- Valuing the market: Are shares cheap?
- Valuing the market by P/E ratio
- The cyclically-adjusted P/E ratio (PE10 or Shiller PE)
- Valuing the market by earnings yield
- The stock market capitalisation to GNP (or GDP) ratio
- Why a higher P/E ratio for the market could lead to higher returns
- Proof that most methods of making stock market predictions don’t work
One way to try to judge whether the stock market is cheap or expensive is to compare the valuation that investors are putting on companies with the output of the economy.
You can do this by comparing the total stock market capitalisation of a country with its Gross National Product ( GNP).
Stock market cap to GNP ratio = (Stock Market Cap / GNP) x 100
Where:
- Stock market capitalisation = The value of all the companies on a particular stock market.
- GNP / Gross National Product = The market value of all the products and services produced in one year by the labour and property of the residents of a country.
Unlike some technical indicators, this ratio makes logical sense to me.
Individual company earnings – or the earnings of entire sectors such mining or retail – can be very volatile. That limits the value of using price to earnings ratios to spot bubbles – even if you smooth them by using a cyclical P/E.
However in a developed economy, it seems intuitive that the economic output of a country and the earnings of its companies – and hence their valuation – should bear some sort of relationship.
That’s where this ratio comes in.
Warren Buffett’s favourite ratio
The stock market / GNP ratio means nothing in and of itself – there’s no rule that the ratio should be X or Y at any point in time.
However by comparing today’s ratio with previous readings over time, we can potentially spot episodes of over- or under-valuation.
I say potentially because the ratio is not perfect, as I’ll explain below.
But it was good enough for Warren Buffett in 2001, who wrote:
The ratio has certain limitations in telling you what you need to know.
Still, it is probably the best single measure of where valuations stand at any given moment.
Buffett was using the measure to explain how he saw the Dotcom bubble developing the late 1990s.
The question is can we use the stock market to GNP ratio to spot new bubbles in advance?
Note on GNP versus GDP: Buffett compares the stock market to GNP. Most analysts use Gross Domestic Product (GDP) instead. In theory this should matter a lot. GDP is defined according to location, being the goods and services produced by a country. GNP is defined by ownership, as I’ve explained above. In practice though the two numbers are very close for the US, and the US is the only country where this ratio seems to be regularly computed. So like others, I use GNP and GDP interchangeably below.
History of the stock market to GNP ratio
When you look back in the history books, most stock market peaks in the US have coincided with an elevated level for the stock market to GNP (or GDP) ratio.
And that makes sense. Excessive profitability for companies tends to revert to the mean, and productivity improves slowly.
Advanced economies as a whole grow at low single digit rates, too.
This all means that big run ups in stock prices over short-periods tend to have more to do with investor sentiment than fundamental changes to the long-term prospects of those companies.
In other words, we’re by turns greedy and fearful. Sometimes we’re like footballers’ WAGs in Harvey Nicks, and no price is too high. At other times we’re haggling for cut-price bargains at a boot sale.
By looking at the ratio of stock market to economic output, you can try to spot the ‘greedy’ times and tailor your strategy accordingly.
For example this article and graph from Smart Money used the ratio to suggest that the US stock market had outgrown the US economy:
Spot the dotcom boom and bust!
Smart Money noted that:
U.S. earnings are near a record high as a share of the economy. Over the past 80 or so years that has tended to mean that earnings are about to shrink.
Thus, the chart. If share prices are constrained by earnings, and if earnings are constrained by the size of the economy, investors might as well compare share prices with the economy directly.
Sounds sensible, but with the benefit of hindsight we can see two glaring problems:
- US company earnings didn’t shrink – they kept growing!
- The US stock market rose over 10% in the subsequent 12 months.
If you that graph prompted you to reduce your holdings of US equities and go to cash, you lost money.
Market timing is extraordinarily difficult, and this is just another reminder.
Problems and limitations with the ratio
Wary though I am of disagreeing with Warren Buffett, it seems to me this ratio is no more useful than most other ways of guesstimating whether shares are expensive or not.
It might give the determined active investor some clues, but it isn’t going to solve the mystery.
Let’s consider a few of the problems.
Little UK data for the stock market to GDP/GNP
It shouldn’t be hard to produce UK data, and no doubt various firms do on a private basis. But I can’t find public sources on the Web.
The ratio has only really been looked at in-depth for the US, which limits its usefulness for the rest of us.
Different countries have very different markets
I think it could be misleading to apply this ratio to other countries too superficially, anyway.
For example, one Seeking Alpha writer wondered in 2011 which stock market was the most expensive by this ratio, and produced the following graphic:
Click to enlarge
The author noted that:
There are several reasons determining the size of a country’s stock market capitalization including, for example, the equity-buying culture of the local population, company use of debt versus equity financing, the country’s regulatory and legal environment as well as how easy it is to list on the local exchange.
In general however, it is probably safe to assume that there should be a positive correlation between a country’s market capitalization and its GDP.
Indeed it turns out there is a very high, 0.96 correlation between the world’s largest economies GDPs and their stock market capitalisations.
I agree with the second paragraph, but I’d question the utility of the third.
The problem is the variations behind that average ratio are so wide. For instance, the author found that at the time of writing, the ratio for Italy was 0.3 and for Germany it was 0.43, whereas for the UK it was 1.5.
Did that mean that Italy or Germany was 4-5 times cheaper than the UK?
I don’t think so.
The UK stock market is home to many companies from around the world. We’re also one of the most capitalistic of the major economies. Our listed market is a bigger deal than the equivalent for Germany or Italy, which are home to far fewer foreign firms and where there are surely more companies in private or even State ownership.
You would need to look at a particular country’s ratio over the long-term to derive any useful insights into its market valuation versus economic output at any particular time.
And comparing one country’s ratio with another seems entirely spurious.
What about overseas sales?
This is an important factor, which most of the articles I’ve cited also mention.
Apple, Google, and many other big US companies earn much of their money overseas. This means that comparing their valuations with US economic output could be misleading. Why should they be limited by US growth?
Of course, global trade is nothing new. I don’t know off-hand how the proportion of earnings generated by US companies overseas has changed over the years.
I know it has changed, though, and it is another factor to consider.
The ratio is often elevated, and thus arguably useless
The biggie. Look at the graph from Smart Money again. The red line is mainly above the blue line, which means the US market has most often tended to look overvalued.
History may yet prove this to be the case, but over long multi-year periods it’s been a pretty poor signal to get out of stocks.
The ratio is even more stretched if you go back to the 1950s:
Click to enlarge this long-term market cap to GDP ratio
(Source: Vector Grader)
This is a really interesting graph, because it suggests the US market might have been very over-valued for a couple of decades.
I have previously explained why I think carefully looking at long-term returns suggests that ain’t necessarily so.
Which view is right? Who knows – history is unfolding before our eyes. My best guess is the answer lies somewhere in the middle.
Clearly the US market was a steal in the 1950s and 1960s (when Warren Buffett started getting rich, incidentally). This was when the ‘Cult of the Equity’ was born, and investors began to shift wholesale from bonds to equities.
However I think globalisation and various other factors have probably changed the ‘fair value’ level for the ratio since the 1950s.
In addition, with interest rates very low – and fears of inflation not abated – it is arguably rational for investors to pay more for the earnings of companies today.
You’ll have to make your own mind up. As always, I certainly wouldn’t use this ratio in isolation from other factors.
Remember, most methods of predicting the returns from equities have a truly dire track record. Be warned!
Comments on this entry are closed.
The Investor,
Another fascinating blog entry.
Monevator seems to be “cooking with gas” at the moment. 🙂
Bob.
I have formed my world view. We’re going to have recession with non-negligible inflation for a few years more; eventually there will be recovery accompanied by higher inflation. So what should I invest in to protect my capital?
I can’t tell if “So like others, I use GNP and GNP interchangeably below.” is deliberate! Either way, it’s pretty funny. 🙂
You have a really good point about globalization and the relevance of Market Cap-GDP ratio.
I used PE10 to determine if the U.S. market is over or undervalued. Do you think it is a better metric?
Mr. Asset Allocation,
Well at least for the US, valuation metrics can be a useful indicator of long-term stock market returns.
That might make them a useful guide for those who are passive investors wanting to tweak asset allocations as they invest a flow of savings for the long-term.
See following chart and article –
http://advisorperspectives.com/dshort/charts/guest/BP/130412-Fig-7.gif
http://advisorperspectives.com/dshort/guest/BP-130412-Valuation-Based-Forecasts.php
The chart suggests that long-term returns in the US stock market will remain poor.
Bob 🙂
@rjack @Bob — Follow my very last link in the last paragraph of the article for more on PE10’s predictive ability. It’s better than most commonly used methods — some are actually worse than guessing! — but it’s not exactly great.
@Greg — Oh dear. 🙁 Embarrassing and fixed now. Let’s hope everyone else thought it was an obscure attempt at wit. 😉
The GDP measure is another useful one but, even though I a big fan and user of these long-term forecasting tools, I think far to many people use them incorrectly.
After reading a bit of an argument in the comments on Greenbackd, on a post on this GDP topic (don’t have the link to hand), I was amazed at seeing people argue over which is ‘best’.
Here’s the point – They (PE10, PE30, Tobin’s Q, this GDP thing) all “work”, but only in the same vague sense that it will be warmer in summer than it is in winter (hopefully).
They are predictive, but with huge uncertainty and within a huge and unpredictable range of outcomes.
For some reason people get obsessed about KNOWING what the market’s future returns will be, when all we can ever really get is a vague idea of what we think a reasonable outcome could look like.
I wrote a post about this showing the frankly colossal range of possible FTSE 100 values over the next 15 years or so, implied by CAPE. All we can say is that most of the time it’s nearer the central value than it is to the extremes. But exactly where it will be, at what point, is simply un-knowable even with something like a plus or minus 20% range of error.
So I’d say pick a long-term measure, use it, realise that they are all very weakly-predictive, and move on to something more worthwhile like picking good investments, or restoring a classic car.
Hello Folks,
Link to the article mentioned above by John.
http://www.ukvalueinvestor.com/2013/02/stock-market-forecast-for-2013-to-2020.html/
http://www.ukvalueinvestor.com/2013/01/an-alternative-to-stock-market-forecasts.html/
http://www.bullbearings.co.uk/traders.views.php?gid=2&id=405
I find John’s fan chart a very useful reality check about what constitutes reasonable expectations for long-term returns.
Bob 🙂
The ratio ignores relativity to alternative investments. Thus in a low interest rate environment like today or the high interest rate environment of the early ’80s, the ratio is out of context.
Hello Folks,
More thoughts on this matter –
(1)
I found this chart of total stock market valuation to GDP ratio for the USA that goes further back in time than the one The Investor has used.
http://www.ritholtz.com/blog/wp-content/uploads/2012/05/4-13-12-Market-Cap-.gif
Observations: Up until 1987 the LEVEL of this ratio was a pretty good indicator of whether one should bias long-term retirement investing towards the stock market – or take a harder look at other assets.
However, after 1987 the “quantum leap” in this valuation ratio would have kept investors out of a very profitable long-term bull market lasting 13 years.
Looking more closely at the chart, rather than using the level of the ratio to make long-term passive investing judgements, it might have been better to look at changes in direction in its trend. This could be done by applying some strong smoothing to the data.
For example, soon after 1968, if somebody was looking to invest a flow of savings over a 40 year period for their retirement, they would notice that the trend in the ratio had turned down. This would indicate to them that the long-term trend of the actual stock market was likely to be sideways and that returns would be poor.
They might then resolve only to passively invest their flow of savings into the stock market when it was lower than 12 months before. This would mean they were investing near the bottoms of the many short bear markets that occur in these long-term sideways eras in the market.
By 1985 the same investor would have noticed that the trend in the market/GDP ratio had turned upwards – and that it was time to be serious about directing their flow of savings in that direction.
Around 2002 they would know to “pull in their horns” again, because a super-smoothed ratio would definitely be trending down.
(2)
I’ve also found this chart, which is similar to one that The Investor uses above, but that goes back further in time.
This might be useful for those interested in long-term trends.
http://greenbackd.files.wordpress.com/2013/03/gdp-wilshire-total-market.jpg
Bob 🙂
Enjoying the discussion everyone. 🙂
@George — I agree, mentioned that towards the end. It’s arguably very rational for investors to bid up assets in a low interest rate environment. (On the other hand it may be irrational because it implies a depressed economy).
@John — Well said.
@Bob — Yes, I saw that chart but I didn’t particularly like the discussion on The Big Picture for that particular post (they almost change their mind from one day to the next on whether the ratio is meaningful) and I didn’t like all the text scrawled all over the graph.
Can we go easy on the links please? I am happy with one or two, and I agree some are useful, but in general I’d rather not see the comments become a link fest, for various reasons.
One plot I’d like to see (which could be done for any share) would be a time-averaged returns plot made as follows:
– Pick a timescale to look at the returns for, e.g. 5 years.
– Pick a timescale fraction to average over, e.g. 20% (1 year in my example)
– Build up a dataset made by looking at what the line graph would be if one had invested at the specified time plus x days, where x ranges from -timescale/2 fraction to + timescale/2 fraction. this would give a whole range of different performances.
– Normalise them to a common start point. (i.e. shift in time and starting value to be the same as the reference line.)
– Plot them all, or perhaps make a fan of line density. (Imagine each time-slice as a histogram of returns viewed from above.) Note that the chart would make less sense when its shorter lines end. (e.g. 6 months before the current day in my example.)
This would probably be quite useful as it would smooth out a lot of the volatility while still showing it as the width of the fan. It would also be a decent thing to compare with the other time averaged observations. e.g. a CAFTSE100 VS CAPE.
Hmm, mow I’ve written it, I don’t think I’ve explained it very well. Perhaps it would need actually doing, then explaining. Not that I’m in a position to comply, time or data-wise.
Greg
As an aside, I’m all for links that back up an argument. If you are worried about SEO, surely having links to relevant article would boost the rating, not lower it?
@Greg — Just briefly (on iphone in bus! 🙂 ) I’m not worried about SEO, all the links in my comments are tagged NoFollow so it does no harm from that perspective. It’s more a general discourse thing.
In any one case a link is fine but added up they overwhelm. There were 5 in very quick succession here. As a reader it starts to look like an ordeal.
In some hands (not here but you see it often enough) you get a sort of link warfare bombardment taking place.
Many sites don’t allow links in comments at all. I’m not at all suggesting that. I’ve just noticed a couple of (very articulate and interesting as it happens) readers putting up an awful lot of links in recent days on a few posts.
Just wanting to ask for moderation on this please. 🙂
Stock market valuation aka tea leaf reading has been shot by more silver bullets than dracula… its in a ditch… Somewhere there is a splendid museum of the failed silver bullets… the coppock indicator is next to the Gilt-Equity Yield Spread… remember that?
The road reaches every place, the short cut only one.
I might be wrong, but I believe the Economist had a graphic on stock to GNP in the UK within the past year. You might try there and I’ll see if I can pull it up… Great insight here, as always.
I wonder if overseas earnings and the weakness of the pound are the most important factors here. According to the latest edition of HL’s “Investment Times” around 2/3 of of the FTSE100 revenues are generated overseas. The fall in sterling versus the US dollar over the last 5 years equates to a 32% increase in earnings when converted to pounds, without any ‘real’ increase in the earnings of the constituent companies.
If GDP doesn’t meet George Osborne’s expectations over the next few years then tax receipts will be insufficient and debt could reach unsustainable levels. This would probably lead to a default/inflation scenario, causing the pound to fall further.
How does this affect the stock market cap/GNP number?
@BeatTheSeasons — Personally I wouldn’t over think any one variable. There are millions at work here, mostly captured in either aggregate GDP or in the valuations put on companies by investors. All those investors know well the risks the pound faces, which have been aired for years now. So you’d imagine that’d be factored in. In my view, unknowns and especially sentiment will play a much larger role I suspect.
@Mike — Would like to see that graph. I’ve had a hunt around, alas to no avail!
@OldPro — I think QE did for the gilt-equity spread indicator. Although having said that, the market has gone up an awful lot since rates came down an awful lot, so while it was failing in 2008, you could say it’s firing now?
Definitions –
Alchemist:
An Alchemist is somebody seeking how to transmute base metal into gold by discovering the appropriate combination of elements.
Investor:
An investor is somebody seeking to transmute base savings into huge profits by discovering the appropriate combination of metrics.
Bob. 🙂
60% of UK profits are declared in dollars implying most profits are earned outside the UK. The relationship between the stock Market and it’s local economy is declining.