Skip to content
All library documents

Deriving Zero-Coupon Bond Prices from Instantaneous Forward Rates

Article Quant Q&A · Author: Frank Swanton

Summary

The document explains how a zero-coupon bond price relates to forward rates in discrete and continuous time. It divides the period to maturity into subintervals and expresses the bond price as the product of discount factors for each interval. With simple compounding, each factor is the reciprocal of one plus the interval forward rate times the interval length.

It then increases compounding frequency within each subinterval and uses the limit that converts repeated compounding into an exponential. As the subintervals become finer, the sum of forward rates multiplied by interval lengths converges to an integral, giving the bond price as the exponential of the negative integrated instantaneous forward rate. This is an intuitive derivation rather than a full treatment of rate conventions: the discrete formula must use discount factors consistent with its compounding assumptions, and the limiting argument presumes suitable regularity of the forward-rate curve.

Key ideas

  • A zero-coupon bond price can be built by multiplying discount factors across successive periods.
  • Increasing compounding frequency converts simple interval growth into exponential growth.
  • Refining the time intervals turns the sum of forward rates times interval lengths into an integral.
  • The continuous-time price is the exponential of the negative accumulated instantaneous forward rate.

Tags

Full text
# zero-coupon bond and forward rate


# zero-coupon bond and forward rate












My understanding, in a discrete-time setting, the relationship between a zero-coupon bond price and forward rates is:

$$p(t,T)=\frac{1}{\Pi_{j=1}^{T-1}f(t,j)}.$$

where $p(t,T)$ represents the price of the zero-coupon bond at time $t$ paying a sure dollar at $T$, and $f(t,S)$ is the forward rate between $t$ and $S$.

In the continuous time setting, my understanding is:

$$p(t,T)=e^{-\int_t^Tf(t,s)ds}.$$

How do we get to this?

Is it:

$$\lim_{k\rightarrow T-1}\frac{1}{\Pi_{j=1}^k f(t,j)}=p(t,T)=e^{-\int_t^Tf(t,s)ds}$$

My Question:

1) Is this correct?

2) If so, how do you prove the third equation?

3) If not correct, how do you get from the first to the second equation?

## Answer by Magic is in the chain (score 2, accepted)

https://quant.stackexchange.com/a/49884

I will try a simplified approach:

Let $P(t,T)$ represent the price at time t of a zero coupon that pays 1 at time T. If you divide the period between t and T into n sub-intervals, assume $F \left( t; t_{i-1}, t_{i}\right)$ represent the simple forward rate at time t for the interval between $i-1$ and $i$, where we assume the length of each interval is equal to $\Delta t$. Then you can write the price as follows:

$P(t,T)=\prod_{i=1}^{n}{\frac{1}{1+F \left( t; t_{i-1}, t_{i}\right) \Delta t }}$

Re-arrange to:

$P(t,T)\prod_{i=1}^{n}{\left(1+F \left( t; t_{i-1}, t_{i}\right) \Delta t \right)}=1$

Now let's say we increase the number of compounding in each interval (these are the sub-intervals of length $\Delta t$, and instead of simple compounding within each of these intervals, we are increasing the compounding frequency. m=1 reproduces the original):

$P(t,T)\prod_{i=1}^{n}{\left(1+\frac{F \left( t; t_{i-1}, t_{i}\right)}m \Delta t \right)^m}=1$

Now let m tend to infinity (continuous compounding within each sub-interval):

$\lim_{m \to \infty} \; P(t,T)\prod_{i=1}^{n}{\left(1+\frac{F \left( t; t_{i-1}, t_{i}\right)}m \Delta t \right)^m}=1$

And recall the basic identity: $e^x=\lim_{m \to \infty} \left(1+\frac{x}{m}\right)^m$

$P(t,T)\prod_{i=1}^{n}{e^{F \left( t; t_{i-1}, t_{i}\right) \Delta t}}=1$

Which then becomes, noting product of exponential is exponential of sum of exponents:

$P(t,T)e^ {\sum_{i=1}^{n}{F \left( t; t_{i-1}, t_{i}\right)\Delta t}}=1$

Now if you let n goes to infinity, then the sum will become integral, and you then just need to shift the exponential term to the right hand side, and you will get the desired formula after applying continuous time interpretation of the instantaneous forward rate:

$P(t,T)=e^{-\int_t^T{f(t,s)ds}}$

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.