Managing stakeholder relations in protracted crisis situations : a discourse analysis of corruption through the Goldenberg and Anglo-Leasing cases of Kenya

Crisis Communication frameworks are essentially restricted to Corporate Communications and Public Relations theory and are mainly developed for use in corporate settings. They therefore do not readily transfer to the public sector where " ... the notion of crisis as process, the transnational dimensions of contemporary crises, the politicization of crises, [and] the increasing importance of the media ... " exacerbates the complexity and longevity of these crises. (Rosenthal, 2003:129). To address this, I have chosen to conduct this study in the public sector, where the increased complexity and mutability of protracted crises and the ensuing impact on stakeholder relations provides an opportunity to extend debates' in Crisis Communication and Stakeholder Relationship Management. To do this, I embark on an analysis of the Goldenberg and Anglo- Leasing corruption Cases of Kenya and explore the impact of the resultant crises on the relationship between the Government of Kenya (GoK), and the Bretton Woods Institutions (BWI), as a keys stakeholder group. In what is a new approach' to a study of this kind, I combine Computer-aided Corpus Linguistic Analysis (CLA) and Critical Discourse Analysis (CDA). Starting off with the CLA, I analyse select' media texts, GoK reports and BWI reports, so as to: identify linguistic patterns within these texts; interpret what these patterns represent; and demonstrate how these representations lead to the emergence and mobilisation of concepts within the data. Secondly, I lift the study to the level of discursive practice, where my objectives are to: analyse the linguistic patterns over time; identify emerging and competing Discourses; explore what these Discourses represent as part of the GoK's sensemaking/giving efforts; and establish the impact on the relationship between the BWI and the GoK over time. Ultimately, based on the findings of this study, I suggest an arena model as an alternative to conduit models of communication where the crisis takes centre stage and the organisation, the public, the media and other stakeholders become active participants ih the crisis resolution process. I also make methodological contributions to Discourse Analysis by demonstrating that CLA can provide a more objective and standardised way of labelling discourses.