In the first half of 2010, I advocated the “Yen Convergence trade” as an alternative to the Yen Carry trade in a G7-wide zero interest rate (short-term) environment. The Yen Convergence trade takes the view that one of two things is likely to happen: a.) America and Europe become more like Japan, and their ageing populations push long-term interest rates permanently below 2% (a relationship described in my earlier post on age and interest rates), or b.) oil or some other external factor sparks inflation in Japan, buoying long-term JGB yields up to western levels. Meanwhile, the trade would benefit from between 1% and 2.5% positive carry between the long western bond / short JGB position. The trade would have paid well in 2011 and 2012, and only fell apart earlier this year when the market response to the Fed taper was clearly “The US is not Japan, the Fed is not the BoJ, and even with a 14% decline in the Yen JGB yields can continue to fall.”
Carry is a generalized form of income, and a top answer to “what makes one wealthy / not go broke?” would probably be “a high income”. In fixed income though, one of the first lessons we all learn in bond math are the two ways to go broke even with a high fixed income: either have inflation eat away your spending power, or similarly by having rising short-term rates eat up and then surpass the fixed rate you locked in. The latter case is more explicitly felt when using leverage (or other names for financing) of the fixed income position, and modern financial systems will calculate the future “bite” of rising expectations of short-term rates second by second and can force years of these losses to be realized immediately in a margin call. Below is a simple example of “bond math that any business student (and hopefully one day, every 5th grader) should be able to do, but few actually bother calculating”:
- At the start of 2013, the 5yr UST yield was around 0.76%, so “buying” a gross income of US$1 million per year would cost about US$132 million to buy the bonds at market price.
- At the same time, the 30yr yield was 3.04%, so you would only need to buy US$32.9 million of bonds to get the same income.
- At those yields, the duration of the 30yr is about 4x that of the 5yr, so assume for comparison that you can finance your purchase for a 9 month fixed term for 5% down / haircut on the 5yr and 20% down for the long bonds. That means you could have financed a gross fixed income of US$1 million / year with only US$6.6 million down using either 5yr or 30yr bonds.
- Yields in Japan of course were much lower, with 5yr JGBs yielding 0.187% vs 30yr at 1.967%. To finance an income of ¥100,000,000 per year would have required about ¥2.7 billion of 5yr JGBs or ¥1.0 billion of 30yr JGBs, or about 2-4x more cash up front.
These leverage ratios are actually somewhat conservative for the two largest and most liquid bond markets in the world, but large enough for the chart below to illustrate the dramatic divergence in JGB and US Treasury performance this year: your ¥1 bio in 30yr JGBs would have doubled in just four months earlier this year when yields there fell as low as 1.28% in April before recovering to a modest +20% gain by now, while either of your US Treasury trades would have lost almost 60%, leaving you less than $3 mio of your original $6.6 mio stake.
The moral of the story is not so much about the risk of leverage (which should be obvious, but is used here to make the story more dramatic), but rather so show how greatly fixed income investors can benefit from global diversification of their interest rate risk. Said differently, this is a dramatic and recent example of how the portfolio manager of just USD fixed rate bonds dramatically underperformed a portfolio manager who bought both USD and JPY fixed rate bonds (with appropriate FX hedges), and how mean-variance optimizations taught in business schools can apply just as well to interest rate risk as to equity risk.
