Most days, each currency pair trades its own story — a rate decision here, an inflation print there. Then a bank fails, a war starts, or a pandemic headline lands, and suddenly every pair on the board is answering one question: do we want risk today, or not? That single-question state is called risk-on, risk-off, and while it is active it overrides almost everything else you have learned about what moves currencies. This sheet explains the regime: which currencies act as shelters, which act as barometers, and why your “diversified” positions quietly merge into one position exactly when it hurts most.
The regime switch: when markets trade fear, not fundamentals
In ordinary conditions, currency prices are a weighing machine for many small stories at once: interest-rate expectations, data surprises, positioning. Risk-on, risk-off is what happens when one story grows large enough to silence the others. In a risk-on phase, investors reach outward — toward growth, toward higher yields, toward assets that pay them for optimism. In a risk-off phase the reach reverses: capital runs home, debts are paid down, and safety is bought at almost any price.
The trigger does not need to involve currencies at all. An equity crash in one market, a credit event at a single bank, a geopolitical shock — any of them can flip the regime, because the response runs through the same global plumbing: funds cut positions everywhere at once, and every cut position has a currency leg. You will see the switch as a spike in volatility across markets that normally ignore each other.
The practical signature is uniformity. On a normal day, EUR/USD and AUD/JPY can disagree. On a regime day they do not: the entire board sorts itself into two benches — currencies bought for shelter and currencies sold as risk — and the interesting question stops being “what is this pair worth?” and becomes “how afraid is everyone?”
The haven bench: JPY, CHF, USD — and why each earns the label
“Safe haven” sounds like a slogan, so it is worth deriving each label from mechanics you can check. The Japanese yen strengthens in stress largely because of borrowing, not optimism about Japan. Decades of very low Japanese rates made the yen the world's favorite funding currency: investors borrow yen cheaply to buy higher-yielding assets elsewhere. When fear hits, those trades are closed — the foreign assets are sold and the borrowed yen is bought back to repay the loan. A wave of forced yen buying looks exactly like a flight to safety, and prices accordingly.
The Swiss franc earns its seat differently: a persistent current-account surplus, political stability, and a long institutional memory of capital arriving in Switzerland whenever Europe worries. The bid has at times been so strong that the Swiss National Bank spent years actively resisting it — a detail worth remembering, because it means the franc's haven behavior comes with a central bank standing on the other side.
The US dollar is the most conditional of the three. It is the world's reserve currency, most global debt is priced in it, and the US Treasury market is the deepest parking lot on Earth — so in a genuine global funding panic, the world scrambles for dollars and the USD rises against almost everything, including against the major pairs. But when the bad news is specifically American, the dollar's haven bid can fade or invert. All three labels are tendencies with mechanisms behind them, not laws: each has failed in specific episodes, which is precisely why this sheet keeps saying check, not believe.
The risk bench: commodity and EM currencies as sentiment barometers
On the other bench sit the currencies that markets buy when they feel brave. The Australian and New Zealand dollars are the classic cases: commodity-exporting economies tied to global — especially Chinese — growth, with interest rates that have historically sat above the funding currencies'. That combination makes them the natural long side of yield-seeking flows, and therefore the natural casualty when those flows reverse. The Canadian dollar and Norwegian krone carry a related exposure through oil.
| Currency | Role | Typical risk-off move |
|---|---|---|
| JPY | Funding currency; borrowed yen gets bought back | Strengthens |
| CHF | Surplus economy, stability premium | Strengthens |
| USD | Reserve currency, global dollar debt | Strengthens broadly |
| AUD, NZD | Growth and commodity exposure, higher yields | Weakens |
| CAD, NOK | Oil-linked exporters | Weakens with energy |
| MXN, ZAR, TRY | High yield, thinner liquidity | Weakens hardest, fastest |
Emerging-market currencies sit at the far end of the bench: higher yields in calm times, thin order books in stressed ones. They are often the first sold and the hardest hit, because the same yield that attracted capital in the calm is the reason the exit is crowded in the storm. A pair like AUD/JPY — risk bench over haven bench — is widely watched as the purest single-number thermometer of the regime.
Reading the regime: many screens, one move
You do not need anyone to announce a regime day; the screens announce it. The risk-off signature reads like this: equity indices fall, government bond prices rise as yields drop, gold catches a bid, and the FX board sorts itself — JPY and CHF up, AUD, NZD, and EM currencies down, the dollar up against most of the rest. Risk-on is the same picture with every arrow reversed. The tell is not any single move but the synchrony: markets that usually disagree all voting the same way at the same minute.
This is also why the regime matters even if you never trade a sentiment headline. A technical setup on AUD/USD, a rate-differential view on EUR/USD — both assume the pair is trading its own story. On a regime day that assumption is suspended, and a chart pattern has no defense against the entire board moving as one. Knowing which regime you are in is not a prediction tool; it is context that tells you which of your other tools currently apply.
Tile the haven pairs against the risk pairs on the next big headline day — the regime is something you recognize, not memorize.
The portfolio consequence: correlations converge when you need them not to
Here is where the regime stops being interesting and starts being expensive. Correlation between pairs is not constant: in calm markets, AUD/USD and NZD/USD move together loosely, and EUR/USD and USD/JPY barely acknowledge each other. In a risk-off episode, those relationships tighten toward lockstep, because every pair is repricing the same single factor — fear.
Work the numbers with the survey's standard anchor, where one standard lot moves about $10 per pip. Suppose you are long one lot of AUD/USD and long one lot of NZD/USD — two trades, two analyses, two stop-losses, each sized to risk $600. A risk-off afternoon takes both pairs down 60 pips together: −$600 and −$600, a $1,200 loss. A risk-on relief day would do the symmetric opposite, +$1,200. Both outcomes tell the same story: you did not have two $10-per-pip positions, you had one $20-per-pip position on global sentiment, and you sized it twice.
That is the trap in the word “diversified.” Diversification assumes the things you hold respond to different forces, and the regime switch is precisely the moment that assumption fails — correlations converge exactly when you need them not to. The defensive habit is to group positions by their regime bench, not by their pair names, and to size the group, not each trade alone. The arithmetic for doing that properly has its own sheet in the risk-management region, on correlation and hidden exposure.
Correlation, volatility, major pairs — the terms this sheet leans on, defined once and linked everywhere.
Where this leads next: the correlation-and-hidden-exposure sheet turns today's observation into sizing arithmetic, and the market-sessions sheet explains why regime moves often land at particular hours of the global day. Read the regime first, though. It is the weather over the whole territory — and no amount of skill at reading individual maps survives ignoring the storm.