Deriving Forward Bond Price Dynamics Under the Vasicek Model
Summary
The document asks how to derive the dynamics of a forward price formed as the ratio of two zero-coupon bond prices in the Vasicek interest-rate model. Its proposed calculation applies Itô’s product rule to the numerator bond price and the reciprocal of the denominator bond price. The accepted explanation identifies the general two-variable Itô formula and shows that the product differential contains both first-order terms and a quadratic covariation term.
The bond volatilities are then related to the forward price’s diffusion under the forward measure. The excerpt does not complete the drift calculation or explain the measure change in detail, so it mainly clarifies the calculus step that prompted the question. Its result depends on correctly accounting for correlated bond-price movements; treating the ratio with an ordinary deterministic product rule would omit that contribution.
Key ideas
- The forward bond price is expressed as the product of one bond price and the reciprocal of another.
- Itô’s product rule includes a quadratic covariation term for the stochastic factors.
- The reciprocal bond price contributes a differential that must be included in the ratio calculation.
- The excerpt focuses on the stochastic calculus step and leaves the full drift derivation implicit.
Tags
Full text
# Forward pricing using Vasicek model
# Forward pricing using Vasicek model
Question:
Vasicek interest rate model: $$dr_t = α(θ−r_t)dt + σdW_t$$
Price at time t of a 0-coupon bond maturing at T is given by: $$dp(t,T) = α_{t,T} . p(t,T)d_t + β_{t,T} . p(t,T)dW_t$$ $$βt,T = −σB(t,T).$$
T-forward price
$$ F_t = \frac{P_{t,S}}{P_{t,T}} $$
Show that the dynamics of the T-forward $F_t$ price with respect to the T-forward measure in the Vasicek model is given by
$$dF_t = σ (B(t,T)−B(t,S))F_tdW_t$$
Solution provided: $$ d(\frac{P_{t,S}}{P_{t,T}}) = \frac{dP_{t,S}}{P_{t,T}} + P_{t,S} d(\frac{1}{P_{t,T}}) + (dP_{t,S}) d( \frac{1}{P_{t,T}} ) $$ $$= \frac{dP_{t,S}}{P_{t,T}} −P_{t,S}\frac{dP_{t,T}}{P^2_{t,T}} + (···)dt$$ $$= (β_{t,S} −β_{t,T}) \frac{P_{t,S}}{P_t}dW_t + (···)d_t. $$
Can anyone explain how to obtain the first line of the solution please. Initially I thought you could just use the product rule here i.e $u (dv/dx) + v (dv/dy)$ but if I use that I don't get the first term $\frac{dP_{t,S}}{P_{t,T}}$
## Answer by NSZ (score 0, accepted)
https://quant.stackexchange.com/a/33382
For the first line of the solution it's just Ito's lemma. \begin{align} d f(X_t,Y_t,t)&= \frac{\partial }{\partial t} f(X_t,Y_t,t)dt+\frac{\partial }{\partial X_t} f(X_t,Y_t,t) dX_t +\frac{\partial }{\partial Y_t} f(X_t,Y_t,t) dY_t \\ &+\frac12 \frac{\partial^2 }{\partial X_t^2} f(X_t,Y_t,t) d\langle X_t,X_t\rangle+\frac12 \frac{\partial^2 }{\partial Y_t^2} f(X_t,Y_t,t) d\langle Y_t,Y_t\rangle\\ &+ \frac{\partial^2 }{\partial X_t\partial Y_t} f(X_t,Y_t,t) d\langle X_t,Y_t\rangle \end{align} In your case you have $X_t=P_{t,S}$ and $Y_t=\frac{1}{P_{t,T}}$ so that the first derivative of $f(X_t,Y_t,t)=X_tY_t$ in $t$ and second derivatives in $X_t$ or $Y_t$ are all 0. So you are left with the first line of the solution. \begin{align}dX_tY_t&=Y_t dX_t + X_t dY_t +d\langle X_t,Y_t \rangle\\ &= \frac{1}{P_{t,T}}dP_{t,S}+P_{t,S}d\left(\frac{1}{P_{t,T}}\right)+d\langle P_{t,S},\frac{1}{P_{t,T}}\rangle \end{align} By the way I suggest you go back to your stochastic calculus class notes because Ito's lemma is a must in quantitative finance.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.