Doctoral thesis (Dissertations and theses)
Sentiment Analytics and Financial Markets
Moreno Miranda, Nicolas
2022
 

Files


Full Text
Thesis_Sentiment_Analytics_Financial_Markets.pdf
Author preprint (2.2 MB)
Download

All documents in ORBi are protected by a user license.

Send to



Details



Keywords :
sentiment analytics; asset pricing; market anomalies; cash flow news; discount rate news; corporate finance; LBO; Glassdoor; Natural Language Processing; Textual Analysis; Earnings Announcements; Reuters News
Abstract :
[en] From major news outlets to social media and the general public, it is common to find mentions of the existence of relationships between narratives and economic outcomes. By definition, those narratives are forms of soft information, which until recently have been difficult to quantify and are often propagated through natural language and text in particular. This thesis seeks to leverage this soft information and harness one key dimension of text in particular: Sentiment. In the context of this thesis, sentiment is defined as the disposition of an entity toward another entity, expressed via a specific medium. In the first three chapters of this thesis, the medium of interest is “News”, specifically news stories published in the financial press. The first paper uses firm-specific news sentiment to understand why market anomalies earn a premium on earnings announcement days. News sentiment shows that this premium for value firms is concentrated on bad news events, which permits us to propose new avenues to understand this market anomaly. The second and third chapters investigate more generally how news can help understand drivers of market anomalies. Market anomalies have played a central role in asset pricing research over the past decades, and numerous competing theories seek to accommodate empirical observations that deviate from the classical model. Chapter two proposes a framework based on cash-flow and discount rate news, allowing us to capture the driving forces behind anomaly returns and disentangle competing explanations for anomalies. The third chapter investigates drivers of anomaly returns and characterizes news of momentum and value stocks, in particular, highlighting the strong negative correlation between the two. It is also the first to link cash-flow news, discount rate news, and news sentiment. The economic outcome of interest in the fourth chapter is to understand how changes in company ownership, especially following leveraged buy-outs, affect employee welfare. We gather millions of online reviews of employees about their employers and investigate the underlying text data to characterize the impact private equity firms have on those narratives. Overall, employees’ satisfaction drops sharper following leverage buy-outs than in other types of ownership changes and we can trace those problems back to specific issues related to lack of management care and fear of cost-cutting and layoffs.
Disciplines :
Finance
Author, co-author :
Moreno Miranda, Nicolas ;  Université de Liège - ULiège > HEC Recherche
Language :
English
Title :
Sentiment Analytics and Financial Markets
Defense date :
08 June 2022
Institution :
ULiège - University of Liège [BE] [HEC], Liege, Belgium
Degree :
Ph.D. in Economics and Management
Promotor :
Lambert, Marie ;  Université de Liège - ULiège > HEC Liège : UER > UER Finance et Droit : Analyse financière et finance d'entr.
President :
Torsin, Wouter ;  Université de Liège - ULiège > HEC Liège : UER > UER Finance et Droit : Financial Reporting and Audit
Jury member :
Ittoo, Ashwin ;  Université de Liège - ULiège > HEC Liège : UER > UER Opérations : Systèmes d'information de gestion
Contreras, Martha Gabriela ;  Université de Liège - ULiège > HEC Liège : UER > UER Finance et Droit ; Radboud Universiteit Nijmegen [NL]
Renault, Thomas;  Université Paris 1 Panthéon-Sorbonne [FR]
Evgeniou, Theodoros;  INSEAD
Available on ORBi :
since 07 September 2022

Statistics


Number of views
255 (16 by ULiège)
Number of downloads
387 (8 by ULiège)

Bibliography


Similar publications



Contact ORBi