Module IX — Direct Mean-Reverting Portfolio Optimisation

Status: outline. Lecture body not authored.

Box–Tiao predictability, portmanteau objectives, crossing statistics, variance constraints, generalised eigenvalue problems, sparse MRP.

Two competing discovery paths:

Path A  Cointegration → candidate space → MR portfolio
Path B  Direct MR optimisation → candidate spread → statistical validation

Unique process — Econometric-First vs Objective-First

Both remain scientific competitors. Neither is the default winner.