G2++

Overview

G2++ is the two-factor additive Gaussian model: the short rate is the sum of two correlated zero-mean Ornstein-Uhlenbeck factors plus a deterministic shift that fits the observed discount curve exactly. The second factor decorrelates rates at different maturities, producing curve twists that no one-factor model can generate — essential for products sensitive to the shape of the curve. Bonds, bond options, and caps/floors remain closed-form; simulation uses exact joint-Gaussian factor transitions. G2++ is equivalent to the two-factor Hull-White model.

Dynamics

Under the risk-neutral measure,

\[r(t) = x(t) + y(t) + \varphi(t)\]
\[dx(t) = -a\, x(t)\, dt + \sigma\, dW_1(t), \qquad dy(t) = -b\, y(t)\, dt + \eta\, dW_2(t), \qquad d\langle W_1, W_2\rangle_t = \rho\, dt\]

with \(x(0) = y(0) = 0\). Fitting the market curve pins the shift:

\[\varphi(t) = f^M(0, t) + \frac{\sigma^2}{2a^2}\left(1 - e^{-at}\right)^2 + \frac{\eta^2}{2b^2}\left(1 - e^{-bt}\right)^2 + \frac{\rho\sigma\eta}{ab}\left(1 - e^{-at}\right)\left(1 - e^{-bt}\right)\]

Bond pricing

Define, for \(\tau = T - t\),

\[V(\tau) = \frac{\sigma^2}{a^2}\left[\tau + \frac{2}{a}e^{-a\tau} - \frac{1}{2a}e^{-2a\tau} - \frac{3}{2a}\right] + \frac{\eta^2}{b^2}\left[\tau + \frac{2}{b}e^{-b\tau} - \frac{1}{2b}e^{-2b\tau} - \frac{3}{2b}\right] + \frac{2\rho\sigma\eta}{ab}\left[\tau + \frac{e^{-a\tau} - 1}{a} + \frac{e^{-b\tau} - 1}{b} - \frac{e^{-(a+b)\tau} - 1}{a + b}\right]\]

Then, with \(B_z(\tau) = \frac{1}{z}\left(1 - e^{-z\tau}\right)\) for \(z \in \{a, b\}\):

\[P(t, T) = \frac{P^M(0, T)}{P^M(0, t)}\, \exp\!\left(\frac{1}{2}\left[V(T - t) - V(T) + V(t)\right] - B_a(T - t)\, x(t) - B_b(T - t)\, y(t)\right)\]

At \(t = 0\) the input curve is reproduced exactly.

Bond options

The log-price of the bond at expiry is Gaussian with total variance

\[\Sigma^2 = \frac{\sigma^2}{2a^3}\left(1 - e^{-a(S-T)}\right)^2\left(1 - e^{-2aT}\right) + \frac{\eta^2}{2b^3}\left(1 - e^{-b(S-T)}\right)^2\left(1 - e^{-2bT}\right) + \frac{2\rho\sigma\eta}{ab(a + b)}\left(1 - e^{-a(S-T)}\right)\left(1 - e^{-b(S-T)}\right)\left(1 - e^{-(a+b)T}\right)\]

and the call price is

\[\mathrm{ZBC}(0, T, S, K) = P(0, S)\,\Phi(h) - K P(0, T)\,\Phi(h - \Sigma), \qquad h = \frac{1}{\Sigma}\ln\frac{P(0, S)}{P(0, T)\, K} + \frac{\Sigma}{2}\]

Caplets and floorlets price by the bond-put replication, as for Hull-White.

Simulation

The pair \((x, y)\) has jointly Gaussian transitions with

\[\operatorname{Var}[x] = \frac{\sigma^2}{2a}\left(1 - e^{-2a\Delta}\right), \qquad \operatorname{Var}[y] = \frac{\eta^2}{2b}\left(1 - e^{-2b\Delta}\right), \qquad \operatorname{Cov}[x, y] = \frac{\rho\sigma\eta}{a + b}\left(1 - e^{-(a+b)\Delta}\right)\]

sampled exactly per step; the shift \(\varphi\) is added at each grid point.

Parameters

Parameter

Meaning

Constraint

curve

market discount curve, fitted exactly

—

\(a, b\)

factor mean-reversion speeds

\(a, b > 0\)

\(\sigma, \eta\)

factor volatilities

\(\sigma, \eta > 0\)

\(\rho\)

factor correlation

\(-1 \leq \rho \leq 1\)

A typical calibrated shape has one slow and one fast factor with strongly negative \(\rho\).

Usage

from interest_rate_models import get_model

model = get_model("g2++", curve=curve, a=0.5, b=0.05, sigma=0.008, eta=0.006, rho=-0.7)
model.bond_price(0.0, 5.0)                 # == curve.discount(5.0)
model.bond_option(1.0, 5.0, strike=0.85)
model.cap_price(strike=0.035, maturity=3.0, tenor=0.5)
times, r, x, y = model.simulate(3.0, 300, 10_000, seed=42, return_factors=True)

References

  • Brigo, D., Mercurio, F. (2006). Interest Rate Models — Theory and Practice, 2nd ed., Springer, ch. 4.

  • Hull, J., White, A. (1994). “Numerical procedures for implementing term structure models II: Two-factor models”, Journal of Derivatives, 2(2), 37-48.