CONVERTIBLE BOND PRICING BASED ON MONTE CARLO SIMULATION

Convertible bonds, as a novel financing instrument, possess dual characteristics of both conventional bonds and options due to the inclusion of clauses with American option-like features such as conversion, redemption, putback, and conversion price adjustment. Therefore, the pricing of convertible bonds is of considerable importance. This study has gathered a substantial amount of relevant industry data. For accuracy considerations, this paper employs the Random Forest algorithm and LightGBM algorithm to predict the probability and price of triggering downward adjustment clauses, achieving a classification accuracy of 60.162% and a regression goodness of fit of 0.757. Subsequently, Monte Carlo simulation was utilized for convertible bond pricing prediction, resulting in the calculation of MAPE values for two convertible bonds of 5.15% and 9.6%, respectively, indicating the model's high accuracy. Finally, a financial analysis of the results was conducted to provide investment recommendations. Leveraging this research outcome, convertible bonds can be better applied and promoted, and the development of the convertible bond market can increase the proportion of debt financing in the capital market, enabling enterprises to flexibly adjust their capital structure. Its development also provides investors with more investment options.