Dynamic Multiplier CPPI Strategy with Wavelets and Neural-Fuzzy Systems

Abstract

Constant Proportional Portfolio Insurance (CPPI) aims to maximize the performance of the portfolio by protecting a determined base value without using any derivative instruments, and by determining the amounts to be invested in risky and risk-free assets with calculations using the risk multiplier and buffer value. Artificial neural networks (ANNs) are mathematical models that are successfully used in studies such as pattern recognition, function estimation, finding the most appropriate value and classifying data by imitating neural networks in the human brain. This study aims to use the advantages of both the decomposition model (Wavelet Transform) and machine learning model (ANN) to predict the future values of stock indices to decide which risk multiplier to use. The dynamic multiplier CPPI yielded better returns than the classic CPPI in all 5 stock market indices analyzed, and both strategies successfully implemented the previously targeted 95% capital protection. It has been observed that predicting future prices of indices using Artificial Intelligence methods and the performance of the dynamic multiplier CPPI strategy applied based on these predictions is more successful than the conventional CPPI strategy with constant multiplier.