“If
you need to use a computer or a calculator to make the calculation, you shouldn't buy it.”

Warren
Buffett

We've already talked about the LTCM debacle
and they were very good with their math. There are entire funds running based
on mathematics today such as Renaissance technologies and other statistical
arbitrage shops. These products are good and probably very few statistical
arbitrage (statarb) shops are able to do this consistently. Even statarb shops
have to sit it out when they are unable to find any opportunities. Then why
does Buffett say don’t do the math?

In the scope of value investing where you
are trying to get as concentrated as possible if the valuation is so complex
that you have to run detailed calculations to prove that its cheap then your
margin of safety is low. Another problem with the math is that the more complex
the math gets the more assumptions get baked into it. These assumptions may or
may not be true. As we have learnt earlier the variables affecting a business
are infinite and cannot be modelled into a spreadsheet.

Let’s run with an example here. If a
company is trading at 10 times earnings with 3 times earnings of debt, with a
return on invested capital of 25% and you think it can double its earnings in 5
years – is it a good buy? Now we can run amok with discounting and prove that
this is probably running at a discount but just for argument sake its do this
with common sense. If it doubles its earnings in 5 years in a linear fashion it
is going to earn approximately 1.5*5 of profits over the period which is 7.5
times its annual profit. This is more than enough to return the entire debt.
Given that its return on invested capital is 25% and it is going to be growing
at approximately 15% (2^(1/5)) which tells you that after returning the debt it
should be paying a dividend of 40% (10%/25%) and after 5 years its earnings
would be 1/5 of your purchase price or equivalent to a bond yielding 20% every
year. This should probably tell you that it’s a screaming buy because you don’t
get bonds that are growing their yield by 2x every 5 years yielding 20% a year!
On the downside if you believe it can at least just generate just its earnings
and you believe they are good capital allocators then the debt can be returned
within 3 years and at the end of the 3 years you could get a dividend yield of
5% to 8% which is also not bad.

Of course there is much research that goes into making the assumption that the company will grow its profits by 2x in 5 years. But those assumptions will have to be subjective and based on the business environment.