泡沫变化过程的动态贝叶斯模型研究

Dynamic Bayesian model for evolution of bubbles

构建了能够刻画资产价格泡沫形成、增长到破灭过程的统计模型.通过估计模型里的时变系数,可以确定泡沫产生的时点和持续时间.该模型也同时刻画了泡沫破灭的概率随着资产价格偏离其基本价值越来越远而变得越来越大的特征.为了对模型进行估计,提出了贝叶斯递归算法和近似算法,并以北京房地产市场为例对模型的应用进行了探讨.

This paper constructs a Bayesian statistical model to describe the formation, growth and burst of bubbles of asset prices. This model shows that, the probability for the bubbles to burst becomes larger and lar- ger as the price of a financial asset deviates from its fundamental value further and further. Based on the time- varying parameter estimates implied by the model, the time for the bubbles to emerge and the duration for the bubbles to accumulate can be determined. A Bayesian method with approximation is developed to estimate the model and is applied to Beijing＇s real estate market as an example