We’re halfway through the year so that can only mean one thing: It’s time to re-calculate fair value for the FTSE 100 and dip a toe into the mystic world of stock market forecasts.
FTSE 100 valuation: Almost at fair value
After an incredibly long and slow bull market, the FTSE 100 is, at last, getting very close to fair value.
“Fair value” is of course a subjective term. Personally, I like to measure it using Robert Shiller’s cyclically adjusted PE (CAPE) ratio.
If you haven’t heard of CAPE before, it’s the ratio between an index’s current price and its average inflation-adjusted earnings over the last ten years.
Fair value occurs when the CAPE ratio is close to its long-term average, which I estimate to be around 16 for the FTSE 100 (based on data from the FTSE 100, S&P 500 and other major indices).
As I write, the FTSE 100’s value is 7,378 index points and its inflation-adjusted earnings have averaged 493 index points over the last decade.
The ratio of price to earnings (7,378 to 493) is 15.0, so:
- The FTSE 100’s CAPE ratio is 15.0
15.0 is close to but slightly under the long-term average CAPE ratio of about 16, which means the FTSE 100 is close to but slightly under fair value.
Fair value, in this case, occurs when CAPE is 16, so:
- Fair value today for the FTSE 100 is 7,896
7,896 is only slightly above where we are today, so clearly the FTSE 100 is not in bargain territory. However, it also isn’t obviously overvalued, despite an eight-year bull market.
This isn’t always the case though. Sometimes an index’s value can reach as high as double fair value (as it did in 1999), or fall as low as half fair value (as it did in 2009).
In the long run though, current value and fair value never stray too far from each other, as the chart below shows:
The chart clearly shows the close relationship between the FTSE 100’s current value and its fair value.
My favourite analogy is that current value and fair value are connected by an invisible elastic band. It goes something like this:
When the market’s current value is far from fair value, there is a strong invisible force pulling it back towards fair value. This force can be overcome for a while by strong animal spirits, such as optimism during booms or pessimism during busts.
But strong animal spirits can only be sustained for short periods of time, perhaps a few years at most. And when those animal spirits weaken or reverse, the invisible force of mean reversion takes over and the market is pulled inexorably back to fair value.
In my opinion, mean reversion is the closes thing we have to gravity in the investment universe, which is why it’s central to all of my market forecasts.
FTSE 100 one-year forecast: Down, but not by much
First of all, I want to make a couple of points clear:
- I have no idea what the stock market is going to do over the next year and neither does anybody else. Rather than trying to actually predict what the market’s going to do (which is impossible), these forecasts are a useful way to visualise the opportunity or threat implied by the gap between the market’s current value relative to its fair value.
- Technically this is a projection, not a forecast. But most people use the term forecast to mean any sort of future-gazing, so I’ll stick with that informal term.
This forecast is based on the assumption that the market will halve the gap between its current value and fair value over the next twelve months. This means that a larger gap will result in a larger correction (either positive or negative), whereas a smaller gap leads to a smaller correction.
I think that assumption is reasonable and approximates what happens in the real world.
We can calculate the FTSE 100’s fair value for mid-2018 by assuming a couple of things:
- That inflation runs at 2% (the Bank of England target) and
- that the FTSE 100’s current earnings of 237 index points (based on its current price of 7,378 and its PE ratio of 31.07) are the same at the end of 2017.
Those two assumptions produce a figure for cyclically adjusted earnings in mid-2018 of 457 index points. That’s slightly down from the 493 we have today.
A forecast for falling cyclically adjusted earnings means that fair value is also forecast to fall.
Why would cyclically adjusted earnings and fair value fall rather than rise?
The reason is that cyclically adjusted earnings are based on a rolling ten-year average of earnings, and ten years ago we were at the height of the credit bubble.
Back in 2007, banks and commodity-related companies were earning vast profits and those two groups make up a significant chunk of the FTSE 100.
In inflation-adjusted terms, FTSE 100 earnings were 690 index points in 2007. Today they’ve fallen to just 237.
That is a massive decline and as the boom-year profits fall out of the ten-year average, they are being replaced by the much smaller profits that are being generated today.
That, in a nutshell, is why cyclically adjusted earnings and fair value are falling.
Given the lower level of earnings, we have this result:
- FTSE 100 forecast fair value for mid-2018 = 7,317
So the forecast is for fair value to fall by 7.3% from today’s fair value of 7,896.
It’s also 0.8% below today’s price of 7,378, so even if the market moved straight to fair value it would still decline in this scenario.
However, my forecast assumes that the market will not completely close the gap to fair value. Instead, it will close the gap by half over the next year.
So rather than CAPE moving from 15 (its current value) to 16 (fair value), it only goes halfway to 15.5.
As a result, we get:
- FTSE 100 forecast value for mid-2018 = 7,077
And so, after much ado, my forecast suggests that a reasonable guess for where the FTSE 100 will be in mid-2018 is 7,077, some 4% below its current value.
Not very exciting, but incredibly useful
I must admit that forecasting a 4% decline is not very exciting. In fact, it makes me think this forecast is kind of useless.
But it isn’t.
This forecast, however dull, contains within it some incredibly useful information:
First, that the market is very close to historically average valuation multiples, so investors should expect historically average returns.
Over the last century that’s been something like a 5% real return from the UK stock market. Or using the dividend growth model, adding the current dividend yield of 3.8% to the long-term real dividend growth rate (which is about 3%) gives a real return of almost 7%, or 9% nominal if we assume inflation remains close to 2%.
I think a 7% to 9% total return sounds reasonable (and perhaps even a little optimistic), so FTSE 100 index-tracking investors should not expect more.
Second, if the FTSE 100 continues to grow by 10% or more per year then it will soon become overvalued. This increases the risk of a bear market and reduces the dividend yield and long-term expected returns.
Third, even though this forecast may be dull, it is the habit of doing a regular comparison between the market’s current value and its fair value that matters.
It matters because one day the market will not be close to fair value. One day, like in 1999 or 2009, it will be very far from fair value, and it’s at moments like those that great fortunes are made or lost, and small ones as well.
John
interesting analysis.
curious to know – do you do this analysis for other main markets – other than the s&p 500?
it seems like a useful perspective – wont be right every time, but a good guide to the prevailing winds/ tides.
best
david
Hi David, I do this for the FTSE 100, FTSE 250, S&P 500 and the UK housing market. You’re right, it’s also entirely applicable to most well-established stock markets, but the cost/benefit ratio gets much worse outside of those core markets (in other words, I could do it for the German stock market, but I doubt many readers come to UK Value Investor for information about the German stock market!).
Actually, why not? I have investments in Europe, Asia, Japan – I’d love a commentary like this on many global markets – I don’t just invest in the UK.
Hi Phil
Okay, good point. Perhaps I (selfishly) mean the cost/benefit ratio is worse for me because I don’t invest outside of UK markets.
The best I can offer is this rather excellent global CAPE ratio tool from Star Capital:
http://www.starcapital.de/research/stockmarketvaluation
Thankyou.
No problem Dawn. I’m here to crunch the numbers so you don’t have to.
Very useful. Thanks.
You’re welcome.
I like the elastic band analogy.
Would you jot simply use the PE10 for the FTSE 100?
Hi Dave, yes PE10 and CAPE are very similar, the only difference being that CAPE is inflation adjusted. Actually a better measure than both is either PE30 or Tobin’s q, but CAPE (and/or PE10) is good enough for me. Having done some back tests with these different ratios the difference is marginal.
John
Great article, as ever.
You state “Or using the dividend growth model, adding the current dividend yield of 3.8% to the long-term dividend growth rate (which is about 3%) gives a real return of almost 7%, or 9% nominal if we assume inflation remains close to 2%”.
This is Gordon’s Growth Model. But why have you assumed this calculates the nominal return, rather than the real return? Surely the 7% is the real return?
Welcome your comments.
Many thanks
Hi Jonathan,
I think we’re getting confused here somewhere.
First, I should have written “long-term real dividend growth rate”, because the real dividend growth rate over the last 30 years has been 3%. So that’s an error which I have now corrected.
Adding that 3% real dividend growth to the current dividend yield of 3.8% gives a “yield plus growth” real total return of 6.8%, or 7% to be imprecise.
Add inflation back in (at 2%) and that gives a nominal “yield plus growth” total return of 9%.
Some of the confusion was probably caused by me missing out the word “real” in the first place, so hopefully this all makes sense now.
Hi John
Many thanks for your reply.
That would make sense mathematically, yes.
However I’m surprised to see a long term real dividend growth rate of 3%. That looks more like a nominal growth rate to me.
What is the source of your data for the long term dividend growth rate please?
Many thanks again
Jon.
Long-term data for the FTSE 100 is quite tricky to come by, but luckily I found some figures for dividends going back to 1986 in the book “How to value shares and outperform the market”. I don’t know where that book got its figures from, but they look pretty plausible to me. The book has a table of dividend data from 1986 to 2011 and the later values also tie in with those on the FT website ( https://markets.ft.com/data/archive ).
For example:
The end of 1986 the nominal FTSE 100 dividend figure is 49.4. I have the average inflation index (CPI from the ONS website) for 1986 at 47.1.
At the end of 2016 we have 260.6 for dividends (from the FT data archive “World markets at a glance” report) and 100.7 for CPI (indexed to 100 in 2014).
Over those 30 years (1986 to 2016) nominal dividends went from 49.4 to 260.6, which is a nominal growth rate of 5.7% (using the RATE function in Excel).
CPI increased from 47.1 to 100.7, which is an annualised inflation rate of 2.6% per year.
5.7% nominal growth minus 2.6% inflation = 3.1% real dividend growth.
That’s how I’ve worked it out anyway. It does seem like a lot, but it’s not far off the historic norm of about 2%. However, trees don’t grow to the sky, so whether we’ll get 3% real growth in the UK or globally for much longer is an interesting topic in and of itself.
Hi John
Many thanks again.
Mathematically your analysis stacks up.
However, I would suggest using the same time interval for each component of the analysis for consistency. So if you use CAPE 10, for example, then analyse the other variables over the same 10 years.
Applying this idea to the dividend growth rate gives 4.3% nominal. My source is the excellent Capita Dividend Monitor, which shows FTSE100 dividends paid in 2008 of £61.6bn and 2017e of £90.6bn, I ignore the difference between regular and special dividends, which is largely cosmetic – cash is cash, after all! (You can quibble that this is only 9 years not 10). To convert this to a real rate we need an inflation assumption, say 2%, so the real dividend growth rate becomes 2.3%. If you think about it, the FTSE100 players pay out much more of their earnings than companies in other major indices (e.g. S&P500), so are investing less in future growth. It therefore makes sense that the growth rate should be lower.
Using Gordon’s growth model, as you have, suggests an expected real return based on the last 10 years data of 2.3% (dividend growth rate) plus 3.8% (current yield) = 6.1%. .
In the current climate, a 6% real return is not to be sniffed at and I think is a more realistic forecast of likely future returns.
Welcome your thoughts, as ever!
I can see your point about using a 10-year period when looking back at inflation and growth, but I still prefer longer-term averages because I think they’re likely to have a higher correlation with future inflation and growth rates.
However, the levels of uncertainty are high enough that in my opinion it doesn’t really matter too much which figures you use. As long as they’re reasonable, i.e. close to historic norms of one sort or another, then and projection or forecast based on them is likely to be as good as any (reasonable) other.
Your estimate of 6.1% real return fits into this reasonable spectrum of projections I think, and is close to the very long-term (>100yr) average UK real return of about 5% (from the Credit Suisse Global Investment Returns Yearbook).
And that makes sense, since the FTSE 100 is close to average valuations, we should “expect” close to average future returns.
Nice article. There is one thing that caught my eye. “In inflation-adjusted terms, FTSE 100 earnings were 690 index points in 2007. Today they’ve fallen to just 237.”
Given that markets are forward-thinking and the British Pound won’t go lower vs. major currencies (especially those who have investments in the UK) shouldn’t we sell the FTSE 100, as 10-year earnings average are going to take a tumble?
Hi Walter, that’s an interesting point. FTSE 100 10yr average earnings have been declining for a couple of years already, However, what happens next is anybody’s guess, so I certainly wouldn’t suggest selling the FTSE 100 on the assumption that they’re going to fall so far as to make the index look expensive.
Personally I don’t expect 10yr average earnings to fall that far, but I could be wrong.
As an American, I’m more bullish on UK equities than US equities due to the big difference in Shiller/CAPE valuation. And I’m invested pretty solidly in the UK for that reason.
The US has a much larger technology sector so a price premium is warranted, but it’s inflated right now while the UK is pretty fair.
Emerging markets look reasonable at this time as well. I like your forecasts. 🙂
Hi Lyn, thanks. Your position is about the same as mine. The US looks pretty expensive (although not in what I would call bubble-territory) while the UK mostly looks fairly reasonable, although its mid-cap index (FTSE 250) is getting a bit on the warm side from a CAPE point of view.
Also, I like your article on CAPE. No valuation metrics are perfect, but other ratios you might want to look at (and probably already know about) are Tobin’s q and PE30 (or CAPE30).
Personally I’m happy with CAPE and its flaws. The market is so volatile that any attempt at forecasting precision is little more than an entertaining folly.
Extremely informative and thought provoking.
Thank You
Hi John,
The problem with CAPE analysis in today’s world is it’s oversight of interest rates. Shiller wrote a revised edition in 2015 presumably because he expected another crash soon (2nd edition 2005, 1st edition 1999) and he talks about psychology as a major factor and on that basis the S&P should have crashed because no body believes it’s value which leads me to think that we will see a massive melt up when everyone realises (intellectually or not) that the markets are acting differently this time – and for good reason.
Professional analysts use discounted cash flows and NPV analysis to price assets and many of the hurdle rates have fallen due to the lower risk free premiums (10-30 year bond rates).
Many asset classes are historically higher than usual for that reason. I read dr. Ed’s blog too which might be a new area of interest for you as he follows this line of thinking – it’s called the FED model. http://blog.yardeni.com/
Buffett say’s stocks are cheap…and as always, I think he’s right! He values businesses in the same way (DCF) and uses 30 year bond yields as the discount rate for his asset purchases to set a benchmark.
The obvious issue here is how long will the environment remain as low inflation and low interest rate(without speculating or second guessing the FED/BOE)? That is where your CAPE comes in handy because we all know where the normalisation of assets sits. The conundrum is not how expensive the market is but when will things normalise?
We only have to look at Japan who has fixed rates for 10 years to know that it might be a long way off… My answer is to keep investing in all markets regardless of CAPE and justify asset prices on the basis of their individual fundamental analysis. If the asset is below fair value in the US with a discount rate of 8-12% then why wouldn’t you buy it? Then look at how that price action moves as you reduce the discount rate to 4-7% which (unfortunately) may technically happen if this extraordinary situation continues.
Admittedly I don’t condone low rates this but I really think you should take this into consideration in your portfolio allocation going forward as you may miss opportunities. I think the FTSE is very cheap in relation to the above and I wouldn’t be surprised to see it become as overvalued based on historical values as the S&P500 in the next few years all things equal.
Cheers
James
Hi James,
Excellent comments, thanks. Here’s my position:
I think a) the US is probably overvalued and the UK is probably undervalued and b) low interest rates are probably helping prop up high valuations.
I think Buffett is probably wrong about US stocks being cheap.
As for CAPE and interest rates, Shiller has apparently looked at CAPE’s predictive power when interest rates are taken into account and they didn’t really make much difference. I assume that’s because interest rates mean revert like most things, although of course the time for rates to mean revert in the future is completely uncertain.
So for now I’m going to stick with plain vanilla CAPE and measure it relative to its historic norms. I’ll let other analysts worry about exactly when interest rates revert to normal, or by how much!
That’s absolutely fine – different ideas make it all so interesting.
I guess the simple way of looking at it is that a CAPE ratio of 16 means that you tolerate returns of no less than 6.25% each year (1/16 = 0.0625) where as other people have toned down their expectations to 3.33% per year in the USA with a hefty CAPE of 30 (1/30 = 0.033).
Natural human responce given the interest rate environment but wholly dangerous. Either way I tend to focus on individual business valuations like you do and invest on that basis. We can’t control the market but we can control our purchase decisions.