Merger-control regimes around the world have for some time now engaged with the theory of harm known as “killer acquisitions.” The idea is simple: an incumbent buys a rival in order to shut down its operations and preempt future competition. Indeed, the original paper that laid out the theory found that 5-7% of mergers may qualify as killer acquisitions.
This thorough analysis of competition and innovation in the pharmaceutical industry led to an explosion of academic and policy interest around the world. But it is in digital markets, which were not covered by the initial empirical research, that the underlying ideas appear to have caught on most with policymakers.
But was this attention warranted? A new working paper, co-authored by one of the builders of the original idea, suggests otherwise.
A Refresher on Killer Acquisitions
Before we get to the new paper, allow me to offer a quick refresher. The primary concern with killer acquisitions is that they may deprive markets of future competition (and ensuing innovation), especially in sectors already marked by monopoly power. They may also shape startup incentives. Rather than pursuing independent growth strategies, startups may orient themselves toward “exit strategies” that involve being acquired and possibly neutralized. At the market level, repeated acquisitions may generate so-called “kill zones,” areas from which entrants steer clear for fear of being bought and shut down.
Several adjacent theories have subsequently been built on this foundation. “Acquisitions for sleep” describe transactions in which the target’s assets are left dormant and unused. “Reverse killer acquisitions” emphasize the risk that incumbents may wind down in-house projects in favor of external acquisitions, thereby substituting away from “organic innovation.” And in recent years, “acqui-hires,” or deals that primarily aim to acquire talent, have been flagged as a way to sidestep competition scrutiny. While these are sometimes all bundled together in policy discussions, they are each unique theories of harm that would arise in different cases.
And yet, much of the empirical evidence to support the concept of killer acquisitions is industry-specific. It has only been documented in pharmaceuticals, where innovation tends to be shaped by long development cycles and regulatory bottlenecks. Other contributions have sought to examine digital markets, payment systems, software, and the broader biotechnology industry, but arguably none with the empirically grounded methodology of the original paper.
But the dynamics of innovation differ sharply across these industries. Slow-moving sectors such as pharma, where the field of potential competitors is comparatively predictable, potentially offer different opportunities to stifle competition than digital markets, where innovation is faster, more experimental, and often complementary rather than substitutive.
Regardless of these uncertainties, the concerns have prompted widespread regulatory change. Several jurisdictions have contemplated “transaction value” thresholds to capture deals involving low-turnover but potentially high-value targets. Some commentators have gone further, suggesting a reversal of the burden of proof for certain mergers.
In the European Union, the most important shift has been the European Commission’s reinterpretation of Article 22 of the EU Merger Regulation, which now allows the Commission to review transactions referred by national authorities even if they fall below turnover thresholds—despite some limitations imposed by the Court of Justice of the European Union (CJEU) in its recent Illumina/GRAIL ruling. In parallel, the Digital Markets Act (DMA) has imposed a notification obligation on gatekeepers, with some proposing that firms should also be required to disclose the “strategic intent” behind their acquisitions.
Despite its intuitive appeal, the killer-acquisition theory has always been open to several criticisms. First, it risks conflating the number of competitors in a given market with the intensity of competition. Empirical research suggests that innovation pressure is not linearly related to the sheer count of firms. It is therefore hazardous to draw a straight arrow from “fewer competitors” to “less innovation.”
Second, the theory creates serious problems of observability. Post-merger integration often entails the consolidation or redirection of projects. Distinguishing genuine “kills” from ordinary integration is methodologically fraught, and false positives could lead authorities to condemn transactions that actually enhance innovation. The same caveat applies to acqui-hires. The extent to which teams are retained post-acquisition, with what effects on innovation, is unclear.
Third, the “kill zone” argument might be overstated as well. While it is plausible that prior acquisitions by incumbents could deter entry, the picture is likely to be more nuanced. Active acquisition markets may, in fact, stimulate entry, since small firms see acquisition as a viable reward for innovation. What looks like a “kill zone” may, in practice, operate as a “cultivation zone.”
Finally, and most notably, the empirical base has remained thin. The eponymous paper estimated that only 5–7% of pharmaceutical acquisitions could plausibly be classified as “killer.” This figure already aligns with the actual intervention rates by agencies such as the European Commission. By contrast, evidence of systematic killer acquisitions in other markets has been far weaker. Transplanting the theory to industries with very different innovation characteristics therefore risks building law on shaky ground.
Digital Killer Acquisitions?
Against this backdrop, a new working paper provides timely evidence on whether the killer-acquisitions narrative holds across firms, markets, and industries. Co-authored by Florian Ederer—one of the original architects of the theory—the study turns the spotlight from pharmaceuticals to digital markets.
The paper examines 1,200 startup acquisitions by large technology firms—the “GAFAM” firms (Google, Apple, Facebook, Amazon, and Microsoft), as well as Qualcomm, Cisco, and Intel—using economic methods to track patenting activity before and after a startup is acquired. In particular, it links each startup to its portfolio of patents in relevant patent classes (“physics” and “electricity”, in which the majority of innovation activity in digital markets occurs).
The authors also gather information on employee mobility. This complements the changes in patenting activity by allowing an estimation of retention patterns, employee turnover, and the relationship between post-acquisition integration and innovation outcomes.
The dataset provides interesting insights on technology firms’ acquisition activity. For instance, the paper finds that, across all acquirers, a substantial share of startup targets did not hold any patents at the time of the transaction. There is, however, broad heterogeneity among the firms. For example, 70% of Qualcomm’s acquisition targets held at least one patent at the time of their acquisition, compared to just 17% of Meta’s acquired firms. This suggests very different acquisition rationales.
Another finding relates to deal value: acquirers pay significantly more to buy targets that hold patents compared to those that do not. In other words, “deal values are markedly higher when patents are present.”
The paper’s aim is straightforward: to see whether startup acquisitions suppress or stimulate innovation. In that regard, the main finding is striking. As the authors put it:
Our findings complicate the popular narrative that digital acquisitions by large incumbents are predominantly anti-competitive. While we do not rule out the possibility of killer acquisitions in individual cases, our evidence suggests that most deals by GAFAM and related firms are motivated by complementarity rather than suppression. Innovation tends to rise rather than fall in the wake of acquisition, particularly in technological domains where the acquirer has prior experience and continues to make follow-on investments.
This conclusion warrants closer attention. To begin, the paper shows that patenting activity by startups tends to cluster in technological domains where the acquirer is already active. Moreover, innovation activity often increases after an acquisition, which the authors highlight as “a finding that undermines the ‘killer acquisition’ narrative in its strongest form.” In particular, the paper notes that “acquired patents experience greater visibility and influence within the broader innovation ecosystem, with more citations coming from unaffiliated firms.”
Another interesting finding relates to serial acquisitions. The paper shows that the positive innovation effects of acquisitions depend on continued dealmaking. In particular, patents owned by first-acquired targets experience strong increases in citations if the acquiring firm continues to buy startups. Many reasons may contribute to such an outcome: the authors suggest, for example, that firms acquired earlier may “benefit from greater organizational attention or resource integration, which in turn enhances the diffusion of their innovations.” The broader point, however, is this: if post-merger innovation rises when incumbents pursue serial acquisitions, there may be a potential tradeoff between concentration and dynamism.
The paper also investigates the possible effects of employee mobility on innovative activity. Accordingly, citation counts for acquired patents increase relative to comparable ones, indicating that acquisitions raise the quality and reach of the underlying technologies. Workforce churn plays a key role here: nearly half of acquired employees leave the acquirer within a short period, creating spillovers as their knowledge diffuses across the broader ecosystem.
Interestingly, the greater the share of departing employees, the higher the citation impact of the acquired patents. One possible reason advanced by the authors is that higher workforce retention may “embed” innovations within the acquiring firm, thus limiting “external visibility.”
Finally, the paper reinforces the point that killer acquisitions are industry-specific. Digital markets differ sharply from pharmaceuticals: many deals are motivated not by formal intellectual property, but by access to capabilities, products, customers, or personnel. Marc Ivaldi, Nicolas Petit, and I conducted a similar study that examined a smaller set of cases more intensively, and reached similar conclusions. The paper also reveals differences across firms. Hardware-oriented acquirers like Intel and Qualcomm tend to buy in domains where they already hold patents, pointing to cumulative or vertically integrated innovation. By contrast, software-oriented firms such as Facebook or Amazon are more likely to acquire without patents, reflecting priorities around user bases, engineering talent, or data assets.
Killing the Killer-Acquisitions Narrative?
Beyond its academic value, the new empirical evidence carries clear legal and policy implications.
The debate around killer acquisitions has almost certainly been overstated. The narrative captured outsized attention in legal, economic, and policy circles, spurring regulatory initiatives, policy shifts, and heightened merger scrutiny. This latest paper adds weight to the argument that much of the debate was driven more by narrative than evidence. The original theory was grounded in evidence from the pharmaceutical industry, but its application quickly spread to digital markets, where subsequent studies—including this one—fail to replicate the pharma-based findings.
Equally troubling has been the willingness of academics and enforcers to overlook the hard-won lesson that context matters. Merger-control regimes are built on the understanding that transactions must be assessed within their specific legal, economic, and technological context. Formalistic approaches that treat vastly different deals as if they were the same had long been abandoned. The new working paper thus serves as a reminder, and a warning, that law and policy should not be rebuilt around untested ideas, however intuitive they may sound.
