“Platforms! There is a great future in platforms … think about it,” said NYU economics professor Lawrence J. White, reinventing a key line from The Graduate in introducing Hal Varian, Google’s chief economist, at an antitrust conference earlier this summer. White offered good advice. Platforms play a major role in every section of our economy today: Google and its search engine, Facebook with its apps and ad business, Apple and its App store, and Amazon with its e-commerce marketplace. In other spheres of the economy we have ridesharing platforms Uber and Lyft, streaming platform Netflix, house-sharing platform Airbnb, and many more.
“Platform” denotes an inherently passive concept, merely facilitating interaction between two or more active entities. This certainly appeals to Big Tech, allowing them to dissociate themselves from any agency. Instead, they portray themselves as big, dumb pipes of information without nefarious intent. It was not Facebook that promoted fake news, but rather its users.
People and governments are now slowly realizing that tech giants like Amazon are in fact extremely vigorous and assertive entities, managing their walled gardens with tenacity and ruthless precision. Politicians have gone on the attack, albeit in very narrow forms. Earlier this year, Elizabeth Warren called for breaking up platforms that engage in unfair competition. For example, she noted that Amazon collects information from its third-party sellers to identify successful products, and launches “Amazon Basics” versions that compete with them. Warren argued that Amazon had to choose if they wished to “have a team” or be the “umpire.”
This allegation is not theoretically incorrect. In fact, the U.S. has a long tradition of mandating antitrust remedies such as platform separation. But Warren’s analysis did not fully acknowledge and comprehend the full scope of Amazon’s behavior on its platform. Even without Amazon Basics, Amazon has never been a dumb pipe or a neutral umpire.
Amazon doesn’t employ a single strategy to leverage power over its marketplace. Sometimes it intervenes with its own products, but more often it uses its control to squeeze sellers and brands. It relentlessly seeks out either the most profitable or beneficial strategy for every economic transaction, all of which can cause harm to its partners and even the consumers it claims to be “obsessed” with. It is worth stepping back to understand Amazon’s conduct in a broader context.
IN A PAPER published earlier this year, I argued that Amazon might have achieved its dominant position in e-commerce by engaging in predatory pricing—selling items below cost to acquire market share. I speculated that Amazon could recoup losses from predatory pricing by expanding the role of third-party sellers on the platform and downsizing its own direct sales. By doing so, Amazon could effectively shift warehousing and shipping costs to third parties, and also charge these third parties for every transaction on its platform.
Crucially, this theory of recoupment relied on the fact that Amazon is able to leverage its dominance as a platform. In the current e-commerce environment, sellers see little alternative to paying Amazon access fees, in exchange for the unparalleled access to millions of consumers shopping on the platform. In more abstract terms, it appears that Amazon uses its “horizontal” or platform power to squeeze entities situated on the “vertical” axis of the supply chain.
It is necessary to recognize at this stage that Amazon is a monopsonistic entity. It possesses sufficient “buyer power” to dictate terms of business to merchants who wish to use its services. We can think of Amazon as the Brooklyn Bridge, the first bridge ever built into Manhattan. While Brooklynites could still use a ferry to access the island or even sail by boat, it is clear that as the only bridge with the easiest and most convenient path to Manhattan, a significant toll could be levied. Even if Amazon engaged in predatory pricing in the past in order to build its “bridge,” it no longer needs to engage in this conduct, because its position as a monopsonist is now fully established.
Even a brief study of the history of American monopsonists reveals the next logical step for Amazon: total control of the “vertical” supply chain, from the factory floor to the customer’s door. Inspiration can be drawn from how J.P. Morgan leveraged his monopoly over the railways to achieve total control over the steel industry, after purchasing Andrew Carnegie’s business and creating U.S. Steel in 1901. If history is any guide, we can assume that Amazon’s primary objective right now is not to maximize profit. In fact, profit maximization should only be viewed as a secondary objective, and the anecdotal evidence collected thus far fits neatly into this theory.
Instead, the primary goal for Amazon is to gain control over the brands sold on the platform. That dominance will pay far more dividends than short-term profits. This means that the analysis should not focus on the sellers and vendors but rather on the brands themselves. While the sellers played a crucial role in the recoupment theory, at this stage the sellers should be viewed only as instrumentalities that serve that primary goal of gaining vertical control over the supply chain.
Amazon does not treat all brands sold on the platform in the same manner. Broadly speaking, it appears that treatment is determined by two main factors: first, the strength of the brand in question in the product market and total volume of sales, and second, the amount of directbrandcompetition that Amazon itself has on the “horizontal” axis with other supply channels such as Walmart or Target.
If a small- or medium-sized brand’s main channel of distribution is Amazon, then Amazon will probably analyze seller data to determine the brand’s profit margins and where to place it in the platform. If profit margins are high due to consumer demand or absence of competition (a good example would be licensed goods for children), then Amazon can source from the brand directly. At times, even if the brand started its life on the platform as a third-party seller, Amazon will exercise a little-known clause in its agreement and force the brand to sell to Amazon directly. When this happens, vendors report that sometimes Amazon will raise the brand’s consumer prices due to the absence of competition on other channels, and will simultaneously decrease the amount the brand itself receives per item sold.
In a more recent phenomenon, Amazon will offer small brands better marketing support, product reviews, and prominent placement—but only if brands sign a deal permitting Amazon to buy them out for $10,000. The program, called Amazon Accelerator, reveals how Amazon contours its strategies to a particular brand—either sourcing from it, competing with it, buying it outright, or just milking it with third-party seller access and fulfillment fees.
On the other hand, if the brand power is weaker and profit margins are low due to abundant competition, it would be placed by Amazon in the marketplace. The reasoning is that the allure and therefore the profit margin of these types of brands is too weak to justify the costs of purchasing directly and managing sales on the platform. These types of brands were infamously purged earlier this year, after Amazon had been selling them directly to consumers for years. Amazon even had a name for them: “Can’t Realize a Profit,” or CRaP.
Although competition from different third-party sellers hawking these brands reduces prices, Amazon’s fees do not go down proportionally as a result. It just enjoys a greater percentage of the overall sale. Brands have little control over the Amazon purchase price; they are effectively disciplined by third-party sellers who acquire their goods.
Things become even more complex when the brand is sold at other outlets, such as Walmart. From conversations with senior former Amazon employees and merchants, it appears that Amazon has a number of ways to slowly but surely gain control over these types of brands, once it realizes that they are popular and have robust demand.
One tactic involves Amazon sourcing directly from the brand and then starting to sell the brand’s products below acquisition cost, if Amazon detects that a product is sold at lower price elsewhere, even for a limited time. This conduct is of course extremely damaging for the brand’s independence, since it effectively dries up transactions that take place at other retailers. When brands complain to Amazon, they are told that Amazon will stop selling below acquisition cost only after the brand can demonstrate that the product is no longer offered at a lower price anywhere else. In some cases, Amazon will later send the vendor an invoice and request that the vendor pay for what it terms “lost margins”; in other words, the vendor or brand must reimburse Amazon for its below-acquisition-cost pricing.
In antitrust law, predatory pricing is traditionally associated with efforts to expand control over consumer choices. For instance, an airline will heavily discount its fares on a given route in order to knock out competition and gain exclusive control over the route. But the primary goal of Amazon’s pricing tactics in this case is not to harm other outlets—that is only a welcomed side effect. Rather, it is to increase Amazon’s power over the brand by becoming the brand’s dominant buyer.
If Amazon offers the brand’s products at the lowest price, it will inherently capture the majority of that brand’s sales volume and make the brand increasingly reliant on Amazon. This also pushes brands to operate with razor-thin profits, since there is no longer robust competition from multiple buyers. Subsequently, these “captured” brands become much more susceptible to sale volume fluctuations and therefore more cooperative with Amazon. So in this case the predation is used to expand Amazon’s control up the supply chain.
Amazon would probably say it engages in these pricing tactics to benefit the consumer. But if Amazon’s true goal was to offer a given product at the lowest price possible, it could do so by implementing a price-matching policy, or other means. Instead, it decided to penalize its vendors and sellers. Thegoal is not to offer a given product at the lowest price possible, but only to be the cheapest available option, even if in reality other outlets could offer—and have been offering—a particular product at a lower price.
That explains how Amazon can harm consumers while offering the cheapest option. In addition, predatory pricing also affects innovation in this case: A supplier or merchant will not improve its products if Amazon is squeezing their profits so much that substantially fewer resources can be put into investment.
Since Amazon cannot deploy the same pricing tactic in its marketplace, where third-party sellers determine the sale price, it has developed other tools that help ensure brands offer their products at the lowest available price on the platform, at the expense of general consumer welfare. According to sellers I’ve spoken with, Amazon will manipulate the “buy box,” suppressing a seller’s opportunity to get prime placement in front of customers unless they match prices at other outlets. Some have speculated that Amazon also manipulates ad placement, though there’s no hard evidence of that.
Crucially, on other outlets, shipping is calculated separately. Because Amazon Prime’s shipping is free to the customer, the matching price sellers must conform to on Amazon will ultimately be much lower for their bottom line. Although Amazon customers pay an annual fee for their Prime subscription, this fee does not cover the costs associated with shipping. Sellers still have to pay Amazon fulfillment fees, while struggling to match a price that doesn’t take shipping into account. Sellers who sell to multiple outlets admit that the prices of the products they are selling could have been lower in those outlets, but due to Amazon’s price-matching scheme they are higher. And with the ever-increasing fulfillment fees levied by Amazon on third-party sellers, this situation is only further exacerbated.
Walmart and other outlets can theoretically offer products at lower prices because shoppers can buy a product at the store, meaning that those outlets do not incur any additional shipping costs. But Amazon’s disciplinary efforts effectively remove this option, making non-Amazon consumers pay more.
Recently, Amazon amended its agreements with sellers to reflect this policy. The “fair pricing policy” corroborates seller claims, explicitly stating that Amazon “regularly monitors the prices of items on our marketplaces, including shipping costs, and compares them with other prices available to our customers.” Amazon even threatens to discipline sellers that “harm customer trust” by removing the buy box, removing listings, suspending shipping options (including independently fulfilled offers), and even terminating the ability to sell through the marketplace. Conduct that harms customer trust includes selling products on Amazon at higher prices than those available in other outlets.
Antitrust scholars and other critics who have rushed to Amazon’s defense routinely point out that regulation and enforcement should protect competition and not competitors. In addition, they argue that while Amazon’s conduct might harm competitors, this conduct is legitimate since it benefits consumers. Even if we are to accept their basic proposition regarding the role of antitrust regulation and enforcement, it is very hard to argue that raising the prices that customers have to pay for same products in outlets like Target is beneficial to anyone but Amazon.
If the goal of Amazon is not to maximize profit by exploiting sellers’ data, but rather to use sellers to exert control over brands, investigators and regulators should focus on the tools and conduct that Amazon is engaging in to achieve this dominance. Only by recognizing Amazon’s end goal, its multiple monopsonisticstrategies to get there, and the effect on consumers will policymakers be able to design effective policies that will truly rein in this extreme use of market power.