Date of Degree
PhD (Doctor of Philosophy)
Thomas S. Gruca
Consumers are spending more time online and their involvement in social media is also growing. Furthermore, consumers truly trust the information they find online. Therefore, I expect that positive social media mentions of a given brand will influence a consumer’s awareness, attitudes, affection, etc. towards that brand. The brand value chain model suggests that such a change in consumer mindset should translate into improved marketplace performance and, ultimately, better firm financial performance.
Previous researchers have studied the relationship between online user generated content and firm performance. They find that various metrics, (e.g. user ratings, comment volume or valence) impact firm performance. However, the extant research focuses on a single online platform (e.g., CNET), type of online posting (e.g., blog posts), or industry. In this study, I focus on social media sentiment expressed across multiple platforms for 180 monobrand firms spanning 10+ industries. I use total comments, total positive comments, total negative comments, proportion of positive comments, and proportion of negative comments as my social media variables.
First, I use the portfolio sort method to determine if firms with higher social media comment volume or higher positive (negative) comments generate higher (lower) abnormal returns, as determined by the Fama French 4 factor model. Using monthly and daily returns data over a period of more than 2 years, I find no significance differences between the returns earned by the top and bottom 20% of the firms as ranked by various social media metrics. Contrary to prior research, this result suggests that social media sentiment is already fully priced into stock returns.
I then examine the possible relationship between social media metrics and firm financial performance by analyzing whether social media sentiment improves forecasts of a firm’s quarterly cash flow. I modify the Lorek & Willinger (1996) multivariate time-series regression model to include social media comment volume and sentiment information to predict future cash flow. Using the Mean Absolute Percentage Error (MAPE) a guide to forecast accuracy I find that utilizing social media information does not provide any improvement in the prediction of future quarterly cash flow forecast. I further analyze the relationship between social media comment volume & sentiment metrics, and firm quarterly cash flow by utilizing a cross sectional regression model. I find no significant effect of social media comment volume and sentiment information on the ability to predict future firm quarterly cash flow. Panel data estimation of both the cash flow model also does not find any significant effect of the social media metrics on quarterly cash flows.
Firm Performance, Marketing, Social Media
ix, 121 pages
Includes bibliographical references (pages 111-121).
Copyright 2016 Chanchal Bahadur Tamrakar