The Austrian Supreme Court (OGH) recently upheld the unconditional clearance of a 4-to-3 merger on the EEA market for couplings for passenger trains (16 Ok 11/25s, 26 January 2026). With combined shares of 42–48%, both applicable presumptions of dominance were formally triggered. Both competition authorities, the Bundeswettbewerbsbehörde (Federal Competition Authority, BWB) and the Bundeskartellanwalt (Federal Cartel Prosecutor, BKAnw), applied for prohibition. The parties did not offer commitments. A fact pattern like that usually carries a high risk of prohibition – or at least a price, in the form of commitments. So how was this transaction cleared unconditionally?
The reason in short: despite having high combined market shares in a market with four suppliers, the parties were not close competitors. The acquirer was not considered an actual or potential competitor at all, because it did not win a single new order in the EEA in the previous five years.
Where the acquirer’s market share actually came from
Wabtec, the acquirer, held roughly 5–6% of the EEA-wide rail coupling market for passenger trains and Dellner, the target, 37–42%. The market leader, Voith, held around 50%; the fourth supplier, Knorr-Bremse, less than 5%.
A pretty difficult position from which to start the competitive analysis. However, when zooming into the competitive dynamics and the closeness of competition, the picture – largely established by the court-appointed expert – looked different.
Market structure. The relevant market for passenger rail couplings was a tendering market with infrequent, high-value tenders and very long lock-in effects. A distinction was therefore made between new orders, where competition actually took place, and follow-up orders, which were almost inevitably awarded to the player that won the initial tender.
No new tender wins. When asking where the 5–6% market share of the acquirer came from, it showed that between 2020 and 2024 Wabtec bid in 66 tenders for new train projects and won none, although the authorities tried to show that some of the acquirer’s tender wins should qualify as “new projects” (see “Between the lines” below). It won 17 of the 19 follow-up tenders it entered in the same period (for the two it did not win, the acquirer had to withdraw its offer for technical reasons). Its entire share consisted of follow-up orders on existing vehicle types – awards that, because of technical lock-in and switching costs, go to the incumbent almost automatically and are not genuinely competed for at all.
The target’s own record made the same distinction. Of the 354 tenders Dellner bid for, it won 97% of the follow-up orders and 52% of the new orders. The split between contested and uncontested awards was visible in a healthy competitor’s data too.
No close competitors. Wabtec bid in only 41 of Dellner’s 354 tenders (roughly one contest in nine). That is a distant-competitors finding in the strict sense, and it is the mirror image of Siemens/Alstom (COMP/M.8677), where the same bidding methodology showed the parties to be each other’s closest rivals and the Commission prohibited the deal. Same method, opposite facts, opposite result.
No competitive pressure. The expert found, based on interviews with competitors, that they did not consider the acquirer an effective competitor at all; it was seen as having fallen behind in recent years.
Customer concerns unfounded. When speaking with customers, the expert was told that while they feared price increases as a consequence of the concentration, they held long-term framework contracts at fixed prices, regarded the other players as viable alternatives and generally cared more about reliability than price.
The judgment: what the courts made of it
On this basis, the OGH fully upheld the first-instance Cartel Court’s judgment (28 Kt 6/25d) that adding Wabtec’s share to Dellner’s would not strengthen Dellner’s dominant position. As the OGH put it, market shares “are not in every case a suitable indicator for the economic position of an undertaking and therefore for assessing the economic effects of a merger”. Nor was Wabtec a potential competitor: theoretical ability to win a future tender was not enough, and the OGH required “tangible circumstances” making that prognosis probable – which neither authority could name.
It is worth contrasting Wabtec/Dellner with the Facebook/Giphy case, the last comparable Austrian fight. There too both applicants sought prohibition, but the Cartel Court cleared the transaction subject to conditions – non-discriminatory access to Giphy’s GIF library, among others (28 Kt 8/21t). Wabtec/Dellner is the rarer and harder outcome: the parties won the facts outright and did not offer commitments.
Between the lines: what to take away
The closeness of competition assessment, largely built by the expert on the basis of the parties’ own (tendering) data, won the case. For merging parties with significant overlaps facing an Austrian filing, this means:
Build the record before you file. The evidentiary sequence that decides a tendering-market case runs: participation, qualification or shortlisting, runner-up placings, bid gaps, head-to-head overlap and customers’ own identification of credible alternatives. Wabtec/Dellner was won on the first, second and last. In a reverse-burden jurisdiction like Austria, the evidence you cannot produce is the point you cannot make.
Own your data. The BKAnw tried to rebut the nil win rate with a list of at least two new-order wins. The court rejected it because the customer on whose statement the BKAnw relied had itself given the expert an order list showing no genuinely new orders, and the acquirer’s own records showed none either. The lesson: a well-kept data record does not merely defend you – it can destroy the authority’s best point.
The court-appointed expert’s opinion is crucial. Testable data is what convinces an economist. A coherent tender record was what carried the closeness-of-competition analysis here. Note too the expert’s technique with the customer who feared price rises: rather than record the concern, he asked what the customer would actually do if prices rose. It could not answer without disclosing the fixed-price framework contracts and the alternatives that undercut its own complaint. This shows that complainants, too, must substantiate their concerns with testable data and a credible counterfactual.
Assume the facts lock at first instance. The OGH fully upheld the Cartel Court’s first-instance judgment, largely because factual findings can be attacked on appeal only in a very limited way. In practice, the first-instance record is the record. Plan accordingly.
The case gives a rare insight into how the Austrian courts assess a difficult merger. But it also points to a few procedural particularities. Who exactly is this court-appointed expert? Why was the case largely dealt with by the courts rather than the competition authorities? We will give an overview of the procedural side of Austrian merger control in one of our next posts.
Disclosure: This article was written the way we work. The topic, the analysis, the arguments and the final wording are ours. We use AI as research and drafting tool along the way. The result is meant to speak for itself.