The failure mode does not announce itself. A trader holds a position in US equities. Another holds an emerging-market bond fund. A third has bought cryptocurrency. None of them thinks they own the same trade. Then one morning — August 5, 2024 — all three positions sell off simultaneously. The Nikkei 225 falls 12.4% in a single session, its worst day since Black Monday 1987. The VIX records the largest single-day spike in its history. Assets with no apparent fundamental connection collapse together. The cause is not a recession, not an earnings miss, not a geopolitical shock. The cause is a funding structure — the yen carry trade — and the forced unwinding of leveraged positions that had quietly linked these assets for years without any of their holders realizing it.

By the end of this article you will be able to explain what the carry trade is and why it persists despite theoretical reasons it should not, describe the exact mechanism by which carry unwinds create correlated crashes across unrelated assets, trace the August 2024 episode with verified dates and figures, and recognize the behavioral signature that distinguishes a carry-driven volatility spike from a fundamentally-driven bear market.

What the Carry Trade Is and Why It Works

A carry trade has a simple structure: borrow in a currency with a low interest rate, convert the proceeds into a currency or asset with a higher yield, and pocket the difference — the "carry." The funding currency is typically one where a central bank has held rates near zero for an extended period. The yen has been the dominant example for most of the past three decades, with Japan's policy rate near zero while rates elsewhere were meaningfully higher.

Standard economic theory predicts this should not be profitable. Uncovered interest parity (UIP) holds that a high-yield currency should depreciate over time by exactly the amount of the interest differential — eliminating any gain. If you earn 4% more in dollars than in yen, the dollar should fall 4% against the yen, and the trades should cancel out. This is a clean, logical prediction. It is also empirically wrong. Since Fama (1984) documented the anomaly, it has been repeatedly confirmed: high-yield currencies do not depreciate as predicted. They tend to appreciate slightly on average, making carry profitable in the aggregate — until, periodically, they do not. Brunnermeier, Nagel, and Pedersen (NBER Working Paper 14473, 2008) documented the full profile: small, steady gains that build for months or years, then abrupt, severe losses. Market practitioners describe this as picking up pennies in front of a steamroller — steady returns masking catastrophic tail risk.

The return distribution is what makes the carry trade dangerous in ways that average performance numbers hide. The negative skewness is real and documented: the losses, when they arrive, are not proportional to the steady gains that preceded them. A year of small positive returns can be erased in days. And because the positions are leveraged — you are borrowing to fund the trade — the losses amplify faster than the gains ever did.

The Hidden Cross-Asset Mechanism

The structural feature that makes carry unwinds distinct from ordinary market drawdowns is the cross-asset correlation they create. Here is the mechanism, stated plainly.

A carry trader borrows yen at near-zero cost, converts it to dollars, and deploys those dollars into whatever offers a higher yield: US equities, emerging-market bonds, corporate credit, cryptocurrency. The yen loan and the foreign asset are now linked. The trader does not need to disclose this link anywhere. The asset — say, a US technology stock — has no label on it identifying it as yen-funded. To any observer, it looks like a straightforward long position in a technology company.

Now the yen strengthens sharply. The loan that funded the position is suddenly more expensive in dollar terms. The trader faces a margin call — or simply needs to reduce the yen liability before it grows larger. The only way to do that is to sell the foreign asset, convert back to dollars, convert back to yen, and repay the loan. The fundamental case for the technology stock is irrelevant. The quality of the emerging-market bond does not matter. The trader must sell, not because anything changed in the underlying asset, but because the funding structure demands it.

Multiply this by the scale of the yen carry trade. BIS data as of Q1 2024 showed yen-denominated bank claims on non-bank borrowers outside Japan reaching $880 billion. Estimates of total carry positioning, depending on the methodology, range from hundreds of billions to multiple trillions — the range is wide because what counts as "carry trade" depends on how broadly you define yen-funded foreign investment. What is not in dispute is that the positions are large, leveraged, and widely distributed across asset classes. When forced unwinding begins, the sell orders are not concentrated in one market. They arrive simultaneously across every market where carry funds had been deployed. Assets that share no fundamental connection collapse together because they share a funding connection.

This is the hidden leverage. It is not visible in any single position's description. It only appears when the funding currency moves.

August 5, 2024: The Anatomy of an Unwind

On July 31, 2024, the Bank of Japan raised its short-term policy rate to approximately 0.25% from a range of 0 to 0.1% — a 15-basis-point hike, by a 7-2 vote, that markets described as a surprise. The BOJ also announced plans to halve its Japanese government bond purchase program. Both moves signaled a tightening trajectory that the market had not fully priced.

The yen, which had reached a 38-year low against the dollar (approximately 161.59 USD/JPY) on July 3, 2024, reversed sharply. In the week of July 29 to August 5, the yen strengthened approximately 6% against the dollar. As the funding currency appreciated, carry traders who had borrowed yen to fund foreign positions faced mounting losses on the currency leg of their trade. The deleveraging began.

On August 5, 2024, the results were visible across every market where carry funds had been deployed:

The Nikkei 225 fell 12.4% — 4,451.28 points — closing at 31,458.42, its worst single-day percentage decline since Black Monday in October 1987.
The S&P 500 fell approximately 3% on closing, with intraday losses reaching approximately 4.3%.
The VIX recorded an intraday pre-market high of 65.73 — the largest single-day spike in the index's history, confirmed by Cboe — though this peak occurred approximately one hour before regular US market hours.

None of these events had a common fundamental cause. The US economy had not entered recession overnight. Japanese corporate earnings had not collapsed. There was no new geopolitical shock. The trigger was a funding structure unwinding under the pressure of yen appreciation.

The recovery confirmed the diagnosis. The Nikkei bounced approximately 10% on August 6 — the following session — as the most urgent forced selling cleared. US markets recovered most of their losses within approximately one week. The Nikkei 225 did not return to its July 31 pre-crash level until approximately September 27, 2024 — roughly seven weeks after the crash — but the trajectory was continuous recovery, not the serial deterioration that characterizes a fundamental bear market. Carry-driven crashes can reverse sharply because the cause is mechanical, not fundamental. Once the deleveraging pressure clears, the underlying assets reassert their own value signals. Nothing had changed about the companies or economies that carry funds had been invested in.

The 1998 Echo: LTCM and the Same Structural Failure

The August 2024 episode was not the first time this failure mode appeared at scale. In 1998, Long-Term Capital Management collapsed through the same structural mechanism, though the strategy was different in form.

LTCM ran convergence and fixed-income arbitrage trades — exploiting pricing differences between related bonds, betting that spreads between securities would narrow. This is not the FX carry trade, and conflating the two would be inaccurate. But the structural characteristic is identical: use leverage to earn thin, steady spreads that collapse violently when the positions need to be liquidated simultaneously. LTCM had approximately $4.7 to 4.8 billion in equity, had borrowed over $124 billion — a leverage ratio exceeding 25:1 — and held derivative positions with a notional value of approximately $1.25 trillion. Its partners included Robert C. Merton and Myron Scholes, who had shared the 1997 Nobel Prize in Economic Sciences for their work on derivative pricing — awarded nine months before the collapse.

On August 17, 1998, Russia defaulted on its domestic-currency sovereign debt and declared a 90-day moratorium on foreign debt repayment while simultaneously devaluing the ruble. The shock caused a global flight to safety and a widening of the exact spreads LTCM was betting would narrow. The fund lost $4.6 billion in less than four months. On September 23, 1998, the Federal Reserve Bank of New York brokered a $3.625 billion private-sector recapitalization of LTCM, contributed by a consortium of 14 financial institutions. The Fed organized the rescue but did not contribute its own funds.

The lesson from LTCM is not that the strategies were wrong in theory. The lesson is that any trade structure that uses leverage to earn thin spreads shares the carry trade's core risk: when the trade goes against you, all the positions need to be liquidated at once, and the liquidation itself worsens the move against you. Brunnermeier, Nagel, and Pedersen describe this as a liquidity spiral — forced liquidation leads to further price moves, which lead to deeper forced liquidation, which accelerates the unwind.

What Carry Unwinds Cost — and Why They Are Hard to Avoid

The harm from carry unwinds falls into two categories: direct harm to carry traders and collateral damage to everyone else.

Carry traders absorb the direct loss: the gap between the interest rate differential they were harvesting and the capital loss from the forced currency move and asset liquidation. At scale, this can be catastrophic, as LTCM demonstrated. The leverage that amplified the gains amplifies the losses at the same ratio.

The collateral damage is less obviously connected. A trader who holds no carry position, has never heard of the yen carry trade, and is long a fundamentally sound business can watch that position fall sharply on August 5 not because anything changed about the business, but because carry funds needed to sell everything to repay yen loans. This is the mechanism by which a funding structure becomes a systemic risk: it imposes correlated selling on assets that have no fundamental relationship to each other or to the trade that is unwinding.

The difficulty in avoiding the collateral damage is that it is invisible until it materializes. The correlation between your position and the carry trade does not show up in any standard correlation matrix during normal conditions. It only appears under stress, at exactly the moment when you cannot protect against it.

Recognizing the Regime: Carry-Driven vs. Fundamentally-Driven Volatility

The diagnostic question when a sharp, correlated selloff occurs is: what would force selling right now regardless of any fundamental view? If the answer involves a funding currency move, a margin call cascade, or deleveraging pressure across diverse assets with no common fundamental driver, you are likely looking at a carry-driven event. If the answer involves new information that changes the fundamental outlook for the underlying assets — an earnings collapse, a recession signal, a policy shift with real economic consequences — you are more likely looking at a fundamentally-driven move.

The behavioral distinction matters for process. In a fundamentally-driven bear market, the deterioration is typically sequential and reinforced by worsening data. The recovery, when it comes, is gradual and tied to improving fundamentals. A carry-driven crash tends to be sudden and correlated on the way down, then sharp on the reversal once the forced selling clears. The August 5–6, 2024 sequence — a 12.4% crash followed by a 10% bounce in the next session — is the pattern. The Nikkei's full recovery within seven weeks, with no intervening recession or earnings deterioration, is the confirmation.

This does not mean carry-driven crashes are safe to hold through. Forced selling can push prices to levels that trigger stop-losses, margin calls, and secondary liquidations before the reversal arrives. The volatility is real even if the cause is mechanical. But recognizing the regime changes what you look for: in a carry event, the diagnostic is not "how bad are the fundamentals" but "how much forced selling is still in the pipeline."

The Discipline

Four process habits reduce exposure to carry-driven volatility as a distinct risk source.

Know your correlation during stress, not during calm. The correlations between your positions that show up in normal conditions understate the correlations that appear during forced deleveraging. Before sizing any position, ask whether the assets you hold could all become sell candidates simultaneously if a major funding currency moved sharply. This is not a quantitative exercise for most traders — it is a qualitative one. The question is: is there a plausible funding stress scenario that would force selling across my positions regardless of their individual fundamentals?

Track the yen as a risk signal, not just as an FX pair. A sharp, rapid yen appreciation is a visible early warning that carry positions are being unwound. USD/JPY is freely available. When it moves more than 2 to 3% in a short period without an obvious domestic Japanese catalyst, the probability of carry-related forced selling in foreign assets rises. This is not a trading signal — it is a diagnostic signal that the regime may be shifting.

Distinguish the volatility spike source before acting. When the VIX spikes and markets fall together, the first process question is not "should I buy the dip" or "should I cut everything." It is: what is the source? Fundamental deterioration requires a different response than funding stress. Funding stress can clear faster than fundamental problems, but it can also cascade further if the deleveraging is not yet complete. Identifying the source is the precondition to any sensible decision.

Size for the possibility of simultaneous correlated losses. A diversified portfolio that shares a hidden funding connection is not diversified under stress. If you cannot rule out that your positions are partly funded by carry trades — directly or indirectly through institutional flows — then your true position size under a carry unwind scenario is larger than it appears. The sizing framework needs to account for this possibility, even if the probability seems low in any given period.

Simulator Exercise

This drill runs in Abu Terminal's Speed Run and trains the single most important diagnostic habit: distinguishing what is forcing selling from what would change a fundamental view.

Setup. Choose any Speed Run scenario that includes a sharp, sudden multi-asset selloff. The August 2024 volatility event (event h363 in the Speed Run) is the direct analog. You can also use any scenario where the VIX spikes and equities fall by more than 2% in a single session while a currency pair moves sharply.

The diagnostic drill. When the selloff appears in the Speed Run, pause before making any decision and answer these three questions in sequence:

What fundamental information changed? Was there new earnings data, a policy announcement with real economic consequences, or a deterioration in macro indicators? If yes, name it specifically. If you cannot name it, note that.
What would force selling right now regardless of the fundamental view? Are multiple asset classes falling simultaneously? Is there a currency move that could trigger margin calls on leveraged positions? Are the falling assets fundamentally unrelated, suggesting a common funding link rather than a common fundamental driver?
What is the likely reversal pattern? If the selling is fundamental, the recovery is tied to when the fundamentals improve — potentially months. If the selling is forced and mechanical, the recovery can begin the moment the deleveraging pressure clears — potentially days.

Work through this diagnostic for every scenario where a sharp correlated selloff appears, not just the carry-specific events. The goal is to build the reflex of asking "what is forcing this" before asking "what should I do." In volatility regime shifts, the answer to the first question determines whether the second question has a different range of correct answers than normal.

After completing the Speed Run, review your decision trail. For the sessions where you correctly identified the selling as mechanically-driven rather than fundamentally-driven, check whether you sized your response differently — smaller, more cautious about catching the bottom, more focused on letting the forced selling clear before acting. That calibration is the skill being developed: not prediction, but regime recognition.

A companion drill for overnight risk: before any position that will be held through a session close, note explicitly whether the position is in an asset class that could be a carry trade target. Yen-funded carry tends to concentrate in high-yielding assets and risk-on positions. If your overnight holding is in a category that carry funds favor, your overnight risk is not just the asset's own volatility — it includes the possibility that a currency move will trigger forced selling in your position before the next session opens.

The Carry Trade and Why It Unwinds Fast