However, valuation models are poorly understood. So, first, it might help to state what we might sententiously define as the four truths of all valuation models:
■ There are no right or wrong answers in a valuation model;
■ There are only sensible or silly inputs;
■ The major purpose of valuation models is to expose the assumptions in market prices; and from that it follows...
■ Valuation models are only a means to thinking critically about an investment proposition.
If that is true of all valuation models, then it must be true of the most ubiquitous of all - the simple price-earnings (PE) ratio. This one is so much used that we forget it's a model (albeit a little one). The perception is it's just a company's share price divided by some measure of earnings per share. Nothing comes simpler; nothing is more widely used; and maybe nothing comes more misunderstood.
Yet even the common-or-garden PE ratio can be put to good use if investors understand the assumptions that underpin it. To explain, let's take the data for specialist electronics engineer Renishaw (RSW) that we used in last week's models. Once again, it will help if readers have open the spreadsheet -
- on the website.The table contains two little models that show where a PE ratio comes from. The first is derived from a well-known formula - the constant-growth dividend discount model, which capitalises a company's dividend given a specified rate of return required by an investor and an estimate for the long-term growth rate in dividends. If the next full-year dividend that Renishaw pays is 48p, the constant-growth model tells us that the share price should be just 823p if the required return is 8.5 per cent and the long-term growth rate is 3 per cent.
What's important is the gap between the required return and the growth rate and it doesn't take a genius to work out that the gap between the two is the yield needed to sustain the required return. In this case that's 8.5 minus 3, or 5.5 per cent. Given Renishaw's likely dividend for 2015-16, that yield only permits a share price of 873p (48p capitalised at 5.5 per cent). The more that the gap narrows, the lower the sought-after yield and the higher the permitted share price. So the question becomes: what are the reasonable figures for these rates? If 8.5 per cent is a given for all equity investment, then how high can one push the long-term growth rate in dividends without straining credibility?
The table also tells us that the growth rate implied by Renishaw's actual share price is 5.8 per cent - is such a pace feasible over the long term? Alternatively, a required rate of just 5.7 per cent justifies the share price - but is that attractive?
Renishaw
| Inputs | Results | ||
| Share price (P) | 1,775 | Constant-growth value | 873 |
| Dividend per share (D) | 48 | where value = D/(k - g) | |
| Required return (k) | 8.50% | Constant growth PE ratio | 8.4 |
| Growth rate % (g) | 3.00% | where PE = (D/E)/(k - g) | |
| Earnings per share (E) | 104 | Rate of return PE ratio | 14.8 |
| Rate of return on new investment % (r) | 15.98% | where PE = (1 - g/r)/(k - g) | |
| Market ratings | Rates implied by market price | ||
| PE ratio | 17.1 | Implied growth rate | 5.8% |
| Dividend yield | 2.7% | Implied required return | 5.7% |
This discussion is about dividend growth, but it can easily be turned into a debate about the appropriate PE ratio, basically by adding 'E' - for earnings - to both sides of the equation. Thus the PE ratio becomes the function of a company's payout ratio (dividend divided by earnings) and the dividend yield. The more the company pays in dividends and/or the faster the long-term growth rate, the higher the PE ratio. So Renishaw's current PE ratio - 17 times - assumes either a 5.8 per cent growth rate or a 5.7 per cent required return.
An alternative approach is to forecast the 'correct' PE ratio for a company's shares given three parameters - in addition to the required rate and the long-term growth rate, there is the estimated rate of return on new investment, 'r' in the table. Using the 16 per cent cash return on investment that fell out of last week's model, this formula justifies a share price of 1,539p (that's 14.8 times 104p), which is not a million miles from last week's exercise that estimated a fair value of 1,460p.
This model is logical to the extent that it rewards companies for a high return on new investment (effectively, a high return on capital). That said - and much like the constant growth model - its results are increasingly sensitive to the assumed long-term growth rate. Best, therefore, to keep that rate anchored to reality and always to remember that valuation models are there to test the market rather than to tell you what to pay.
