Longstaff–Schwartz (LSM) Algorithm
Quant ResearchLevel: Intermediate
Longstaff–Schwartz (LSM) Algorithm
Hints
- Identify the financial object, available information and source of risk.
- Separate theory from empirical validation with costs and constraints.
Answer
Simulate pricing paths forward, then work backward through permitted exercise dates. Regress discounted future cash flows under the current policy on basis functions of the current state to estimate continuation; exercise when immediate payoff exceeds that estimate. Evaluate the fitted policy on independent paths to avoid in-sample look-ahead effects. An admissible out-of-sample policy gives a lower bound in expectation, not a guaranteed bound for each noisy estimate. Check basis sensitivity, sample size, discount timing and exercise-grid bias.
Source and licence: Quantitative Finance: From Mathematical Foundations to Trading Desk Reality — zcoderz · MIT · answer: Quantoros editorial