being statistically significant. Unfortunately he does not show the data or charts for these cast-aside measures.
Personally I still like the yield as a high yield like we have in the FTSE 100 today can provide a powerful floor underneath the market price, as long – of course – as the dividend level is maintained.
As for Russia, I guess it would depend on the long-run data, but yes personally I would only take that sort of international approach if I had a basket of 20 or more geographically CAPE-based funds.
I haven’t tried to do anything like equal weighting. I think in the long-run it’s unlikely to make much difference, although I agree that an equally weighted portfolio is theoretically better, as long as the additional trading costs don’t eat up the extra return.
John