This repository illustrates and provides code for the estimation procedure developed in the paper Jump Contagion among Stock Market Indices: Evidence from Option Markets written by Peter Boswijk, Roger Laeven, Andrei Lalu and Evgenii Vladimirov. The latest version of the paper is available via this link.
The illustration is based on simulated data from the bivariate option pricing models proposed in the paper. The simulated data are stored as a MATLAB .mat file in data. The state vectors (stored in the variables mY, mV and mLambda) are simulated using an Euler discretization and the option prices (variables mOptPriceIV1 and mOptPriceIV2) are computed using the COS method. For further details concerning the simulations, see Section 4 of the paper and Appendix C.4 of the Online Appendix.
The code is written in MATLAB version R2018b. The main.m file reads the simulated options data and estimates the parameters of the bivariate model proposed in the paper based on the partial-information C-GMM procedure. The estimation procedure minimizes the criterion function ‘./code/mSVhatHJ_crit_inst4.m’, which in turn involves the implied state procedure (function ‘./code/mSVhatHJ_ImpIntens.m’) and four numerical integrations of criterion functions based on the marginal states (function ‘./code/mSVhatHJ_int_inst4.m’). Given the estimated parameters, the standard errors are calculated using the function ‘./code/mSVhatHJ_std4.m’. The estimated parameters and the figure with implied intensities are displayed as the result of the optimization. The estimation results and the figure are provided in the accompanying pdf file for reproducibility.
The simulation and estimation procedures are described in Section 4 of the paper.