Generals: High-Level Valuation Analysis

Generals: High-Level Valuation Analysis

September 23, 2016


  • Recent PPS: $804.70
  • Shares Out: 490 million
  • Market Cap: $394,303
  • Net Liabilities: $5,579
  • Enterprise: $399,882


Amazon is an unbelievable business case study; and I would put the compilation of Bezos’ annual letters up there with Warren Buffett’s as 110% required reading for anyone and everyone interested in business and investing. Likewise – as in the early days of Buffett’s Berkshire Hathaway, where a masterful owner/operator was staring down decades of virtually unconstrained growth potential, A) it does not pay to bet against the stock, and B) one should have a working fair value/entry point on hand at all times for when the market begins to doubt the long-term outlook.

At the moment, however, the market is far from doubting the long-term outlook for AMZN; and I believe the expectations currently embedded in AMZN’s stock price are likely to disappoint recent shareholders for the foreseeable future. Again – I will never bet against the stock; but as a potential long-term owner I want to be cognizant of the expectations embedded in the stock price before taking a position.

This is a very short high-level look at AMZN’s valuation designed to provide context…which in my opinion is sorely lacking. In other words, too much in the way of ephemeral TAM estimates, indefensible EBITDA multiples (indefensible for lack of a better term…I mean most EBITDA multiples are just assumed, and not tied back to a DCF-backed valuation), and nonsensical definitions of free cash flow.

17M AMZN Analysis September 2016


Yesterday I enlisted the help of “Finance Twitter” via four valuation input polls (I have my own assumptions, but wanted to gauge the distribution of expectations):

  1. AWS Fair Value: $200B, $150B, $100B
  2. Amazon Retail ‘Intrinsic’ EBIT Margin: 6%, 5%, 4%
  3. Amazon Retail 5-Year Sales CAGR: 25%, 20%, 15%
  4. Amazon Retail Terminal PE in 5 Years: 35x, 30x, 25x, 20x

On a weighted-average basis the results look as follows:

  1. AWS Fair Value: $144 billion
  2. Amazon Retail ‘Intrinsic’ EBIT Margin: 4.82%
  3. Amazon Retail 5-Year Sales CAGR: 18.3%
  4. Amazon Retail Terminal PE in 5 Years: 24.95x

Using the poll-based inputs + a 25% tax rate + an 8% discount rate + ignoring interest expense/excess cash/debt/etc, I arrive at a total fair value of $674 per share (details provided in PDF above).

While at last closing price AMZN is only 19% this poll-based fair value, I believe the AWS fair value is likely on the high side…and after discussing with some folks on Twitter yesterday, ‘intrinsic’ EBITM is likely south of 4%…


My personal fair value is approximately $558. I use a 20% Retail Sales 5-year CAGR, a 22.5x 5-year Retail terminal PE, 30% tax rate, 10% discount rate, an ‘intrinsic’ EBITM of 4%, but leave the AWS fair value the same at $144 billion.

If AWS is ‘only’ worth $100 billion today, my fair value falls to $468.


“Free cash flow” is, IMO, one of the most misused terms/phrases in the market. Not to be picky – but Buffett’s original definition of “FCF” (using a newspaper company as an example, I believe) was a simple series of adjustments to the income statement in order to arrive at normalized level of distributable earnings power. In other words, assuming an appropriate maintenance level of capex and normalized working capital that keeps earnings power flat in perpetuity, what can a company fully distribute to shareholders assuming no reinvestment in growth capex/working capital. That’s the pure definition of FCF, on which you base a fair value price/earnings ratio.

The next-best definition of FCF is currently distributable FCF. So for a mature company such as Pepsi that is growing at say 6% a year, assuming a 30% ROE Pepsi’s currently distributable FCF would be 80% of earnings (to generate 6% growth at a 30% ROE, Pepsi would only need to retain 20% of normalized earnings power). Currently distributable FCF is best used in a FCF yield calculation (if distributable FCF is paid out 100% as a dividend, the calculation is easy – the dividend yield).

* * * * * * * * * * * *

Jeff Bezos is (clearly) not dumb. He knows that the marketplace A) loves easily calculable metrics (because *it’s* lazy), and B) loves free cash flow…in all of its forms (because FCF *=* earnings power). As such, Mr. Bezos uses a highly manipulated definition of free cash flow in order to tick both boxes:

Free Cash Flow = Operating Cash Flow – Cash Capital Expenditures

Of course for good measure, he throws in FCF less principal repayments and FCF less principal repayments less assets acquired under capital leases.

Bezos’ baseline FCF definition is absurd on two fronts – and he knows it: 1) material stock comp expense is included in operating cash flow, and 2) capital lease-financed capex is reported ‘below’ the capex line. Stock comp expense is a capital markets transaction, and should be deducted from operating cash flow; and capital lease-financed capex is no different than cash capex financed via public market bond issuance.

But the market doesn’t really care because both stock comp and capital lease-financed capex are “NON CASH”. Yay.

More than anything this is simply a rant. I want Bezos to retain every single dime of ‘intrinsic’ earnings power, as the PV of AMZN’s growth opportunities is far higher than anything investors could produce themselves by reinvesting distributed cash flow back into the capital markets. Plus it’s more tax efficient. But for goodness sakes, enough with the absurd FCF definition(s).

As found in the PDF above, AMZN’s cleaned up FCF generation (‘currently distributable’ definition) looks as follows:

  • 2013: -$2.2 billion
  • 2014: -$5.5 billion
  • 2015: -$844 million
  • LTM 2Q16: -$1.1 billion


In simple summary, in my opinion there is too much optimism currently priced into the stock. While the growth runway is enormous, it is very difficult to handicap the fundamental business value with a high degree of precision, and I believe those looking to enter the stock (speaking for myself) are best served to wait for when there is widespread doubt around various aspects of the business.

Hindsight is 20-20, and I didn’t take a position thus what I am about to say is next to moot…but in late 2011 I did a valuation similar analysis (at the time it was only Amazon Retail, for the most part) using a 4% ‘intrinsic’ EBITM, 25% growth rate and a 15x terminal PE and arrived at a fair value materially higher than where the stock was trading. Again, I did not take a position – but the point is that the time(s) to buy the stock is/are when it is trading materially below an easily defensible fair value estimate.

AWS valuation and the cost of distribution to meet Prime-led demand growth are the two key areas I will be watching for doubt to begin creeping into the stock price.


One thought on “Generals: High-Level Valuation Analysis

Leave a Reply

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

You are commenting using your 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 )

Google+ photo

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

Connecting to %s