How does the carry trade affect emerging market currencies?

The short answer

A carry trade exploits interest rate differentials. An investor borrows yen at near-zero cost, converts the proceeds into Brazilian real or Mexican peso to earn 8-10%, and pockets the spread as long as the FX rate stays stable. The strategy is simple in concept and devastatingly leveraged in practice.

For emerging market currencies, this creates a paradoxical relationship with global volatility. Periods of low volatility encourage carry positioning, which itself supports EM currencies and makes them appear stable. The build-up of leverage is invisible until something forces an unwind, at which point the strategy’s offsetting capacity in the FX market disappears all at once.

The systemic feature most participants underweight is that the unwind is essentially a left-tail event compressed into days. Carry returns drip in slowly during good times and reverse violently. The August 2024 yen episode is the cleanest recent illustration of this asymmetry.

→ New to FX dynamics? FX markets and monetary regimes

What the data shows

The August 2024 unwind documented by the BIS provides the cleanest contemporary case study of how carry leverage amplifies FX moves.

The numerical context (BIS Bulletin No 90, BOJ, FRED, 2024):

• The Bank of Japan raised its policy rate from 0.10% to 0.25% on July 31, 2024, a small move on paper
• The Nikkei 225 fell approximately 20% between July 31 and August 5, 2024, the worst stretch since 1987
• The U.S. 10-year Treasury yield fell 55 bp between July 24 and August 5 as risk-off compounded the move
• The yen and Swiss franc — both classic funding currencies — appreciated most, while the offshore renminbi unusually appreciated too, suggesting funding role expansion

The exception that nuances the rule: BIS analysis showed most large EM currencies stayed remarkably stable through the episode, with only the Mexican peso and Australian dollar depreciating sharply versus the yen. EM rates actually rallied despite the risk-off tone, because positioning had been more concentrated in developed-currency carry pairs than in EM carry pairs.

→ Dataset: VIX volatility index dataset

Why it happens — the macro mechanism

The carry trade reshapes EM currency dynamics through three interlocking channels.

The compression channel. When carry positions accumulate, they generate persistent buying pressure on high-yield EM currencies, suppressing their volatility and pushing realized FX rates above what fundamentals alone would justify. Empirical work on uncovered interest parity documents that carry returns persist precisely because the rational FX adjustment that should offset the rate differential consistently fails to occur in real time.

The leverage channel. Carry trades are typically implemented synthetically through FX forwards, swaps, and options, all booked off-balance-sheet. The BIS estimates the aggregate size of yen-funded carry exceeded one trillion dollars in mid-2024, but the off-balance-sheet portion is impossible to measure precisely. This is the angle conventional sovereign analysis misses entirely — leverage builds up in derivatives markets that authorities cannot easily track.

When margin calls hit, the deleveraging is mechanical rather than discretionary, which is why the unwind moves so fast.

The contagion channel. Once an unwind starts, traders close positions across all carry pairs simultaneously to reduce overall exposure. This means a shock originating in one specific pair — yen versus dollar, say — propagates to apparently unrelated EM currencies. The August 2024 episode showed Bitcoin and Ethereum dropping 20%, suggesting retail margin calls forced sales of unrelated assets.

Synthesis by regime: in a low-volatility regime with a stable funder (2010-2021 yen, much of post-2009), carry compresses EM funding costs and supports EM currencies, building hidden leverage. In a funder-rate-hike regime (2022 Fed cycle, 2024 BOJ), positions begin to unwind selectively, with the most leveraged pairs moving first. In a full unwind regime (1998 LTCM, August 2024), volatility spikes across all asset classes simultaneously as deleveraging dominates, and the apparent diversification benefits of EM currencies evaporate within days. The transition between regimes typically occurs when the funder’s policy rate moves above the breakeven implied by the carry differential plus expected FX volatility.

Carry returns drip in slowly. Carry losses arrive all at once.

→ Framework in view: FX markets and monetary regimes

What it means for different economic actors

Allocators in EM debt funds need to recognize that quoted yields embed a meaningful carry premium that compensates for tail risk, not just credit risk. The Sharpe ratio of EM local-currency strategies looks attractive in stable regimes and unattractive in unwind regimes — averaging across both is the only honest measure.

EM corporate borrowers who fund themselves in dollars or yen for cost reasons are themselves participating in the carry trade implicitly. The 2024 episode showed how quickly that funding can become more expensive than the local-currency alternative they avoided.

Pension funds and insurers in funder countries — particularly Japan — have increasingly relied on unhedged FX exposure to boost returns. The August 2024 unwind was partly driven by these long-term holders being forced to sell U.S. equities to meet margin requirements on currency hedges, demonstrating how carry positioning can drive supposedly long-term flows.

A common analytical error is to view carry trade unwinds as one-off events tied to specific catalysts. The data suggests they are recurring features of the FX system, with the timing unpredictable but the eventual occurrence essentially certain whenever positioning has accumulated enough.

Practical observation

Go deeper

Related questions

• Why are emerging markets more vulnerable to dollar cycles?
• What is carry trade and why is it dangerous?
• What causes the Japanese yen to act as a safe haven?
• How does dollar funding stress manifest in global markets?

Frequently asked questions

How is the carry trade different in 2025 versus the pre-2008 era?

The strategy is identical in mechanics, but execution has shifted decisively to derivatives. Where 1990s carry positions involved cash borrowing in yen and physical FX conversion, modern positions use FX forwards and options, making the aggregate exposure invisible to balance sheet measures. This is why BIS estimates of carry trade size carry such wide uncertainty bands — the data infrastructure has not kept pace with the off-balance-sheet expansion.

Why did most EM currencies hold up in August 2024?

BIS analysis suggests carry positioning had become more concentrated in developed-market funder pairs than in classic EM pairs after 2022. The Mexican peso and Brazilian real had attracted carry inflows but smaller than 2010-2014 episodes. The episode was therefore largely a Japan-versus-United States story, with EM acting as bystander rather than epicenter — a notable departure from the 1998 pattern.

Can central banks prevent carry-related stress?

EM central banks build FX reserves precisely as a buffer against unwinds, and several have intervened directly in carry-sensitive episodes. But the BIS notes that the sheer scale of off-balance-sheet positioning means individual reserve buffers are unlikely to absorb a synchronized unwind. The 2008 dollar swap line architecture between major central banks remains the most credible backstop, and it does not extend systematically to EM funding stress.

Last updated — 12 July 2026