American Put Option pricing using Least squares Monte Carlo method under Bakshi, Cao and Chen Model Framework (1997) and comparison to alternative regression techniques in Monte Carlo

TL;DR

This paper compares least-squares Monte Carlo with neural networks and gradient boosting for pricing American put options under a complex stochastic model, showing how alternative regressions perform in calibration and out-of-sample testing.

Contribution

It introduces and evaluates neural networks and gradient boosting as alternative regression methods in American option pricing under a comprehensive stochastic model.

Findings

ANN and GBM outperform LSM in certain market conditions

Alternative methods improve pricing accuracy and calibration

Neural networks capture complex option dynamics better

Abstract

This paper explores alternative regression techniques in pricing American put options and compares to the least-squares method (LSM) in Monte Carlo implemented by Longstaff-Schwartz, 2001 which uses least squares to estimate the conditional expected payoff to the option holder from continuation. The pricing is done under general model framework of Bakshi, Cao and Chen 1997 which incorporates, stochastic volatility, stochastic interest rate and jumps. Alternative regression techniques used are Artificial Neural Network (ANN) and Gradient Boosted Machine (GBM) Trees. Model calibration is done on American put options on SPY using these three techniques and results are compared on out of sample data.