All Boxed Out

Market turbulence hit Japanese interest rate markets in March, leaving hedge funds and banks with losses in the billions of dollars. What went wrong? By Christopher Jeffery


Relative-value fixed-income hedge funds, bank proprietary trading desks, local securities dealers and other large Japanese government bond (JGB) underwriters suffered billions of dollars - perhaps tens of billions of dollars - of mark-to-market and real losses in March, as turbulence struck the country's once-well-behaved interest rates market.

Seven-year JGB futures - the most liquid government-linked debt instruments in the country and consequently the cheapest to deliver - saw yields move 7 basis points in a day in March. This led a number of Japanese securities dealers to describe the market as "broken" that month. "I've never seen such a huge move in the market," says a senior fixed-income trader at one of the country's biggest securities dealers.

Endeavour Capital, reputedly lost 27% of its value, or $810 million, as losses caused by market dislocation triggered bank covenants that forced the London-based fund to slash its leverage from around 18 times to virtually nothing. It was not alone - most relative-value fixed-income funds were exposed to the market moves, as were foreign bank proprietary trading desks and local securities dealers and banks. Losses are expected to range between $100 million and $400 million apiece for major market participants.

The development was all the more surprising given that the Japanese interest rate markets have behaved benignly for several years. The rapid repricing of interest rate curves for both cash and synthetic markets put tremendous pressure on hedge funds and proprietary trading desks that had engaged in the so-called 'box trade' and similar strategies.

"It was one of those events where things start to go wrong and everyone cuts their positions," says a senior fixed-income salesman at a large European bank in Tokyo. "Initially, people were selling fairly discreetly." Asked for names of those scrambling for the exits, he said: "Our dealers were purposefully trying to be conservative on pricing these bonds, so they didn't get to see the flows."

Bets on creditworthiness

These box trade-type transactions are ultimately linked to bets on the relative creditworthiness of the Japanese government versus swaps counterparties - that is, the creditworthiness of banks. While some parties make outright bets on the direction of that spread - something Japanese counterparties are more likely to take on, say some market participants - until the events in March, it was perceived as less risky to do a trade based on a so-called 'spread of spreads'. This is effectively a transaction based on one spread against the others.

"You may not know if the government credit is improving or getting worse relative to bank credit," says the fixed-income salesman. "But a five-year spread shouldn't be that much different from a seven-year spread. If you think one spread is rich or cheap relative to something of a different maturity, it is considered a less risky position than an outright bet on the overall direction of these spreads."

For example, spreads between seven-year and 30-year asset swaps have traded in a rough range of plus or minus 19bp since around December 2000 (see figure 1). There are many reasons why the spread moves around - for instance, it could be linked to changing perceptions about whether the Japanese government has more of a short- or long-term debt problem.

Relative-value funds assume there is a form of mean reversion associated with such positions. This means the fair value should trend to zero - although it tends to be lower than zero, due to concerns about the long-term implications of Japan's debt mountain. So when 30-year asset-swap spreads rise relative to seven-year spreads, there is an expectation that 30-year spreads will narrow and seven-year spreads will widen, and vice versa.

The box trade is described a 'smart money' trade - although, as in the past, it looked like the smart money was highly correlated in this instance. Essentially most fixed-income desks will move into the trade once there is a big differential from historic spreads. While it is done in asset-swap form and in a cross-currency format - the latter by hedge funds based outside Japan - the most popular approach to the trade is a box-trade format.

Traders do this because there is perceived to be better liquidity in the underlying instruments compared with asset swaps and cross-currency swaps. This makes the trade - which requires high leverage of up to 50 times to make significant profits on the small spread moves - cheaper and usually cheaper to exit.

For example, a hedge fund could buy a long-dated JGB, typically with a maturity of 20-30 years. It then shorts swaps in the corresponding year to create the first part of the spread trade. At the same time, hedge funds sell seven-year JGB futures - because they are more liquid and cheaper than seven-year JGBs - and receive seven-year swaps.

What should then happen is that the basis reverts back to the mean and the fund makes money as it unwinds its positions. This is reinforced in Japan, as large local pension funds and insurers - which need to buy long-dated bonds to help them match their assets to their liabilities - will spot that long-dated bonds have cheapened and pile into the market. Their buying activity subsequently helps push the differential towards equilibrium, and the highly leveraged investors thus make significant profits.

The transaction has worked to mutual benefit for many years, with word catching on in the investment community to the point where hedge funds based outside Tokyo also put on the trades and foreign real-money investors, like their peers in the pension and life business in Japan, also buy the relatively cheap long-dated JGBs.

But then it all went wrong. What appears to have sparked the collapse of the relative-value - at least for a short disruptive period - were external factors. The timing came as Bear Stearns was on the verge of collapse, falling eventually into the hands of JP Morgan; Lehman Brothers was looking precarious; and Morgan Stanley seemed far from in robust financial health. This caused liquidity to freeze in the global markets and dealers shied away from taking risk, forcing hedge funds to unwind positions in a relative liquidity vacuum.

Real-money players

So what happened to the Japanese real-money players? "The domestic players are not stupid," says the fixed-income salesman. "If they see this capitulation, they will come in and buy some. This was a once-in-a-lifetime opportunity to buy government paper. Funds bought some. But they weren't going to take every trade that came through. They were going to wait a bit and then come in. It was the same story in all markets. The real-money cash guys had several opportunities of which leveraged guys could not take advantage."

Meanwhile, international real-money funds that tend to buy the bonds and swap back into their own currency appeared to have better options. That's because the market dislocation in the US had made AA paper incredibly cheap, with even AAA paper being available at cheaper levels than AA JGBs (see figure 2). "When long-end asset swaps were cheap enough, they were very big backstop support for the market," says the salesman. "But the global credit market completely took that support away

"Pretty much anyone who has done this trade since 2006 hasn't done well," he adds. "People that had this on were long-term traders and weren't going to be pushed out by a little move, otherwise they would have been pushed out already. But we were surprised how large positions people had on."

The move may have also hit JGB underwriting banks, say some market participants. The largest underwriters of JGBs include Mitsubishi UFJ, SMBC, Mizuho and Nomura. Traditionally, JGB underwriters have hedged their exposures to long-dated JGBs by using shorter-maturity JGBs or futures. But more recently, they have adopted the global practice of using a mix of government bonds and swaps, say market participants.

"I imagine the domestic bond desks would have had on that position or some [variation of it just through their natural hedging," says one market participant who requested anonymity. "So when things started to go wrong, they had to get out of them pretty quickly."

However, a risk manager at a big US investment bank in Tokyo, says the impact would probably be muted. "You want to hedge with an extremely liquid instrument, and my guess is that in Japan that it is probably bond futures," he says.

In the underwriting area, non-primary dealers put in bids at a level they believe they will be filled at, adds the risk manager. "It would only cause problems if you happened to have taken down a huge chunk of the auction at one point in time and are carrying that on your book when the curve flattens dramatically and hedges don't work. So it is a joint probability event."

Meanwhile, the box trade has reverted back to its normal trading pattern. This was, in part, due to the restoration of confidence in global financial markets following the US Federal Reserve-orchestrated bail out of Bear Stearns and its propping-up of US securities dealers by opening its discount window to them for the first time. In addition, after the end of Japan's fiscal year on April 1, pension funds and life insurers - which do not alter their portfolios in the immediate run-up to the end of the year - entered the market and bought up cheap bonds.

The result is that anyone who was in a position to hold their nerve during March or bought into the capitulation did very well.

Only users who have a paid subscription or are part of a corporate subscription are able to print or copy content.

To access these options, along with all other subscription benefits, please contact or view our subscription options here:

You are currently unable to copy this content. Please contact to find out more.

You need to sign in to use this feature. If you don’t have a account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an individual account here