Asset classes · Forex
Currency carry trades: how they pay and why they crash
How a currency carry trade earns the rate gap, a worked P&L with leverage, and what the 2024 yen unwind and the Mexican peso and Brazilian real show about its risk.
On this page
A currency carry trade borrows money in a currency with a low interest rate and invests it in a currency with a higher one. The investor keeps the gap between the two rates, the "carry", for as long as the position is open. If exchange rates stood still, it would be one of the simplest trades in finance: a steady income from a difference in central-bank policy.
Exchange rates do not stand still. The carry is small and arrives slowly; a currency move against the position can be large and arrive in days. That asymmetry gives carry trades their reputation: gains go up the stairs, losses come down the elevator. Central banks around the world: Fed, ECB, BOJ, BoE and more shows a short yen example; this lesson goes further, with financing, leverage, the July–August 2024 yen unwind on daily Federal Reserve data, and the emerging-market version of the trade.
How a carry trade pays
Every currency comes with a short-term interest rate, set largely by its central bank. Holding a currency means earning (or paying) that rate. A carry trade has two legs:
- The funding currency is the one you borrow or sell. It has the low rate: the yen and the Swiss franc for most of the past three decades, the euro from 2014 to 2022, and the dollar itself in 2009–2015 and 2020–2021.
- The target currency is the one you buy and hold. It has the high rate: the dollar after 2022, the Australian and New Zealand dollars before 2008, and emerging-market currencies such as the Mexican peso and the Brazilian real.
Retail traders rarely borrow yen at a bank. They hold a long position in a pair such as USD/JPY or AUD/JPY with a forex dealer, and the dealer pays or charges a daily financing adjustment (often called the swap or rollover) that reflects the rate gap, minus its own spread. Institutions do the same thing with forward contracts: buying a high-rate currency forward is cheaper than buying it today by roughly the rate gap, so a rolled forward position earns the carry (see FX forwards and interest rate parity: how forward rates are set). Currency futures on CME embed it the same way. The mechanics differ; the economics do not.
The worked profit and loss
The breakeven is the key number. The investor is betting that the funding currency will not rise by more than the rate gap. Economic theory says that, on average, it should rise by exactly that much (uncovered interest parity, explained in FX forwards and interest rate parity: how forward rates are set). The trade exists because, over long stretches, it has not.
Leverage turns income into risk
Carry traders rarely put up the whole amount. Suppose the investor above holds only $10,000 of their own money against the $100,000 position, a leverage ratio of 10 to 1 (within the US retail limit for major pairs; see Currency pairs and pips: quotes, lots, pip value and leverage).
| USD/JPY after a year | Yen move | Profit or loss | On $10,000 of equity |
|---|---|---|---|
| 147 | −4.8% | +$9,286 | +92.9% |
| 140 | 0% | +$4,500 | +45.0% |
| 133 | +5.3% | −$789 | −7.9% |
| 126 | +11.1% | −$6,667 | −66.7% |
A 4.5% income on the position becomes a 45% return on equity, and an 11% yen rally takes two-thirds of the equity. In practice a dealer's margin call would close the position before the year ends, at whatever the rate is that day, which is why carry losses tend to be realized at the worst moment.
Why carry trades crash
The pattern repeats across decades: a long climb in the target currency while the rate gap holds, then a fast reversal. Three forces make it so.
- Crowding. When many investors fund in the same currency, they all need to buy it back to close. A first move against them forces margin calls, the margin calls force buying, and the buying extends the move.
- A change in the rate gap. A carry trade is priced on the expected gap. A surprise rate rise in the funding currency, or a cut in the target one, cuts the income and invites a rush for the exit at the same time.
- Risk-off episodes. In a sell-off, investors cut risk everywhere. High-rate currencies tend to fall and low-rate funding currencies tend to rise, so carry losses coincide with stock losses (see Safe-haven currencies: the yen, Swiss franc and dollar in a crisis and risk-on, risk-off).
Academic work finds the same shape. Brunnermeier, Nagel and Pedersen (2008) showed that carry-trade returns are negatively skewed: many small gains, occasional large losses, and the losses cluster when market volatility jumps.
The yen unwind of July–August 2024
The yen was the cheapest major currency to borrow for most of 2016–2024. The Bank of Japan held its policy rate at −0.1% until March 2024, then set it at 0–0.1%, while the effective fed funds rate was 5.33% from August 2023 through August 2024. A gap that wide pulled money out of yen and into dollars, pesos and other high-rate assets.
On the Federal Reserve's daily H.10 noon rates the dollar rose from ¥103.19 on January 4, 2021 to ¥161.73 on July 10, 2024: 56.7% over three and a half years, much of it the slow climb a carry trade enjoys.
Then it reversed:
- July 11, 2024: ¥158.58, after Japan's Ministry of Finance bought yen (see Currency intervention: how central banks move exchange rates).
- July 31, 2024: ¥150.38. The Bank of Japan raised its rate to 0.25% and set out a plan to slow its bond buying. A softer US jobs report followed on August 2, and markets began to price faster Fed cuts. The gap was now expected to narrow from both ends.
- August 5, 2024: ¥143.95. The yen had gained 12.4% against the dollar since July 10, and 4.5% in the three trading days after the BOJ decision. Stock markets in Japan and elsewhere fell sharply that day; the BIS's September 2024 review describes how the unwinding of yen-funded positions amplified the turmoil.
- September 16, 2024: ¥140.79, a 14.9% yen gain from the July peak.
Set that against the income. At a gap of about 5.2 points, two months of carry earned about 0.9% of the position. An unhedged dollar position funded in yen lost roughly 13% of its value in the same two months. The rate gap had not closed: the fed funds rate was still 5.33% in August. The move came from positioning and expectations, not from the carry itself disappearing.
The unwind did not end the trade for good. By November 18, 2026 the Finabase snapshot of USD/JPY stood at ¥159.37, near the 2024 peak, as the dollar again paid far more than the yen (the BOJ's rate had reached 1.25% by October 2026; see Central banks around the world: Fed, ECB, BOJ, BoE and more).
Emerging-market carry: the peso and the real
Emerging-market currencies offer bigger gaps. Banco de México held its target rate at 11.25% from March 2023 to March 2024, and Brazil's central bank held the Selic at 13.75% from August 2022 to August 2023, against a fed funds rate that peaked at 5.33%. The income is several times larger than in the yen trade, and so are the moves.
From the Fed's monthly averages:
- 2020: the shock. The peso went from 18.81 per dollar in January 2020 to 24.18 in April 2020 as the pandemic hit, a fall of more than a fifth in the peso's dollar value in three months.
- 2023–2024: the "super peso". With Mexican rates above 11%, the peso strengthened to 16.77 per dollar in March 2024, its strongest monthly average since November 2015. Carry investors earned both the gap and a rising currency.
- Mid-2024: the reversal. The peso fell to 20.26 per dollar by December 2024, a 17.1% loss of dollar value from April. Mexico's June 2024 election result and the August yen unwind both hit it; a year of carry at a gap of about six points did not cover that.
- The real. Brazil's currency went from 4.15 per dollar in January 2020 to 6.10 in December 2024, a 32% fall in its dollar value, even though Brazilian rates were far higher than US rates for most of that time. By October 2026 it was at 5.13, still 19% weaker than in January 2020.
Over the whole period the peso ended close to where it began: October 2026's average of 18.45 left it 2% stronger than in January 2020. A carry investor who held through every swing collected about six years of a wide rate gap. One who was forced out in April 2020 or August 2024 by a margin call did not.
Pitfalls
- Income is not return. The rate gap is visible and feels safe. The currency move, which is not, can be many times larger.
- The crash risk is in the tails. Monthly volatility understates carry risk because losses come in rare, sharp episodes. A short history without one flatters the trade.
- Correlation with stocks. Carry losses tend to arrive in sell-offs, the moment other holdings also fall, so the trade adds less diversification than its average return suggests.
- Financing terms change. A dealer's rollover rate is the rate gap minus a markup, and dealers can widen it. On a small gap, the markup can leave nothing.
- Liquidity in emerging markets. Spreads widen and gaps open in stressed markets; capital controls can trap money, as in several past crises (see Currency pegs and exchange-rate regimes: floats, bands and boards).
Carry compared with related trades
| Trade | What earns the income | What causes the loss |
|---|---|---|
| Currency carry | Short-term rate gap | Funding currency rises |
| Bond carry (borrow short) | Long yield above short yield | Long yields rise |
| Futures roll in backwardation | Futures below spot | Curve flips to contango |
| Selling options | Premium collected | A large move in the underlying |
All four share the shape: steady small gains in calm markets and occasional large losses when conditions change. For futures roll yield, see Contango, backwardation and roll yield: why USO lags oil.
How Finabase does it
The forex page lists 51 pairs, including the standard carry crosses: yen crosses such as AUD/JPY, NZD/JPY and ZAR/JPY, and dollar pairs against high-rate currencies such as USD/MXN. Each has a symbol page with a chart, where you can see the long climbs and short drops this lesson describes, and returns from one week to two years.
Finabase shows exchange rates, not interest rates by currency: it does not display rollover or swap rates, forward points or a carry calculator, and it does not execute forex trades. The Treasury page gives US yields for the dollar side. The figures here are ours, built from Federal Reserve data.
Frequently asked questions
What is a carry trade in simple terms?
Borrowing in a currency with low interest rates and holding one with higher rates to earn the difference. It profits as long as the borrowed currency does not rise by more than the interest gap.
Why is the Japanese yen used for carry trades?
Because Japanese interest rates were the lowest among major economies for decades, including a negative policy rate from 2016 to March 2024. That made the yen cheap to borrow and plentiful to sell.
What caused the August 2024 carry trade unwind?
A narrowing of the expected US–Japan rate gap from both sides: the Bank of Japan raised rates on July 31, 2024, and weak US data raised expectations of Fed cuts. Investors closed yen-funded positions at once, and the yen gained 12.4% against the dollar from July 10 to August 5.
Do carry trades make money over the long run?
Historically they have earned positive average returns across many currencies, but with large, sudden losses. Whether the average compensates for that crash risk depends on the period, the currencies and the leverage.
Can I do a carry trade with ETFs?
Some funds have offered currency exposure, but most retail carry trades are leveraged forex positions that receive or pay daily financing. Fund fees and the absence of leverage change the economics, and none of this is a recommendation.
Summary
A currency carry trade earns the interest-rate gap between a low-rate funding currency and a high-rate target currency, and loses whatever the funding currency gains. In the worked example, a 4.5-point gap paid $4,500 a year on $100,000, and a 4.5% yen rise erased it; with 10-to-1 leverage the same moves became swings of tens of percent of equity. The yen trade shows the pattern on real data: a 56.7% climb in USD/JPY over three and a half years, then a 12.4% yen rally in under four weeks in mid-2024 that dwarfed two months of carry. Emerging-market carry pays more and swings harder, as the peso's 17% fall in late 2024 and the real's 32% slide from 2020 to 2024 show.
Check your understanding
4 questions on this lesson. Each answer comes with the reason, and nothing is scored.
An investor borrows yen at 0.5% to hold dollar bills paying 5.0% for a year. Ignoring costs, roughly how far can the yen rise against the dollar before the year’s carry is wiped out?
- About 0.5%
- About 4.5%, the size of the rate gap
- About 10%
- Any rise causes a loss
Show the answer
B. About 4.5%, the size of the rate gap The trade earns the 4.5-point gap. A yen rise of about the same size makes the yen loan that much dearer to repay in dollars, which cancels the income. Beyond that, the trade loses money.
Why do carry-trade losses tend to come in sudden, large moves rather than gradually?
- Because interest is paid only once a year
- Because many investors hold the same position, and a move against it forces them to close together, which extends the move
- Because central banks forbid gradual moves
Show the answer
B. Because many investors hold the same position, and a move against it forces them to close together, which extends the move Crowded positions and margin calls make the exit a stampede. Closing a yen-funded trade means buying yen, and when many traders do it at once the yen jumps, triggering more closing.
In August 2024 the fed funds rate was still 5.33%, yet yen-funded carry trades lost heavily. What changed?
- The rate gap had already closed
- The expected future gap narrowed (a BOJ rate rise and expected Fed cuts) and crowded positions were unwound
- Japan banned foreign borrowing in yen
Show the answer
B. The expected future gap narrowed (a BOJ rate rise and expected Fed cuts) and crowded positions were unwound Carry trades are priced on the expected gap. The BOJ’s July 31 rise and weaker US data cut the expected gap from both ends, and the rush to close positions drove the yen up 12.4% from July 10 to August 5.
A trader runs a $100,000 carry position on $10,000 of equity. The position loses 7% in a currency move. What happens to the equity?
- It falls 7%
- It falls 70%, and a margin call is likely before that
- Nothing until the position is closed
Show the answer
B. It falls 70%, and a margin call is likely before that At 10-to-1 leverage every point on the position is ten points on equity. Losses are marked daily, so the dealer will usually demand more margin or close the position long before the year ends.
Exercise
One task to do in Finabase. Nothing is graded or saved.
Open the AUD/JPY chart on a five-year range. Find the peak in July 2024 and the low in early August 2024, and measure the fall in percent. The Reserve Bank of Australia’s rate was about 4 points above the Bank of Japan’s at the time. How many years of that carry did the drop cost? Then look for other sharp drops on the same chart and compare their size with the slow climbs between them.
Open the AUD/JPY chartWhat you should find
The July–August 2024 fall in AUD/JPY was several times a year of carry, and it took weeks while the climb before it took years. Expect the same asymmetry in other drops: slow rises while the gap holds, fast falls when positions are closed together.
Sources and further reading
- Federal Reserve Board (H.10) — Japanese yen per US dollar, daily noon buying rates (FRED series DEXJPUS), downloaded 2026-11-18
- Federal Reserve Board (G.5) — Mexican pesos and Brazilian reais per US dollar, monthly averages (FRED series EXMXUS, EXBZUS), downloaded 2026-11-18
- Federal Reserve Board — effective federal funds rate, monthly (FRED series FEDFUNDS), downloaded 2026-11-18
- Bank of Japan — Statements on Monetary Policy (March 19, 2024 and July 31, 2024 decisions)
- Banco de México — monetary policy target rate
- Banco Central do Brasil — Selic rate history
- Bank for International Settlements — BIS Quarterly Review, September 2024: the market turbulence of early August 2024
- Brunnermeier, Nagel and Pedersen, “Carry Trades and Currency Crashes,” NBER Macroeconomics Annual 2008, vol. 23
