Japan's last cheap-money trade

Money-market swaps put a 90 per cent chance on a 25-basis-point rise at the Bank of Japan's 19 December meeting, and currency desks are getting ready. If the odds are right, the yen carry trade, for decades one of the most profitable positions in finance, is nearing its end. Borrowing yen to lever up elsewhere is about to stop being easy money.

How the trade works

The carry trade is simple in outline. An investor borrows in yen at a near-zero rate, converts the money into a higher-yielding currency and buys an asset there, pocketing the difference in interest rates for as long as the position is held. The profit depends on the yen staying weak or at least stable, because a rise in the yen raises the cost of repaying the loan. Most of the time the interest income dribbles in steadily; occasionally the currency moves against the position and wipes out years of gains in days.

The trade takes many forms. Hedge funds borrow yen directly or sell it forward; foreign banks fund positions through cheap yen loans; Japanese households themselves have poured savings into foreign equity funds through the tax-free NISA accounts expanded in January 2024, a slower-moving version of the same bet. All of them rely on Japanese rates staying far below those abroad, and all of them add to selling pressure on the yen while they grow.

The rate gap narrows

For more than twenty years hedge funds, corporate treasurers and sovereign wealth managers borrowed billions of yen at zero or negative rates and put the money into higher-yielding assets, from US Treasuries to Mexican pesos and tech stocks. Now the Federal Reserve is cutting US rates towards 3.50 per cent while the Bank of Japan lifts its own towards 0.50 per cent. The spread between the two is shrinking faster than it ever has.

The Fed has cut three times since September, most recently last week, while the Bank of Japan has been moving the other way since it ended negative rates in March 2024. Each step on either side takes income out of the carry. More important than the level of the gap is its direction: a trade that pays a known spread is tolerable while the spread is stable, but once markets expect it to keep narrowing, the incentive is to close positions before others do.

Earlier unwinds

The history of the trade is a record of slow gains and sudden losses. In October 1998, as the collapse of Long-Term Capital Management forced leveraged funds to cut risk, the yen rose by more than 10 per cent against the dollar in a matter of days. In 2008 the global crisis produced a longer and larger rally in the yen as carry positions across currencies were abandoned. Most recently, the Bank of Japan's rate rise in July 2024 coincided with weak American jobs data, and the combination triggered a violent unwinding in early August that sent the Nikkei to its worst day since 1987.

The common thread is that the unwinding happened fastest when a second shock hit at the same time; a change in rates alone was rarely enough. The 2024 episode also showed that estimates of the trade's size tend to be too low until it starts to unwind, because much of the borrowing is hidden in derivatives and in the balance sheets of global banks. That uncertainty is itself a reason for caution now.

When the hedge costs more than the carry

Once the spread falls below the cost of hedging the currency, there is little reason to keep leveraged positions offshore. A rush to buy back yen could force selling of assets across global credit markets.

The logic runs through the forward market. An investor who hedges the currency risk on a yen-funded position pays away roughly the interest differential in the forward price, which is why fully hedged carry earns close to nothing by construction. The unhedged investor keeps the spread but bears the currency risk, and the spread is the only cushion against a rise in the yen. As it shrinks, the yen move needed to wipe out a year's income gets smaller, and so does the investor's tolerance for holding on.

Who pays in an unwind

The assets that gain most from yen funding would suffer most as it leaves. High-yielding emerging-market currencies such as the Mexican peso have been favourite destinations, and a scramble to repay yen loans would mean selling them. Crowded equity positions, particularly in large technology stocks, would face selling from leveraged holders who need to cut exposure regardless of valuation. Credit spreads would widen as funds sell what they can rather than what they want.

Inside Japan the effects cut the other way. A stronger yen would lower the price of imported energy and food and ease the pressure on households, while cutting the value of overseas earnings for exporters and the yen value of foreign assets held by pension funds and NISA savers. The government has intervened to support the yen in recent years; a rapid move in the opposite direction would bring relief at the Ministry of Finance, though the stock market would probably not share it.

Still a wide gap

The strongest argument against an imminent unwinding is arithmetic. After a December rise, Japanese rates would be at 0.75 per cent, still close to three percentage points below the Fed's range, and Japanese inflation would still be above the policy rate. The yen weakened after the March 2024 exit from negative rates rather than strengthening, because markets judged the pace of tightening too slow to change the calculation. On that view, carry traders will keep collecting a large spread for another year.

That argument is right about the level and wrong about the risk. The trade survives on low volatility as much as on the spread, and every step by the Bank of Japan raises the chance of a sudden move in the currency. The signals to watch are the bank's guidance on how far rates will rise from here, speculative positioning in yen futures reported by the Commodity Futures Trading Commission, and the pace of Fed cuts in the first half of next year.

Cheap yen funding has propped up returns on leveraged portfolios for a generation. With it gone, those portfolios will have to earn their keep, and the ones that cannot will probably be unwound in a hurry. The unwinding is most likely to come in bursts rather than a steady retreat, triggered by the next shock to American growth data rather than by the Bank of Japan's own decisions.