【作者】 李铮；

【导师】 熊熊；

【作者基本信息】 天津大学 ， 管理科学与工程， 2020， 硕士

【副题名】基于对数周期幂律模型

【摘要】 资产价格泡沫问题不仅是金融实业界最关心的话题,也是金融学术界最重要的研究课题之一。1636年荷兰的郁金香泡沫、1716年的密西西比股市泡沫、19世纪40年代英国的铁路狂热泡沫、1929年的美国由股市崩盘而引起的大萧条、1985年日本房地产和股市抛锚、1987年美股的“黑色星期一”、20世纪90年代的互联网泡沫、2007年的美国的房地产泡沫与次贷危机,历史上每一次资产价格经过狂热上涨后必然会积累泡沫,而每一次泡沫的破裂都会伴随着居民财富的大量缩水、企业的破产、同时伴随着经济的衰退。随着我国金融改革的不断推进和金融开放的不断扩大,在经济和金融发展的过程中,必然会出现资产价格偏离、甚至远远背离其基本价值的情况。因此是否能够正确的认识泡沫、研究泡沫,并由此而发现应对泡沫的方法就显得尤为重要。本文在行为金融理论中投资者行为偏差的理论基础上,基于目前广泛应用的对数周期幂律(Log-Periodic Power Law,LPPL)理论对我国金融市场中铁矿石期货、苹果期货、焦炭期货以及股票市场指数和比特币价格进行了泡沫累积过程的实证分析;基于对数周期幂律理论对我国股票市场创业板指数以及中国资本市场全指数进行了预测分析。研究发现,交易限制越少、交易连贯性越好并且没有政策反复扰动的资产更适合LPPL模型;投资者结构稳定情况下更适宜使用LPPL模型;并不是所有大幅上涨与大幅下跌都适用于LPPL模型;泡沫形成后崩溃并非唯一结果,通过改变基本面可以起到“坐实”泡沫效果。基于本文实证和预测结果,本文建议监管者完善做空机制、增加政策预期管理、完善退市机制、增加对冲工具,建议投资者建立理性的投资思维、合理化资金管理。更多还原

【Abstract】 The asset price bubble problem is not only the most concerned topic in the financial industry,but also one of the most important research topics in the financial academic community.1636 Dutch tulip bubble,1716 Mississippi stock market bubble,British railway crazy in 1940 s,Great Depression caused by the stock market crash in 1929,Japan real estate and stock market breakdown in 1985,US stocks markets crashes in 1987’s“Black Monday”,the Internet bubble in 1990 s,the US real estate bubble in 2007 and the Credit Crisis.Each time asset price in history has accumulated a bubble after a feverish rise,and each timethe asset price plummeted,which leadedthe massive shrinkage of wealth,the bankruptcy of enterprises,and the economic recession.With the continuous advancement of China’s financial reform and the continuous expansion of financial openness,in the process of economic and financial development,there will inevitably be a situation in which asset prices deviate or even deviate seriously from their basic values.Therefore,it is particularly important to be able to correctly understand the bubble and study the bubble,and thus discover the way to deal with the bubble.Based on the theory of investor behavior bias in behavioral finance theory,this paper is based on the widely used Log-Periodic Power Law(LPPL)theory for iron ore futures,apple futures,coke futures and stocks market indices in China’s financial market.The market indices and bitcoin price are empirically analyzed by the bubble accumulation process andbased on the LPPL theory,the stock market indices in China stock market and the index of China’s capital market are forecasted and analyzed.Thispaperfinds that the fewer transaction restrictions,the better the transaction coherence and the lack of policy repetitive assets are more suitable for the LPPL model;the LPPL model is more suitable for assets with stable investors structure;not all bubble crashes apply to the LPPL model.The collapse of the bubble is not the only result.After the application of dozens of major assets around the world,it is found that changing the fundamentals,you can achieve a "sit-on" bubble effect.Based on the empirical and predictive results,this paper suggests that regulators improve short-selling mechanism,increase policy expectation management,improve delisting mechanism,increase hedging instruments,and suggest investors to establish rational investment behaviors and rationalize fund management.更多还原