Research On Risk Management In Behavior Finance With Crowd Psychology Embedded

Uncertainty is rooted in complexity of the objective world and the contradiction between human being's bounded rationality and irrationality. It might arise from the objective world, the human's subjective world or the relation between the two worlds. In the case of the security market, the so-called uncertainty from the human's subjective world is investors' bounded rationality and irrationality in investment decisions. Current financial theories, including behavioral financial theories, suggest that investors are of complete rationality or bounded rationality, and deny the possibility of irrational investment behavior. In practice, current financial theories cannot give a reasonable explanation to the violent fluctuations of security price involving collective behavior, let alone to say how to conduct risk management. On the other hand, crowd psychology is precisely the discipline attempting to interpret human's irrational behaviors. For this reason, this paper, releasing the basic hypothesis of rational person in traditional finance, starts from the simple philosophical dispute whether people are rational, and attempts to introduce the findings of crowd psychology to the research of the financial market based on behavioral finance.Then how is this paper to confirm the existence of crowd behavior? To empirically demonstrate the existence of crowd psychology, this paper first confirms the existence of crowd behavior with evidences, as an investor's psychology plays a decisive role in his/her behavior. We then prove whether the crowd behavior is caused by crowd psychology through exclusion procedures and questionnaire. There are quite a few procedures to confirm the existence of crowd psychology with empirical evidence. After much analysis and comparison, we decide to adopt HM-method. With Shanghai stock market as a sample, we confirm an extensive existence of crowd behavior in China's security market. To confirm the existence of crowd psychology, we first study the difference in the two kinds of crowded behavior caused by the second and the third reasons: the crowd behavior caused by crowd psychology arises from suggestions worked upon an investor when he is subject to unconsciousness; while herd behavior is a rational decision an investor makes with complete rationality. Freud made a definition of: I would like to make a point of differentiate the influence of suggestion from those of other psychologies, for example, a piece of information or an instruction. Their difference lies in that, as far as a suggestion is concerned, when an idea is aroused in the mind of another person and accepted thereafter without running a check on its origin, as if it arosespontaneously from his own mind. We confirm through questionnaire that the crowd behavior we measured is performed by a multitude of investors under unconsciousness, that is, the crowd behavior is caused by crowd psychology. This confirms the existence of crowd psychology.What is the specific influence of crowd psychology upon the security price? As crowd psychology has the three basic characteristics: impulsiveness, submissiveness, and extremeness, it is bound to cause violent fluctuation of security price, making the price deviant from the real value. A direct representation of this phenomenon is the fat tail of the distribution of security market return rate relative to a normal distribution. We measure the tail thickness of the return rate distribution with Hurst Index. When Hurst Index gets bigger, the tail of the distribution gets thicker; conversely, the smaller Hurst Index is, the more the tail of the return distribution approximates Gaussian Distribution. We divided the sample into 2 sub-samples: the first sub-sample is from December 26,1997 to January 25, 2002. The crowd psychology is rather low (2.81) in this period of time, which represents a high rate of crowd psychology. The other sub-sample is from Feburary 1, 2002 to March 26, 2004. The crowd behavior is quite high (4.65) in this period of time, which signifys a low rate of crowd psychology in the market. Due to the statistical significance of the crowd behavior of the two sub-samples, we can know the influence of the crowd behavior on return rate distribution by analyzing the tail thickness of the return rate distributions of the two time periods.After calculation, we find that the Hurst Index of the weekly return rate is 0.75226 in the time period of December 26,1997 to January 25, 2002, and the counterpart for Feburary 1, 2002 to March 26, 2004 is 0.63347. With statistical testing, we find that the Hurst Indexes are strikingly different at confidence level of 95%. This means crowd behavior can really result in the thick tail of the return rate distribution. Then how should we conduct effective risk management when crowd psychology is present? The considerations of risk management are rather extensive, this paper attempts to research into the measurement of risk-the core of risk management. This paper finds that the risk measurement method Prob[X≤E(X)] in behavioral finance is essentially in perfect agreement with the mainstream VaR method of risk measurement, and it is a risk measurement method with a better focus on investment loss. When the subject considers the safety value of the objective function is the mean of the variants, the risk measured by Prob[X≤E(X)] is equivalent to risk measured by the relative VaR method; when the subject thinks that the safety value of the objective function is capable of maintaining the wealth level of the present utility, the risk measured by Prob[X≤E(X)] is equivalent torisk measured by the absolute VaR method; but when crowd psychology is present in security market, this method is unable to measure the risk brought forth by crowd psychology, as it gauges only a percentile point of the return rate distribution and fails to take the tail thickness of return rate distribution into account. Meanwhile, integrating the latest development of risk management, this paper proposes a new risk measurement method- Cohesive Value at Risk: the expected value 1-aon the left tail of the return distribution in T time and at a confidence level of a. Finally, this paper constructs a mathematical model of this method, and develops several algorithms of uniformity risk value with a few kinds of distributions and an investment decision model based on Cohesive Value at Risk.This research involves a further research, which, due to the constraints of the structure of this paper, cannot be incorporated into this paper. I put it into the section of conclusion. For example, the problem of whether information arrives in the market in a random way is discussed. We develop a new variant t, and convert that problem of information random arrival into a distribution testing of the new variant t. With empirical evidence, we find that t doesn't conform to normal distribution, with the Hurst Index at 0.7. This signifies that information arrives the security market with a skewed manner. This conclusion poses a still more powerful challenge to the existing financial theory, because the traditional financial theory has a rigorous hypothesis: information arrives in the market in a random way.