In a previous article (‘Stop using the IRR!‘), we explained why the Internal Rate of Return is a completely misleading and useless measure of finance performance; and recommended to use the ROE (Return on Equity), the ROIC (Return on Invested Capital), as well as the peak funding requirement, the payback period and the NPV (Net Present Value).

There is another die hard in business: the P/E ratio for a company listed on the stock market. This is the ratio of the share price to its earnings after tax.

The average P/E ratio of US equities was 14 in the 20th century, and roughly the same in other countries.

So now, this is how the story goes in the media: “P/E ratio lower than 10 – BUY! The stock is a bargain!”. Or “P/E ratio higher than 20 – SELL! The stock is expensive”.

The truth is, buying stocks on the basis of their P/E ratios is like buying stocks when the temperature is below 10°C, and selling them when the thermometer shows 20°C and more. But as much as there is no correlation whatsoever between the price of a share and the current outside temperature, there is no correlation whatsoever between the current P/E ratio of a share and its 1-year or 10-year return.

The proof is brought to you by Ken Fisher, the investment guru, using simple statistical analysis. Take any stock, or any index like the S&P 500, and take 1, 10 or even 100 years of historical data e.g. from finance.yahoo.com. Put the P/E of the stock in one column, and the stock return over 1 year or 10 years in the next column. Now run a simple linear regression, calculate the R2, and look at the results.

If you don’t have time to do that yourself, then trust INVESTAURA, or Ken, and read Chapter 1 of his book ‘The Only Three Questions that Count‘ (Amazon).  Here are the results:

• Taking the S&P 500 over the period 1872-2005, the P/E ratios at the start of each year, and the subsequent 10-year return, the R2 is 0.2. This tells us that P/Es explain only 20% of 10-year returns. Not much really, and certainly not a strong causality. 80% is explained by something else. An R2 of 20% is as good a random.
• If you are interested in a short-term view instead, because you don’t have the patience to wait 10 years after buying a stock, look at 1-year returns rather than 10-year returns.  The R2 is 0.03.  Statistically, this is completely random i.e. P/Es have no correlation whatsoever with 1-year returns.

Furthermore, using the same data for the period 1872-2005, Ken brilliantly shows that:

• 68% of the ‘monster’ drops in the stock market (more than 10% down in one year) happened when the P/E was lower than 16
• when the P/Es were higher than 20, the market ended the year down in 30% of the years; but when the P/Es were lower than 20, the market also ended the year down in 30% of the case! So high P/Es are not more risky then low P/Es.

In fact, P/Es tell us nothing about how the future price and earnings for stocks will develop. Low P/Es can be low because the ‘E’ is high but the company experiences low or negative growth, putting pressure on the price. So why buy a stock when its earnings are going down? But low P/Es can also come from high price and low earnings, but with strong earnings growth potential.

So why are P/E ratios so durable? Old habits die hard. But there is one rational explanation none-the-less.  For a business with constant growth (g) in sale and constant profit margins, one can show that the Entreprise Value (EV) of the business is given by the following formula:

EV / EBIT = (1-g/ROIC) x (1-TaxRate) / (WACC-g)

The Entreprise Value is the sum of the equity value and debt.  EBIT is earnings before interest and tax. So the formula above states that an EBIT Multiple can be use to value the entreprise, and this EBIT multiple is related to: 1. the growth rate to perpetuity; 2. the Return on Invested Capital; 3. the tax rate; 4. the cost of capital WACC.

Let us talk a look, assuming tax = 30% and WACC=10%. The table below shows the EV/EBIT multiples for a range of ROIC and g.

So if you have a business that is sufficiently mature and stable, you can use an EBIT valuation to estimate its entreprise value. And then deduct the value of debt to estimate the value of the Equity.

If there is no debt in the business, the relationship above becomes, after transferring the (1-TaxRate) term to the left side:

Equity Value / Earnings = P / E = (1-g/ROIC) / (WACC-g)

So here they are your P/Es. But note that they can only be used in mature and stable businesses, certainly not high-tech stock, and even less young companies or companies with negative earnings.

The table below gives us the P/E value as a function of the growth rate to perpetuity and the ROIC (assumptions: no debt, WACC=10%).

Mature and stable businesses? None of them high-tech stocks? With low debt? These are exactly the kind of business that Warren Buffett love. But if you invest like Warren, you shouldn’t buy stock with a 1-year or 10-year horizon, you should buy stocks and hold them for the next 20-50 years. Are you really this type of investor?

Warren Buffett does not use P/E ratios, he turns them on their head and looks at E/ P ratios instead.  If a stock has a P/E of 10, then it gives you earnings after tax of 1/10 = 10%. Or 15% if the tax rate is 30%.

Not bad, you say. Especially when bonds only give you 5% per annum pre-tax. So time to buy?

Wait a minute. Warren only buys stocks after careful historical analysis of the business, its economics, its earnings and their historical growth, as well as countless other key parameters. And only if he has reason to believe that future earnings growth in the next 20 years will also be strong. After all, when you buy a stock, you buy its future cashflow and earnings, not what happened 5 years ago. If you want to learn reading financial statements like Warren does, read further.

Warren looks at a share as an ‘equity bond’ i.e. its pre-tax earnings (equivalent to the bond coupon / interest payment), the historical growth of earnings and their relation to the current share price. When Warren bought Coca-Cola in the late 80s at USD 6.5 a share, he bought shares in a company that was essentially generating a pre-tax earning of 11% on his USD 6.5 investment, and had been growing its earnings at an annual rate of 15% historically. 20 years later, Coca-Cola generates an EBIT per share of more than USD 4, so the company is paying Warren a pre-tax yield of 60% (post-tax 40%) on his initial investment. Do you know a bond that increases its coupon from 11% to 60% over 20 years?

Whatever the current stock market price of Coca-Cola is, the Coca-Cola ‘equity bond’ was a fantastic investment for Warren. And whatever the ups and downs on the stock market, the market eventually recognises this kind of performance and increases the price of the shares to a level that reflects their true value. And the value of a Coca-Cola share is around USD 76 in mid-July 2012.

Summa summarum: forget P/E ratios, look at E/Ps, and do a deep dive into the business that you plan to buy. Analyse its historical financial statements, and develop a sense for where the business is going. When P/Es are low, E/Ps might be high, but this is of no value to you if the earnings will decrease in future years!