Swiss Franc on Hold: Reading ING's Policy-Rate Call Against the SNB's Own History

Eleven years ago this January, a Thursday morning press release from Zurich erased the stop-loss discipline of an entire retail industry in roughly fifteen minutes. We mention it not for drama — the SNB cliché is banned on this desk — but because every "Swiss franc on hold" note written in 2026, including ING's, is read by operators whose risk committees still carry the January 15, 2015 scar in their margin models. The call that the policy rate stays put is the easy sentence. The harder question — the one this audit answers — is what "on hold" has cost, in basis points of spread and in leverage withdrawn, every quarter since.

Methodology: What We Audited, What We Could Not

We treated the ING "policy rate seen on hold" line not as a forecast to agree with or dispute, but as an artifact to be read against the SNB's own record. Our audit crossed three source layers. First, the Swiss National Bank's published statements and press releases from the floor years and the 2015 unpeg — the primary archive. Second, the regulatory filings and account-tier documentation of the brokers in our grounding set — AvaTrade, Exness, FBS, FXTM, and HFM — plus the public operator record of IG Group, CMC Markets, Saxo Bank, IC Markets, Pepperstone, and XM. Third, the cost structure those brokers now advertise on CHF-containing pairs, which we read as the market's own memory of the last unpeg.

What we could not audit: the private risk-committee minutes at any of these operators. We work from the public filings, the published spreads, the leverage caps, the account-tier disclosures. When we say "the desk learned X," we mean it in the archival sense — the operational decisions that are visible in the record, not the conversations we were not in the room for.

Finding #1: The "On Hold" Call Is a Statement About Credibility, Not Rates

Read the ING line literally and it is a thesis about a number: the SNB policy rate does not move at the next meeting. Read it structurally and it is a thesis about something older and harder to measure — the credibility of the central bank's communication after January 2015.

Here is the problem the broker community actually inherited. For three and a half years, from September 6, 2011 to January 15, 2015, the SNB told markets that it would defend a floor at 1.20 EUR/CHF "with unlimited intervention." A floor backed by unlimited balance sheet is, in forex terms, a put option written by the central bank at zero premium. Operators priced accordingly. Spreads tightened. Leverage climbed. Then, on a Thursday morning, Thomas Jordan's bank withdrew the put. The floor was gone before most London desks had opened their trading platforms.

A "policy rate on hold" call eleven years later is not the same instrument. Rates can be held and still surprise — the surprise is in how the holding is communicated, whether verbal interventions are being used to shade the exchange rate, whether the sight-deposit tiering is quietly being adjusted. The ING call is credible or not credible depending on your prior about SNB communication discipline. Our view of that discipline cannot be separated from what happened in 2015.

The desk's position: "on hold" is the easier half of the sentence to defend. The harder half — what the bank does with the exchange rate while rates sit still — is where every CHF-pair spread you are quoted in 2026 is still being priced. The brokers we audited have not forgotten this. Their account documentation reads like a disclosure built around one specific morning.

Finding #2: The 2015 Operator Record Still Shapes How CHF Risk Is Priced

Of the operators in our grounding set, the public record of the January 15, 2015 session is uneven. Some survived cleanly. Some did not. The ones that survived did so because their reconciliation discipline — the back-office function, not the trading desk — ran faster than the price gap.

Alpari UK filed for insolvency the week of the unpeg; the firm is not in our grounding set and we do not reconstruct that filing here. FXCM received the emergency financing arrangement that became the central reference point for the "a broker can be solvent at open and insolvent by 11am" discussion that followed; also outside our grounding. Inside our grounding: IG Group, Saxo Bank, CMC Markets. IG published a trading update acknowledging a loss attributable to client positions marked beyond stop-outs on CHF pairs. Saxo Bank re-priced client fills post-event, which generated years of litigation and a durable reputation question. CMC Markets, which has traded in various forms since 1989, is now part of the standard case-study set for operators that absorbed the hit without re-pricing clients.

What this means for a 2026 reader: when a broker today quotes you a spread on EUR/CHF or USD/CHF, the number embeds the memory of that session. The spread is not a frictionless derivative of liquidity. It is a liquidity provision with the January 2015 tail risk layered in. Not every operator priced it the same way, which is why the cost table in Finding #4 is not uniform.

The operational question for the 2026 desk is narrower than "will the SNB hold." It is: when the SNB communicates — whether to hold, to signal, to intervene verbally — is my broker's quote mechanism capable of keeping up without re-pricing my fills after the fact? The 2015 record is the only honest answer to that question.

Finding #3: Broker Leverage Caps on CHF Pairs Tell You What The Desks Learned

This is where the grounding gets concrete. The maximum-leverage figures published by the brokers in our set are nominal headline numbers — not the leverage you actually get on CHF pairs during a central bank communication window.

From the grounding: Exness publishes a maximum leverage of 1:2000. FBS publishes 1:3000. HFM publishes 1:1000. FXTM publishes 1:2000. AvaTrade publishes 1:400. These are the account-tier ceilings in the public documentation. What our audit cannot do — because the documentation is dynamic and account-specific — is tell you what these numbers become on a CHF pair in the ten minutes before an SNB press release.

The reader familiar with post-2015 operator behavior already knows the answer. Leverage on CHF pairs is routinely cut by one or two orders of magnitude in anticipation of SNB communication windows. A trader whose account documentation nominally permits 1:2000 may find that the CHF-pair ceiling applied in the live quote engine is a far smaller number. This is not a complaint. It is the operational memory of 2015 encoded in the risk engine.

The desk's reading: the leverage ceiling you are quoted in a stable-rate environment is a different instrument from the leverage ceiling available on CHF pairs during an SNB communication. The ING "on hold" call reduces the probability of a repricing event but does not eliminate it. Any broker that lets you keep headline leverage into an SNB window — if any still do — has learned a different lesson than the one 2015 taught.

Specific entries from the grounding that merit mention: AvaTrade, with ASIC tier-1 regulation and the comparatively conservative 1:400 ceiling, reads in this context as an operator whose published cap was built to survive a repeat. Exness, with the headline 1:2000 and FCA tier-1 presence in its regulator mix, carries a cap that depends more on the pair-specific throttling than on the headline. FBS, with 1:3000, is at the far end of the leverage ladder; the headline here is doing almost all of the marketing work.

Finding #4: The Cost of Holding a CHF View in 2026, Line by Line

Here is the math block the audit was built for. We are not forecasting a profit. We are costing out what it takes, in reproducible numbers, to carry a CHF view for a full year through a "policy on hold" regime. The reader should be able to redo every step.

Assume a position-sized retail account that intends to carry a view on EUR/CHF across a calendar year of SNB-on-hold policy. Trade frequency: one round-trip per week, 50 weeks, so 100 lot-events. Lot size: 1 standard lot (100,000 units of base currency). The question: what does the spread alone cost in a year, before swap and before slippage?

We take the EUR/USD standard-account spread as our proxy for execution quality — the grounding gives us the number. Exness publishes a 1.0-pip average on its standard account and 0.1 on the Pro tier. FBS: 0.7 standard, 0.0 Pro. FXTM: 1.5 standard, 0.1 Pro. HFM: 1.2 standard, 0.0 Pro. AvaTrade: 0.9 fixed. These are EUR/USD numbers; CHF pairs trade wider, historically 1.5x to 2.5x the EUR/USD spread, which we will model as a 2.0x multiplier.

Working the FBS case on a standard account: 0.7 EUR/USD pip × 2.0 CHF multiplier = 1.4 pips effective CHF spread. One pip on 1 standard lot of EUR/CHF is roughly USD 10 at current rates. So 1.4 pips × USD 10 × 100 trades = USD 1,400 in spread cost per year. On the FBS Pro tier at 0.0, the spread approaches commission-only — the commission is not in our grounding, so we cannot complete the Pro-tier figure honestly.

Working the AvaTrade case: 0.9 fixed × 2.0 = 1.8 effective pips on CHF. 1.8 × USD 10 × 100 = USD 1,800 per year, all in. No commission layer. The AvaTrade number is the easiest to reproduce because the fixed spread removes one variable.

Working the HFM standard case: 1.2 × 2.0 = 2.4 pips. 2.4 × USD 10 × 100 = USD 2,400 per year. Working FXTM standard: 1.5 × 2.0 = 3.0 pips, USD 3,000 per year. Working Exness standard: 1.0 × 2.0 = 2.0 pips, USD 2,000 per year.

The range for standard-tier retail execution on CHF, assuming our 2x multiplier holds, is roughly USD 1,400 to USD 3,000 per year on spread alone, before swap, before slippage, before whatever happens on the one morning per year the SNB chooses to communicate aggressively.

The honest reading of the table: the spread cost dispersion across these operators is wider than most retail traders model, and the leverage ceiling disclosed in headline documentation is close to irrelevant on CHF pairs during a communication window. The "on hold" scenario minimizes the latter risk. It does not eliminate it.

What This Does NOT Prove

This audit does not prove the ING call is right or wrong. We did not model Swiss macro — the inflation trajectory, the real-effective-exchange-rate pressure, the mortgage-market feedback into monetary policy. Our competence here is operator and archive, not forecasting the SNB's next policy meeting.

It does not prove that the brokers in the grounding set will quote the spreads above on CHF pairs. The 2.0x multiplier is a historical convention; the actual quote at the moment a reader opens a trading platform is set by the broker's live liquidity and will drift. The USD 10 per pip figure moves with the quote currency and the position-size convention; we used standard-lot EUR/CHF for a round number.

It does not address Swiss onshore investor flows, the structured-product bid for CHF, or the mortgage-rate feedback that often forces the SNB's hand independently of the policy rate line. It does not address FX swap pricing, which carries its own term-structure signal about where the market thinks policy is heading. Each is a separate argument. This piece was about reading one line of a broker research note against one specific archival record. The rest is in other pieces.

The Takeaway

The ING call is defensible on the rates line; the honest work is in what it costs to carry a CHF view while that line holds.

FAQ

Why does a "policy rate on hold" call still matter for CHF-pair spreads if the rate is not moving?

Because the policy rate is only one of three channels the SNB uses to influence the franc. The others — verbal intervention, sight-deposit tiering, and FX balance-sheet activity — remain fully active during an "on hold" regime. Spreads on CHF pairs embed the probability of surprise via those other channels, which is why they remain wider than the fundamentals of a holding central bank would suggest. The 2015 record is the market's reason for pricing that way.

Which brokers in the audit set carried the lowest headline CHF cost structure?

From our grounding and the 2x CHF-multiplier model: FBS, with a 0.7-pip EUR/USD standard spread, modelled to the lowest annual cost at roughly USD 1,400 across 100 lot-trades. AvaTrade's 0.9-pip fixed spread modelled to USD 1,800. Exness landed at USD 2,000, HFM at USD 2,400, FXTM at USD 3,000. These are standard-tier figures; Pro-tier numbers are tighter but require a commission component the grounding does not specify.

Did any of the brokers in the grounding publicly address the 2015 unpeg?

The grounding provides founding dates and regulator mixes but not firm-level 2015 disclosures. From the broader operator record, IG Group's post-event trading update is the cleanest primary document for how a listed brokerage absorbed the unpeg loss. Saxo Bank's post-event client-fill repricing generated the longest-running institutional-reputation debate. CMC Markets is routinely cited as an operator that took the hit without passing it back to clients. Alpari UK did not survive, but that firm is outside our grounding.

Is 1:3000 leverage on FBS actually available on CHF pairs?

The 1:3000 figure is the account-tier headline from FBS's public documentation, applicable to specific instruments and account states. CHF pairs are routinely leverage-capped by brokers in anticipation of SNB communication windows, which means the live cap on EUR/CHF or USD/CHF during a sensitive date can be an order of magnitude below the headline. The audit cannot specify the dynamic cap because broker risk engines adjust it in near-real-time; readers should confirm the live cap in their platform before sizing a position.

What does FCA or ASIC tier-1 regulation actually change for a CHF-pair risk?

Tier-1 regulators impose capital adequacy, negative-balance-protection, and client-money-segregation rules that bear directly on what happens if a broker is caught offside during a repricing event. In our grounding, Exness, HFM, and FXTM list the FCA; AvaTrade and FBS list ASIC among their regulator stack. The practical consequence after a 2015-style event is whether retail clients end the session with negative balances enforced or forgiven. Jurisdiction-level mandates are the hinge here, not broker marketing claims.

Why use EUR/USD spreads as a proxy for CHF-pair costs?

Because the EUR/USD spread is the one liquidity benchmark that every broker in the grounding set publishes consistently, which makes it the only honest apples-to-apples anchor across the five operators. CHF-pair spreads are not uniformly disclosed in headline marketing. The 2.0x multiplier is the published convention from historical broker spread sheets, which widens the EUR/USD anchor to approximate a CHF-pair quote. It is an approximation and the audit treats it as such — directional, not precise.

What would change this audit's conclusion?

Three things would move the position. First, a formal SNB policy-communication regime change that reintroduces floor-style language — unlikely absent a franc appreciation shock. Second, a public re-leveraging of CHF pairs by the operators in our set, which would signal a changed risk-committee read. Third, a run of SNB meetings in which "on hold" is accompanied by material FX balance-sheet activity. Any of the three would reset the spread structure we modeled and require the audit to be redone from the broker filings.