A yield spread stops predicting a currency pair the moment the market stops trading on interest rate differentials and starts trading on something else entirely - risk sentiment, a growth shock, a political event, or a central bank surprise that hasn't been priced in yet. This isn't a fixed point in time you can mark on a calendar. It's a regime change you have to spot using price action and positioning data alongside the spread, not the spread in isolation.
Once that regime change happens, leaning on the yield spread alone will get you offside, sometimes for weeks at a time.
Why do yield spreads work for currency pairs in the first place?
They work because capital tends to flow toward the higher-yielding currency when risk appetite is stable and the carry trade is attractive.
In calm markets, a widening 2-year yield spread in favor of, say, the Australian dollar over the US dollar gives traders an incentive to hold AUD and fund it in USD. That demand shows up as a slow, grinding move in the exchange rate that tracks the spread fairly closely. This relationship tends to hold best in quiet, trending macro environments where nothing else is competing for attention.
What makes a yield spread stop working?
The spread stops working when something bigger than carry takes over price action - usually a risk-off shock, a repricing of future rate expectations, or an unwind of crowded positioning.
Common triggers include:
Risk sentiment flips, where safe-haven flows dominate regardless of where yields sit
Markets pricing in future rate moves faster than actual yields catch up, so the spread lags what traders already expect
Extreme positioning unwinding sharply, overwhelming the slow carry story
Structural shocks like a fiscal crisis, war, or a sharp commodity move that changes the whole risk calculus for a currency
How do I know the spread has stopped predicting the pair in real time?
You'll see the spread and price action pulling in opposite directions for several consecutive weeks, and that divergence usually gets confirmed by positioning data before it shows up in headlines.
- Plot the yield spread against the currency pair on a weekly chart.
- Watch for the spread widening or narrowing while price does the opposite for three or more weeks.
- Cross-check CFTC positioning data for the pair to see if speculative positioning is already stretched.
- Check the economic calendar for upcoming central bank meetings or data that could explain why price is front-running the spread.
- Decide whether this is short-term noise or an actual regime change before adjusting your bias.
The Macro Compass is built for exactly this kind of weekly check, since it lines up yield spreads, central bank signals and futures flow in one place so you're not guessing from a single chart.
Does COT positioning help explain the breakdown?
Yes - extreme speculative positioning is often the real reason a yield spread stops predicting the pair, because the carry trade becomes overcrowded and vulnerable to a sharp unwind regardless of what the spread is doing.
When large speculators are heavily long a currency because of favorable carry, that position eventually has to be reduced or reversed. If that unwind happens while the spread is still widening in the currency's favor, you get a divergence that has nothing to do with interest rates and everything to do with crowded trades resolving. Weekly COT positioning data is the clearest way to see this building before price confirms it.
Which market regimes break the yield spread?
The spread is most reliable in calm, risk-on markets and least reliable during shocks, surprise pivots and crowded-position unwinds.
| Market regime | Yield spread reliability | What drives price instead |
|---|---|---|
| Risk-on, stable growth | High | Carry / rate differential |
| Risk-off shock | Low | Safe-haven flows |
| Central bank surprise pivot | Low until repriced | Rate expectation repricing |
| Extremely crowded positioning | Low | Positioning unwind |
| Commodity-currency shock | Mixed | Commodity price, not rates |
Example
Say AUD/USD and the Australia-US 2-year yield spread have tracked closely for several months. The spread widens in AUD's favor and the pair climbs from 0.6500 to 0.6700, matching expectations.
Then the spread widens another 20 basis points over two weeks, but AUD/USD stalls at 0.6700 and starts drifting down to 0.6600. Checking COT data shows speculative AUD long positioning at its highest level in several months. That combination - spread still favorable, price falling, positioning stretched - is a reasonable sign the spread has stopped predicting the pair for now, and the crowded long trade is starting to unwind instead.
This is hypothetical and meant to show the pattern, not an actual trade recommendation.
What should I actually do when I see this divergence?
Reduce size or step aside rather than doubling down on the yield-based thesis, and let price structure confirm direction before re-entering.
Keep your risk at your normal cap, typically around 0.5% per trade, rather than widening it to win back what the spread already cost you. A divergence is a reason to trade smaller, not bigger, until price and the spread agree again.
If you want to check the spread itself, the free Yield Spread Radar shows 2-year spreads for 15 pairs, how far price has drifted from them, and an honest history check, including the periods where the spread said nothing at all.
Trading forex and gold carries a high risk of loss, and no yield spread or positioning report removes that risk.
Frequently asked questions
How long can a yield spread and a currency pair disagree?
There is no fixed length; a divergence can last a few days around a data shock or several weeks when crowded positioning is unwinding, which is why a weekly check works better than a daily one.
Should I trade against the spread when it stops working?
Not automatically; a broken relationship tells you the spread is not the driver right now, not which way price goes next, so wait for price structure to give you direction.
Which yield should I use for the spread?
Most traders use the 2-year government yield because it tracks near-term central bank expectations, as explained in why the 2-year yield matters more than the 10-year.

