This thesis presents a theory and analysis of the role of imagination in organizations. More specifically, the work exposes how central traditions of interpretative research within organization studies have overlooked the phenomenon of imagination and, thus, it aims at narrowing this gap, by empirically investigating the role of imagination in the risk management practices of a Brazilian conglomerate. With a long tradition in the sugar-ethanol sector, this company – here called Group – has been diversifying its business in order to position itself as a global player in the energy and infra-structure sector. This change involves the articulation of idiosyncratic notions of risk and opportunity within different levels of the conglomerate. Using a qualitative research strategy and an interpretative method, the thesis explores a narrative of risk management within the three levels of the conglomerate: its corporate level, a business here called Energy and a business named Land. Accordingly, the constitutive and reproductive aspects of risk management in these levels were highlighted. Constitutive refers to the moments in which the signification of risk led practitioners to engage in transforming their own practices; reproductive refers to the routinization and dissemination of such practices within the company. By casting light on the intersections between these aspects and imagination, the thesis describes risk management as an imaginary institution. This description adopts a dialectical phenomenological approach to institutions developed from the works of Cornelius Castoriadis (1987). Following this perspective, risk is a social imaginary signification working as an institutional resource that managers use in order to introduce new regimes of practice. Findings point to three emergent imaginaries of risk management in the conglomerate, along with, respectively, different logics: risk management as the exploration and exploitation of new uncertainties; risk management as a cultural collusion; and risk management as a techno-logical production of images. Implications of these findings to both organizational and risk management theory are explicated. First, the findings imply that rather than logics being a central mechanism of organizing (Thornton et al., 2012), organizations have at their core imaginary significations which enable and determine multiple logics. This assumption challenges the dominant understanding that institutional logics shape action in organizational fields, because logics stem, in this case, from imaginary institutions. Second, the thesis shows that, while enacted by domains of practice, risk is necessarily deferred within discourses and practices. Through this deferral, the imaginary of risk incorporates novel and unpredicted elements into itself contributing to changing the very way risk is understood within the conglomerate. This leads practices of risk management to be revised, changed or supplemented. Risk management assumes, in this case, a less monolithic aspect. As the thesis makes explicit, the managers and analysts of the Group and the other two businesses are not being colonized by risk (Rothstein et al., 2006), but are instead colonizing through and within the imaginary signification of risk. The thesis draws, finally, two associated conclusions regarding the role of imagination in organizations. On the one side, managers use imagination to produce figural orderings. These refer to the many forms of accounting in organizations (i.e. through figures, images, sets, etc.) that are constructed and differentiated in order to give support for action. Rationality emerges here out of the practices, techniques, language and tools used to organize action around spatial images and the recombination of imaginary locations. On the other side, however, imagination provokes (re)presentations in organizations, that is, imagination pushes practitioners to constantly re-interpret established images and to construct new representations beyond available logics of organizing. Individual imagination then intertwines with the practices, objects and technologies constructed in and for organizations, through the non-technical aspects involved in organizational action and change. Practices in organizations are, according to this view, always fragmentary and must rely on the continuous expansion of action within action itself. Organizations are, therefore, socially and imaginarily constructed through the putting in place of particular techniques and languages – the logics of talking and doing things – according to interests and projects that underpin and are underpinned by social imaginary significations.