Path integral for equities: Dynamic correlation and empirical analysis

This paper develops a model to describe the unequal time correlation between rate of returns of different stocks. A non-trivial fourth order derivative Lagrangian is defined to provide an unequal time propagator, which can be fitted to the market data. A calibration algorithm is designed to find the empirical parameters for this model and different de-noising methods are used to capture the signals concealed in the rate of return. The detailed results of this Gaussian model show that the different stocks can have strong correlation and the empirical unequal time correlator can be described by the model’s propagator. This preliminary study provides a novel model for the correlator of different instruments at different times. ► Novel model for unequal time correlator between different equities. ► Fourth order Lagrangian to provide an unequal time propagator. ► Gaussian kernel smoothing. ► Fitting R2 greater than 0.85.