Multimodal Data Analytics for Mergers and Acquisitions Performance Analysis

基於多模式數據分析的併購績效研究

Student thesis: Doctoral Thesis

View graph of relations

Author(s)

Related Research Unit(s)

Detail(s)

Awarding Institution
Supervisors/Advisors
Award date14 Sep 2020

Abstract

Mergers and Acquisitions (M&A) is a major type of corporate investment strategy for continuous growth and development. The rapid development of the online network is radically transforming how acquirer firms identify targets, senior executives communicate with key stakeholders, and investors retrieve valuable and previously inaccessible information. However, the literature on M&A is limited to the traditional media. Besides, multimodal data such as images, emojis, and videos are now prevalent on online networks. Whether this kind of multimodal information can complement text information remains unclear. Therefore, this dissertation utilizes multimodal data analytics to examine how the changes of different parties brought by the online network impact the decisions and performance of firms’ M&A. It consists of three studies. 

The first study, “Do the Public’s Emotions Predict Post-M&A Performance? A Machine Learning-based Multimodal Empirical Study,” develops a machine learning-based multimodal analytics methodology to derive the public’s emotions from their textual and image-based Twitter posts, and explore the association and predictive value of these emotions on the post-M&A performances of both the acquirer and target firms. The empirical results show that the surprise emotion has significant, negative correlations to the target and acquirer firms’ post-M&A performances across both textual and image-based Twitter posts, while the emotions of fear and joy are found to be only significant in image-based posts. In addition, we find that the public’s emotions of surprise, fear, and joy expressed through their image-based posts is an effective predictor of corporate post-M&A stock returns and has stronger predictive power than those emotions expressed through textual posts. Furthermore, the study unveils that the negative relationship between the surprise emotion expressed through public multimodal Twitter posts and corporate post-M&A performance is weaker under a friendly M&A deal than an unfriendly deal. 

The second study, “The Impact of Social Executives on Firms’ Mergers and Acquisitions Strategies: A Difference-in-difference Analysis,” focuses on the group of social executives, top managers, who disclose work-related or personal information to a broad set of stakeholders through their personal social media accounts. This study explores the role of the presence of social executives on acquirers’ preference for M&A investments versus internal investments and the impact of emotions generated from their Twitter postings on firm value. Empirical results from a difference-in-difference analysis suggest that the presence of acquirer-side social executives improves the likelihood and numbers of corporate M&A transactions and subsequent financial performance, but reduces internal investments (i.e., capital expenditures). Furthermore, there is evidence that a senior executive with higher social media engagement is more likely to engage in M&A decisions that lead to greater gains and reduced capital expenditures. Finally, the study finds that managerial emotions provide valid information about firms' prospects. Using Plutchik’s wheel of emotions theory, the study further finds that social executives’ fear emotion can effectively predict firms’ one-quarter-ahead and half-year-ahead stock returns.

The third study, “Do the Interplays Between the Personality traits of CEOs and CFOs Influence Corporate M&A Intensity? A Machine Learning-based Empirical Study,” focuses on two pivotal figures in the top management team—CEO and CFO. This study examines how the interplay between the personality traits of CEO and CFO enables or restricts mergers and acquisitions (M&A) intensity based on a machine learning enabled econometric analysis methodology. Earnings call transcripts were used to detect CEO/CFO personality from linguistic and syntax cues. Based on the M&A historical data of S&P 1500 firms, our empirical results show that the openness personality trait of CEOs is positively associated with corporate M&A intensity, while CEOs’ personality traits of consciousness and neuroticism are negatively associated with corporate M&A intensity. We also find that the impacts of CEOs’ openness and consciousness on corporate M&A intensity are more pronounced when CFOs have similar personality traits as those of CEOs. The impact of CEO and CFO personalities on M&A intensity is further confirmed by specific pairwise combinations of CEO-CFO personality traits. Finally, this study reveals that CEOs that partner with more powerful CFOs are more likely to undertake M&A than those with less powerful CFOs.

The contribution of this dissertation is multifaceted. First, this study proposes a multimodal analytics methodology, which shows its effectiveness in deriving and evaluating emotions expressed through multimodal social media content. The study’s document of the predictive value of image emotions for corporate strategic outcomes is novel. Practically, practitioners can apply the proposed methodology to effectively extract online users’ emotions from their multimodal social media posts and then utilize the emotions of surprise, fear, and joy to predict firm’s or the counterpart’s post-M&A performance.

Second, the study contributes to the upper echelons theory by documenting the importance of social media presence of senior executives and exploring its impact on firms’ M&A decisions and financial performance. Previous studies have proved fruitful theoretical approaches to verify or enrich the upper echelons theory by examining the roles of executive background characteristics, experiences, and values on corporate strategic choices and performance. Living in the social media era, this study explores a new heterogeneous characteristic of senior executives, that is, social versus non-social executives, and sheds light on how the presence of social executives impacts a firm’s preference for M&A investments versus internal investments. The study is one of the first to document the predicting power of senior executives’ social media emotions on firms’ value. The presence of social executives improves M&A transactions but reduces capital expenditures; employers can use this new criterion when recruiting top managers.

Third, this thesis is the first successful research of applying a machine learning enabled econometric analysis methodology to empirically examine the impact of the interplays between the personality traits of CEOs and CFOs on corporate M&A outcomes. Our study lays a solid foundation for IS researchers to apply machine learning enabled econometric analysis methodologies to explore the influence of executives’ personal characteristics on firms’ business decision-making. Besides, the managerial implication of our research is that business managers can apply the proposed methodology to automatically extract senior executives’ personality traits from their textual comments, and then utilizing these traits to support various business decision-making such as M&A target selection, corporate hiring, institutional financial investment, and so on.

    Research areas

  • Fintech, Mergers and Acquisitions, Multimodal Data Analysis, Machine Learning