How tone is used across CEO letters : the impact of financial crisis evidence from the UK

• Main Author: Ferreira, Filipa Freire
• Publication Date: 2021
• Format: Master thesis
• Language: eng
• Source: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Download full: http://hdl.handle.net/10400.14/36870
• topic: Financial crisis; Natural language processing; Tone; Uncertainty; Causality; Impression management; Crise financeira; Tom; Incerteza; Causalidade

We study the tone present in CEO Letters of firms listed in London Stock Exchange (LSE), more specifically, by defining two alternative approaches to assess tone: the CEO Letter as a whole and the weighted approach, where the percentages chosen for Sections 1, 2 and 3 are, 10%, 80% and 10%, respectively. By differentiating these two models, we intend to analyse if differences exist between the whole text tone and the tone present in each section. Secondly, we aim to understand if there are differences in the tone means within sections for the same CEO Letter. Additionally, we study the impact of financial crisis by analysing the changes in some textual features of CEO Letters, more specifically, tone, uncertainty, and causality. Furthermore, we also analyse if there is a significant change in the relationship between firm’s current performance and its reported tone in crisis versus non crisis settings that might help us conclude which scenario is more likely for firms to engage in impression management. By using several linear regression models, we conduct an analysis of the textual features in CEO Letters as a whole and in its sections, individually. There are indeed differences between the whole text tone and the tone of each section and, differences in the tone means within sections. Tone is explained by the financial crisis. While uncertainty words increase with Financial Crisis, we do not find any relevant link in the use of causal words. Regarding current firm’s performance and tone relationship, we do not observe any significant change in this relationship from crisis to non-crisis years.