Estimation of Sentiment Effects in Financial Markets: A Simulated Method of Moments Approach

摘要: We take the model of Alfarano et al. (J Econ Dyn Control 32:101–136, 2008) as a prototype agent-based that allows reproducing main stylized facts financial returns. The does so by combining fundamental news driven Brownian motion with minimalistic mechanism for generating boundedly rational sentiment dynamics. Since we can approximate herding component among an ensemble agents in aggregate Langevin equation, either simulate full at micro level, or via law motion. In simplest version our model, only three parameters need to be estimated. explore performance simulated method moments (SMM) approach estimation this model. As it turns out, sensible parameter estimates obtained if one first provides rough “mapping” objective function extensive grid search. Due high correlations estimated parameters, uninformed choices will often lead convergence any large number local minima. also find efficiency SMM is relatively insensitive size sample over range sizes and converge their GMM counterparts sizes. believe feature due limited available univariate asset pricing models, sensitivity present specification estimator could carry many related models markets well similar diffusion processes mathematical finance.