Buffett and Coca-Cola, 1988.

5 04 2010

In his post here, Joe Ponzio trys to understand Buffett+Coke in terms of the F Wall Street model. He attaches the PDF of his analysis, showing that if Buffett had assumed a 21.8% growth rate for the first 10 years, the F Wall Street model shows a value of $22.3 billion. However, when I do the same analysis, I get a value of $18.7 billion. I can’t see his spreadsheet, so I really don’t know why there is a difference between our numbers. When I do the same analysis using my own model, I get a final value of $17.4 billion. Not a big difference, however, let’s look at what happens when we use a different discount rate- say 9%.

Because the final stage of my model is based on cash flows rather than equity, my model accurately reflects a change in the discount rate across all three stages. Joe’s, however, is only affected in the first two stages. Using the same growth rates and a 9% discount rate, my model shows a value of $44 billion (2.5X more than the 15% discount rate!). With the 9% discount rate, Joe’s model shows only $34 billion, or 22% less than mine.

So which version did Buffett see when he looked at Coke in 1988? We know that Buffett uses the long-term treasury yield for his discount rate because it’s risk free (he compensates for risk by using a margin of safety). The average treasury yield for 1988 was 8.6% so we’ll use that. In my opinion, Joe’s 21.8% growth rate is pretty ridiculous. The nature of his model, along with his high discount rate, require that you assume an absurd growth rate to make the KO investment seem sensible. I think that Buffett probably used a more conservative 12%ish growth rate for the first ten years. Using these parameters, Joe’s model says KO is worth $19.7 billion–only a 23% discount from the average 1988 market price for KO. My model shows KO to be worth $23.3 billion–a 35% discount from the average market price of 1988 and well above the “25% margin of safety rule” for big companies.

Note that if Coca-Cola had less economic goodwill (ie. required more equity to produce the same cash flows), Joe’s model would actually add to the valuation. That is, a crappy company would have a higher valuation than a great company. My model, on the other hand, isn’t affected by the company’s return on equity.

To make a short story shorter, Joe’s model overvalues companies with low returns on equity and undervalues companies with high returns on equity. Since Joe is a great thinker and analyst, I’m sure that he can see through the relatively minor difference between his model and the real value of the company. However, why use a model that obfuscates the evaluation?




3 responses

8 04 2010

Hey Adam! first time to your site, and I have to say that I’m enjoying myself =)
Anyways, as an avid reader of Fwallstreet, it was weird for me to find someone saying that he’s mistaken in his valuation method (as interesting as it really is).

so, to understand the difference between you and him fully, i’d love to ask some questions:
1. why use the discounted value of the 20th cash flow instead of SE? I understand that it turns out quite similar at most valuations, but I just can’t seem to understand the logic behind it. What’s the relation between the money accumulated/invested in the company until this moment (SE) and the cash flow in 20 years discounted back to today?
2. looking at joe’s valuation, I actually don’t understand fully how he calculated the 21.8% from the multi year components, so I’ll check it out once I’ll get home and post another message.
3. why did you use 12%? is it due to the difference in valuation or is it just a random number?
4. you wrote: “Note that if Coca-Cola had less economic goodwill (ie. required more equity to produce the same cash flows), Joe’s model would actually add to the valuation.” – isn’t this shown through the use of percentage (in FCF)? (if it used higher amount of equity, still it would need to earn more to get the same %) or through the CROIC? -I’m not sure about this, but it does require some clarification.

sorry for the many many questions, but i’m the kind of want-to-know-it-all person =)

to be clear – right now I’m studying Economics in my university, hoping to live my life value investing.

Thanks upfront! (so much =)


8 04 2010

Hey Ziv,

I know I said it in the post, but I have to reiterate again that my model and the F Wall Street model are barely different. I love the F Wall Street blog/book and the vast majority of my strategy is from those two resources.

Now, to answer your questions…
1.) I use the discounted value of the 20th year’s cash flow projected out to infinity rather than adding the value of the company’s equity. Why? Because it’s consistent with the axioms of value investing. I wrote a blog post about those here: https://reasoninvestor.wordpress.com/2010/02/12/value-investing-axioms/

Basically, the value of a company is the discounted value of the cash it will produce-not necessarily how much cash it has.

2.) Joe calculated the 21.8% by taking the median of the 5 and 7 year growth rates.

3.) Like JNJ, I chose 12% because it is just a much more realistic number than 21.8%. If you look at Joe’s PDF, you can see that the lowest growth rate for a 5 or 7 year period was 13.1%. If we had used the 13% growth rate, the company would appear even cheaper!

4.) Let me illustrate with an example: Take two companies, both earning $200 a year in cash. For simplicity’s sake, we will assume they amount they earn doesn’t change from year to year (this doesn’t affect the principles involved with the example).

Company A has $1,000 dollars in equity meaning it earns a 20% return on equity. Company B has $2,000 dollars in equity meaning that it earns a 10% return on equity. You can see that Joe’s model will produce a higher value for Company B than Company A even though they are producing the same amount of cash, and company B actually has a worse return on equity.

21 04 2010

Thanks for the swift reply! it was great.
But, I still don’t understand the logic behind using discounted value of the 20th year’s cash flow projected out to infinity rather than adding the value of the company’s equity. It just seems that there are two different views – some think that the intrinsic value of the business is it’s past and future cash flows and some think that it’s only the future ones.

The main question here is why do the 2 different valuation methods come out in the end with similar numbers?

as for point 4) – it’s true that the valuation will be higher, but the CROIC will have to be lower – same FCF and more SE. I guess it’s still safer to use a valuation method that deals with the problem directly…

It’s just all comes to the question above.

Thanks again!


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: