• John Hawver

Amazon, Parbuckled

Updated: May 25, 2019

Berkshire’s purchase of Amazon stock is intriguing. Warren has moved more into technology with his purchases of Apple and Red Hat stock. And at the 2018 meeting, he said he would have purchased Microsoft stock had it not been for his friendship with Bill Gates. Yet the purchase of Amazon strikes me as just different.

Let’s first compare Amazon to his current portfolio, which I’ve taken from Guru stocks. I’ll use a technique I’ve applied before in a previous post, K-means analysis. In the plot below I’ve grouped the Berkshire portfolio by PE and Market Cap. You can see Amazon in the upper right corner. Amazon is a definitely an outlier compared to the “older” parts of Warren’s portfolio, but so are Apple and Red Hat (currently being acquired by IBM).

The question becomes, why did Berkshire buy Amazon stock? What does Warren (or Ted Weschler and Todd Combs) find appealing?

First, let’s look at Amazon’s valuation. With a PE of 80 it doesn’t seem cheap, so let’s try to do a proper valuation. I do valuations slightly differently than the norm, let me explain why. If you read books on valuation, like Investment Valuation by Aswath Damodaran (great book), you find that for every individual company you have to make a variety of assumptions. These assumptions tend to vary by company, sector, etc. Good analysts who follow these techniques become like artisans, painting very detailed story-filled pictures. I take a different approach.

Being a quantitative researcher, I don’t want to spend my evenings valuing every company in the investable universe. I’d rather be with my family. Or play hockey. Or do a lot of other things rather than being a spreadsheet monkey. So, I take a “Henry Ford” approach and automate all valuation models by sector and industry, normalizing assumptions based on each company’s characteristics. In this way, I can do multiple valuation techniques, not just discounted cash flow, but Graham, ratio comparisons and a few others. What I lack in precision I make up for with volume. At the final step, I take the results of all these models and regress them versus current market prices to determine which weights I should apply to each model. Clear as mud?

And here’s what I get for Amazon: $1625 per share.

Now, that’s just my number and, as described, it could be wildly off. Using Graham’s technique I get (pause) $164. I doubt Mr. Graham envisioned the internet, so maybe we shouldn’t apply much weight to that (my regression doesn’t). Amazon doesn’t have a dividend, so we can’t use a dividend discount model either.

And then there is DCF, Discounted Cash Flow. If we apply some liberal assumptions to Amazon's EPS growth rates, we can get a valuation as high as $2975. And if we get more conservative, that valuation can be as low as $625 per share. I prefer the “average” assumptions, which place it at $1890. As you can see, It all depends on the “artistic” assumptions that you make.

All valuations are relative. Either relative to how other similar assets are priced or relative to future cash flows. Given that, what should Amazon be compared to? It’s a deceptively hard question. Book stores? I think Jeff Bezos has grown past that. Grocery stores? Ebay? I’m not really sure. Amazon is an outlier on a few dimensions so let’s compare some of the financial metrics to the big boys, Google and Microsoft.

From the above plots, you can see that Amazon compares favorably. It is a cash generating machine. Based on its steadily growing book value per share, revenue, and gross margin it has a very solid economic moat.

An interesting feature of a stock’s price are the bounds; how high and or low it can go. Every company can go bankrupt and if you’re an equity holder the value of your shares can easily go to zero. Conversely, the share value can also grow, theoretically, to infinity. Bonds, on the other hand, while also bound on the downside at zero, have a limited upside; when you invest in a bond, if you hold to maturity, you know your exact future returns.

Amazon ain’t no bond. The valuation question really revolves around its EPS growth rate, can that be sustained? 102% growth rate over the last 5 years is extremely impressive. If the next five years are as good, Berkshire made another great investment.