基于JLS模型的股市崩盘点研究

发布时间：2018-10-31 12:54

【摘要】：泡沫以股票为载体出现在资本市场之中,而股票市场成为泡沫积累的主要场所,导致泡沫经济与金融危机的重要因素又是股市的泡沫。所以,对股票以及股市泡沫进行深入的研究十分关键。为有效的对股市泡沫进行调控,相关理论的提出与完善是迫切需求的,只是因为中国目前的股市相对世界其他成熟股市还存在很多缺点与不稳定因素,更何况即便是成熟的国际市场也是颇有瑕疵。本文首先在引言中回顾了历史上几次重要的崩盘事件：1637年“荷兰郁金香事件”；1929年“黑色星期二”；1987年“华尔街历史上最坏的日子”。通过鲜明的案例说明崩盘对于整个金融和经济系统的影响和危害性。然后介绍了进行股票市场崩盘研究的时代背景,结合历史事件给出崩盘的研究意义。其次在前人的研究基础上,介绍了通过不同角度看崩盘问题的各种模型。国内外学者将股市崩盘现象的主要理论模型分为流动性缺乏模型,多均衡与太阳黑子模型,块状信息聚集模型,羊群效应变形模型和泡沫破裂模型五类。在五类模型对股市崩盘现象的研究中,JLS模型是最近几年比较认可的模型。然后给出JLS模型适合的条件及其应用前提,给出JLS模型中应用的对数周期幂函数(LPPL)的推导过程,对数周期幂函数有如下形式： JLS模型的未知参数估计是一大重点,本文中我们提出了一种新的参数估计算法,该算法通过改进JLS模型中所应用的LPPL方程形式,将线性参数由3个转变为4个,从而减少了非线性参数的个数,降低了非线性拟合的参数空间。因此,大大降低了拟合过程的复杂性,使其稳定性大大提高。其中,泡沫最终崩溃的时点即崩盘点t。作为最关键的非线性参数,我们将另外两个非线性参数看做是t。的函数,从而进一步降低了搜索过程的复杂性。我们所提出的新的算法与原来的算法相比,“启发式搜素”是不需要的,可以仅仅采取严格的局部搜索算法,从而拟合过程的效率大大增加。最后通过中国股市以及黄金市场的数据,用该模型进行了实证研究。结合中国股票市场国情,给出中国股市泡沫崩盘的原因,指出深人研究崩盘现象的形成机制在我国显得尤为重要,开展这一领域的深入研究必将推动现代金融理论的发展,并给政府金融监管带来重要的启示。
[Abstract]:The bubble appears in the capital market with the stock market as the carrier and the stock market becomes the main place where the bubble accumulates. The important factor leading to the bubble economy and financial crisis is the stock market bubble. Therefore, the stock and stock market bubble in-depth study is very important. In order to effectively regulate the stock market bubble, it is urgent to put forward and improve the relevant theories. It is only because China's current stock market has many shortcomings and unstable factors relative to other mature stock markets in the world. Moreover, even mature international markets are flawed. In the introduction, this paper reviews several important events in history: the tulip in the Netherlands in 1637, Black Tuesday in 1929, and the worst day in Wall Street's history in 1987. The impact and harmfulness of the crash on the entire financial and economic system are illustrated by clear cases. Then it introduces the background of stock market crash research and gives the significance of stock market crash research combined with historical events. Secondly, on the basis of previous studies, this paper introduces various models of crash problem from different angles. Scholars at home and abroad divide the main theoretical models of stock market crash into five categories: lack of liquidity model, multi-equilibrium and sunspot model, block information aggregation model, herding deformation model and bubble burst model. In the study of stock market collapse in five kinds of models, JLS model is a more acceptable model in recent years. Then, the suitable conditions and application premises of JLS model are given, and the derivation process of logarithmic periodic power function (LPPL) applied in JLS model is given. The logarithmic periodic power function has the following forms: the estimation of unknown parameters of JLS model is an important point. In this paper, we propose a new parameter estimation algorithm. By improving the form of LPPL equation used in JLS model, the linear parameters are changed from three to four, thus reducing the number of nonlinear parameters. The parameter space of nonlinear fitting is reduced. Therefore, the complexity of the fitting process is greatly reduced and the stability is greatly improved. The point at which the bubble finally collapses is the point where the bubble collapses. As the most critical nonlinear parameter, we regard the other two nonlinear parameters as t. This further reduces the complexity of the search process. Compared with the original algorithm, the new algorithm proposed by us does not need the heuristic search element, it can only adopt a strict local search algorithm, so the efficiency of the fitting process is greatly increased. Finally, through the data of Chinese stock market and gold market, we use this model to carry on the empirical research. According to the situation of China's stock market, this paper gives the reasons for the bubble collapse of China's stock market, and points out that it is particularly important to study the formation mechanism of the phenomenon of crash in China, and to carry out in-depth research in this field will certainly promote the development of modern financial theory. And bring important enlightenment to government financial supervision.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：O212.1;F832.51

【相似文献】