I have already addressed how Morgan Stanley's (MS) bullish thesis for Amazon.com (AMZN) keeps changing. First Adopter also called into question the ebook estimates in the latest Morgan Stanley research report titled "The Kindle Franchise: A Strategic Profit Pool" on Amazon.com. However, these are not the only problems with this latest research report. While refining my own model, I have come across a further serious and easy to see flaw in that same report.
As we all know, Amazon.com pays a substantial part of its employees' compensation in stock, using RSUs (Restricted Stock Units) which vest over time. It then recognizes the cost of these units as they vest. In its latest quarter (Q4 2012), Amazon.com recognized $235 million in stock based compensation, and over 2012 it recognized a full $833 million. Through this continuous stock issuance, naturally Amazon.com's stock is being diluted over time. Indeed, at the end of 2005 Amazon.com had 416 million shares outstanding, and 438 million shares outstanding including unvested RSUs, while at the end of 2012 this has gone up to 454 million shares outstanding or 470 million including unvested RSUs.
Analysts, including Morgan Stanley, have taken a liking to ignoring the costs these shares represent, and instead calculating EPS numbers without the stock based compensation, while at the same time computing operating and free cash flow including the cash that is saved by paying with shares instead of cash. We could say this is an option, except when analysts do like Morgan Stanley did.
Indeed, as we can see in Morgan Stanley's report, Morgan Stanley has considered the continuance of Amazon.com's practice of paying compensation with stock. That much is implied in Morgan Stanley's "Annual Income Statement" as well as the "Annual Cash Flow Statement", both of which have the relevant sections reproduced below as proof (the red box highlight is mine):
Morgan Stanley then also includes the impact of share based compensation in the "Annual Balance Sheet". This increases book value. All fine and dandy, I reproduce the relevant section below (the red box highlight is mine):
So what's the huge flaw? We have the favorable effect on earnings when ignoring share based compensation. We have the favorable effect on cash flow from paying with shares. We have the favorable effect in book value from issuing a boatload of shares every year … now let's see how many shares Morgan Stanley considered Amazon.com as having over time, we can find this at the bottom of the "Annual Income Statement" (again, the red box highlight is mine):
Opppss … where are all those shares which Morgan Stanley considered were being used to pay the employees? They're nowhere to be found. Worse still, Amazon.com's current diluted share count is already at 461 million shares, so why does Morgan Stanley start considering 455 million going forward? Perhaps Morgan Stanley considered that Amazon.com would be buying back stock over time to keep the share count stable? Let us see the corresponding line in the "Annual Cash Flow Statement" … (again, the red box highlight is mine)
Nope, that ($960) million is relative to 2012. From 2013 onwards Morgan Stanley didn't consider any repurchase of stock, so the share count not going up can only be attributed to "magic", since Morgan Stanley is considering Amazon.com will be paying employees with stock, is considering that the corresponding paid in capital will hit equity, but that no shares will be issued and no cash will be spent to keep the share count constant.
This, in turn, inflates the cash Morgan Stanley calculates, as well as the EPS, the present value of the shares, and mostly everything else. Above all, it shows that Morgan Stanley took care to reflect every positive it saw, and left out obvious negatives such as share dilution.
And the impact isn't minor. Every $265 million means 1 million shares, and for instance in 2020 Morgan Stanley is considering $3.22 billion in share based compensation. That year alone would have to add more than 12 million shares to the share base.
Morgan Stanley's recent research report is flawed. It considers the supposedly positive effect of paying employees with shares in the income statement, the cash flow statement and the balance sheet. But it does not consider the corresponding implied dilution of the stock, and it does not account for any repurchasing of the shares to eliminate such dilution. This kind of careless research cannot be relied upon and is little more than a marketing ploy (to market the shares).
Disclosure: I am short AMZN.