Here is a valuation one you might appreciate, dated 8/30/2019
TMF: Re: Fair value conundrum / Berkshire Hathaway
"It’s great to really understand the math of this kind of calculation.
But personally, I recommend never using a DCF model for any equity investment decision or value calculation.
It’s correct (for the variations without impossible assumptions like perpetual growth), but it’s far too easy to get wildly varying results with tiny changes to inputs which are hard to choose, so the results generally contain less information than you started with.
My conclusion is that valuation rules of thumb work better.
Here’s one that tries to cover all the bases:
A stock is a good buy if you pay less than 10 times the average real annual EPS (measured in today’s dollars) 5-10 years from now.
It’s an OK buy up to 12 times that number.
This handles everything from slowly fading cash cows to moon shot growth companies.
But you have to be 90% sure that the real EPS will be as high as you’ve estimated.
Examples of 10 and 12 multiples:
A firm with real EPS falling at 2%/year (flattish in nominal terms) might be worth a current P/E of 7.8 to 9.4
A firm with real EPS flat might be worth a current P/E of 10 to 12
A firm with real EPS rising 5%/year might be worth a current P/E of 14.8 to 17.8
A firm with real EPS rising 10%/year might be worth a current P/E of 21.6 to 26.0
Figures above that become moot, as it’s pretty much impossible to be 90% sure that any firm’s real EPS will grow at something like 15%/year for that long.
Incidentally, this rule of thumb contains a lot of hidden subtleties that I think are useful.
For example, using the average in a five year time range emphasizes that it’s pretty likely there will be one bad year in the average.
I think the distance in the future is about optimal, especially in terms of cancelling out the “high growth is worth a premium” factor.
(high growth is indeed worth a premium, but one should never assume above-average growth rates further into the future than this model does)
There is no temptation to think about whether market multiples will be high or low at any particular time in future, nor whether the stock price is rising or falling right now.
It eliminates from consideration any firm whose earnings trajectory isn’t (for you) reasonably predictable for that time frame.
Inflation matters, sometimes a lot.
Ignore current year earnings, except to the extent it helps you estimate the future. Earnings might be unusually good or unusually bad at the moment.
It doesn’t have a varying quantity like current bond interest rates as an input.
There is also no terminal multiple explicitly assumed…though it implicitly assumes above average growth beyond the 10 year mark is so uncertain that it is given no value.
The same calculation can be done for things you own. If you need to sell stock to raise cash, sell the one with the highest multiple calculated this way.
Exercise for the alert readers:
What multiple of average real EPS 5-10 years from now does today’s price ($304800 or $203.20 per B) represent for Berkshire? Is it under 10-12?
What multiple of average real EPS 5-10 years from now does today’s price ($1193) represent for Alphabet? Is it under 10-12?
For a high growth firm it’s very tough to come up with a number you’re 90% sure of, but an interesting exercise nonetheless.
You can do it backwards.
With Alphabet’s price around $1200, the average real EPS 5-10 years out has to be around $100 for the multiple to stay under 12.
To hit that figure from today’s earning rate of around $55/year, real EPS would have to grow at a
rate of 7.6%/year or better to make an “OK buy” today according to the 12x rule of thumb.
They’d have to manage earnings growth of inflation + 10.1%/year to meet the “good buy” threshold of 10x."
