A normal Thursday—until one sentence removed the floor
For more than three years, EUR/CHF had carried an unusual promise. The Swiss National Bank said it would not tolerate a rate below CHF 1.20 per euro and was prepared to buy foreign currency in unlimited quantities. Traders could see the floor, code around it and build strategies that treated the downside as politically defended. That did not make the trade risk-free, but it made the distribution of risk look far more comfortable than it was.
On 15 January 2015 the SNB discontinued that minimum exchange rate without a transition period. It simultaneously cut the interest rate on qualifying sight-deposit balances to −0.75% and moved the target range for three-month Libor deeper below zero. The same institution that had promised unlimited defence now said the defence was no longer justified.
The result was not one clean candle. Quotes disappeared, spreads widened by orders of magnitude and different venues printed different lows. Market orders and stop-losses waited for the next executable counterparty. When that counterparty appeared far away, positions closed at prices clients had never imagined—and sometimes after the entire account balance had already vanished.
This investigation separates three records that are often mixed together: the SNB policy decision, observable market prices and the losses later reported by brokers. The charts use official ECB daily reference rates. They prove the scale and persistence of the repricing, but not the lowest tradable tick. That limitation is part of the story: in an over-the-counter market under extreme stress, there was no single universal price.
Why 1.20 looked like a wall
The SNB introduced the minimum rate on 6 September 2011. It described the franc as massively overvalued, an acute threat to the Swiss economy and a source of deflation risk. The commitment was unusually direct: the central bank would enforce CHF 1.20 per euro with the utmost determination and stood ready to buy foreign currency in unlimited quantities.
That language changed behaviour. Exporters gained a planning boundary. Banks and funds could hedge around an explicit policy level. Retail traders saw a famous asymmetric trade: buy EUR/CHF near 1.20, place a stop just below the floor and expect the central bank to stand between the position and a large fall. The small potential upside encouraged large leverage because the recent historical range looked compressed.
But a policy floor is not a put option issued to every trader. The SNB never guaranteed continuous liquidity at 1.20, never promised a notice period and never promised to compensate anyone if policy changed. It controlled its own interventions, not the solvency of every dealer or the execution price of every client order.
By early 2015, monetary policy across major currency areas was diverging. The euro had weakened substantially against the dollar and expectations of European Central Bank quantitative easing were intensifying. Maintaining the floor meant absorbing increasing pressure into the SNB balance sheet. In its announcement, the bank said the franc’s overvaluation had diminished and enforcing the floor was no longer justified.
The official ECB series shows how thoroughly the commitment had pinned the pair: daily reference rates clustered just above 1.20 for years. The stability was real, but it was policy-produced. When the policy disappeared, the historical volatility used by many margin systems was no longer a useful description of the next minute.
EUR/CHF · 2011–2015
1 Aug 2011–27 Feb 2015 · daily ECB reference rate
The first hour: policy decision becomes market accident
The promise
SNB sets a minimum exchange rate of CHF 1.20 per euro and says it is prepared to buy foreign currency without limit.
The last calm reference
The ECB reference rate is 1.2010. The floor still looks intact; positioning and broker margin rules reflect the recent low-volatility regime.
The release hits
SNB discontinues the floor and cuts the relevant sight-deposit rate to −0.75%. There is no staged exit and no new lower floor.
Bids vanish
EUR/CHF falls through levels where stop-losses were waiting. Dealers receive fewer usable quotes; spreads and venue-to-venue differences explode.
Forced liquidation
Margin engines try to close losing positions, but the next available price can be far below the stop or margin threshold. Some accounts become negative before execution completes.
A new regime
The ECB reference rate is 1.0280—14.4% below the previous day. Intraday prints on individual venues were lower, but no single tick represented the whole fragmented market.
A stop-loss is an instruction, not a guaranteed price
In a liquid market, a stop order feels almost continuous: a trigger is reached, the order becomes marketable and the fill arrives nearby. That experience creates a dangerous mental shortcut. The stop price is normally the trigger, not a contractual guarantee of the fill. Between trigger and execution there must still be a willing counterparty.
On the SNB morning, the order book did not merely become thin. At many dealers it became discontinuous. Liquidity providers paused, widened quotes or protected themselves while trying to discover where the franc should trade without the central-bank floor. A client selling EUR/CHF found that the next credible bid could be thousands of points lower.
Retail platforms showed different histories because Forex is decentralised. One broker’s chart represented its own liquidity relationships and filtering rules; another showed a different composite. Trades executed at distinct prices could all be genuine. This is why screenshots of a single spectacular low cannot establish one official market bottom.
The leverage transformed a currency move into a solvency event. A €100,000 EUR/CHF position controlled with €1,000 of margin at 100:1 leverage loses roughly CHF 1,000 on a one-cent fall before conversion and costs. A fall from around 1.20 to near 1.00 is not a routine drawdown: it can exceed the posted collateral many times over before a closing order finds a price.
The broker then faces a second-order loss. If a client owes more than the cash in the account, the shortfall becomes a receivable from that client. Thousands of simultaneous receivables are not cash, may be legally difficult to collect and immediately weaken the broker’s regulatory capital. The clients’ gap becomes the dealer’s balance-sheet gap.
EUR/CHF · shock window
1 Dec 2014–27 Feb 2015 · daily ECB reference rate
The shock travelled from price screens to company balance sheets
The numbers below are not estimates of total industry damage. They are publicly reported examples showing four different transmission paths.
Alpari UK
Large client losses exceeded account equity, leaving negative balances that the broker could not absorb. Weekend rescue talks failed and administrators were appointed.
FXCM
FXCM reported about $225 million in negative client balances and fell below regulatory capital requirements. A $300 million Leucadia financing kept the firm operating.
Interactive Brokers
The firm said several customers suffered losses beyond their collateral, producing an aggregate shortfall of about $120 million—material, but absorbable relative to its capital.
Retail clients
Some customers lost more than their deposits because liquidation occurred only after the market reopened at executable prices. Contract terms and jurisdiction determined who ultimately carried the debt.
Same market, different outcome: what separated survival from collapse
The event did not prove that every retail broker secretly traded against its clients, nor that an agency model automatically removes broker risk. A broker can pass a client position to a liquidity provider and still remain exposed to timing differences, rejected or repriced hedges, collateral calls and a client who cannot pay a negative balance.
Capital mattered. Interactive Brokers could recognise a large client deficit without threatening the company’s existence because the loss was small relative to its net worth. FXCM needed emergency financing to restore regulatory capital. Alpari UK did not secure a rescue. The same external shock produced different corporate outcomes because buffers, concentrations, client mix and funding access differed.
Margin policy mattered too. Low margin attracted activity but allowed a relatively small client deposit to support a position whose gap loss could be enormous. Dynamic margin can reduce exposure before a known event, but the SNB decision was not scheduled. A model calibrated to the previous three years saw stability precisely because the central bank had suppressed the volatility it was supposed to measure.
Negative-balance language mattered after the trade, not during it. Where protection was contractual or later imposed by regulation, a retail client’s loss could be limited to account funds and the provider would absorb the rest. Where it was absent, the provider could pursue the deficit. The legal outcome depended on the entity, contract, client classification and local law.
Client-money segregation addressed another question: whether unencumbered customer money was held apart from the firm’s own funds. It did not eliminate trading losses, stop a company entering insolvency or guarantee immediate access during administration. Segregation is essential protection, but it is not insurance against market gaps.
The most revealing failure was conceptual. Traders, brokers and risk systems treated the absence of recent volatility as evidence that a violent move was improbable. In reality, the calm existed because risk had accumulated behind a public-policy barrier. The floor did not remove the energy; it postponed price discovery.
How a €1,000 account could owe money in minutes
A simplified example shows the sequence. It is not a reconstruction of one client statement and excludes financing, conversion and recovery costs.
1. Position
A client deposits €1,000 and buys €100,000 of EUR/CHF near 1.2000. At 100:1 leverage, the margin looks sufficient while the floor holds.
2. Trigger
The SNB removes the floor. A stop at 1.1950 is triggered, but it only releases an order into a market with few or no bids near that level.
3. Gap
The next executable price may be near 1.05 or 1.00. A 0.15-franc move on €100,000 implies roughly CHF 15,000 of market loss before conversion and costs.
4. Deficit
The deposit is exhausted. The remaining loss becomes a negative account balance—and, from the broker’s perspective, an unsecured client receivable.
5. Contagion
When this happens across many accounts at once, the broker must fund liquidity providers and regulatory capital immediately while attempting to recover client debts later.
What changed—and what did not
The event accelerated a reassessment of retail leverage, margin buffers, liquidity-provider concentration and negative-balance policies. Some brokers forgave deficits; others sought payment. Regulators and firms stress-tested gap scenarios more explicitly and raised margin around concentrated political or central-bank risks.
In the European Union, ESMA’s 2018 CFD intervention later introduced leverage limits, account-level margin close-out and negative-balance protection for retail clients. The Swiss-franc shock was part of the history that made these weaknesses impossible to ignore, but it would be too simple to claim that one morning alone produced the entire reform package.
A protected retail balance is not the same as a guaranteed execution price. Negative-balance protection can cap a client’s aggregate loss at the funds in the CFD account; it does not make a stop fill at the requested level, prevent the position from being closed or protect professional clients under the same terms.
Central-bank credibility also survived in a more complicated form. The lesson was not that policy statements are meaningless. The 1.20 commitment influenced the market for years. The lesson was that a credible policy can end abruptly when the institution decides its costs exceed its benefits. Credibility applies to the policy while it exists—not to a trader’s assumption that it must exist tomorrow.
The modern parallel is any market whose apparent stability depends on one actor defending a threshold: a currency peg, managed band, volatility-control product, stablecoin or liquidity backstop. The instruments differ, but the analytical question is identical: what happens to price, execution and counterparties if the defender steps away before positions can be reduced?
USD/CHF · cross-market repricing
1 Dec 2014–27 Feb 2015 · derived from ECB reference rates
Eight checks before trading near a defended price
- Treat a peg, floor or intervention promise as a reversible policy—not as a guaranteed stop.
- Model a discontinuous gap of 10–30%, not only the worst candle in the recent chart.
- Read whether stops are guaranteed, how slippage is handled and when orders may be rejected.
- Confirm negative-balance protection for your exact legal entity and client classification.
- Compare margin requirements with the notional position and calculate the loss per full price unit.
- Check broker capital, regulator, client-money rules and dependence on a small liquidity-provider set.
- Avoid concentrating several strategies in the same hidden macro bet or defended threshold.
- Keep enough uncommitted capital to survive platform outages, widened spreads and delayed execution.
The franc did not break Forex. It exposed what Forex had hidden.
15 January 2015 is remembered as the day a currency moved by an extraordinary amount. Its deeper significance is that the event made the market’s contractual plumbing visible. A price chart showed the fall; insolvencies and emergency loans showed who had promised liquidity, who had provided leverage and who ultimately owned the gap.
The floor created genuine stability, but it also encouraged positions sized for a world in which the floor persisted. When policy changed, liquidity disappeared before risk systems could translate a new regime into orderly exits. A stop was not a price, margin was not a maximum loss and a broker label was not a capital guarantee.
The useful question is therefore not “Could the SNB shock happen again in exactly the same pair?” It is “Where does today’s calm depend on a promise that can be withdrawn?” That question belongs in every risk model built around a peg, a backstop or an unusually narrow historical range.
Questions the event still answers
Was EUR/CHF officially pegged at 1.20?
The SNB called it a minimum exchange rate and committed to defend it with foreign-currency purchases. It was a policy floor, not a legal guarantee of execution for private trades.
Why do charts show different lows?
Spot Forex is decentralised. Liquidity providers and brokers received and filtered different quotes during a market vacuum, so there is no single universal intraday low.
Could a stop-loss fail?
The trigger could work exactly as designed while the fill occurred far away. A normal stop becomes a market order after triggering; only an explicitly guaranteed stop promises a price under stated conditions.
Can retail CFD clients still go negative in Europe?
ESMA introduced account-level negative-balance protection for retail CFD clients in 2018. Coverage depends on product, jurisdiction, entity and client classification, so the exact contract still matters.
How a stop-loss executes: trigger price, slippage and quotation gaps
Risk note
Leveraged Forex and CFD positions can lose capital rapidly, especially during gaps and liquidity disruptions. This investigation is educational and is not personal investment advice.
Sources and methodology
We use a compact source set: the two SNB decisions establish the policy record, ECB reference rates anchor the charts, public reporting documents the broker losses, and ESMA records the later retail-CFD protections.
ECB EXR.D.CHF.EUR.SP00.A · daily reference rate at 14:15 CET · USD/CHF derived from matching CHF/EUR and USD/EUR observations