Don't Be Fooled by Amazon's Cash Flow

Retail giant Amazon.com (AMZN 0.42%) seems to be on a roll. The stock surged following a better-than-expected earnings report last month, and the company's cash flow figures are exploding. Over the past 12 months, Amazon has produced a staggering $8.98 billion of operating cash flow and $4.37 billion of free cash flow. Both dwarf the $188 million net loss the company posted over the same period.

Earlier this year, I argued that Amazon's free cash flow figures are deceptive. The company finances billions of dollars of additional capital spending using capital leases, and this greatly inflates Amazon's free cash flow. Including the $4.7 billion of assets Amazon acquired under capital leases over the past year, its adjusted free cash flow is actually negative.

I also argued separately that the nature of Amazon Web Services, where servers and computing equipment get depreciated over just a few years, is largely responsible for the vast increases in Amazon's operating cash flow. Even a structurally unprofitable cloud infrastructure business can produce rapidly growing operating cash flow as long as cash keeps getting poured into it.

Today, I want to ask a simple question. If Amazon decided it was big enough and stopped investing for growth, how much cash could be sustainably pulled out of the business each year, after all necessary expenses are paid? This is similar to Warren Buffett's concept of "owner earnings," which he first introduced in his 1986 letter to shareholders of Berkshire Hathaway.

The cash flow numbers that Amazon reports suggests that the answer to this question is measured in the billions of dollars. But a big chunk of Amazon's cash flow is unsustainable, and a deeper look at Amazon's cash flow is required.

The search for sustainable cash flow
Amazon's $8.98 billion of operating cash flow over the past 12 months can be separated into a few different sources:

All values in millions. "Other operating expenses" and "other expenses" aren't included.

Buffett defines owner earnings as net income plus depreciation, amortization, and certain non-cash items, minus the average amount of capital expenditures necessary to maintain long-term competitiveness. Amazon spent a total of $9.32 billion, both directly and through capital leases, on capital expenditures over the past 12 months, but only a portion of that is required to maintain the business.

Estimating maintenance capex is difficult, and there are very few companies that explicitly break down capital spending into maintenance spending and growth spending. The annual depreciation charge is typically in the ballpark of the maintenance capex required to maintain a business, although that's an imperfect estimate. It's possible that Amazon could end up spending less than depreciation on maintenance capex, which would create a source of cash flow, although likely not a very big one. It's also possible that Amazon would need to spend more, particularly for AWS, in order to maintain its competitive position.

What we can safely say is that whatever sustainable cash flow that can be generated through the difference between deprecation and maintenance capex will be a fraction of the total depreciation charge. Most of that cash flow will be balanced out by capital expenditures necessary to keep Amazon competitive in the long run. 

Working capital is typically a drain on cash flow as a company grows, but Amazon operates with a negative cash conversion cycle: It collects payments from customers before it pays suppliers. As Amazon grows, this creates a source of cash flow each year.

Ultimately, this extra cash flow is a good thing, but it's not sustainable. It exists only as long as Amazon keeps growing. The moment Amazon stops growing, the $541 million of cash flow derived from changes in working capital over the past 12 months goes away. This cash flow is just the result of the timing of payments and thus can't be sustainably pulled out of the company.

Amazon Prime, which charges customers $99 per year for free shipping and other perks, is another source of cash flow for Amazon. When Amazon receives a Prime subscription payment, it books it as deferred revenue, which is then recognized as revenue over the course of the year. As the Prime membership grows, the increasing deferred revenue balance creates cash flow for the company.

Deferred revenue rose by $1.065 billion over the past 12 months, mostly because of the growth of Prime, which is billed annually, but also in part because of the growth of AWS, which is billed monthly. It should be obvious that this isn't a sustainable source of cash flow; the moment the Prime membership base stops growing, most of this operating cash flow disappears. Prime, in fact, doesn't even need to be profitable for it to produce cash flow. It only needs to grow.

The last item is stock-based compensation. This $1.755 billion of operating cash flow is sustainable as long as employees continue to accept stock as compensation, but it dilutes existing shareholders.

This isn't the kind of non-cash charge that Buffett was talking about. Amazon would need to pay employees more cash if it stopped handing out stock as compensation, or it would need to spend cash on buybacks if it wanted to keep the share count constant. For the same reason that selling additional shares and treating the proceeds as sustainable cash flow doesn't make sense, stock-based compensation shouldn't be treated as a source of sustainable cash flow.

Not exactly a cash cow
Amazon's trailing-12-month net income is currently negative, but if the company stopped investing in growth, presumably its utilization of its existing fulfillment centers would rise, since the company is likely adding them ahead of actual demand. In a no-growth scenario, Amazon's net income would likely rise as well, producing a source of sustainable cash flow for the company. Exactly how much cash flow this would generate is hard to say; I doubt that Amazon would be able to raise its prices by very much, since that would allow a competitor to undercut it on price.   

A big chunk of Amazon's operating cash flow, coming from changes in working capital, changes in deferred revenue, and stock-based compensation, is unsustainable in the long run. This amounts to about $3.36 billion over the past 12 months. Improvements in net income and the difference between deprecation and maintenance capex would act as sources of cash flow in a no-growth situation, but I think it's safe to say that the amount of cash that can be sustainably pulled out of Amazon if the company stopped investing for growth is very likely less than the free cash flow Amazon reported for the past 12 months. 

Free cash flow is supposed to represent the cash flow left over after investments in growth. In light of this, it shouldn't be surprising that Amazon's debt, including outstanding capital leases, has ballooned over the past few years as it's invested in growing the business:

Source: Amazon 10-Ks.

Amazon's cash flow numbers certainly look impressive, so much so that operating cash flow and free cash flow are the first two figures mentioned in a typical Amazon earnings press release. But Amazon isn't the cash flow machine it claims to be. Cash flow numbers are certainly useful to investors, but they shouldn't be blindly accepted. Buffett had something to say about this in his 1986 letter to shareholders:

Why, then, are "cash flow" numbers so popular today? In answer, we confess our cynicism: we believe these numbers are frequently used by marketers of businesses and securities in attempts to justify the unjustifiable (and thereby to sell what should be the unsalable).

Amazon has a great sales pitch, but its cash flow numbers aren't quite what they seem.