Merger, she wrote (Part 2): How to run an investigation – Market power, theories of harm and the Innovation Shield

• The Guidelines describe a market power assessment that is holistic and forward-looking. Structural indicators remain the starting point, but dynamic competitive potential carries its own analytical weight, and a firm with no current market share can still be an important competitive force whose removal the Commission will treat seriously.
• The taxonomy of eight theories of harm codifies two decades of enforcement practice into a single reference framework, while adding chapters that were not previously there: investment and expansion competition as a standalone theory, freestanding entrenchment applicable across all sectors, commercially sensitive data as a harm without foreclosure, labor market monopsony as a recognized parameter, and algorithmic coordination as an emerging enforcement focus.
• The Innovation Shield will be a useful tool and may reassure those concerned about Europe’s approach to letting start-ups, scale-ups and nascent researchers grow via M&A. However, the Shield comes with a number of strings attached and will, in many cases, require a careful and factually challenging analysis.

Part 3 of this series will turn from harm to benefit: the Guidelines’ restructured efficiencies framework, the new category of dynamic efficiencies, resilience and sustainability as verifiable consumer benefits, and the balancing exercise that determines whether a merger goes through when harm and benefit are genuinely in play. It will also address Article 21 EUMR, which governs when Member States can intervene in cross-border deals and what happens when they overstep – a merger control provision that has attracted growing attention in recent years and was not featured in the old guidance at all.

Authored by Christoph Wünschmann, Christian Ritz, Florian von Schreitter.

As already noted in part 1 of our series, the consolidated framework retires the horizontal/non-horizontal architecture and replaces it with a unified taxonomy of eight theories of competitive harm, extending the Commission's analytical reach into territory that the older guidelines, HMG (2004) and NHMG (2008), did not formally map.

The first four share a common logic: the elimination of a competitive constraint.

• Loss of head-to-head competition remains the baseline: where the combined entity would hold high or very high market shares, the Commission requires substantial evidence pointing away from a SIEC before it will clear the deal; at lower or moderate shares, substantial corroborating evidence is required before it can find one (para. 127). A specific scenario of head-to-head competition codified for the first time in the formal EU framework is labor market monopsony, where a merger creates or strengthens employer market power, leading to lower wages and reduced worker mobility, with downstream effects on product markets (paras. 160-162).
• Loss of investment and expansion competition is also new to the guidance: a merger may create a SIEC by altering incentives to invest going forward, even if the merged entity intends to proceed with all pre-merger plans, because the SIEC turns on the change in incentive itself (para. 170).
• Loss of innovation competition distinguishes between specific innovation competition, where overlapping R&D projects reduce the merged entity's incentive to pursue both due to cannibalization, and general innovation competition, where overlapping capabilities weaken early-stage competitive pressure even before any specific product overlap is identifiable (paras. 180, 186).
• Loss of potential competition retains its three cumulative conditions: actual competitors fail to constrain the merged entity; the potential entrant already exerts an actual constraint or has the capacity to exercise a future one; and no other potential competitor can step in (para. 194).

The remaining four categories reach into structural and forward-looking harm.

• Foreclosure (input, customer and conglomerate) retains its three conditions of ability, incentive and effect, but the recognized mechanisms now expressly include degraded interoperability and impeding rivals’ pipeline products from reaching the market. Notably, diagonal mergers receive express treatment for the first time at guideline level: combinations between parties without pre-existing commercial relationships, where one controls a product accessed primarily by the other's rivals, are flagged as particularly prone to foreclosure harm and unlikely to be offset by integration efficiencies (para. 251).
• In other developments, entrenchment of a dominant position is formalized as a standalone theory: the combined firm gains control over assets that structurally raise barriers to entry and expansion, deterring future rivalry without requiring any conventional foreclosure finding (para. 252). The theory builds directly on the Booking/eTraveli case (M.10615) and tracks the 2023 United States DOJ and FTC Merger Guidelines, with implications for parties planning cross-border transactions. It applies in principle across all sectors, not only digital markets, although it requires dominance in one or more closely related markets.
• Reaching terms of coordination in a given market is materially tightened as a theory of harm. A paragraph added in the final draft, absent from March’s leaked version, states that the mere existence of a credible deterrence mechanism satisfies the third Airtours condition (and thus supports a finding of anti-competitive coordinative effects), with no need to prove an actual threat or its use (para. 280). Para. 278 reinforces the same logic by confirming that the prospect of reverting to normal competitive conditions is itself, in most cases, a sufficient deterrent. Unsurprisingly, AI-enabled monitoring is specifically called out as a new coordination facilitator (para. 267 et seq.).
• Other anticompetitive effects cover access to commercially sensitive information as a standalone harm, without requiring a foreclosure finding (paras. 282-286), and portfolio effects, which reach beyond the conventional bundling-and-tying analysis to capture bargaining-power dynamics where a broader product range strengthens the merged firm's negotiations with retailers (paras. 287-290). Notably, minority shareholdings below 5% are in principle not treated as standalone concerns absent additional rights or links; common ownership through institutional investors is acknowledged as a factor that may cause standard concentration measures to underestimate effective market consolidation (paras. 165 et seq.).

Reaching straight into the Draghi Report’s toolbox, the Guidelines also introduce an Innovation Shield, to which they devote an entire sub-section. While the Shield stops short of a clear-cut safe harbor, it sets out five specific scenarios in which the Commission will in principle not find any harm to competition (para. 192). These concern transactions involving small innovative companies (including start-ups) or R&D projects with “dynamic competitive potential.” The Commission identifies five such scenarios, organized by the type of overlap the transaction creates:

• no overlap of any kind;
• overlap between one party's R&D project and the other party's existing activities, provided that the merging parties do not (individually or combined) exceed a 40% market share, and that there are at least three independent competitors with comparable R&D competitive potential;
• overlap between the parties' R&D projects, provided that the “three competing R&D projects” criterion is met;
• overlap between R&D capabilities at the innovation-space and industry level, provided that the merging parties do not exceed a 25% combined share in either the relevant innovation space or in R&D activities at industry level;
• and overlap between one party's R&D project and the other party's activities in a vertically or otherwise closely related market, provided that the 40% share threshold on that market is not exceeded individually or by both parties combined.

As if this were not complex enough, the Guidelines stipulate both an exception and a counter-exception specifically for that last rule. In scenarios where one party already has a market presence, a further carve-out applies to acquisitions of start-ups with an R&D project: the Shield still applies even if the thresholds are exceeded – i.e., even if the acquirer is a large company – provided it is neither the largest firm in the relevant market nor a gatekeeper under Article 2 DMA. For gatekeepers, that counter-exception was hardened. The leaked version of the draft framed it by reference to acquisitions in “the digital sector” and, within that sector, to markets that were not just closely related to the target’s but also comprised products and services in respect of which the gatekeeper was actually designated. In other words, the gatekeeper counter-exception was limited to acquisitions close to activities that actually concerned the “core platform services” for which a gatekeeper falls within the DMA’s ambit. The final draft strips the Innovation Shield from gatekeepers simply for being gatekeepers, provided they exceed the 40% market share threshold on a market “closely related” to the target’s R&D.