There comes a time in every columnist’s life when he ponders a thinkpiece about whether it is time to break up Amazon. Back in 1995, we never had to worry about the question of whether Amazon was growing too large: it was just a friendly looking, disarmingly cheap online bookstore which appeared to stock everything under the sun.
Today, Amazon seems to have its tentacles in everything, from healthcare to cloud computing and the Internet of Things. The books bit seems almost tangential. Now, Amazon stands accused of gutting out the high street and driving much-loved retail favorites out of business. It is a hydra-headed monster which is gobbling up everything in its wake.
Some, like Nick Srnicek, have argued that Amazon is a “platform monopoly,” and should be nationalized. According to Srnicek, Amazon “leads in e-commerce” and its “dominance is undermining old department stores.” But this does not add up to an argument for Amazon being a monopoly power.
If we run the numbers, over the last 12 months, Amazon has racked up $150bn in revenue, not all of which have come from its online retail business. Around 10% of Amazon’s revenue came from its cloud services business, AWS, which means that Amazon’s retail business brought in closer to $135bn.
That seems like a large number, but total U.S. retail and food services sales last year were around $5.5tn, so Amazon is around 2% of that. Okay, so Amazon doesn’t sell cars or restaurants. If we strip those out, annual retail revenue in the US is $3.7tn. Even then, Amazon only makes up 3.6% of the total. Amazon isn’t even the largest retailer in the US. That distinction still belongs to Walmart, which brought in around $500tn in revenue last year.
Neither is there much evidence that this is being used for ill. Srnicek argues that “at the heart of platform capitalism is a drive to extract more data in order to survive.” But this is to ignore the innovation unleashed by AWS. Twenty years ago, running your own online store would have required buying your own server and $15,000 worth of kit from IBM (not to mention lots of headaches about what you did when your site got popular). Now you can rent a server and the software for a few cents an hour from AWS. The move to cloud computing which Amazon pioneered has reduced the startup and maintenance cost for online businesses, which means that VCs can take risks on more companies.
Neither does there ever seem to be much detail on which parts of Amazon’s businesses should be broken off, or what social good it would serve. Breaking off Amazon’s book business might save a few mom-and-pop retailers by preventing Amazon from undercutting the competition on books, but supermarkets have been using chart books as loss-leaders for years anyway. Nor is it going to do much to harm Amazon’s dominance in other areas. The problem isn’t Amazon, it is that the bookstores have an uncompetitive business model.
Likewise, hiving off Amazon Web Services would have little benefit. Aside from the fact that Amazon would continue to plow large amounts of money into AWS, as its largest customer, Amazon would then be unable to subsidize its platform computing business with its retail business. This might be good for the likes of Microsoft and Google, but is hardly going to draw new players into the utility computing space, which requires enormous resources to be competitive in.
Most fantastical of all is the idea that we nationalize Amazon. Because, well, why not? According to Srnicek, “natural monopolies like utilities and railways that enjoy huge economies of scale and serve the common good have been prime candidates for public ownership.”
In other words, oh that looks nice, I think I’ll have that. Not only is Amazon fundamentally not a monopoly, there is no reason to think why the government would manage it efficiently, or in the interests of the consumer. What kind of signal would it send out to businesses if government owned one of the largest retailers?
Any convenience that customers enjoyed with Amazon would disappear, leaving government holding the baby. The government failed fundamentally with its IT modernization of the NHS and resisted the move to public cloud bitterly: why are we to believe that they could manage Amazon’s server business?
If you want to halt innovation in its tracks and use the power of the state to prop up failing and inefficient retail businesses, then there is no better way than to nationalize or break up the large tech businesses.
There are legitimate concerns about the power that businesses like Amazon, Facebook, Netflix, and Google wield, including privacy concerns, worker conditions, and how they address their tax affairs. That doesn’t mean we should reach for the most drastic tool in the closet.