investment analysis and performance of hotel industry- a research of international chain hotel company’s investment strategy
ratio analysis and graph needed
8 reference minimum
The Impacts of Social Interaction on the Stock Market – An Empirical Analysis Based on Sina Finance Blog
MSc Accounting and Financial Management
Total word count: 2972
1.1. Background of the study……………………………………………….…1
1.2. Research aim and objectives………………………………………….…2
1.3. Scope and limitations………………………………………………….….2
2. Literature review…………………………………………………………….3
2.1. The meaning of social interaction and its measurement methods…..3
2.2. The impact of Weibo and blog on the stock market ..…………………4
3. Research methodology and design……………………………………….6
3.1. Sample selection and data source………………………………………6
3.2. The variable construction description……………………………………7
3.3. Analysis and model …………………………………………………….…9
3.4. Robustness tests………………………………………………………….10
4. Expected results…………………………………………………………….11
1.1. Background of the study
For individual investors, stock investment is a complex decision, and individuals often face limitations of lack of knowledge and incomplete information. In this case, social interaction becomes an important way for investors to exchange information and express opinions. In recent years, there has been considerable literature discussing the impact of social interaction on financial decision-making behavior. Duflo and Saez (2003) confirmed that social interaction has an impact on the decision-making behavior of individuals purchasing pension products; studies by Hong et al. (2004) and Li Tao (2006) showed that social interaction can significantly promote residents’ participation in stock market investment; Karlan (2005) provided evidence of individual credit decisions that influence social interaction through social interaction.
With the rapid development of the Internet, investors can not only obtain information on the Internet, but also actively publish information and communicate with other investors. Online communities such as blogs, Weibo, and forums have become an important platform for interaction between investors. In the Internet age, the scale, frequency, and speed of investor social interaction have all undergone revolutionary improvements. Internet-based social interaction is a double-edged sword for the development of the securities market. On the one hand, the Internet platform has the characteristics of timely and wide coverage. The social interaction between investors helps to improve the transparency of market information, thereby improving the information validity of the securities market; on the other hand, it is contrary to the authority of traditional media. In comparison, the Internet platform has the characteristics of anonymity and grassroots, which can easily become the main channel for rumor communication and subjective emotional catharsis. The social interaction between investors will lead to overreaction of market prices, thus affecting the stability of the market. Studying the influence of Internet-based investor social interaction on the stock market is theoretically beneficial to explore the new characteristics of asset pricing in the Internet era. In practice, it helps to provide a policy basis for promoting the standard operation of online media and the stable development of financial markets.
1.2. Research aim and objectives
The major aim of this proposed study is to use the online interactive platform- Sina Finance Blog, as an example to explore the impact of social interaction on the stock market. However, some specific objectives have been set to help achieve the research aim:
1. To examine the relationship between the related variables of Sina blog and the yield, volatility and volume of the stock market (CSI 300, Shanghai Composite Index, GEM) from June 2017 to August 2018;
2. To establish models to use the Stata software to empirically test the data.
1.3. Scope and limitations
This proposed study is an attempt to analyze the impact of social interaction on the stock market between June 2017 and August 2018, as there was a clear period of sustained ups and downs during this period. The limitations of the study include:
a) This article may have some slight deviations in data when using text mining technology to obtain blog information, but it will not affect the conclusion.
b) This paper mainly reveals the macroscopic effects of social interaction on the stock market, but the influence of social interaction on specific individual decision-making behavior cannot be distinguished.
2. Literature review
The literature related to the topic of this article includes two aspects, one is social interaction; the other is the influence of the online interactive platform represented by Weibo and blog on the stock market.
2.1. The meaning of social interaction and its measurement methods
Social interaction comes from the concept of sociology. Buck (1984) argued that social interaction is a reaction of one or more parties, depending on the extent to which the other party said, and the social context changes with this degree. Durlauf and Ioannides (2010) extended the connotation of social interaction from sociology to economics. They believed that social interaction refers to the interdependence of individuals. Under these interdependences, a social and economic people with behavioral characteristics, their preferences, beliefs, and the budget constraints they face are directly affected by the characteristics and choices of others.
In recent years, economic research has gradually begun to discuss the impact of social interaction on economic behavior. The introduction of the concept of social interaction has also brought a new perspective to economic research. Duflo and Saez (2003) used experimental methods to study the impact of social interaction on individual decisions to purchase pension products and the conclusions show that this effect is significant. Hong et al. (2004), after controlling the level of wealth, race, education, and risk tolerance, found that those who interacted with neighbors and attended churches were more likely to participate in stock market investments, and the impact of such social interactions are even more pronounced in areas where stock market participation is higher. Hong et al. (2004) and Karlan (2005) also provided evidence that influences fund managers’ shareholding behavior and influences individual credit decisions through social interactions. Li (2006) studied the relationship between social interaction and stock market participation in China. Through questionnaire survey, the two indicators of “the total number of people who visited relatives, friends and people in various forms during the Spring Festival in 2004” and “residents’ subjective evaluation of their level of interpersonal interaction in society” were used to measure social interaction. The results showed that social interaction promotes residents to participate in the stock market. Zhu et al. (2014) used the data from the China Household Finance Survey in 2011 to find that one of the channels of influence on the positive promotion of family stock market participation is social interaction. They divided the community into high, medium and low participation rate communities and set them as dummy variables according to the community’s stock market participation rate. They found that the cross-items of high-interaction rate communities and relationships were significantly positive, while the cross-items of low-participation communities and relationships were significantly negative, which means that in communities with higher stock market participation rates, social interactions will have a greater role in promoting family participation in the stock market.
2.2. The impact of Weibo and blog on the stock market
The research on the influence of Weibo and blogs, which is expanded from the field of media studies, mainly includes marketing and finance. In the field of marketing, Mishne and Glance (2006) studied the relationship between blog sentiment and the box office of the movie. The results showed that before and after the film was released, the discussion of the film and the box office income of the movie were significantly positively correlated. Positive emotions are an effective predictor of movie box office success. Liu et al. (2007) constructed a model that uses blog sentiment information and past sales performance of commodities to predict future sales of goods. Taking the movie box office income as an example, the accuracy and effectiveness of the model have been verified.
The research direction in the field of finance is mainly the influence of Weibo and blog on the stock market. Zhang et al. (2011) analyzed the emotions contained in each Twitter, divided into positive emotions and negative emotions. The ratio of the number of Twitters containing emotions to the total number of Twitters was used as an independent variable. The ratio was found that significantly negatively correlated with the Dow Jones Index, Nasdaq Index and the S&P 500 Index，but was significantly positively correlated with the Chicago Board Options Exchange Volatility Index. Similarly, Bollen et al. (2011) used Twitter’s emotional mining tools OpinionFinder (OF) and Google-Profile of Mood States (GPOMS) to measure text sentiment based on Twitter. OF can quantify the intensity of positive and negative emotions on the current network in real time; GPOMS can subdivide emotions into six categories (Calm, Alert, Sure, Vital, Kind, and Happy). The empirical results showed that the ratio of positive and negative emotions only had a significant impact on the Dow Jones index change of lagging one day, while only the calm in the six-dimensional emotional sequence had a significant impact on the Dow Jones index change of lagging two to six days.
Choudhury et al. (2008) specifically analyzed the characteristics of the four company blogs from January to November 2007 for four US technology companies, including the number of blogs, the number of comments, the average length of comments and the time of reply, and the role of different comments respondents, etc. The results showed that after quantifying the characteristics of these blogs, they were significantly related to the daily rate of change of the company’s stock price, 87% accuracy in the direction of stock price movements, and 78% accuracy in the scale of stock price changes. Ruiz et al. (2012) not only focused on the changes in stock prices but also studied the changes in trading volume and found that the relationship between Twitter’s own characteristics and stock trading volume is more significant than the relationship between stock prices.
Cen et al. (2014) used the Q&A data of the “Huyibao” platform of the Shenzhen Stock Exchange, which is similar to the Weibo website, to identify the number of questions asked by investors in Shenzhen stock market as a measure of investor attention. They found that the higher the stock returns, the lower the stock volatility risk and the liquidity risk, indicating that the interactive platform between the investors and listed companies can significantly reduce information asymmetry, stabilize the market and protect Investor rights. Zhang and Han (2015) provided information on the timeliness and clarity of the management of listed companies to the subjects based on the statistical data of the “Huyibao” platform of the Shenzhen Stock Exchange. The experimental results showed that the timelier the management response and the higher the degree of response, the higher the investor’s investment potential.
The research on social interaction in the literature has analyzed the influence of social interaction on individual decision-making behaviors, such as the influence of stock investment participation, personal credit decision-making and purchasing pension product decision-making. The research in this paper tests the impact of social interaction on stock market yield, volatility and volume and reveals the macro effect of social interaction on the stock market.
3. Research methodology and design
3.1. Sample selection and data source
The four sections of the Sina Finance blog that are most directly related to the stock market and have the highest relevance are “exclusively watched”, “large market trend”, “street/single stock” and “securities market” are selected as blog samples. The time span is determined from June 1, 2017 to August 1, 2018 and includes a period in which the stock index gradually enters a rising phase, the price continues to expand, and then falls sharply. This research uses text mining to obtain the text information of each blog, the number of blogs read, the comments/forwarding/likes/collections, the popularity of bloggers, etc., and uses the texts classification to obtain the signals or Emotions transmitted by each blog, thus constructing indicators that reflect social interaction on a daily basis. Stock market variables include the rate of return, volatility and volume. The data are all from the Wind database. The rate of return is the same as the CSI 300 index. The volatility is selected from the Shanghai and Shenzhen 300 Index for nearly 26 weeks. According to the 26-week window rolling calculation, the formula is , where is the log-return, is the average rate of return.
3.2. The variable construction description
In order to obtain relevant data of Sina Finance Blog, this study uses text mining technology to extract information on web pages. The specific method is to use the Gooseeker Web Crawler Software to scan the underlying code of the blog site and grab the original data through the html tag. After the original data is cleaned and sorted, 964 bloggers and 142,579 blog posts are obtained. The variables of the blogger data include the blogger ID, the blogger’s homepage, the blogger’s nickname, the popularity of the blogger, etc. The main variables of the blog post data are the blog ID, the blog title, the blog post time, the number of readings/the number of comments/likes/number of favorites/ Forwarding number, etc.
This study builds social interaction variables on a daily basis. The variable construction steps are as follows:
In the first step, calculate the number of daily blogs (N), the number of daily readings (R), the number of daily comments/forwarding/likes/collections (L) and the number of daily popular bloggers (I).
Daily readings = Daily sum of readings for all blogs / Number of blogs per day
Daily Comment/Forward/Like/Favorites=Total number of comments, forwards, likes, favorites/ Number of blogs per day
The calculation method of the proportion of daily high popularity bloggers is to rank 964 bloggers according to the popularity of the people and select the top 100 bloggers who are concerned about popularity. This study defines the 100 bloggers as highly popular bloggers, and then counts the proportion of these highly popular bloggers who blog on blogs every day.
The second step is to use the text classification method to determine the daily mood index (E).
142,579 blogs are classified into three categories: “positive”, “neutral” and “negative” according to the signals they transmit. Since the body of the blog is too long, this study only classifies the blog title. The classification process is mainly as follows: First, a small part of the entire blog sample is randomly selected as a “training data set” for manual classification. The computer then builds the computer’s own classification model by learning the results of the training data set. Finally, this study will use Weka which is a data mining open source package to apply the trained classification model to the training data set for verification.
After emotional classification of 142,579 blogs, this study used Antweiler and Frank (2004) to construct a blog sentiment index. Assuming that the number of all blogs that pass the “positive” signal in a day is set to , and the number of all blogs that pass the “negative” signal in a day is set to , the daily blog sentiment index is constructed as follows:
It can be seen that the larger the index, the more positive the signal or emotion transmitted by the blog on the day, and whereas the more negative it is.
The third step is to construct an interactive variable that contains emotional features.
Combine the number of daily blogs, the number of daily readings, the number of daily comments/forwarding/likes/collections and the number of daily popular bloggers calculated in the first step, and the daily blog sentiment index built in the second step, respectively. Multiply the product to determine the final daily social interaction variable that contains emotions.
Blogger’s interactive appeal (EN) = Daily Number of blogs * Sentiment Index
Blogger’s influence (EI) = Daily High popularity of bloggers * Sentiment Index
Interactive coverage (ER) =Daily Readings Number* Sentiment Index
Degree of interaction (EL) = Daily Comment / Forwarding / Like / Collection number * Sentiment Index
3.3. Analysis and model
The relevant variables of the stock market in this paper are presented in time series, so the GJR-GARCH (1,1) model is used to analyze the impact of social interaction on the stock market. First, the ADF stationarity test is performed on each interaction variable and stock market variable sequence using Stata measurement software. Then, the least squares method OLS is used to estimate the original equation, and then the residual square sequence is tested by Q test to test whether the model has ARCH effect. If there is an autocorrelation, the GJR-GARCH (1,1) model is established such that there is no autocorrelation of the model. Based on the Shanghai and Shenzhen 300 Index Yield, the regression model is estimated to have 12 groups:
3.4. Robustness tests
In order to further examine the impact of social interaction on the stock market and to test the robustness of the above results, this study followed the Shanghai and Shenzhen 300 index regression model, replacing the dependent variable for the Shanghai Composite Index and the GEM index.
Then, the regression models are established for the continuous rise period before January 2018 and the subsequent continuous decline period, and finally further conclusions are obtained.
4. Expected results
This paper expects that the blogger’s interactive appeal, blogger’s influence, interactive coverage and interaction will have a positive impact on the market index’s yield and volume, as well as have no significant impact on volatility. With the apex of the stock index in January 2018 as the dividing line, the influence degree of social interaction on the rate of return is greater when the index is in a down period than when it is rising, because people are more sensitive to falling stock prices. Compared with the Shanghai and Shenzhen 300 Index and the Shanghai Composite Index, the GEM Index yield is more affected by social interaction.
This article uses Sina Finance Blog as a way of social interaction to empirically test the impact of social interaction on the stock market. From the perspective of the main body of interaction – bloggers and investors, text mining technology is used to analyze the emotional tendency of blog texts from June 2017 to August 2018, and four variables are constructed to describe and measure investors’ Social interaction. It includes the blogger’s interactive appeal, the influence of the blogger, the level of interaction and the degree of interaction. The analysis is then performed using the Stata tool, which is expected to draw conclusions from the previous section, and these conclusions will provide evidence for the impact of social interaction on the stock market.
Antweiler. W. & Frank. M. Z. (2004). Is all that talk just noise? The information content of internet stock message boards. The Journal of Finance, 59(3): 1259-1294.
Bollen. J., Mao. H. & Zeng. X. (2011). Twitter mood predicts the stock market. Journal of Computational Science, 2(1): 1-8.
Buck (1984). Social Psychology. Tianjin: Nankai University Press.
Ceng. W., Li. S. & Tong. N. (2014). The Impact of Investor Attention on Stock Returns and Risk——An Empirical Study Based on Data from Shenzhen Interactive “Easy-Easy” Platform. Securities Market Herald, (07): 40-47.
Choudhury. D. M., Sundaram. H. & John. A. (2008). Can blog communication dynamics be correlated with stock market activity? Proceedings of the nineteenth ACM conference on Hypertext and hypermedia. ACM, 55-60.
Duflo. E., & Saez. E. (2003). The Role of Information and Social Interactions in Retirement Plan Decisions: Evidence from a Randomized Experiment. The Quarterly journal of economics, 118(3): 815-842.
Durlauf. S. N., & Ioannides. Y. M. (2010). Social Interactions. Annual Review of Economics, 2(1): 451-478.
GUO. S.& LIANG. P. (2014). Social Interaction, Information Channels and Family Stock Market Participation——An Empirical Study Based on 2011 China Household Finance Survey. Economic Research, (S1): 116-131.
Hong. H., Kubik. J. D.&Stein. J. C. (2004). Social interaction and stockmarket participation. Journal of finance, 59(1): 137-163.
Karlan. D. S. (2005). Using experimental economics to measure social capital and predict financial decisions. American Economic Review, 1688-1699.
Li. T. (2006). Social Interaction, Trust and Stock Market Participation. Economic Research, (01), 34-45.
Liu. Y., Huang. X. & An. A. (2007). ARSA: a sentiment-aware model for predicting sales performance using blogs. Proceedings of the 30th annual international ACM SIGIR conference on Research and development in information retrieval. ACM, 607-614.
Mishne. G. & Glance. N. S. (2006). Predicting Movie Sales from Blogger Sentiment. AAAI Spring Symposium: Computational Approaches to Analyzing Weblogs, 155-158.
Ruiz. E. J., Hristidis. V. & Castillo. C. (2012). Correlating financial time series with micro-blogging activity. Proceedings of the fifth ACM international conference on Web search and data mining. ACM, 513-522.
Zhang. J. & Han. D. (2015). The Timeliness, Clarity and Investment Decision of Management in Network Interactive Platform Communication-An Experimental Evidence. Management Review, (10): 70-83.
Zhang. X., Fuehres. H., & Gloor. P. A. (2011). Predicting stock market indicators through twitter “I hope it is not as bad as I fear”. Procedia-Social and Behavioral Sciences, (26): 55-62.
Zhu. G., Du. Z. & Zhang. L. (2014). Relationship, Stock Market Participation and Stock Market Returns. Economic Research, (11): 87-101.