How the flight to par effect can be used to create long option positions without paying premium.
Accounting regulations impact the pricing of bonds in many countries, as investors seek low coupons, high coupons or at-par issues. In Japan, this effect is especially pronounced. We shall therefore use the Japanese example to describe in the following article how the predictable premiums and discounts for certain issues can be exploited to replace long over-the-counter options with cash positions, which do not require premium payment. However, the results can be transferred to some other markets when adjusting for their specific accounting regulations.
It turns out that Japanese investors have a pronounced preference for at-par issues versus both below- and above-par bonds. For example, JGB 205, with a price of 99.476, trades on an asset swap spread 14 basis points through JGB 207 (price of 91.40), and 18bps through JL 8 (price of 128.56), although the latter spread also is influenced by differences in liquidity.
The main reason behind the premium for close-to-par bonds is the accounting constraints of the major players in the JGB market that encourage them to avoid accumulation and amortization. While the capital gains from below-par bonds cannot be realized before maturity, the capital losses from above-par bonds have to be amortized on a yearly basis. Par bonds, by contrast, both yield a coupon income reflecting market conditions and avoid accounting problems. Since the preferences of the major investors drive the pricing, this explains the general observation, that the closer to par, the higher the relative richness of a JGB. This fact is illustrated schematically in the first of the following charts1.
For investors not affected by these accounting regulations, the obvious trade is therefore always to sell those bonds quoting around 100. Subsequently, we shall show how to extend this basic idea to create synthetic long option positions.
Buying a bond with a low coupon, say 1%, versus selling a bond of the same BPV with a high coupon, say 2%, exhibits the following behaviour in a sell-off. As long as yields are below 1%, both bonds lose the same amount of above-par discount when yields rise and their spread remains constant. Once yields increase above 1%, the low-coupon bond trades with a rising below-par discount, while the high coupon bond continues to appreciate by coming closer to par, and the spread trade loses money. After the interest rate level passes 2%, both bonds are below par and cheapen in parallel with the spread remaining constant. This theoretical pattern, illustrated in the picture above, holds true in the market. The next chart depicts the underperformance of JGB 207 0.9% 12/08 versus JGB 206 1.5% 09/08 during the recent sell-off. At yields between the two coupons (0.9% to 1.5%), when JGB 206 lost its above-par discount and JGB 207 was quoted with an increasing below-par discount, there was a high correlation between the spread of the two bonds and the yield level (R2 = 84%, beta: 18bps spread increase per 1% sell off). Outside this range, we observe the predicted sideward movement of the spread, whereas the R2 is lower.
The predictable richness of at-par bonds allows the construction of long option positions using high and low coupon bonds. Unlike OTC bond options, these substitutes require no premium payment other than the cash and repo bid/ask spread.
Bull spread2. Sell a bond at par, buy a bond below par (hedged against yield curve shifts). In a bull market, the below-par bond will move towards par and outperform the current at-par bond, which underperforms as it rises above par. In a bear market, both bonds are below par and fall in parallel, leaving the spread unchanged. An example of such a synthetic option is a short JGB 206, long JGB 207 position.
Bear spread2. Sell a bond at par, buy a bond above par. An illustration is provided by a long JGB 206 and short JGB 207 position at a yield level of 0.9%. The history presented above suggests a 15bps performance potential in a sell-off, while in the brief rally below 0.9% the bond yield spread was stable.
Butterfly. Sell a bond at par, buy a bond below and above par (again, hedged against curve moves, e.g. via PCA weightings). This creates a long gamma option position, exploiting yield volatility for free.
1 Though the below-par discount is currently higher than the above-par discount, we use the same slope for both. Note that the payoff profile of the combination depicted in the second graph is independent of the ratio between below- and above-par discount.
2 The payoff profile approaches that of a call or put, respectively, when the bond bought is very far below or above par.
This week's Learning Curve was written by Christian Schaller, global relative value research, at Deutsche Bankin Tokyo.
