Beyond negative and positive: Exploring the effects of emotions in social media during the stock market crash. Issue 4 (July 2020)

• Record Type: Journal Article
• Authors: Ge, Yidi; Qiu, Jiangnan; Liu, Zhiyong; Gu, Wenjing; Xu, Liwei
• Is Part Of: Information processing & management. Volume 57:Issue 4(2020:Jul.)
• Publication Date: 2020-07
• Subjects: Stock market crash -- Emotion in social media (ESM) -- Market cognition -- HMM; Information storage and retrieval systems -- Periodicals; Information science -- Periodicals; Systèmes d'information -- Périodiques; Sciences de l'information -- Périodiques; Information science; Information storage and retrieval systems; Periodicals; 658.4038
• DOI: 10.1016/j.ipm.2020.102218
• Languages: English
• ISSNs: 0306-4573

Abstract

Highlights: This paper investigates the effects of emotions in social media (ESM) on the stock market during the market crash using a cognition-based framework. It is verified that ESM firstly influence the market cognition, and then influence the stock market during the market crash. High arousal ESM increases the probability of market cognition of a crash state both in-crash and post-crash periods. Positive valence ESM adjusts the market cognition to the stable state only in the later stage of the market crash. "Fear" and "Disgust" are the two significant sub-categorical emotions correlated with the market cognition during the market crash. Abstract: Recent studies indicate that the stock market is influenced by emotion in social media (ESM) which are embodied in user-generated content. However, the relationship between ESM and the stock market in the event of the market crash has not been fully explored. This study thus explores the effects of ESM on the stock market during the market crash by empirically validating the proposed cognition-based framework of "Emotion-Cognition-Market". A three-component model is constructed for the framework, which uses sentiment analysis to calculate emotions from two-dimension of valence and arousal, uses Hidden Markov Model (HMM) for market cognition mining, and uses ordered logistic regression for relationship establishment between ESM and market cognition. More than 280, 000 Weibo of 34 listed companies during the market crash (the second half-year of 2015 in China) are used. It is confirmed that the impact process of ESM on the stock market is actually the result of constantly changing market cognition influenced by ESM. This study goes beyond the commonly used positive and negative (polar) emotions or sub-categorical emotions, discovering the salient effects of arousal dimension in the market crash, and also quantify the effects. The results show that the increase of one unit high arousal ESM significantly increases the probability of the cognition of a crash state by 9.99 and 17.41% during the in-crash and post-crash period, of which "Fear" is the main risk factor. In addition, positive valence ESM is the driving force of restoring the stability of the market cognition only in the later stage of the market crash. This paper calls on market participants to pay attention to high arousal emotions in emergency situations in advance.