Saturday, December 9, 2006

PEG is a rediculous valuation metric

I have never understood why people use PEG (PE ratio divided by the growth rate) as a valuation metric. It really makes no sense. Here are a few reasons.

First of all, if you knew what the future earnings were going to be, you could value the company with discounted cash flow (DCF). That's the real way to value stocks. Lets give a few examples. We will assume a discount rate of 10% and assume that earnings are to be thought of as free cash flow.

First example. G=30% for 10 years followed by 20 years of 4% growth. The correct DCF PE is 90 and so the correct PEG is 3.
Now if instead we had only 5 years of 30% growth followed by 25 years of 4% growth the correct PE is 38.6 and so the correct PEG is 1.3. Finally lets consider two years of 30%, 8 years of 10% and then 20 years of 4%. That gives PE = 30.1 and so PEG=1.

So obviously the right PEG has everything to do with how long the company will grow its earnings at that initial rate. In most cases the recommendation of buying stocks with PEG < 1 is actually too conservative. In retrospect buying Walmart with PE=100 would have been a very profitable thing to do since it grew quickly for 2 decades.

But there are other things wrong with PEG. It ignores things like return on equity (ROE) and quality of earnings. Maybe companies can grow earnings but will never develop into very profitable companies. Maybe they are growing earnings so quickly only because their initial earnings are so tiny (compared to say invested capital or total assets). For companies that have positive equity and fast growing earnings but very low ROE, it is useless to look at PEG. You should instead analyse the business model and figure out what ROE they will eventually obtain and how long it will take to get there.

Another reason to hate PEG. For slow growth companies, it makes no sense either. Perpetually slow growing companies can be valued with P/E = 1/(DR-G) where DR is the discount rate and G the perpetual growth rate. This requires G < DR.

inverting this E/P = DR -G or DR = E/P + G

Your expected return will be the discount rate DR = E/P + G. So you can see that what matters is E/P + G not PEG=(P/E)/G. You could choose to write this DR = E/P * (1+ G/(E/P)) for your expected return. Now you see that what matters is the
valuation E/P and the ratio of growth to E/P not its inverse P/E. In other words the right PEG is 1/(G*(DR-G)) which still obviously depends on G and DR.

So PEG makes no sense for slow growers and for fast growers it depends critically on how long the fast growth will continue. So why it is used at all?