Brink's-Atleos should go to phase two, not be settled with a promise

A strange price for a dying business

Six and a half billion dollars is a lot to pay for an industry everyone agrees is shrinking. Brink’s, described in the trade press as a cash management giant, has agreed to buy the ATM operator NCR Atleos for $6.6 billion. Britain’s competition watchdog now says it will take the deal into an in-depth second-phase investigation — unless the parties do something first to address its concerns.

That conditional is not a throwaway. UK merger control works in two stages. The first is a screening exercise on a tight statutory clock: the authority decides whether it has a realistic prospect of a substantial lessening of competition. If it does, it must refer the deal for a full inquiry — unless the parties offer undertakings in lieu of reference, commitments usually involving the sale of a business, which the authority can accept on the strength of that first-pass view. Phase two takes months and produces evidence. Undertakings in lieu take weeks and produce a promise.

In most deals that trade-off is sensible. In this one it is not.

The savings are the harm

Cash logistics is a density business. Armoured vehicles, cash centres and machine estates are fixed costs spread over a falling volume of notes. As volume drops, unit costs rise, and consolidation becomes the rational response: one network instead of two, one route instead of two, one cash centre serving a region instead of a pair running half empty. Anyone paying a premium for a business in a declining market is paying for the cost that can be taken out of it.

Which means the efficiency case and the competition concern are the same fact seen from two directions. The acquirer’s synergy model is a list of duplicated infrastructure. A competition remedy that keeps a rival alive on paper does nothing about that list. Divest a contract book, and the vans still get rationalised. Accept a behavioural promise on pricing, and the price of a service is capped while the availability of the service is not.

That is the structural reason to prefer the long inquiry here. The question is not what the combined firm will charge a supermarket next year. It is what the physical network looks like in five years, and whether anyone other than the owner gets a say.

Why the timing makes it worse

Cash has quietly become a resilience asset rather than a payments product. Last week the Bank of England’s Financial Policy Committee recorded that rapid advances in artificial intelligence have increased cyber and operational resilience risks, and that interconnected vulnerabilities are more likely to crystallise together. The Reserve Bank of Australia’s financial stability review the same week flagged critical service provider disruption alongside geopolitical risk. The whole architecture of payments resilience assumes that when one rail fails, another is reachable.

Cash is the rail that does not depend on anyone’s cloud. But its reachability is entirely a function of physical coverage: whether there is a machine, whether it is filled, whether a shop can bank its float. None of that is captured by the competition test, which asks about rivalry, not about coverage. A regulator can clear a deal, correctly, on its own statute, and still have waved through a reduction in the country’s ability to pay for things when the card networks go dark.

The answer is not to pretend merger control is resilience policy. It is to use the stage of the process that can actually gather the evidence — including whether cash volumes are falling as fast as the parties will claim, and what the combined firm’s own internal plans say about capacity.

The case for settling it quickly

The strongest argument the other way is proportionality. Phase two is expensive, slow and chilling; the authority has been pressed to weigh growth and investment in how it polices deals. If the overlap is genuinely narrow — two service lines in defined places — a sale fixes it in weeks and spares everyone a year. And in declining markets, blocking consolidation can produce exit rather than competition: a sub-scale operator that cannot cover its fixed costs does not become a vigorous competitor, it becomes a withdrawal.

All of that may be true. But proportionality governs the size of the remedy, not the depth of the analysis. A screening test cannot establish how fast a market is really declining, and the failing-market argument is precisely the kind of claim that needs testing rather than asserting. If it is right, it survives phase two — and the parties should want it to. An unconditional clearance after a full inquiry is worth more, commercially and politically, than a divestment nobody believes will hold. The deals that most need the long process are the ones where the acquirer’s best defence is that there is no future worth competing over.

What would settle it

Only one document matters now, and it is not the press summary: the authority’s own phase one decision, with its theory of harm named. If that theory is a narrow overlap, sell the overlap. If it is control of the chain from cash centre to machine, no promise should be accepted and the reference should be made.