近年来，社交媒体已经成为投资者特别是中小投资者获取信息、发表观点、表达情绪的重要平台，这些社交平台如何影响微观交易决策和宏观股票市场成为重要的研究问题。以往研究主要集中于社交平台对股票市场的影响，对社交互动如何影响投资者异质信念与微观交易决策尚无充分的讨论。本文旨在围绕社交互动对投资者异质信念、交易决策和股票市场的影响进行全面分析。本文构建了投资者异质信念理论模型，借鉴了Pedersen（2022）根据信念形成过程划分不同类型投资者的设定，同时调整了信息公布方式、引入了投资者结构参数，以理解在社交互动存在时，投资者异质信念、交易决策和资产价格的作用机制。在此基础上，本文对2012年至2022年雪球平台上近4万用户发布的300余万条帖子数据进行文本分析，根据发帖语调前后一致性和内容的具体性将投资者划分为天真投资者、狂热投资者和理性投资者。同时，本文改进了投资者异质信念和社交互动影响力衡量指标，在一定程度上解决了以往指标存在的问题。接下来，本文从三个方面进行了实证检验。首先，本文研究了不同类型投资者异质信念的形成与传递过程，以及社交互动所发挥的作用，发现狂热投资者和理性投资者的信念会对下一期天真投资者的信念产生影响，这种影响是单向的，并且随着狂热投资者和理性投资者社交互动影响力的增加而增大。其次，本文检验了社交互动如何影响异质信念与交易决策，结果表明：第一，投资者的异质信念会反映在各自的交易决策上；第二，不同类型投资者之间表现出交易决策的效仿性，而基于外生冲击的研究发现，交易决策的效仿性来源于异质信念的传递；第三，相比于天真投资者和狂热投资者，理性投资者获得了更高的收益，这种高收益更多来自于其择时能力而非选股能力。最后，本文检验了社交互动如何影响异质信念与股票市场，结果表明：第一，股票市场散户交易量反映了天真投资者的信念；第二，投资者异质信念分歧程度越大，未来股票交易量上升越多；第三，投资者异质信念对股票收益率有显著的预测作用，社交互动影响力越大、非理性投资者占比越高，预测效果越好；第四，非理性投资者占比越高、社交互动影响力越大，未来股价波动率越大。
In recent years, social media has become a crucial platform for investors, especially small and medium-sized ones, to access information, express opinions, and convey emotions. How these social platforms influence micro trading decisions and macro stock markets has become an important research question. Previous studies have mainly focused on the impact of social platforms on the stock market, with insufficient discussion on how social interaction influences investors‘ heterogeneous beliefs and micro trading decisions. This paper aims to conduct a comprehensive analysis of the effects of social interaction on investors‘ heterogeneous beliefs, trading decisions, and stock markets.The paper constructs a theoretical model of investor heterogeneous beliefs, drawing on Pedersen (2022) for the categorization of different types of investors based on the belief formation process, while adjusting information disclosure methods and introducing investor structure parameters to understand the mechanisms of heterogeneous beliefs, trading decisions, and asset prices in the presence of social interaction. Based on this foundation, the paper conducts text analysis on over 3 million posts published by nearly 40,000 users on the Snowball platform from 2012 to 2022. Investors are categorized into naive investors, fanatic investors, and rational investors based on the consistency of posting tones, as well as the concreteness of content. Furthermore, the paper improves measures of heterogeneous beliefs and social interaction, to some extent addressing issues with previous indicators.Next, the paper empirically tests three aspects. Firstly, the paper examines the formation and transmission process of heterogeneous beliefs among different types of investors and the role of social interaction. It is found that beliefs of fanatic investors and rational investors influence beliefs of naive investors in the subsequent period, with this influence being unidirectional and increasing with the social interaction influence of fanatic investors and rational investors.Secondly, the paper examines how social interaction influences heterogeneous beliefs and trading decisions. Results indicate that investors‘ heterogeneous beliefs are reflected in their respective trading decisions. Additionally, different types of investors exhibit imitation among themselves in trading decisions, with evidence suggesting that the imitation of trading decisions originates from the transmission of heterogeneous beliefs. Moreover, rational investors achieve higher returns compared to naive and fanatic investors, primarily due to their market timing ability rather than stock selection ability.Finally, the paper examines how social interaction influences heterogeneous beliefs and the stock market. It is found that the trading volume of retail investors in the stock market reflects beliefs of naive investors. Furthermore, the greater the degree of heterogeneity in investor beliefs, the more significant the increase in future stock trading volume. Additionally, investor heterogeneous beliefs significantly predict stock returns, with stronger social interaction and a higher proportion of irrational investors leading to better predictive performance. Moreover, a higher proportion of irrational investors and stronger social interaction lead to greater future stock volatility.
