In this Repository is the code I have written for my Masterthesis in Economics. It is the code for a Baysian Synthetic Likelihood estimation (BSL), which is used to find parameters of small scaled Agent-based models (ABM) for financal markets. The Goal is to find parameters for the ABM which replicated key features of the S&P 500 for monthly iterations (from January 1871 to January 2021). It did not work proparly. The used Kolmogorov-Smirnov-test dismissed the simularity between the simulated time series and the empirical data.
This is the first try to used a BSL on financal ABMs, I am aware of. Futher reaserch maybe can get better results.