Forward Rate Agreement¶
Payer/Receiver Forward Rate Agreement¶
where
Forward Rate from FRA¶
If we have a traded FRA, meaning \({\bf\text{FRA}}(t, T_{k-1}, T_k, N, K)=0\) or \(l_k(t)=K\), then we can extract the forward rate from this market data:
Libor Rate l_k(t) is a T_k-Martingale¶
With tenor \(\tau_k=T_k-T_{k-1}\)
is a \(\mathbb{T_k}\)-martingale.
Proof
FRA Valuation via Change of Numeraire¶
The fair value \(K\), which makes \(V(t)=0\), is given by
Proof
Since \(l_k(t)=\frac{1}{\tau_k}(\frac{P(t,T_{k-1})}{P(t,T_k)}-1)\), we have
Therefore,
Exercises¶
Exercise 1. A payer FRA is entered at time \(t = 0\) with reset date \(T_0 = 0.5\), payment date \(T_1 = 1.0\), notional \(N = 1{,}000{,}000\), and fixed rate \(K = 3\%\). If \(P(0, 0.5) = 0.985\) and \(P(0, 1.0) = 0.968\), compute the forward rate \(l_1(0)\) and the FRA value at inception.
Solution to Exercise 1
Forward rate. With \(T_0 = 0.5\), \(T_1 = 1.0\), and \(\tau_1 = T_1 - T_0 = 0.5\):
The forward rate is \(l_1(0) \approx 3.5124\%\).
FRA value at inception.
The payer FRA has a positive initial value of approximately \(\$2{,}480\). This is positive because the forward rate (\(3.51\%\)) exceeds the contractual fixed rate (\(3\%\)), so the payer (who receives float and pays fixed) expects a net positive cash flow.
Exercise 2. Show that the fair fixed rate \(K\) of a FRA (the rate making its initial value zero) equals the forward rate \(l_k(t)\). Starting from \(\text{FRA}(t, T_{k-1}, T_k, N, K) = N\tau_k(l_k(t) - K)P(t, T_k) = 0\), derive \(K = l_k(t)\).
Solution to Exercise 2
Starting from the FRA valuation formula:
Setting the value to zero:
Since \(N > 0\), \(\tau_k > 0\), and \(P(t, T_k) > 0\), the equation reduces to:
The fair fixed rate is:
This result is intuitive: the fair fixed rate in a FRA equals the market's implied forward rate for the same period. If \(K > l_k(t)\), the fixed payer would overpay relative to the forward, giving the FRA negative value to the payer. If \(K < l_k(t)\), the payer underpays, and the FRA has positive value. Equilibrium requires \(K = l_k(t)\).
Exercise 3. Prove that the simply compounded forward rate \(l_k(t) = \frac{1}{\tau_k}\left(\frac{P(t,T_{k-1})}{P(t,T_k)} - 1\right)\) is a martingale under the \(T_k\)-forward measure. Identify the numeraire and explain the economic intuition.
Solution to Exercise 3
We need to show that \(l_k(t)\) is a martingale under the \(T_k\)-forward measure \(\mathbb{Q}^{T_k}\), whose numeraire is the zero-coupon bond \(P(t, T_k)\).
Proof. The forward rate can be written as:
The numerator \(P(t, T_{k-1}) - P(t, T_k)\) is the price of a portfolio long a \(T_{k-1}\)-bond and short a \(T_k\)-bond, which is a traded asset. Dividing any traded asset price by the numeraire \(P(t, T_k)\) yields a martingale under \(\mathbb{Q}^{T_k}\) (by the fundamental theorem of asset pricing). Since \(l_k(t) = \frac{1}{\tau_k} \cdot \frac{P(t,T_{k-1}) - P(t,T_k)}{P(t,T_k)}\) is a constant multiple of such a ratio, it is also a \(\mathbb{Q}^{T_k}\)-martingale.
Formally, for \(s < t\):
Economic intuition. The \(T_k\)-forward measure corresponds to using the \(T_k\)-maturity zero-coupon bond as numeraire. Under this measure, all asset prices expressed in units of this bond are martingales. The forward rate \(l_k(t)\) determines the FRA payoff at \(T_k\), and since the FRA payoff is settled at \(T_k\) (the maturity of the numeraire bond), the natural pricing measure for this payoff is precisely \(\mathbb{Q}^{T_k}\). The forward rate being a martingale under this measure means that its current value is the best (unbiased) predictor of its future value---no drift adjustment is needed when pricing FRA-related derivatives under \(\mathbb{Q}^{T_k}\).
Exercise 4. At the reset date \(T_{k-1}\), the payer FRA payoff is \(N\tau_k(l_k(T_{k-1}) - K)\) paid at \(T_k\), or equivalently \(\frac{N\tau_k(l_k(T_{k-1}) - K)}{1 + \tau_k l_k(T_{k-1})}\) paid at \(T_{k-1}\). Verify that these two expressions are consistent by discounting the \(T_k\) payoff back to \(T_{k-1}\) using \(P(T_{k-1}, T_k) = \frac{1}{1 + \tau_k l_k(T_{k-1})}\).
Solution to Exercise 4
At the reset date \(T_{k-1}\), the floating rate \(l_k(T_{k-1})\) is observed. The payer FRA payoff at \(T_k\) is:
To express this as a payment at \(T_{k-1}\), we discount using \(P(T_{k-1}, T_k)\). At \(T_{k-1}\), the simply compounded rate for the period \([T_{k-1}, T_k]\) is \(l_k(T_{k-1})\), so:
The present value at \(T_{k-1}\) of the \(T_k\) payoff is:
This confirms that the two expressions are consistent: the payment \(N\tau_k(l_k(T_{k-1}) - K)\) at \(T_k\) has the same economic value as \(\frac{N\tau_k(l_k(T_{k-1}) - K)}{1 + \tau_k l_k(T_{k-1})}\) at \(T_{k-1}\).
We can also verify using the FRA valuation formula at \(t = T_{k-1}\):
This is the market-standard "discounted FRA" settlement: rather than waiting until \(T_k\), the FRA is settled at \(T_{k-1}\) by discounting the \(T_k\) payoff at the just-observed floating rate.
Exercise 5. Given a set of traded FRA rates \(K_1 = 2.5\%\), \(K_2 = 2.8\%\), \(K_3 = 3.1\%\) for consecutive 6-month periods starting at \(T = 0.5, 1.0, 1.5\) respectively, extract the corresponding continuously compounded forward rates \(R(t, T_{k-1}, T_k)\) using the conversion \(R = \frac{1}{\tau}\ln(1 + \tau K)\).
Solution to Exercise 5
The conversion from simply compounded FRA rate \(K\) to continuously compounded forward rate is:
with \(\tau = 0.5\) (6-month periods).
Period 1: \([0.5, 1.0]\) with \(K_1 = 2.5\% = 0.025\):
So \(R_1 \approx 2.4845\%\).
Period 2: \([1.0, 1.5]\) with \(K_2 = 2.8\% = 0.028\):
So \(R_2 \approx 2.7806\%\).
Period 3: \([1.5, 2.0]\) with \(K_3 = 3.1\% = 0.031\):
So \(R_3 \approx 3.0762\%\).
In each case, the continuously compounded rate is slightly lower than the simply compounded rate. This is because continuous compounding generates more frequent interest accrual, so a lower rate is needed to produce the same discount factor over the period: \(e^{-R\tau} = (1 + K\tau)^{-1}\). The difference is small for short tenors and low rates but grows with both.
Exercise 6. A portfolio contains a payer FRA and a receiver FRA on the same reset and payment dates but with different fixed rates \(K_1\) and \(K_2\) (\(K_1 < K_2\)). Show that the portfolio value is \(N\tau_k(K_2 - K_1)P(t, T_k)\), independent of the forward rate. Interpret this as a deterministic cash flow and explain why it can be perfectly hedged with a zero-coupon bond.
Solution to Exercise 6
The portfolio consists of a payer FRA with fixed rate \(K_1\) and a receiver FRA with fixed rate \(K_2\), both on the same dates and notional.
Payer FRA value:
Receiver FRA value:
Portfolio value:
The forward rate \(l_k(t)\) cancels completely. The portfolio value depends only on the fixed rates \(K_1\), \(K_2\) and the discount factor \(P(t, T_k)\).
Interpretation. The portfolio generates a deterministic cash flow of \(N\tau_k(K_2 - K_1)\) at time \(T_k\), regardless of where the floating rate fixes. Since \(K_2 > K_1\), this cash flow is strictly positive.
Hedging. This deterministic payoff is equivalent to holding \(N\tau_k(K_2 - K_1)\) units of a zero-coupon bond maturing at \(T_k\). The portfolio can therefore be perfectly replicated (and hedged) by purchasing this quantity of the \(T_k\)-maturity zero-coupon bond at cost \(N\tau_k(K_2 - K_1)P(t, T_k)\), which is exactly the portfolio value. No exposure to future floating rates remains---all interest rate risk has been eliminated by the offsetting floating legs.