While there is no one single ratio to truly define a share markets valuation, the price earnings (P/E) ratio is, however, one of the most popular valuation tools. This ratio has been around basically since stock markets have been around. It is simply the current price divided by the last 12 months’ earnings, referred to as “trailing earnings”. When you then compare this ratio against some historic average, you can get an idea of the markets current valuation.
A share price, in theory, is supposed to give you the current valuation of all the future earnings (into perpetuity) discounted at an appropriate (risk-adjusted) discount rate. If the market had perfect foresight, a share price should increase relatively constantly as earnings materialise and time progresses. This is clearly not the case.
While (again in theory) a share price is supposed to take into account all future earnings, the reality is that anticipated earnings over the next 12 to 24 months are all that really matter. So if you could forecast earnings (accurately) over the next 24 months, you would have a very accurate idea if these earnings were discounted in the current price.
Now the real problem with a “straight” P/E ratio (based on trailing earnings) is that the current P/E ratio is calculated on historic earnings. Therefore the current P/E ratio may look very expensive (compared to some sort of historic average), but if earnings over the next year or two are expected (or actually are) well above the last year’s earnings, the share could indeed be cheap.
The net result of this inherent problem is that, unless you estimate future earnings (accurately) and use this in a valuation assessment, you could make a poor investment decision, based only on historic data.
Therefore you can get a better idea of current valuation by using a forward P/E ratio. You take an estimate of future earnings, calculate the forward P/E ratio, and compare this against some sort of average. This method at least takes into account what is anticipated to happen.
Another name for this measure is cyclically-adjusted P/E Ratio (Cape).
This is less important for companies with reasonably stable and predictable earnings, but is critical for the heavily cyclical companies.
Inherent problem with the solution to cure the original inherent problem
The problem with a forward P/E ratio is that you have to estimate the future earnings. If you get this materially wrong, this method does not help you at all. We all know how difficult (impossible) it is to forecast the future!
However, I suppose that trying to estimate the future is a better solution than to just completely ignore the future. I certainly believe that this method gives you a better idea about current valuation levels.
All Share Index
In this exercise I have used market consensus earnings forecasts. Hopefully these are reasonably accurate. I have also used historical earnings, lagged by either 12 or 24 months, in calculating the historical data for the forward P/E ratios. This is important to understand: in a 12-month forward earnings calculation, the last actual 12 months earnings should have been discounted by share prices 12 months ago. The “historical” 12-month P/E ratio in a forward P/E analysis ends 12 months ago. Again, in theory, the share price movements over the immediate year are discounting future earnings. So in a 12-month forward, the history (to get average) ends 12 months ago and obviously in a 24-month analysis, the history ends 24 months ago.
24-month forward analysis
It’s quite clear from the above; the market was very expensive at the end of 2014, based on earnings that actually materialised in 2015/2016. Put another way, the market did not anticipate the extent of the resource earnings collapse that actually occurred. Or put even another way, maybe the market did anticipate resource earnings collapsing, but was prepared to accept this, as earnings could not stay at that level and had to recover. Or put even another way, resource shares could not go negative. They can only fall to zero.
The historical data is not all that relevant, because it is historic. You cannot backdate an investment decision. The 24-month forward P/E ratio is around 14.3 (based on market consensus earnings). The historical average is 12x. While this is on the expensive side, it is at least within a reasonable tolerance level. Also if you think that market consensus is too low and you use your forecast numbers, the market could look cheaper. However, if you think that market consensus is very high (strong rand, commodity prices falling) you may become very cautious about current valuation levels.
Where was the market definitely wrong?
In 2008, for example, you can see the extent of the market over valuation (if you could forecast the actual earnings accurately). The market was clearly anticipating earnings that did not materialise. The actual historic P/E was far too high for the earnings that did indeed happen.
You may ask: But the level at the end of 2014 was much higher than in 2008? Could another collapse happen? Well, we hope not. The level has already dropped from +-23x to +-20x on actual earnings and is anticipated to drop to +-14x in two years’ time, bringing the valuation down to more acceptable levels. So as long as the earnings forecasts are accurate, we should be ok. For what it is worth, I think that the earnings estimates are reasonably accurate.
Oops! We could not locate your form.