Skip to content
Fixed Income and Interest-Rate Models / The LIBOR Market Model

BGM Dynamics and the LFM under Multiple Measures

Under its own TkT_k-forward measure QTkQ^{T_k}, the forward LIBOR rate FkF_k is driftless and lognormal — this is the content of the last lesson and gives Black's caplet formula exactly. But caplets are single-rate payoffs. Pricing a swaption requires the joint dynamics of Fk,Fk+1,…,Fk+nF_{k}, F_{k+1}, \ldots, F_{k+n} simultaneously, which means working under a single common measure where all rates are alive at once. When you shift FkF_k from its natural QTkQ^{T_k} measure to any other measure — the spot LIBOR measure, the terminal measure, or the QTjQ^{T_j} measure with j≠kj \neq k — the rate acquires a non-zero drift. This drift is a sum of pairwise covariance terms between FkF_k and the adjacent rates. The resulting system of SDEs under the spot LIBOR measure is the BGM model (Brace-Gatarek-Musiela 1997). Solving for these drifts via the change-of-numeraire toolkit is the central technical step of the LFM; the result is a tractable system for Monte Carlo simulation.

In this lesson

    1. The Multi-Measure Problem
    1. The Drift Under a Forward Measure (Proposition 6.3.1, IRM)
    1. The Spot LIBOR Measure Drift Recursion
    1. The Terminal Measure Dynamics
    1. Connecting Adjacent Forward Measures
    1. The Full LFM System Under QdQ^d
    1. Where This Goes Next

Before you start

The full lesson, with worked examples and exercise solutions, is for members.

Read the full lesson

New to SelfMadeQuant.com? See membership options or browse the Fixed Income and Interest-Rate Models syllabus.