Bayesian estimation of geometric mean-reversion
Economic theory and lots of empirical results indicate that commodity prices and exchange rates are more believed to revert to some level associated with marginal production costs than geometric Brownian motion and thus mean-reversion process is more popular in econometrics. In the literature there are several ways to model the mean-reversion process and among them the geometric mean-reversionprocess seems to be the most important. Unfortunately, to our best knowledge, no results are seen in the literature concerning how to estimate the parameters in this model and on the contrary, the same problems on arithmetic mean-reversion model are well studied. By the time-discretization approach, this paper derives a stochastic difference equation from the geometric mean-reversion process and then a nonlinear regression model is established. By this way, the paper obtains the distribution and estimation for each parameter by Bayesian inference. The results appear interesting from experiments by Monte Carlo simulation with artificial data.
Geometric mean-reversion Bayesian inference Monte Carlo simulation
Jinqiang Yang Zhaojun Yang
School of Economics and Trade,Hunan University,Changsha 410079,P R China
国际会议
International Symposium on Financial Engineering and Risk Management(2008年金融工程与风险管理)
上海
英文
113-117
2008-06-01（万方平台首次上网日期，不代表论文的发表时间）