It is often said that Amazon’s cloud services (which operate under the flag of Amazon Web Services) is a terrific profit generator for Amazon. The retail side of Amazon might be weak, with sky-high valuations on the expectation that this will change for the better in the future, but Amazon’s cloud services are already massively profitable, as most investors in the American tech-company would argue. Amazon is without any doubt the market leader in cloud services and reaps the rewards.

At least, so goes the story.

Moreover, this story is backed by “hard numbers.” The free cash flow that Amazon is able to generate, allegedly mostly through its cloud services (so far, Amazon has been surprisingly reluctant to disclose much information about the specific source of their revenues, a red flag for investors), is enormous. In other words, after years of profitless growth, Amazon suddenly has turned into a profit monster.

Yet Amazon does exactly what many companies did during the dotcom bubble of the late 90’s: hide debt.

You can derive the free cash flows of a company by taking its operating cash flows and subtract any investments in capital goods (capital expenditures, or: capex). In the case of Amazon, capex currently amounts to a few billion dollars. But also in the case of Amazon, this capex does not include a large portion of its (absolutely necessary) annual capital investments.

Amazon buys its servers, indeed, through financial leases. It does not simply buy these servers directly with out of pocket cash. In essence, Amazon pays every year a part of the total purchase amount of the server, and when it gets to the last payment, it acquires legal ownership over the server.

And these financial leases are not included in capex.

If we would adjust capex for these capital investments through financial leases, then Amazon’s capital expenditures would double and, surprise surprise, the alleged profitability of Amazon disappears as snow before the sun.

This implies that the free cash flow of Amazon is heavily overstated. It should not surprise us that it is precisely founder Jeff Bezos who points the market to the idea that Amazon’s free cash flow is wonderful, while net earnings no longer count. After all, net earnings are scant compared to its market cap.

The free cash flow of Amazon is a pie in the sky, although a legal and often applied accounting method, and does not reflect the capital investments that Amazon is currently making and will have to make in the foreseeable future.

A growing business normally suffers cash problems. That sounds like a paradox (a successful, growing business with cash problems?), but it is actually completely true. The reason is simple: average businesses incur costs and generate revenues at a later point in time. That is why these fast-growing businesses suffer from a lack of working capital that is needed for their expansion.

But Amazon has a negative “cash conversion cycle.” In laymen terms, clients pay Amazon before Amazon pays somebody else. And that leads us to the contrary effect.

First, let us take a look at the Amazon Prime-subscriptions. Clients pay a year ahead ($99), to acquire perks that represent future outlays for Amazon. Amazon Prime means money in the pocket, but in accounting terms these subscription fees are spread out over the year in which Amazon is supposed to “deliver the goods.” This has a very positive effect on operational cash flows when the number of subscriptions is growing, but that is only a temporary effect because they imply negative cash flows (or negative outlays) in the future. Growth in subscriptions, even if the subscriptions are not profitable, will lead to huge swings in cash flows and will result in enormous positive cash flows, even when the subscriptions might turn out to be not profitable after all!

If every American would take a Prime-subscription, then Amazon would receive, by way of example, $31.5 billion dollars. Are these revenues, or positive cash flows, counterbalanced by negative cash flows? No, at least not straight away. Only when subscribers during their subscriptions make use of Amazon Prime’s services, negative cash flows related to the subscriptions arise. In the first few months, we appear very profitable, no doubt. But only at the end of the year it becomes clear whether we were profitable or loss-making, and how much the profits or losses were.

In the case of AWS, the “cash conversion cycle” is also negative. And on the retail side, too, Amazon is accustomed to get paid first by their retail clients before passing the money on to suppliers.

That also means that, as long as Amazon grows, operational cash flows will seem enormous. The contrary, however, is true: net earnings are a far better indicator of profitability in the case of Amazon because of its negative “cash conversion cycle.”

And with our good, old-fashioned yet (in this case) more reliable net profit, we can see that the optimism regarding Amazon is highly overstated and should remind us of the insane valuations of internet companies during the dotcom bubble. And we all know how that ended …

As long as everything is awesome, at least for the moment, nobody will care about some accounting wizardry. Only when the crisis strikes and investors start losing money, they will call this fraud an outrageous scandal.