TQBTHEQUANTBATEMAN
TQB/ learn/ rates/ lmmEN · DARK
IRfront-officemodel

LMM

Model a family of market forward rates under linked measures.

Reviewed 2026-08-10TheQuantBateman ResearchReading note
01Intuition

Start from the observable dynamics.

Separate state variables, dynamics and valuation measure before looking at a calibration. Model a family of market forward rates under linked measures. A fitted surface is evidence about today's prices, not proof of tomorrow's dynamics.

ONE-LINE DEFINITION

Model a family of market forward rates under linked measures.

02Mathematics

Write the state process and pricing map.

dLi(t)=μiQ(t)dt+Li(t)σi(t)dWtQdL_i(t)=\mu_i^{\mathbb{Q}}(t)dt+L_i(t)\sigma_i(t)dW_t^{\mathbb{Q}}
Notation and units

Decimal rates and volatilities, year-fraction time and continuous compounding unless stated otherwise.

03Assumptions

Identify what the model cannot represent.

01

The state dynamics and valuation measure are stated independently of the calibration instruments.

02

Parameters are treated as deterministic over the pricing run unless the model says otherwise.

03

A calibration fit does not validate out-of-sample dynamics or hedge performance.

“An unstated convention is a future reconciliation break.”— THEQUANTBATEMAN
04Market use

Separate calibration fit from dynamics.

LMM is used to translate liquid IR calibration instruments into prices and sensitivities. Residuals, parameter stability and hedge behaviour must be reviewed together.

Intuition→Mathematics→Implementation→Desk risk
05Desk view
FRONT OFFICE VIEW

Challenge the hedge outside the fitted slice.

A good IR calibration explains today's instruments; the hedge reveals whether the assumed dynamics survive tomorrow's move.

Ask Bateman about this topic →
06Related

Compare the adjacent model family.