Registered users can save articles to their personal articles list. Login here or sign up here

Are JSE shares too expensive?

And are we asking the right question?

Now that we have dodged the ratings downgrade bullet, at least for now, and the US stock market has survived Trump (indeed, confounded the prophets of doom – the DJI is up around 7% since the election) we can return to more conventional ways of looking at stocks.

Some commentators are saying that the local market is due for a major correction.  The JSE has been in a broad sideways trend since June 2014.  It’s currently (December 2 2016) around 10% off its April 2015 high (and about 8% above its January 2016 low), but some think the market remains vulnerable.  One of the oft-heard arguments is that JSE P/Es are very high, that such high P/Es are never sustained and presage a fall.

There is never a time in the market when we are without optimists and pessimists.  It doesn’t matter how high or low share prices go, there are always investors on both sides of the fence.  When the market is in a trough, the pessimists argue it will go lower and the optimists talk of recovery.  When the market is soaring the pessimists warn the next crash is around the corner and the optimists buy more shares.

These claims are seldom unsubstantiated.  Optimists and pessimists alike marshal statistics and charts to support their positions.  Which makes it hard, as an investor, to know what to do.

The tail end of 2016 is no different to any other time in history.  On the one hand we have doomsayers telling us that the JSE is dangerously expensive; on the other we have optimists telling us that this time it’s different.  Look at Naspers, they say, a top three JSE stock with a P/E above 70 is unprecedented – it distorts the index.  (In fact, on a free-float basis Naspers is the biggest index contributor.)

The JSE All Share P/E is a key number in these arguments.  Currently (late November), the JSE puts the All Share index P/E at around 23.  This is expensive historically.  The pessimists say that these levels often signal an approaching crash, or at least a major correction.  It is certainly true that, over time, market P/Es return to their long-term averages (which for the JSE, over the last two decades, is somewhere between 14 and 16 depending on the period you use).  When the All Share P/E is above 20 one of two things happen to get it back down to acceptable levels – either strong corporate earnings bolster the earnings side of the ratio, or the market corrects to reduce the price side of the ratio.

No one is holding out much hope for strong corporate earnings in the foreseeable future. Even the optimists agree that profits are likely to remain muted for a while.  Which means, say the pessimists, that share prices have to fall – the current market P/E is not sustainable.  

The counter-argument is that the current P/E is skewed by certain factors and cannot be compared with historical P/Es.  But can this claim be supported?

A market P/E is not necessarily a simple calculation.  An index P/E is usually weighted by market cap, which means you can’t simply average the P/Es of the shares that make up the index.  An unweighted average has some merit – it gives an idea of how much investors are prepared to pay across the board – but it’s not the correct number for historical comparisons.  There’s also an argument for weighting shares based on the level of trade in each share, which is similar to (but not exactly the same as) the free-float principle used by FTSE/JSE in all their index calculations, but again, this number is not available historically.  For the earnings side of the ratio, should you use total attributable income or earnings from continuing operations?

The point of this article, however, is not to argue for one method over another, but simply to present some numbers.  You, dear reader, can pick the numbers you like best.  Based on an analysis of numbers provided by ProfileData for the end of November:

  • A simple average of the P/Es of the 162 shares that make up the JSE All Share index is 18.8, significantly lower than the FTSE/SE figure of 23.
  • A weighted market cap P/E excluding Naspers  is 20.4, a full 11% lower than the P/E including Naspers.  That’s a significant impact, but it still leaves the index looking historically expensive.
  • Excluding the ten shares with the highest P/Es drops the index P/E to 20.0, not a significant change (most of the other very high P/E shares are not large caps).
  • The Alsi Top 40’s 23.2 P/E drops to 20.3 if Naspers is excluded.
  • Of the 162 shares in the index, 16 have negative P/Es and 35 have P/Es above 20.
  • If you exclude loss-making companies (negative P/Es), the weighted P/E drops to 18.2 (15.9 if you also exclude Naspers).

The last point is an interesting one.  Shares with negative P/Es have a significant impact – they contribute to the price side of the P/E (because they still have market value), but they reduce the earnings side of the ratio.  This is largely academic, however – there are always loss-making companies and, as far as we can establish, they have always been included JSE index P/E calculations.  A P/E based only on profitable companies is therefore not a good basis for historical comparisons.

In the same way, there have always been shares with high P/Es potentially skewing the index.  The last time the JSE All Share P/E reached current levels, for example, was in April 1998.  It was only briefly above 20 – by September the market had fallen over 40%.  According to the ProfileData archives, in the run-up to the peak of April 1998 SAB’s P/E went over 40 and Didata’s P/E went well over 100.

As usual in the stock market, we have to go beyond the numbers to get perspective.  Why, after all, is the index P/E important?  What does it tell us?

The theory is that markets see-saw around fair value. High P/Es suggest that investors are paying too much, that assets are over-valued, and low P/Es suggest bargain prices.  It’s important to remember that what causes high P/Es is competition for stock.  Market bubbles are caused by buying frenzies; crashes are caused by selling frenzies.

The JSE’s movements over the last two years are certainly not indicative of a buying frenzy – the market has essentially been in a broad sideways movement since mid-2014.  The All Share P/E averaged 18 from July 2014 (when the last bull phase ended) until March this year (when the index P/E established itself above 20).  Because earnings have been poor and prices have been fairly flat, the market P/E has been rising – since August it has mostly been above 23.

Panic selling is not always foreshadowed by a buying frenzy, even though this is a common pattern.  Selling pressure can and does arise after a prolonged sideways market if investors lose faith.  The JSE’s lacklustre performance since mid-2014, in other words, is not a valid argument against a fall.  But there is little evidence that negative sentiment is linked to high P/Es.  The average All Share P/E in the six months before the crash of 2008 was 15.2.  The average P/E in the months before the forty percent market retraction that started in May 2002 was 12.8.

The fact is that JSE market movements are more closely correlated with the major US indices than our local P/E.  The JSE fell in 2002 and 2008 not because local stocks were particularly overpriced, but because foreign investors precipitated a wave of selling.

The question, therefore, is whether US stocks are expensive.  A number of market commentators, including Bank of America Merrill Lynch, are warning that they are.  The P/E of the S&P 500 index is at its highest level since 2001.  Investors are advised to keep an eye on the US markets – unless history fails to repeat itself, any major decline is likely to start there.



The JSE All Share index (blue line) against the DJIA since July 1979.  Although the correlation is not perfect, it is clear the JSE follows the movements of the US market.  The period from September 1998 to April 2003 might appear to break the pattern, but the smaller movements remain in sync.  The JSE performance looks good, but only because the index has not been adjusted for rand weakness (see second chart).



Translated into US dollars, the JSE All Share (blue line) has dramatically underperformed the US market over the last twenty years.

Nic Oldert is an active equity investor and freelance writer. He co-founded the Profile Group and served as MD for 20 years.

Oops! We could not locate your form.


To comment, you must be registered and logged in.


Don't have an account?
Sign up for FREE

Two factors that should be considered in an analysis of this kind. Firstly, interest rates: the long-term return expected by investors has historically included compensation for the real interest rate that was available risk-free. This has been compressed in this low-growth environment, which perversely, has the effect of pushing up share prices (signalling lower returns in future). Secondly, the average market P/E cannot be calculated by working out the average market P/E. It should be calculated by working out the average earnings yield (ie the inverse), which can then be converted into a P/E. The results differ substantially: the average market P/E based on the average earnings yield is considerably lower than the average market P/E.

No more Comments, leave a reply.


ZAR / Euro



Follow us:

Search Articles:Advanced Search
Click a Company: