Are US stocks cheap or expensive? The most common reference point is the price-to-earnings (P/E) ratio. A quick noodle on the internet and I discover that the S&P 500 index has a P/E of 26.3…or 24.6…or 23.7…or 20.9. So which is the “right” P/E ratio?
Those numbers above all purport to be something that’s known as the “current” or “trailing” P/E. You get it by taking the current market capitalisation (market value of all shares at the current price) and dividing by post-tax profits for the last 12 months. This is the most common measure of P/E, and provides a snapshot of where stocks are priced in relation to recent earnings.
Current P/E has its drawbacks. If earnings plummet more than stock prices crash – such as during the global financial crisis – it can actually rise. In normal times a high P/E indicates expensive stocks. But in May 2009 the S&P 500’s P/E spiked to 124, which was actually when stocks were near their March 2009 low, and relatively cheap. So you have to be careful about using this ratio in isolation.
One way to get around this is to use what’s known as the “forward” P/E. This uses some kind of estimate of earnings for the next 12 months – or an average of many analyst’s estimates – for the “E” part. But obviously this is flawed too. Stock analysts at big brokers and investment banks – the “sell side” – tend to be eternal optimists, and no one has a crystal ball.
Another approach is something various called the “P/E10”, “Shiller P/E” or “CAPE” (short for Cyclically Adjusted P/E). In this case the “E” takes earnings for the past 10 years (specifically earnings per share, or EPS), adjusts each for inflation up to the present, adds them together and takes an average.
The idea is to smooth out the business cycle to get to some measure of average, underlying earnings power. That average, inflation adjusted EPS is then divided into the stock price to get the CAPE.
CAPE has its drawbacks too. Most obviously it relies on the accuracy of modern government inflation statistics, which probably understate reality by 1-2% a year. And historically speaking – over the long haul since 1900, mostly when stock buybacks were minimal – EPS has tended to grow around 1% faster than inflation. Finally, where there are lots of net stock buybacks, such as the present day USA, past EPS with more shares is understated in relation to current EPS with fewer shares.
Nonetheless, I estimate that each of those effects would only add about 5% to the average “E” used in CAPE, meaning about 15% in all (maximum). According to multpl.com the current S&P 500 CAPE is 28.9. Even reduced by 15% it’s still 25, which in turn is 55% above the 16.1 median (distribution midpoint) level since 1881. Without the adjustment it’s 80% above median. Either way, it looks expensive relative to history.
The current P/E depends on who you ask
But putting CAPE aside, let’s get back to current (trailing) and forward P/Es. Let’s start with the current P/E. Here are the sources for those differing S&P 500 figures that I used at the start:
- 26.3 according to multpl.com
- 24.6 according to the Wall Street Journal
- 23.7 according to Advisor Perspectives
- 20.9 according to State Street Global Advisors, for their SPDR S&P 500 ETF (NYSE:SPY)
If I looked more I’m sure I could find other figures. The point is that the highest figure of this selection is 26% above the lowest figure. Depending which you believe it gives a hugely different signal as to whether the index is cheap, reasonable or expensive.
The higher the valuations today, the lower the future returns that an investor can expect to make. Also, as valuations creep higher so do the risks of a big correction or crash, although you can never predict the timing.
The price “P” isn’t the problem here. It’s easily observable in the market. The S&P 500 index is at 2,334 at the time of writing. It’s the earnings “E” that messes things up.
Should you use reported earnings, after all “one offs” and “extraordinary items”? You know, settlements for litigation, fines for misbehaviour, asset write downs and the like. Or should you use “normalised” or “adjusted” earnings which ignore all that, hence boosting the “E” and reducing the P/E?
At the individual company level it can make sense to strip out most of the one offs if a company just went through a particular set of issues. But at the index level it makes little sense. Someone, somewhere, will always be having a bad patch or screwing up. “One off” charges have a habit of being hardy perennials when it comes to a large group of companies.
In other words, at the index level it makes sense to place more trust in the calculations of P/E that give higher results. These aren’t necessarily pessimistic. They’re just more realistic.
Fortune tellers and forward P/Es
Then we get to forward P/Es. I’m as suspicious of these as a prediction of my personal prospects that’s been laid out by a fortune teller. Let me give some examples of why.
The Wall Street Journal says the current P/E of the S&P 500 is 24.63 and the forward P/E is 18.32. Since the “P” is the same in both cases, this means that the “E” would have to grow 34% when comparing profits in 2017 to those of last year. It’s a similar story for the MSCI USA index, which – according to MSCI – has a current P/E of 23.9 and forward P/E of 18.0, implying 33% growth.
Over the very long run, since 1900, nominal EPS growth for US stocks was a little under 5%, or a little under 2% after stripping out inflation (real growth). Of course that’s an average through all ups and downs. But are there really grounds for believing that this year, 2017, will see profit growth that’s 6-7 times faster than average? I very much doubt it.
Even if President Trump’s promised corporate tax cuts happen – which is far from guaranteed – it’s a massive stretch. Most analysts, including myself, seem to think they might give a short term boost to post-tax corporate profits of 5-10%. That would still leave things way short.
None of the P/E measures is perfect, and the same goes for other valuation ratios as well. Those include: price-to-book, price-to-cash flow, Tobin’s Q, EV/EBITDA, dividend yield, market capitalisation-to-GDP or anything else you care to mention. Your best bet as an investor is to look at multiple measures.
If they’re all pointing in the same direction then it’s a much stronger signal than one measure taken in isolation. Pretty much everything says US stocks are currently expensive (for example see here and here, and note that things have become more extreme since they were written).
There’s no perfect way to value stocks. But investors are more likely to come out smiling if they take a prudent approach.
For this reason I stick to the current/trailing version of P/E most of the time and use the sources quoting the higher, more realistic figures (especially at the index level). Then I compare them with their own history, and make sure I’m looking at other measures too. I recommend you do the same.
Stay tuned OfWealthers,
Our goals are simple. We want to help private investors do two things:
Build wealth. Invest with success.
Twice a week we help thousands of subscribers who share these goals with our free publication, the OfWealth Briefing.
As well as everything published on our website, subscribers receive additional exclusive comment and analysis that is unavailable anywhere else.
New subscribers will also receive several free special reports as soon as they join up. Click on the link below for more details and to start receiving your premium OfWealth content.