Safe Haven Currencies: What Makes One and When It Fails

Three currencies are named as safe havens in almost every explanation of the term: the dollar, the yen and the franc. The list is stable, the reasons given for it are not, and the label is usually presented as something a currency is rather than something a currency has done.

That distinction decides whether the idea is usable. A property holds until it changes; a behaviour observed in past episodes holds only while the conditions that produced it are still in place.

What the Safe Haven Label Actually Claims

No authority publishes a list of safe haven currencies. There is no test a currency passes, no committee that admits it, and no threshold above which it qualifies. The label is applied by market participants and by commentary, after the fact, to currencies that strengthened while riskier assets were being sold.

What it claims, stated plainly, is a correlation: when a particular kind of shock hits, this currency has tended to appreciate. That is a statement about a sample of past events, and it inherits every limitation of one.

Read that way, the useful question changes. Not whether a currency is a safe haven, but which mechanism produced the appreciation in the episodes the label was drawn from, and whether that mechanism is still operating today.

The label also sits inside a wider framework of risk-on and risk-off behaviour, which our page on intermarket analysis covers as a whole. This page deals with one part of it: the currencies themselves, and what has to hold for each to behave the way the label says.

The Three Currencies That Carry the Label

The dollar, the Japanese yen and the Swiss franc are the three that appear on essentially every list. A wider list sometimes adds the euro or sterling, and gold is usually mentioned alongside them even though it is not a currency in this sense.

All three are freely floating, deeply traded and issued by economies with functioning legal systems and no capital controls. Those are the shared conditions, and they are necessary rather than sufficient: several currencies in the G10 group meet all of them and carry no haven reputation at all.

The three are often treated as interchangeable versions of one idea. They are not. Each reaches the label through a mechanism the other two do not share, which is why they can move apart in the same week and why a portfolio holding all three is less diversified than it looks.

The mirror image of the group is the set of currencies that fall in the same conditions, which are usually the higher yielding and the commodity linked ones.

What the Official Numbers Measure, and What They Do Not

The figure quoted most often in support of the dollar comes from the survey the Bank for International Settlements runs every three years. The comparable pages still cite the 2022 round, which is two rounds old as a benchmark and one round old as a fact.

The April 2025 survey, released at the end of September that year, measured average daily turnover in over the counter foreign exchange markets at 9.6 trillion dollars, a 28 percent increase on 2022. The dollar appeared in 89.2 percent of all trades, against 88.4 percent three years earlier. The yen was little changed at 16.8 percent and the Swiss franc rose to 6.4 percent.

Two things about those percentages are worth stating, because both are routinely lost. They sum to 200 rather than 100, since every trade has two currencies in it. And they measure turnover, which is a liquidity fact.

Liquidity explains why a large flow can move into a currency without dislocating its price, so it is a precondition for haven behaviour. It is not evidence of haven behaviour, and no BIS number establishes that anyone bought a currency for safety.

Why the Yen Rises in a Shock

Japan has run low policy rates for a long time relative to the countries it trades against, and that rate gap has made the yen the standard funding currency for carry positions. A trader borrows in the low yielding currency, holds a higher yielding one, and earns the difference while the exchange rate stays still.

Those positions are short the yen by construction. When volatility rises, the carry trade stops paying for its own risk, and closing it means buying yen back. The buying is mechanical, it happens at speed because everyone unwinds into the same shock, and it produces exactly the appreciation the safe haven label records.

The mechanism has an implication most descriptions leave out. The size of the move depends on how much carry was built up beforehand, not on how safe Japan looks at the time. A shock arriving after a long quiet period with heavy positioning moves the yen further than an identical shock arriving after positions have already been cut.

It also means the direction can reverse without the label changing. Where a shock widens the rate gap in the dollar favour rather than compressing it, the funding incentive strengthens and the yen can weaken through the same episode.

The Franc and the Episode That Gets Misread

The franc case rests on a small open economy with a long record of monetary discipline, no capital controls, and a domestic asset market too small to absorb large inflows without the currency moving. That last point is the operative one: modest flows produce large price effects.

The episode cited most often as proof is January 2015, and it is the weakest available evidence for the haven claim. On 15 January 2015 the Swiss National Bank discontinued the minimum exchange rate it had been defending against the euro and cut its policy rate to minus 0.75 percent, in a single announcement.

What followed was a repricing after a policy floor was withdrawn, not a flight to quality. The distinction matters because the two have different triggers: one is announced by a central bank, the other emerges from investor behaviour, and only the second is what the label is meant to describe.

The years of intervention that preceded the announcement are the more useful part of the record. A currency whose appreciation a central bank has actively resisted is one whose future strength depends partly on policy tolerance, and that is a condition worth knowing about before treating it as a refuge.

The Dollar Is a Haven for a Different Reason

The dollar case is not about stability at all. It is about function: the dollar is the currency most global debt is denominated in, most commodities are invoiced in, and most cross border funding is raised in.

In a stress episode, institutions that owe dollars need dollars, regardless of what they think of the United States. That demand is a settlement requirement rather than an opinion, which is why the dollar has strengthened in shocks that originated inside the American economy itself.

The practical consequence is that dollar strength in a crisis says little about relative economic health, and reading it as a verdict on the American economy inverts the mechanism. Where dollar strength is broad rather than against one currency, the dollar index is the cleaner way to see it.

When the Pattern Fails

The most common failure is a shock whose source is the haven itself. A crisis centred on Japanese policy, Swiss banking or American debt does not send money toward that currency, and the label offers no protection in the one case where the risk is domestic.

The second is a rate driven episode. Where the shock changes interest rate expectations more than it changes risk appetite, differentials dominate and the currency that should have appreciated on the haven story weakens on the rate story instead.

The third is timing. Haven appreciation happens in the first hours or days of a shock and often reverses well before the underlying situation resolves, so a position taken after the headline is a position taken after the move.

The fourth is central bank tolerance, which the franc record illustrates directly. A currency can be a haven and still be capped, and the cap is a policy decision that can be introduced or removed without warning.

Whether any of these is currently in play is checkable rather than assumed, and a broad measure of how much stress is actually being priced is covered in what the VIX measures.

Which of These Applies to You

If you hold positions in higher yielding currencies funded in yen, the yen mechanism is the one that concerns you, and the exposure you actually carry is to a crowded unwind rather than to Japan.

If you are hedging a book against a European or regional political shock, the franc mechanism is the relevant one, together with the policy question of how much appreciation the Swiss National Bank is prepared to tolerate.

If your concern is a global funding squeeze, the dollar is the mechanism that matters, and the thing to watch is dollar borrowing demand rather than any national comparison.

And if none of those describes your position, the label is not a trade at all. It is a piece of context that explains why some pairs move together in a shock, which is worth knowing and is not a reason to hold anything.