mBank is quitting forex while Poland's CFD market grows 50 percent

mBank's brokerage arm will stop letting clients open new forex and CFD positions on 1 October, according to Finance Magnates. That leaves Dom Maklerski BOS and Alior Bank's brokerage unit as the last traditional Polish banking houses still offering the product. The detail that makes this interesting is not the exit. Banks exit product lines constantly. It is that mBank is walking away from a market that grew by half in a single year, which means the decision was not made on revenue.

The numbers a bank board actually looked at

The Polish Financial Supervision Authority counted 369,737 active clients trading over-the-counter derivatives through domestic brokerages in 2025. A year earlier the figure was 246,826. That is a rise of roughly 50 percent, in one year, in a single mid-sized European market. On any conventional reading of a product line, that is growth worth defending.

The second set of numbers explains the exit. The KNF recorded that 72.2 percent of active clients closed 2025 with a loss. That sits inside the 70.6 to 79.1 percent band the regulator has logged every year since 2021, so it is not an anomaly or a bad market. It is the steady state. Combined client losses came to 2.68 billion zlotys, which the regulator noted was close to four times what the winning clients collectively made.

Put those two datasets next to each other and the board paper writes itself. The product is growing quickly, it is profitable for the house, and roughly three in four of the bank's own customers lose money on it every year with a consistency that no amount of education has moved in five years. For a standalone broker that is a business model. For a universal bank that also sells those customers mortgages, deposits and pensions, it is a reputational liability sitting inside the same customer relationship.

Why bank-owned brokers keep losing this argument

The structural problem for a bank-owned CFD desk is that it cannot be evaluated on its own profit and loss. An independent broker measures the desk against its cost of acquisition and its regulatory capital. A bank measures it against the same numbers plus a set of costs an independent never carries: the conduct risk sitting on the group licence, the supervisory attention the product attracts across every other line, and the awkwardness of a retail customer who lost money on leveraged contracts walking into a branch to discuss a mortgage.

European regulators have been steadily raising the cost of the first of those. The Polish regulator fined XTB, the country's largest listed broker, around 5.5 million dollars over CFD marketing rules, which is a substantial penalty in a market of Poland's size and a clear statement about where supervisory attention is pointed. A fine of that shape is survivable for a specialist. Applied to a universal bank, the same conduct finding contaminates a licence that supports several billion in unrelated business.

So the banks leave, one at a time, and the specialists stay. That has been the direction of travel across Europe for most of a decade, and Poland is simply a clean, well-documented example because the KNF publishes the client outcome data that most regulators do not.

• 369,737 active OTC derivatives clients in Poland in 2025, up from 246,826
• 72.2 percent of active clients closed the year at a loss
• The loss rate has stayed between 70.6 and 79.1 percent every year since 2021
• Combined losses of 2.68 billion zlotys, close to four times the winners' gains
• Two bank-owned brokers left in the market after 1 October

What the specialists inherit

The obvious read is that this is good news for the specialist brokers. A growing market with one fewer competitor, and a competitor whose distribution advantage came from an existing banking relationship rather than from a better product, is a straightforward win. mBank's departing clients will go somewhere, and the firms with Polish licences and Polish-language support are the natural destination.

The less obvious read is that the specialists also inherit the scrutiny. Supervisory attention does not evaporate when a bank exits a product. It concentrates on whoever is left. A regulator that has watched a universal bank conclude the product is not worth the conduct risk does not draw the conclusion that the product is fine. It draws the conclusion that the remaining participants are the ones carrying the risk it just watched a bank refuse, and it supervises accordingly.

That dynamic is already visible in how the larger Polish specialists talk about themselves. XTB has spent several years repositioning around licensing breadth and product diversification rather than leveraged volume, and the sponsorship strategy that put its brand on FC Porto shirts is part of the same move: build a brand that survives the product mix changing underneath it. Firms that cannot make that move are the ones with a problem when the regulatory attention arrives.

Why the loss rate never moves

The most striking thing in the Polish dataset is not the level of the loss rate. It is the stability. Between 70.6 and 79.1 percent every year since 2021, through a rate cycle, an energy shock, a war on Europe's eastern border and a 50 percent expansion in participation. Nothing in the market environment has moved it, and neither has five years of mandatory risk warnings, negative balance protection and leverage caps across the European regime.

That stability is the argument the industry cannot answer. If the loss rate moved with market conditions, a firm could reasonably say that clients do badly in difficult years and well in easy ones, and that the product is a tool whose outcome depends on how it is used. A rate that sits in a nine-point band across five wildly different years is describing something structural about the product and the population using it, not something situational. Regulators have noticed, which is precisely why the KNF publishes it.

The industry response has been education, and the data suggests education has not worked, at least not at the scale it has been delivered. That is worth saying plainly rather than defensively, because the alternative reading, that clients simply need more warnings, has now been tested for five years in a market that publishes the results. Firms building a serious answer are moving toward smaller position sizing defaults, genuine suitability gating and revenue that does not depend on client turnover, which are product changes rather than disclosure changes.

What happens to the clients

A practical question gets lost in the strategic framing: where do mBank's leveraged clients go on 2 October. The bank is stopping new positions, which means existing positions can be managed down rather than force-closed, but the client who wants to keep trading has to open an account somewhere else. The Polish market has capable domestic specialists and a deep bench of European brokers passporting in, so there is no shortage of destinations.

The risk sits in the tail of that migration. A client leaving a bank-owned broker is leaving an environment with conservative default leverage, an onboarding process designed by a bank's compliance function, and a complaints route that terminates at a banking ombudsman. Not every destination replicates that. The clients most likely to shop aggressively for replacement are the ones who found the bank's limits restrictive, which is to say the ones for whom those limits were doing the most work.

That is the quiet cost of a bank exiting a retail product on conduct grounds. It removes the most conservative operator from the market and redistributes its clients across operators who are, by construction, less conservative. The regulator gets a cleaner banking sector and a marginally riskier retail derivatives population, which is a trade that makes sense from a prudential seat and looks different from a consumer protection one.

The product is growing quickly, and three in four of the bank's own customers lose money on it every year.

The signal for brokers outside Europe

For firms operating in Southeast Asia and the Gulf, the Polish case is worth reading as a preview rather than a curiosity. The sequence is consistent wherever it has played out: rapid growth in retail participation, a regulator that begins publishing client outcome data, a widening gap between the growth narrative and the outcome data, and then institutional participants quietly withdrawing before the rules tighten rather than after.

The markets in the region that have started down this path are the ones where the regulator has begun collecting and publishing the loss statistics. Once that dataset exists and is public, the argument about whether the product serves the retail client stops being a matter of opinion, and every participant has to answer it with numbers rather than with education programmes. Brokers that have already built a diversified product mix and a verifiable disclosure practice will find that a manageable conversation. Those whose entire revenue is leveraged volume against an undisclosed loss rate will not.

mBank did not leave because the business was failing. It left because a bank that publishes a loss rate of 72 percent alongside a mortgage book eventually has to explain the relationship between them, and no explanation is good enough. The specialists who inherit the market inherit that question too. The difference is that they have no mortgage book to protect, which is an advantage right up until the moment a regulator decides the question deserves an answer anyway.