Ecuador Priced an Acquiring Licence at $500,000. The Expensive Part Is Five Days.
The capital requirement is a headline. The five-day settlement rule is the business model.
The capital requirement is a one-time ticket. The settlement rule is permanent working capital, and it decides who can actually compete.
On 6 November, Ecuador's financial and monetary policy board issued a resolution that rewrites how card payments work in the country. The central bank published it the following day.
It covers credit, debit and prepaid cards, and it does something regional regulation rarely does: it addresses security, settlement, pricing, acceptance and liability in a single instrument, rather than one at a time over five years.
The headline everyone took was the price of entry. A non-bank that wants to be an acquirer in Ecuador now needs USD 500,000 in minimum capital and authorization from the central bank. A payment aggregator needs USD 250,000, and the ones already operating have a year to get there.
That is a clean, quotable number, and I think it is the least important thing in the document.
The expensive part is a different clause, three lines further down: acquirers must pay merchants for ordinary card sales within a maximum of five business days.
Why five days costs more than half a million dollars
Capital is a gate. You raise it once, you park it, you are in.
Settlement is working capital, and it never stops.
An acquirer that owes its merchants settlement within five business days is funding five days of processed volume, continuously, forever. The money leaves before it arrives. That is not a startup cost; it is a permanent balance-sheet position that grows in direct proportion to how successful you are.
Run the arithmetic on a modest book. An acquirer processing one million dollars a day — small by any regional standard — is carrying roughly five million dollars of settlement float at any moment, which is ten times the capital requirement that got it the licence. Grow to five million a day and the float is twenty-five million.
So the structure of the rule is this: the capital requirement decides who may start, and the settlement requirement decides who may grow. The second filter is far more selective than the first, and it favours exactly the institutions the rule might have been expected to open the market against — banks, which fund settlement from deposits they already hold, and large international processors, which fund it from scale.
I do not read this as a flaw or as regulatory capture. Five days is good for merchants, and merchant cash flow in this region is a genuine problem worth legislating about. But anyone reading this resolution as the opening of acquiring to new entrants should look at the second number rather than the first, because that is the one a business plan has to survive.
The question I would put to anybody preparing to use this licence: with half a million immobilised, five days of float to fund, no ability to charge more than cash, and an obligation to accept every brand and issuer, what exactly are you selling the merchant that a bank is not already selling them?
There are good answers to that — same-day reconciliation, software that runs the business rather than just taking the payment, vertical specialisation. There are no good answers that are about price.
The blank card
The most radical clause in the resolution has attracted the least analysis, and it is this: physical cards must now be issued without the card number, the expiry date or the security code printed on them.
A blank piece of plastic. The credential lives in the app.
As a security measure it is close to total against an entire family of attacks. You cannot photograph a card that has nothing on it. You cannot memorise it at a restaurant table, copy it from a hotel reception desk, or lift it from a receipt. Everything that depends on a human eye reading sixteen digits simply stops working, and no amount of tokenisation downstream ever solved that, because the number was always printed on the front.
But look at what else stops working.
Every flow that depends on a human reading the card out loud. Telephone orders. A hotel taking a reservation by voice. A clinic setting up a recurring charge at the counter. A supplier invoiced monthly who has the number written in a file. Mail order and telephone order is not a legacy curiosity in this region; it is live volume in travel, in professional services, in small business billing.
All of that now has to move into an app the cardholder opens to retrieve a number, or into a properly tokenised flow. In the long run that is where everything should be anyway. In the transition it means a forced migration with a real failure mode: the customer who cannot find the number, in the moment they need to pay, in front of a merchant who has no alternative to offer.
And it concentrates something. When the number exists only in the issuer's app, that app becomes a single point of failure for the cardholder's ability to transact at all. An outage used to mean you could not check your balance. Now it means a category of purchases cannot happen. That is a meaningful change in what an issuer's availability actually guarantees, and I would want it reflected in how the app is operated and how its downtime is measured.
The resolution pairs the blank card with a requirement for additional identity verification on online purchases, at no cost to the user, which is the right companion control. It also requires issuers to provide immediate blocking and 24/7 channels for claims and for reporting loss or theft.
A regulator wrote the liability rule
There is a clause in here that I want to hold up next to everything else happening in payments this year.
Issuers must reimburse users for transactions made after a card was reported lost or stolen, and for transactions caused by security failures at the institutions.
Read that second part again. A regulator has written, into binding regulation, a rule that says: when your security fails, you refund the customer. Not a framework. Not an audit trail that helps resolve a dispute. Not a basis for accountability. A rule, with a named party carrying the loss.
Now set that against the question the global card industry has spent eighteen months not answering — who pays when an AI agent buys the wrong thing. Networks have shipped credentials, tokens, intent data, registries and benchmarks. Not one of them has shipped a sentence that short and that clear about who absorbs the loss.
The lesson is not that Ecuador is ahead on agentic commerce; this resolution says nothing about agents. The lesson is structural, and the industry should take it seriously: when the industry declines to write a liability rule, regulators eventually write one. And a rule written by a regulator is drafted to protect the consumer, not to balance the interests of the parties who failed to agree.
There is a version of the agentic liability question that gets settled in a scheme rulebook, negotiated among the parties who understand the economics. There is another version that gets settled in a resolution like this one, by a board that has run out of patience. The industry currently gets to choose which, and the window on that choice is not open indefinitely.
From a Panamá desk
Ecuador matters here more than its size suggests, for one reason: it is the other one.
Ecuador and Panamá are the two fully dollarized economies in the region with meaningful card markets. Same currency, same absence of monetary policy, same inability to devalue, same dependence on correspondent banking to move money that is already denominated in the currency it will arrive in.
That shared structure has a consequence most commentary misses. A central bank without monetary policy still has to act, and the instruments left to it are prudential and conduct rules. So you should expect a dollarized regulator to reach more often for exactly this kind of measure — capital thresholds, settlement deadlines, pricing prohibitions, security mandates — because they are the levers that remain.
Which makes this resolution the closest thing Panamá has to a natural experiment. Not a comparison with Brazil or Mexico, where the monetary context is completely different, but a country facing the same constraint reaching for the same kind of tool.
Three things I would watch over the next eighteen months, and I think anyone operating here should be watching them too.
Whether anyone actually uses the non-financial acquirer licence. The category now exists. If nobody takes it up within two years, that is information about whether the economics work — and it is information Panamá gets for free, without having to run the experiment itself.
Whether five days becomes a regional expectation. Settlement timing is the kind of rule that travels. Merchant associations read each other's regulations. If five days becomes the norm, acquiring economics reprice across the region, and the institutions best placed to absorb it are the ones with deposits.
How the blank card actually lands. This is the one with real operational risk, and Ecuador is about to generate the only evidence anybody has. What happens to MOTO volume, what the call-centre load looks like in month two, how many cardholders cannot complete a purchase because the number lives somewhere they cannot reach. Those numbers will exist in a year, and they are worth more to a regional issuer than any amount of speculation.
My judgment
This is a more serious piece of regulation than its coverage suggests, and the coverage focused on the wrong numbers.
Five hundred thousand dollars is a headline. Five business days is a business model. Blank plastic is an operational migration that will produce a year of friction and a permanent security improvement, in that order.
And the clause nobody outside Ecuador will read — that institutions reimburse users for losses caused by their own security failures — is a better piece of liability drafting than anything the global card industry has managed to produce about AI agents in eighteen months of trying.
That should be uncomfortable, and I mean that about my own side of the industry rather than about anybody else's. We have the expertise, the data and the standing to write those rules ourselves. When we do not, somebody with less context and a different mandate writes them for us — and then we implement them anyway.
Three questions for the people who own this
1. If a five-business-day settlement obligation arrived in your market next year, what would it do to your acquiring book — and do you know, today, what your average settlement float actually is?
2. If your physical cards were issued blank tomorrow, how much of your current volume depends on somebody reading the number off the plastic? If you cannot answer that from data, you cannot assess the migration.
3. Ecuador's regulator wrote, in one sentence, a liability rule for institutional security failures. What is stopping your national association, or your scheme, from writing one that short about agent-initiated transactions — and who would have to start it?
Reply to this email with a line or two. The sharpest answers will shape the next issue.
Sources
Banco Central del Ecuador, 7 November 2026 — the central bank's account of the resolution issued on 6 November by the Junta de Política y Regulación Financiera y Monetaria: USD 500,000 minimum capital for non-financial acquirers with BCE authorization and USD 250,000 for payment aggregators, with one year for those already authorized; settlement to merchants within a maximum of five business days; the prohibition on charging more for card than for cash, including in promotions and discounts; the obligation for merchants operating through acquirers to accept cards of different issuers and brands; physical cards issued without printed number, expiry date or security code; additional identity verification on online purchases at no cost to the user; free real-time notifications through at least two independent electronic channels; immediate blocking and 24/7 claims channels; reimbursement for transactions after a reported loss or theft and for transactions caused by security failures at the institutions; the prohibition on charging interest on interest; and six months for the Superintendencia de Bancos, the Superintendencia de Economía Popular y Solidaria and the BCE to align their regulations. Link
Forbes Ecuador, November 2026 — confirms the two capital minimums as flat requirements. Link
A note on what I left out. Some coverage of this resolution reports additional requirements — PCI DSS certification, a card brand licence, and capital that scales as a percentage of monthly processed volume. I could not confirm any of those three in the central bank's own release or in the Forbes account, so they do not appear above. If they are in the full text of the resolution, the conclusions here about settlement float become stronger rather than weaker. The float arithmetic is mine and is illustrative, not drawn from any published figure.