Cette thèse étudie le rôle des tierces parties dans les problématiques informationnelles autour de la responsabilité sociale et environnementale des entreprises (RSE). Les entreprises revendiquent de plus en plus le fait qu'elles adoptent des pratiques responsables. Cela peut être lié au fait qu'il existe des parties prenantes (e.g. consommateurs, investisseurs) qui sont prêtes à récompenser les comportements responsables. Cependant, la véracité des revendications des entreprises est souvent impossible ou au moins difficile à certifier, ce qui peut inciter les entreprises à manipuler leur communication. Cette dissertation a pour objectif d'analyser la façon dont de tierces parties peuvent tempérer cette asymétrie d'information. Des organisations comme des agences de notations, des organisations non-gouvernementales (ONGs) ou des organismes certificateurs, peuvent examiner les actions réelles des entreprises et transmettre cette information aux parties prenantes. Ces tierces parties ont un comportement informationnel hétérogène: certaines fournissent des informations plus particulièrement sur les entreprises qui se comportent de manière irresponsable, alors que d'autres se spécialisent dans la mise en lumière des entreprises responsables. Nous développons un modèle théorique simple qui explore l'offre de qualité dans des environnements informationnels amicaux, où la qualité est révélée plus souvent lorsqu'elle est haute que lorsqu'elle est basse, et dans des environnements hostiles, où c'est l'inverse. Nous appliquons ensuite ce modèle à deux types de tierces parties: les ONGs et les labels. Nous étudions le choix de production de bonnes ou mauvaises nouvelles afin de comprendre quand et pourquoi les ONGs choisissent d'être hostiles ou amicales. Ensuite, nous cherchons à évaluer quel est le niveau de sévérité des labels qui entraîne le plus d'investissement et le bien être social le plus important. / This thesis studies the role of third parties in tempering informational issues at stake with regard to corporate social responsibility (CSR). Firms growingly claim to be adopting responsible practices, and it may be so because some socially conscious stakeholders (e.g. consumers, investors) are ready to reward such behaviors. Nevertheless, truthfulness of firms' virtue claims is often impossible or at least difficult to ascertain, which creates an incentive for firms to manipulate their communication. This dissertation intends to analyse how third parties can temper this asymmetry of information. Organisms such as rating agencies, non-governmental organisations, labeling institutions, can investigate firms' real actions and convey this information to stakeholders. These third parties have heterogeneous informational behaviors: some of them mostly disclose information on firms that do not behave responsibly while others are specialized in revealing firms that are socially or environmentally responsible. We develop a simple theoretical model that explores the provision of quality by a firm under friendly informational environments, in which quality is more often disclosed when it is high than when it is low, and hostile environments, in which the converse holds. We apply this base model to two types of third party: NGOs and labels. We endogenize the production of good and bad news in order to understand when and why NGOs choose between being hostile or friendly. Then, we seek to assess what is the more efficient level of label's strictness to spur social welfare and investment in high quality.

Since 2005, publicly traded European companies are required to prepare their consolidated financial statements in accordance with International Financial Reporting Standards (IFRS). This is the result of the so-called “IAS-Regulation” of the European Parliament and of the European Council (Regulation (EC) No. 1606/2002) which formulates two objectives directly related to financial reporting: (higher) comparability and transparency of financial statements. With regard to transparency, researchers often rely on proxies to measure the quality of financial statements. In this dissertation, I follow this approach and examine the quality of IFRS financial statements relying on different proxies, such as value relevance, the degree of earnings management, and disclosure quality. The four papers in this dissertation cover research questions related to the determinants and consequences of managerial discretion in the three most important components of financial statements: the balance sheet, the income statement, and the notes. In summary, this dissertation answers important questions concerning the quality of accounting under IFRS which have remained unanswered – and in some cases even unasked – until now. Therefore, this dissertation has a material impact on the understanding of accounting quality under IFRS.:1 Accounting Quality under IFRS – Essays on Value Relevance, Earnings Management and Disclosure Quality: An Overview … 1 1.1 Introduction … 2 1.2 Overview of the Manuscripts … 4 1.3 Principal Research Contributions … 13 References … 16 2 Discretion in the Accounting for Defined Benefit Obligations - An Empirical Analysis of German IFRS Statements … 19 2.1 Introduction … 21 2.2 Literature Review … 25 2.3 Research Approach … 29 2.4 Results … 38 2.5 Conclusion … 45 References … 54 3 The Association between Corporate Social Responsibility (CSR) and Earnings Quality - Evidence from European Blue Chips … 59 3.1 Introduction … 61 3.2 Literature Review and Hypothesis Development … 63 3.3 Research Design … 70 3.4 Results … 79 3.5 Conclusion … 88 References … 101 4 Reexamining OCI Pricing - Empirical Analysis of Reporting Location Changes due to IAS 1 (rev. 2007) … 105 4.1 Introduction and Background … 107 4.2 Literature Review … 111 4.3 Research Approach and Hypotheses Development ...114 4.4 Results … 121 4.5 Conclusion … 130 References … 150 5 Short-term and Long-term Effects of IFRS Adoption on Disclosure Quality and Earnings Management … 155 5.1 Introduction … 157 5.2 Institutional Background: The Development of the German Accounting Environment … 161 5.3 Related Literature and Hypotheses Development … 164 5.4 Research Design … 175 5.5 Results … 186 5.6 Conclusion … 195 References … 209