Rational or Irrational? A Comprehensive Studies on Stock Market Crashes
27 Pages Posted: 21 Jun 2017
Date Written: June 20, 2017
This study attempts to illustrate the contributing factors for different patterns of crashes. In addition to the fundamental macro-economic factors, this paper argues that the existence of herding behavior as well as the level of investor attention are also important factors affecting the pattern of stock price fluctuations. By differentiating the rational component and irrational component of these behavioral factors, more insight concerning financial crisis can be drawn.
Patterns of crashes are defined by three dimensions, which are the cumulative decline, the speed of decline, as well as the duration of the crash. Innovative measures and comprehensive analyses are conducted based on three sets of explanatory factors: macroeconomic factors, market microstructure factors and behavioral factors. Results of partial R2 show that behavioral factors are the most influential factors explaining the magnitude as well as the duration of crash; while the speed of decline is mainly related to market microstructure factor. Our results show that investors' irrational behavior is more important than fundamental factors in explaining or predicting market crashes.
The contribution of this study are threefold: First, crashes in 40 markets are defined, measured and categorized into eight types of crash patterns, providing interesting statistics for international market crashes. Secondly, we differentiate between rational and irrational components of behavioral factors in explaining the causes of market crashes, which are largely neglected in past literatures. Thirdly, threshold of each explanatory variable of market crash are estimated. The results of this paper can provide policy makers, fund managers and investors valuable information in risk management and pre-warning system.
Keywords: stock market crash, macro-economics, market microstructure, herding behavior, investor attention, rational and irrational components, heterogeneous agent model, network analysis
JEL Classification: G01, G02, G14, G15
Suggested Citation: Suggested Citation