The Dollar Smile Theory, Why the USD Can Rise in Both Good and Bad Times
The US dollar sometimes strengthens when the global economy is struggling, and also when the US economy is booming. The dollar smile theory tries to explain this pattern.
Imagine a smile-shaped curve:
Left side (global stress or risk-off): investors seek safety and buy US dollars, supporting it.
Bottom of the smile (moderate global growth): when the world economy is doing steadily well and risk appetite is high, capital flows to higher-returning markets outside the US, so the dollar may weaken.
Right side (strong US growth and rising rates): when the US economy outperforms and rates are higher, the dollar can strengthen because of capital inflows.
How to use the idea:
1. Identify which part of the smile the market seems to be in.
2. Check risk sentiment, growth data, and interest rate expectations.
3. Compare with how the dollar index and major pairs behave.
Cautions:
- It's a conceptual model, not a precise tool
- Reality is messier, and multiple forces act at once
- Don't force current conditions into a framework
The main lesson: the dollar can respond to different drivers at different times. Avoid assuming that "good news for the world" always means a weaker dollar, or that "bad news" always means a stronger one.
Use models for perspective, and let price confirm what's happening.
The Dollar Smile Theory, Why the USD Can Rise in Both Good and Bad Times
-
LondonNewsTrader
- Posts: 421
- Joined: Mon Sep 21, 2026 9:30 am
Re: The Dollar Smile Theory, Why the USD Can Rise in Both Good and Bad Times
Two clear real-world examples of both ends of the smile. In March 2020, during the worst of the Covid panic, the dollar index went from roughly 95 to over 102 in about ten days while global markets were collapsing. That is the left side. Through 2022, when the Fed was hiking aggressively and US growth held up, it climbed to around 114 by September. That is the right side.
The middle is harder to spot in real time, which is your caution about forcing a framework. What I watch in practice is how the dollar reacts to US data surprises. If good US numbers lift the dollar and bad global news also lifts it, you are probably on one of the edges. If US data barely matters and the dollar drifts lower on good news elsewhere, you are nearer the bottom.
It's a context tool rather than a trading signal, as you say.
The middle is harder to spot in real time, which is your caution about forcing a framework. What I watch in practice is how the dollar reacts to US data surprises. If good US numbers lift the dollar and bad global news also lifts it, you are probably on one of the edges. If US data barely matters and the dollar drifts lower on good news elsewhere, you are nearer the bottom.
It's a context tool rather than a trading signal, as you say.
Re: The Dollar Smile Theory, Why the USD Can Rise in Both Good and Bad Times
Easiest way I've found to use the smile in practice is to ask which side of the curve we're on this week. Is the dollar rising because the US data is strong, or because the rest of the world is scared? The answer changes which pairs I'd want to be in. Fear-driven dollar strength usually comes with yen strength too, so USDJPY may go nowhere while AUDUSD falls hard. Growth-driven dollar strength often lifts USDJPY because US yields rise. A quick check is the stock market. If US stocks are rising along with the dollar, you're probably on the right side of the smile. If both stocks and yields are falling, it's the left side.
It’s Fairman 
Re: The Dollar Smile Theory, Why the USD Can Rise in Both Good and Bad Times
It's a simplification, to be honest. There are periods where the dollar falls even in risk-off, especially if the stress starts in the US. I'd use it as a way to ask the right question, not as a rule that always holds. In March 2023, during the US regional bank stress, the dollar didn't get the usual haven bid, partly because the problem was in the US and markets expected the Fed to slow hikes. The yen and gold did better. So when a scare hits, the first question is where the problem started, then whether the dollar is the haven or the source.
It’s Fairman 
Re: The Dollar Smile Theory, Why the USD Can Rise in Both Good and Bad Times
Worked example of the middle of the smile. When US growth looks average and the rest of the world is doing fine, money tends to flow out of the dollar looking for better returns elsewhere. That's the phase where EURUSD and AUDUSD tend to grind higher without much drama. The second half of 2017 is a decent example, with eurozone growth surprising on the upside while US data was steady and EURUSD climbing steadily. In that phase, pullback buys on the majors against the dollar tend to work better than breakout shorts. Recognising which part of the smile you're in helps choose the setups.
It’s Fairman 
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: The Dollar Smile Theory, Why the USD Can Rise in Both Good and Bad Times
The two examples above also show that the ends of the smile move at very different speeds. Going from roughly 95 to 102 in ten days is about 7 percent, while the 2022 climb to 114 started near 96 and took most of the year, about 19 percent over eight or nine months.
That difference matters for a scalper. Left side moves arrive as gaps and spikes where stops slip, and right side moves arrive as a grind where pullback entries keep working.
A rough gauge for the regime is the rolling 20 day correlation between the dollar index and US stocks. Strongly negative, with stocks down and the dollar up together, points to the left side, while clearly positive readings fit the right side.
Readings near zero fit the bottom of the smile, but they also show up during a scare born in the US, like March 2023, when neither rule holds. So a reading near zero is a reason to trade smaller, not a signal in itself.
The yen check from the earlier post sharpens it. If USDJPY falls while the dollar index rises, the bid is fear, and the AUDUSD short is the cleaner way to express it. Does anyone here track a correlation like that?
That difference matters for a scalper. Left side moves arrive as gaps and spikes where stops slip, and right side moves arrive as a grind where pullback entries keep working.
A rough gauge for the regime is the rolling 20 day correlation between the dollar index and US stocks. Strongly negative, with stocks down and the dollar up together, points to the left side, while clearly positive readings fit the right side.
Readings near zero fit the bottom of the smile, but they also show up during a scare born in the US, like March 2023, when neither rule holds. So a reading near zero is a reason to trade smaller, not a signal in itself.
The yen check from the earlier post sharpens it. If USDJPY falls while the dollar index rises, the bid is fear, and the AUDUSD short is the cleaner way to express it. Does anyone here track a correlation like that?
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: The Dollar Smile Theory, Why the USD Can Rise in Both Good and Bad Times
One caution when reading the smile off the dollar index is how lopsided the index is. The euro alone carries close to 58 percent of the weight, and the yen, sterling and the Canadian dollar account for most of the rest.
So a euro story alone can move the index a long way. Say EURUSD falls 1 percent on weak eurozone data while the other pairs barely move, then the index rises about 0.6 percent with no change in global risk or US growth.
On the index that looks like dollar strength and could be read as either end of the smile. It is really a euro move, and a regime call made from it would be wrong.
A quick cross check is the dollar against a few currencies the index leaves out, such as the Australian dollar, the Mexican peso or the Chinese yuan. If the dollar is up broadly across those as well, the smile reading is on firmer ground.
An equal weighted basket of five or six majors against the dollar does the same job in a spreadsheet. Does anyone here use a broader dollar measure instead of the index for this kind of call?
So a euro story alone can move the index a long way. Say EURUSD falls 1 percent on weak eurozone data while the other pairs barely move, then the index rises about 0.6 percent with no change in global risk or US growth.
On the index that looks like dollar strength and could be read as either end of the smile. It is really a euro move, and a regime call made from it would be wrong.
A quick cross check is the dollar against a few currencies the index leaves out, such as the Australian dollar, the Mexican peso or the Chinese yuan. If the dollar is up broadly across those as well, the smile reading is on firmer ground.
An equal weighted basket of five or six majors against the dollar does the same job in a spreadsheet. Does anyone here use a broader dollar measure instead of the index for this kind of call?
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: The Dollar Smile Theory, Why the USD Can Rise in Both Good and Bad Times
Three currencies form a triangle of prices. If EUR/USD, GBP/USD and EUR/GBP are quoted, the third rate should equal the ratio of the other two. When it does not, a cycle of conversions, such as dollars into euros, euros into pounds and pounds back into dollars, could in theory return more than the starting amount with no directional risk.
In practice, such gaps are tiny and short-lived. Automated systems inside banks and specialist firms monitor these relationships continuously and trade away any inconsistency within fractions of a second. Each leg also incurs spread, and the bid and offer sides of every quote mean that apparent discrepancies on a screen often disappear once the costs are included.
Latency matters as well. By the time a retail platform displays a mismatch, faster participants have usually acted on it. Quotes shown on a screen may be stale, and execution at the displayed prices is not assured.
The concept is still valuable for understanding market structure. It explains why cross rates track their component pairs so closely, why correlated moves ripple across related pairs in near unison, and why claims of easy retail arbitrage deserve considerable skepticism. Efficiency at the top of the market is largely maintained by speed that individual traders cannot match.
In practice, such gaps are tiny and short-lived. Automated systems inside banks and specialist firms monitor these relationships continuously and trade away any inconsistency within fractions of a second. Each leg also incurs spread, and the bid and offer sides of every quote mean that apparent discrepancies on a screen often disappear once the costs are included.
Latency matters as well. By the time a retail platform displays a mismatch, faster participants have usually acted on it. Quotes shown on a screen may be stale, and execution at the displayed prices is not assured.
The concept is still valuable for understanding market structure. It explains why cross rates track their component pairs so closely, why correlated moves ripple across related pairs in near unison, and why claims of easy retail arbitrage deserve considerable skepticism. Efficiency at the top of the market is largely maintained by speed that individual traders cannot match.