Your landlord and your landlord's rival may take advice from the same program

A revenue-management product already common in commercial property would, in a scenario set out by competition lawyers, become standard across the residential lettings of a single city. Landlords would subscribe individually, upload their own vacancy and rent figures, and receive a daily recommendation. That recommendation would occasionally be to leave a flat empty, on the grounds that accepting less would damage the price of every other flat the model is tracking.

Each landlord would be acting alone. The scenario depends on it. There would be no meeting, no message, no understanding and no agreement — only several hundred subscribers taking the same advice from the same system in the same hour, and a market that behaves exactly as it would if they had all sat down in one room.

A tenant would notice only the effect: a listing that holds its price for weeks, a viewing cancelled without explanation, three flats on one landing dark for a month, and a rent that refuses to soften in a soft market. A landlord would notice only the dashboard, which would say hold, and which would be right often enough to be trusted and wrong in a way nobody has to sign.

The vendor's position would be narrow and durable. The software recommends; it does not set. Any subscriber may decline. That most of them would not decline would be described as a measure of quality rather than of market power, and the acceptance rate would appear in the sales material as a strength.

Competition law would be left holding a definition. Concerted practice requires coordination between undertakings; a cartel requires an agreement; both were written for people. A market can align to within a euro without a single undertaking having coordinated with another, because the coordination would sit in the middle, sold as a subscription, accountable to its customers and to nobody else.

A city council would respond the way councils do, with a local ordinance banning the use of such recommendations in setting residential rents. Enforcement would be the hard part. A ban on advice is difficult to police when the advice arrives privately, calls itself a forecast, and is followed by people who would tell a tribunal, accurately, that they had simply reached the same conclusion.

The scenario is not a prediction of collusion. It is a description of what collusion would look like if nobody did it: a market moving in step, with the mechanism in plain view, a subscription fee, a support line, and no person at any point who chose the outcome.