The European Commission told Amazon in November 2023 of its preliminary view that the company’s proposed acquisition of iRobot restricted competition in the market for robot vacuum cleaners (RVCs) and could hamper rival RVC suppliers’ ability to compete effectively. The deal, the Commission asserted, would give Amazon incentive to foreclose iRobot’s competitors by engaging in several anticompetitive strategies, including delisting rival RVCs from its marketplace, reducing their visibility, limiting access to certain widget or attractive product labels (“Amazon’s Choice,” “Works with Alexa”), and/or raising rivals’ sales and advertising costs.
Recently, and in anticipation of a negative clearance decision from the Commission, Amazon and iRobot jointly announced they had terminated their acquisition agreement.
The Commission’s position is built on a series of legal and economic fallacies. Amazon has earned its reputation as a valuable retail platform through the varied selection of affordable products. By de-listing, reducing visibility, limiting access, and/or raising the prices of products sold on the Amazon platform, the company would effectively ruin their marketplace and their credibility as a provider of goods, pushing consumers to alternative marketplaces (which, yes, includes brick-and-mortar shops).
This strategy would only make sense under an extremely skewed vision of reality in which Amazon is more interested in monopolizing a marginal vacuum-cleaner market than in preserving the goose that lays the company’s golden eggs: its marketplace platform. Further, the notion that Amazon would even be able to monopolize the sale of RVCs in the European Union necessarily ignores the plethora of other venues through which RVCs are sold on the continent, including other online stores, websites, and physical shops. Finally, the Commission’s approach in this case recalls its misguided crusade against vertical integration more generally, which refuses to see the many legitimate and procompetitive reasons that a platform and a seller might choose to merge.
Missing the Forest for the Trees
Amazon is often labeled an “online superstore market” thanks to its enormous variety of products, fast shipping, and large consumer base. But this classification can serve to obscure the competitive pressure that a company like Amazon faces from other sales channels that are technically not “online superstore markets” (here). For example, despite the Commission’s assertions to the contrary, Amazon’s acquisition of iRobot won’t prevent other RVC companies from reaching consumers, either online or through physical stores. Nor will it prevent rational consumers from choosing different sales channels if they feel they offer better products or services at better prices. Amazon may be an “important” sales channel for RVCs, but it is simply false to suggest it is or that it will be the only one.
If Amazon were to delist or raise the prices of iRobot’s rival RVCs, consumers could take their business to another online store, like Aliexpress, Mediamarkt, El Corte Inglés online, or Worten online, or to a brick-and-mortar shop like Carrefour, Costco, Fnac, or El Corte Inglés. They also could buy directly from RVC sellers like Cecotec or Roboroc, among others. Note that this doesn’t account for the new shops that are guaranteed to mushroom if Amazon were to decide not to offer a strong selection of RVCs on the marketplace.
Similarly, concerns about Amazon self-preferencing by “reducing visibility of rival RVCs in both non-paid (i.e., organic) and paid results (i.e., advertisements) displayed in Amazon’s marketplace” fail to understand the structure of Amazon’s marketplace. If Amazon decided to increase their own products’ visibility over that of better and/or better-priced alternatives, rational consumers are likely to scroll down to find those alternatives (88% of users scroll through two or more pages on Amazon) or to explore other online marketplaces. Why would Amazon trade the quality and credibility of its primary service offering to monopolize a marginal market that, in all likelihood, it could not monopolize without buying out all competing retail channels? (Even then, there would still be a strong potential for new market entrants.)
The Commission’s statement of objections missed this important dimension, and rested on a simple (and simplistic) equation: that, because selling more Amazon RVCs would be better for the company, the acquisition must therefore be anticompetitive, as Amazon would have an incentive to foreclose. But what about Amazon’s stronger incentive not to prioritize inferior products (or, at least, products that its users do not want) in order to preserve the attractiveness of its online marketplace? The same could be said, mutatis mutandis, about “delisting.”
As for denying iRobot’s rivals the opportunity to qualify for “commercially attractive” labels, why would Amazon undercut sales from its own platform, even if these sales come from products that compete with its own label? Amazon has an interest in bolstering its marketplace, not monopolizing the peripheral markets for nails, batteries, smart doorbells or, in this case, RVCs. Where Amazon’s products compete with third parties, Amazon often allows those third-party products to keep the labels.
The Procompetitive Benefits of the Merger that Will (Probably) Never Be
There are many legitimate and/or procompetitive reasons that Amazon might have for acquiring iRobot. It could have chosen to buy iRobot simply because it truly believes their RVCs are the best. Amazon General Counsel David Zapolsky asserted in a company statement that the Amazon team “have always been fans of iRobot’s products, which delight consumers and solve problems in ways that improve their lives.”
Additionally, iRobot (IRBT) has been a pioneer in the robotic vacuum market and has been a top seller on Amazon marketplace. This is consistent with the deal’s rationale. If Amazon didn’t believe that iRobot made the best RVCs out there, why did it buy iRobot, and not another RVC seller? It certainly has the financial muscle to take its pick.
This turns the logic of anticompetitive self-preferencing on its head in a way that is becoming increasingly difficult for competition authorities to grasp. On this view, self-preferencing is a natural consequence of acquiring the best downstream company in a market; and not the last step in an anticompetitive scheme. In other words: self-preferencing is the symptom of a rational business decision, not an indicator of foreclosure.
If this is the case, then Amazon also has an interest in making the purchase worthwhile by improving iRobot’s products. Indeed, Amazon could have wanted to buy iRobot to bring it to greater scale and/or integrate its products into Amazon’s pipeline.
Take the example of Amazon’s acquisition of Ivona in 2013. In just a few years, Amazon scaled that company beyond the founders’ wildest dreams. Ivona went from 13 to 1,000 (highly paid) workers, and served as a basis for Alexa. Alexa, in turn, enabled a range of Amazon products that today’s consumers love, including the Amazon Echo Smart Speaker, Echo Dot, and Tap speakers. As of 2018, 10,000 employees worked on Alexa and Alexa-related products. As of 2023, Amazon had sold more than 500 million Alexa and Alexa-enabled devices.
Could iRobot have been next? Amazon spokesperson Alexandra Miller declared that the company could “offer a company like iRobot the resources to accelerate innovation and invest in critical features while lowering prices for consumers.” This would help iRobot compete in the global marketplace of RVCs and create better products. As Zapolsky reflected on the deal’s termination:
Amazon and iRobot were excited to see what our teams could build together…This outcome will deny consumers faster innovation and more competitive prices, which we’re confident would have made their lives easier and more enjoyable. Mergers and acquisitions like this help companies like iRobot better compete in the global marketplace, particularly against companies, and from countries, that aren’t subject to the same regulatory requirements in fast-moving technology segments like robotics.
This is also consistent with recent findings in the literature that examined acquisitions by large tech firms, finding that acquired products are often not killed, but scaled; that post-merger industry output demonstrably increases; and that the relevant markets remain dynamic post-transaction.
The Vertical Integration Delusion that Refuses to Go Away
The Commission’s hard-nosed approach to the Amazon/iRobot merger arguably stems from undue hostility to vertical integration across the board. Increasingly, competition authorities see vertical integration as suspect (here). The misguided crusade against self-preferencing is largely to blame for this.
But insofar as the ability to self-preference is one of the primary reasons for firms to vertically integrate (here), removing the option to do so is likely to dampen incentives for vertical mergers and destroy the many benefits that flow from them. Indeed, vertical integration is far from the anti-competitive boogeyman some make it out to be.
One of the main purposes of vertical integration is to transfer technology between firms and reduce the inefficiencies and discoordination that can occur in supply chains. This benefits consumers “through a number of mechanisms that allow for reduced costs and better product quality.” As Alden Abbott has written on the case at hand, “Amazon’s acquisition of iRobot would likely promote efficiencies, raise welfare, and enhance competition.”
Exit Strategy for Startups
Acquisition is a key pathway to exit for entrepreneurs. Many startups are created by founders with the explicit plan of being acquired by a larger tech company. European startups have traditionally looked to Big Tech’s deep pockets as a way to maximize their growth plans (here).
If this avenue is foreclosed, where do startups go from there? How can they scale without such opportunities? Do they all have to go public? According to a recent Financial Times article, the Commission’s response to the Amazon/iRobot deal has faced criticism from the startup community:
Some entrepreneurs are concerned that if Amazon can’t buy a maker of vacuum cleaners, it sends a signal that it will be difficult for Big Tech to buy anything at all — and that might be a blow for their exit strategies and for innovation as a whole.
Further, precluding acquisition by larger tech companies stifles an important exit strategy for founders and, by extension, an important incentive to invest in startups in the first place. In an interview for that same FT article, Stefan Mortiz of the lobbying group European Entrepreneurs said:
It’s a bad sign if the EU intervenes so heavily…in the long run nobody will want to be an entrepreneur, many companies will shut down or be bought if they have any remaining valuable assets.
Perhaps this is what the EU really wants? German Member of European Parliament Andreas Schwab has stated that:
It’s good for the economy that startups should not rely on a few Big Tech players but that we push innovative companies with new products to penetrate the market by themselves, thereby diversifying institutional channels.
But it is not good for the economy if founders are discouraged from creating startups because a vital exit strategy is cut off by regulators. Harold Demsetz—one of the most important regulatory economists of the past century—coined the phrase “nirvana fallacy” to critique would-be regulators’ tendency to justify policies on the basis of the discrepancy between the messy, real-world economic circumstances they see and the idealized alternatives they imagine. Wishful thinking, in other words (here).
Schwab’s comments fall into the nirvana fallacy of assuming that, without the possibility of acquisition, startups would be able to grow, scale, and challenge incumbents organically. But it is much more likely that many of these startups would not exist or would fall by the wayside due to an inability to secure funding.
Conclusion
We recently praised the Commission for taking market realities into account when it exempted iMessage from gatekeeper status, despite meeting the Digital Markets Act’s (more-or-less arbitrary) quantitative thresholds. But in the Amazon/iRobot case, the Commission somehow completely manages to ignore those same market realities. The assumption that the Amazon/iRobot merger would lead to a harmful outcome for other suppliers of RVC’s was based on flawed premises and an unlikely theory of harm that failed to see the bigger picture.
When firms like Amazon abandon procompetitive deals, the effects ripple through the rest of society, well beyond the comparatively narrow confines of antitrust and the optics of consumer welfare. Jobs, workers, and entrepreneurs are all affected—largely for the worse. While this is not a concern of direct relevance to merger control, those who defend broader conceptions of antitrust’s goals—who, somewhat ironically, tend to include many of those celebrating the abandoned iRobot deal as a victory against “big tech”—should care.