Abstract:
This paper extends the previous work on parametric mixed copula models and implements a nonparametric approach, i.e. a local maximum likelihood estimationprocedure, to estimate the time-varying correlation between returns of the world major market indexes such as S&P, NIKKIE, FTSE and DAX. We check the robustness of our model using the Monte-Carlo simulation with different sample sizes and our modeling strength is confirmed.
