Information exchange between competitors can infringe competition law on its own — no price-fixing agreement required — when it removes strategic uncertainty about future prices, volumes or plans. Hub-and-spoke arrangements extend the risk vertically: competitors coordinating through a shared supplier, retailer, consultant or algorithm can be liable as cartelists though they never met. The 2023 EU horizontal guidelines and the algorithmic-pricing cases have redrawn this frontier.
Information exchange and hub-and-spoke collusion are where cartel law reaches conduct that feels innocuous from inside a business: benchmarking, market colour, category-management chats with a common retailer, adopting the same pricing software as everyone else. This guide maps the legal tests, the leading cases and the modern algorithmic frontier — closing the cartel pillar of our Competition & Antitrust hub.
When is information exchange illegal by itself?
When it reduces strategic uncertainty between competitors — above all exchanges of future prices or quantities, which the EU treats as by-object infringements. One meeting can suffice, and passive receipt without public distancing counts as participation.
What is a hub-and-spoke cartel?
Horizontal coordination achieved through a vertical intermediary: competitors (spokes) align conduct via a common supplier, customer, platform or adviser (hub) that relays intentions — each spoke knowing the hub is doing the same with rivals.
Where does algorithmic pricing fit?
Three risk layers: using algorithms to implement an explicit cartel (clearly illegal); many rivals feeding data to one pricing provider that aligns them (hub-and-spoke — the RealPage theory); and autonomous algorithmic alignment without human agreement (the contested frontier legislators are studying).
Why can merely sharing information be a cartel-level offence?
Because competition presupposes independent decision-making under uncertainty. Article 101, Section 1 of the Sherman Act and Türkiye’s Article 4 all condemn “concerted practices”: knowingly substituting practical cooperation for competition’s risks. When rivals tell each other what they will charge next quarter, each prices with the rival’s intentions in hand — the market outcome converges on coordination without any promise being exchanged.
The law’s machinery reflects this. Exchanging future pricing or volume intentions is treated as restrictive by object — illegal without effects analysis. Exchanges of current or recent granular data are assessed by effect, weighed by market concentration, data age, aggregation and frequency. And the Anic presumption does the evidentiary work: an undertaking that receives competitor information at a meeting is presumed to have used it unless it publicly distanced itself — silence in the room is participation, a rule that decides most contested cases.
What do the leading information-exchange cases hold?
T-Mobile (EU, 2009): a single meeting where Dutch mobile operators discussed reducing dealer commissions sufficed for a concerted practice — frequency is irrelevant when the subject is strategic. UK Agricultural Tractor Registration Exchange: even a genuine statistics system infringed where it identified individual rivals’ sales in a concentrated market. Bananas (EU): weekly pre-pricing calls about quotation-price factors were a by-object infringement though no price was ever agreed.
More recent enforcement extends the line: the Commission’s car-emissions decision fined coordination on the scope of technical development (AdBlue tank sizing) — an exchange about innovation parameters, not prices; financial-sector cases (bond-trading chatrooms) treat trader information flows as cartel conduct; and Türkiye’s Rekabet Kurulu has fined pure information-exchange arrangements in retail and labour markets, applying the same by-object logic to wage benchmarking among employers. The perimeter, in short, now encloses any systematic channel through which rivals learn each other’s forward-looking conduct — the compliance rules for which are collected in our competitor-contact guide.
The UK’s replica football kit and Hasbro/Argos/Littlewoods toy cases built the modern hub-and-spoke template: retailers and a common supplier exchanged retail-pricing assurances — A tells hub it will price at X if rivals do; hub relays to B; B prices accordingly. Liability attached horizontally to the spokes although they never communicated directly, on the state-of-mind test later refined in case law: a spoke is liable where it may be taken to have intended its information to reach rivals and to influence their conduct, and where rivals used it on that basis. The doctrine now polices category management, resale-price ‘policing’ complaints and platform-mediated pricing worldwide.
How does hub-and-spoke liability actually attach?
The test triangulates intention and use: spoke A discloses future conduct to the hub in circumstances where it foresees relay to rivals; the hub relays; spoke B receives and uses it, knowing its provenance. Each element is provable by inference from context — category-management meetings where a retailer reports “the market’s” pricing intentions, a supplier assuring each distributor that others will follow the increase, a consultant circulating a “forecast” assembled from clients’ confidential plans.
The hub is liable too: as cartel facilitator, on the AC-Treuhand principle (the Swiss consultancy that organised heat-stabiliser cartel meetings was fined though it sold nothing in the market). That principle reaches modern hubs directly: platforms, data vendors, benchmarking consultancies and software providers whose service consists in aligning competitors. For businesses at the centre of vertical information flows — distributors, franchisors, marketplaces — the design rule is strict unidirectionality: what one competitor tells you must never be retransmitted to another in identifiable, forward-looking form.
Where does algorithmic pricing cross the line?
Sort the scenarios. Algorithm as implement: rivals agree and let software execute — a plain cartel (the Amazon-marketplace posters case ended in convictions). Algorithm as hub: many competitors feed non-public data to one provider whose engine recommends aligned prices — the theory in the US RealPage litigation over rental pricing and in parallel hotel revenue-management suits, and the configuration EU and UK regulators have flagged as hub-and-spoke. Autonomous alignment: independent self-learning algorithms converging on supra-competitive prices without any agreement — currently outside most prohibitions, intensely studied, and the object of early legislative drafting.
Compliance posture for the middle scenario is already clear: diligence any shared pricing or revenue-management vendor for whose data trains the model and whether recommendations are built from rivals’ non-public inputs; contract for data segregation; and treat “everyone in our industry uses the same tool” as a risk statement, not a comfort. Authorities have said in terms that outsourcing your pricing to a common brain does not outsource liability.
How should companies run benchmarking and market intelligence safely?
Through structure. Industry statistics: independent aggregator, five-participant minimums, no cell where one firm dominates, data at least several months old, outputs in aggregate form only. Benchmarking clubs: written protocols, no pricing or capacity forwards, counsel review of questionnaires. Trade-association dashboards: same rules plus agenda discipline. Buying “market colour” from consultants: provenance warranties in the engagement letter.
Internally, train the receivers: the salesperson who is emailed a competitor’s price list by a customer, the analyst offered a rival’s forecast at a conference. The safe drill is documented non-use — do not circulate, notify legal, respond distancing where the source is a competitor. Most information-exchange liability is contracted passively, by companies that never sought the data but visibly kept it in the decision loop. The broader architecture — training, monitoring, audits — belongs to the compliance program covered in our compliance pillar.
How do platforms and booking systems create spoke liability?
The Eturas case supplies the template: a Lithuanian online travel-booking platform messaged member agencies that discounts would be technically capped at 3%; agencies that saw the notice and kept using the system without objecting were, the EU Court held, presumptively participants in a concerted practice. A system administrator’s broadcast plus passive continuation equalled collusion — no meeting, no reply, no bilateral contact.
The doctrine transfers to every shared commercial infrastructure: marketplace seller tools announcing ‘recommended’ price floors, franchise systems relaying outlets’ pricing intentions, booking engines synchronising member rates. For platform operators the design duty is to avoid becoming a coordination device — no broadcast of members’ future commercial parameters, no rule-changes that align competitors’ prices announced through the tool. For members, the Eturas drill applies: object visibly, in writing, through the same channel — silence subscribes you to whatever the system just organised.
How are concerted practices proven without documents?
Through convergence of circumstantial strands, each innocent alone. Communications metadata: calls and meetings clustering before synchronized price moves. Economic evidence: parallel conduct inexplicable under independent incentives — price increases against falling demand, stable shares in volatile markets, margin patterns breaking at contact dates. Behavioural tells: announcements phrased for rivals rather than customers, freight surcharges identical to the decimal across carriers.
Courts police the inferences — parallelism alone never suffices, and defendants win where an independent explanation (common cost shocks, price leadership followed unilaterally) fits the data as well. But the evidentiary economy has shifted with digital exhaust: expense systems, calendar data, message-app records and pricing-system logs give authorities reconstruction abilities the classic cases never faced. The practical compliance conclusion is uncomfortable and clarifying: assume conduct will be legible in retrospect, and manage the contacts themselves — the subject of our competitor-contact rules — rather than trusting the absence of a smoking gun.
What does a defensible information-governance policy look like?
Written, specific and audited. It inventories every channel through which competitor information can arrive — association statistics, benchmarking vendors, category managers, JV boards, new hires from rivals, customers relaying quotes — and assigns each a rule: permitted form (aggregated, historical, public), prohibited content (forward prices, volumes, capacity, bids, wages), owner, and escalation path. New hires from competitors sign onboarding attestations and quarantine periods for sensitive knowledge; sales teams get scripts for the customer who offers a rival’s quote.
Enforcement makes it real: periodic audits sampling actual inboxes and chat channels against the policy, a no-fault reporting route for contamination events, and visible consequences for breaches. Documentation converts the program into evidence — the company that can show trained rules, an audit trail and documented distancing responses defends a concerted-practice allegation from a different posture than one relying on witness memory. Regulators read policies cynically and audit trails seriously; invest accordingly.
Does signalling through public announcements count?
It can. Unilateral public price announcements made well in advance of effect, addressed in substance to competitors rather than customers, have been treated as invitations to collude — the US container-shipping and airline cases, the EU’s liner-shipping commitments and cement-sector scrutiny all police ‘cheap talk’ that structures rivals’ expectations. Where announcements are followed by rivals’ matching statements and then by aligned implementation, authorities read the sequence as concertation achieved in public.
The compliance line: announce prices when commercially necessary, as close to effect as practicable, in terms serving customers (final prices, dates, products) rather than rivals (percentage intentions, ‘industry needs’, conditional language inviting response). Investor-relations calls need the same filter — earnings-call statements about ‘pricing discipline’ and capacity intentions have drawn enforcement attention on both sides of the Atlantic, and scripts deserve counsel review in concentrated markets.
Frequently Asked Questions
Is it illegal to look at competitors’ public prices?
No — monitoring public prices, scraping public websites and reacting unilaterally is competition working. The line is non-public, forward-looking or systematically exchanged data — and public *announcements* aimed at rivals as signals can still be problematic.
Can we discuss the market with competitors at conferences?
General, backward-looking, public-information conversation is survivable; anything touching your or their future prices, capacity, bids, customers or strategy is not. The safe rule: talk about the industry’s past, never anyone’s future.
Is receiving one unsolicited competitor email really liability?
It can be, under the presumption of use, if you stay silent and keep pricing. The documented response — distancing reply, legal notification, non-circulation — converts the same email into evidence of your compliance instead.
Are wage and HR benchmarks caught too?
Yes — compensation is a competitive parameter in labour markets. Wage-fixing and no-poach cases in the US, EU states and Türkiye treat HR information exchange exactly like price exchange; use independent aggregated surveys, never direct salary-sharing with talent competitors.
Discover more from Kurums | Business Intelligence
Subscribe to get the latest posts sent to your email.
