3
$\begingroup$

What is the formula for the forward price of a bond (assuming there are coupons in the interim period, and that the deal is collateralised)

Please also prove it with an arbitrage cashflow scenario analysis!

I suppose it is like fwd = spot - pv coupons) × (1+ repo × T ), I am not certain at what rate to pv the coupons.

$\endgroup$
12
$\begingroup$

Amazingly, there are several different methods for computing bond forward price – the underlying ideas are the same (forward price = spot price - carry), but the computational details differ a bit based on market convention.

Let's start with the basics. Assume between now ($t_0$) and the forward settlement date $t_2$, the bond makes a coupon payment at time $t_1$. Now consider the following series of trades:

  • Today, a trader buys a bond at a price of $P + AI_0$ (spot clean price + spot accrued interest).
  • To fund the purchase, the trader enters into a $t_1$-year term repo agreement at a repo rate of $r$. More specifically, he/she sells the repo by borrowing $P + AI_0$ and delivering the bond as collateral.
  • At time $t_1$ (coupon payment date), the repo balance is $(P + AI_0)(1 + rt_1)$ and the trader receives a coupon payment of $c / 2$ for being the owner of the bond.
  • The trader re-enters into another repo agreement that spans from $t_1$ to $t_2$ on a principal of $(P + AI_0)(1 + rt_1) - c/2$. This new loan, combined with the coupon payment of $c/2$, allows the trader to retire the old repo loan without putting up any additional capital.
  • Finally, at time $t_2$, the trader gets back the bond and repays the repo loan along with interest from $t_1$ to $t_2$: $$ \left((P + AI_0)(1 + rt_1) - \frac{c}{2}\right) \bigl(1 + r(t_2-t_1)\bigr) . $$

These trades are economically no different from buying the bond forward at time $t_2$. Therefore, the forward clean price for settlement at $t_2$ must be $$ F(t_2) = (P + AI_0)(1 + rt_1)\bigl(1 + r(t_2-t_1)\bigr) - \frac{c}{2}\bigl(1 + r(t_2-t_1)\bigr) - AI_{t_2}. $$

The method above is known as the Compounded Method. In the US Treasury market (and most international bond markets), a small approximation is made. Recall for small $rt$, we have $$ (1 + rt_1)(1+r(t_2-t_1))\approx 1 + r(t_1+t_2-t_1) = 1 + rt_2, $$ we therefore have the Proceeds Method: $$ F(t_2) = (P + AI_0)(1 + rt_2) - \frac{c}{2}\bigl(1 + r(t_2-t_1)\bigr) - AI_{t_2}. $$

The Proceeds Method is for all intents and purposes the standard/default way of pricing bond forwards. There's also the "Simple" and "Scientific" methods, but these are rarely used.

$\endgroup$
  • 1
    $\begingroup$ its interesting that theres no ois rate used in your approach. i guess because you assume there is a decent liquid market for repoing this bond for the periods needed.... $\endgroup$ – Randor Sep 9 '16 at 20:19
  • 1
    $\begingroup$ It's customary to finance Treasuries in the repo market. Some on-the-run issues, when in short supply, have been financed at -3% over the past few years, while OIS doesn't go near that level. $\endgroup$ – Helin Sep 9 '16 at 20:22
  • 1
    $\begingroup$ what about for some less liquid , riskier issuer bonds? $\endgroup$ – Randor Sep 9 '16 at 20:23
  • 1
    $\begingroup$ You should use whatever funding rate you and the dealer agree on. On maturity, $F = P$, where P is hopefully par. $\endgroup$ – Helin Sep 9 '16 at 21:31
  • 1
    $\begingroup$ @helin a little light on the accrued interest at time t2 would be nice. Everthing else is fully explained, so it's only fair that accrued interest gets the same treatment :) $\endgroup$ – Yuca May 2 at 15:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.