Amortising securities are structured so that their principal is repaid over the lifetime of the security, rather than as a single lump sum at maturity. Typically, the ratio of interest to principal is designed to change over time so that the borrower can make equal payments throughout the security’s lifetime.
Some types of amortising bonds, such as mortgage-backed securities (MBSs), passthroughs and collateralised mortgage obligations (CMOs), have the additional feature that the borrower can pay the principal back to the lender faster than scheduled. When early repayments of principal are made, such securities can generate paydown return in addition to return from carry, curve and credit shifts:
To see why, suppose that a security is purchased at $90 per $100 face value. The owner is guaranteed to make a profit, irrespective of what happens in the marketplace. All they have to do is to hold it to maturity, wait for the price to converge to its par value, and bank a capital gain. This pull to par effect is described in Chapter 9.
However, this assumption holds only if the face value of the investment stays unchanged at $100, which is precisely what sinking securities do not do. Suppose that the security had half of its face value returned to the purchaser just after it had been bought at $90. The effect of this paydown is to halve the rate of return, because only 5% of the nominal face value will be generated by pull to par, not 10% as previously.
Paydown return is therefore closely related to pull-to-par return, but should be displayed separately as it is not a function of yield. In general, paydown return rp is given by
where:
The following example is from Fabozzi (2012, p. 1675):
Return due to paydown is then
The security was trading at a premium, so its paydown return was negative over the interval.
In general, paydown return should be calculated and displayed for all sinking securities, in addition to carry and market returns, when they allow early repayment of principal. Vanilla bonds such as US Treasuries do not show any paydown return, because their principal remains unchanged throughout their lifetime.
Equation (8.6) shows that the relationship between yield and price is roughly linear, with nonlinearity depending on the security’s convexity.
Convexity is a very desirable property in a bond. If bond yields fall, the price of a bond with high convexity will rise faster than a bond with low convexity. Conversely, if bond yields rise, a high convexity bond’s price will not fall as much as a low convexity bond.
Roughly speaking, convexity varies as the square of the maturity of the bond, so long-dated bonds tend to show the highest convexity. In some cases this extra convexity makes long-dated issues so attractive that bonds at the end of the yield curve may see their price bid up in the marketplace. For this reason, some yield curves occasionally show a modest degree of inversion, or downwards slope, at the long end that is entirely due to convexity effects.
In some cases it can be useful to separate a bond’s convexity return from its duration return. Suppose that a trader owns a bond with the following characteristics (see Table 12.1):
The trader believes there will be significant volatility in the marketplace, but does not know which way the markets will move. Their strategy is to sell the bond and buy two other bonds as follows (see Table 12.2):
Table 12.2 Barbell strategy after restructuring
The purpose of this restructure is to construct a new portfolio with the same modified duration as the original, but with much higher convexity. This has increased the trader’s portfolio’s second-order sensitivity to yield changes. As long as there is a substantial change in yields, the new portfolio will outperform the old one, and this will occur whether the market moves up or down.
The strategy is based on two assumptions:
This is known as a barbell strategy, as it moves the main cashflows of a portfolio outwards so that they lie at opposite ends of the term structure, with the graph of exposures against maturity resembling a barbell.
To assess whether this strategy worked, one needs to know the portfolio’s return from convexity, as distinct from modified duration. A first-principles approach to pricing will not generate a value for modified duration, in which case it will be necessary to calculate this quantity separately as follows.
Suppose that the current YTM is given by y. Given a pricing function p = p(y, t), its numerical derivative may be substituted into equations (8.4) and (8.5) to give the effective duration and convexity:1
where δ << 1.
The modified duration and convexity returns are then given by the second and third terms in equation (8.6), respectively.
Consider a security that has a single cash flow one year in the future. The yield curve is steeply sloped at the 1-year point, but flattens out at longer maturities.
Suppose further that market conditions do not change for a month, so that the level and shape of the yield curve remains unchanged over this interval.
At the end of the month, this one-year bond will have become an 11-month bond, and the yield used to price this security will now be read from the 11-month point rather than the 12-month point on the yield curve.
Since the yield curve is downwards sloping, the 11-month yield will be lower. Since the yield of the security is lower, the price will be higher, and a positive return will have been generated.
This effect is sometimes called riding the yield curve, as it is most effective when a security’s cash flows are positioned at maturities where the curve is most steeply sloped.
Note that this return has not been generated by movements in the market, since we explicitly assumed that market conditions were unchanged. Nor has it been generated by elapsed time because the return is generated entirely by a change in yield. Rolldown is distinct from either source of return, and should be measured separately.
Everything else being equal, it can be worth positioning one’s securities so that the bulk of their cash flows are positioned at the most steeply sloped parts of the curve. However, rolldown return is seldom substantial, and will be overwhelmed by even quite small changes in the level or shape of the curve.
A bond’s change in yield due to rolldown is given by
where Yt is the yield curve at time t, Yt(m1) its value at maturity m1 and Yt(m2) its value at maturity m2 where m2 < m1.
Return generated by rolldown may then be calculated in the same way as carry:
Pull-to-par return is generated even if the yield curve is flat, as it is driven by differences between the price of the bond and its eventual par price. In contrast, rolldown return is driven entirely by the shape of the yield curve.
A security’s price may be affected by liquidity, which is a measure of the ease with which it may be bought or sold without affecting its price. For instance, a US bond that is on-the-run (a reference issue) will be traded in significantly larger volumes than one that is off-the-run, despite being identical in all other respects. The off-the-run bond may trade at a lower price, or a higher yield, because of perceived difficulties in unwinding the position.
Issue size can also affect liquidity, and hence the bond’s yield. For instance, a corporate issue that has identical cash flows and creditworthiness to a highly liquid sovereign issue can trade at a substantially higher yield, because of perceived difficulties in unwinding a position in that security. The effect may be especially pronounced if the amount on issue of the corporate bond is small.
Liquidity considerations can generate returns in two ways:
1 Effective duration measures yield sensitivity when optionality is present. If a bond has no options, its effective duration is equal to its modified duration, and similarly for convexity.