Economic theories as well as global financial governance (international financial institutions such as the IMF and the World Bank) and public policies have been confronted with several global crises (e.g., the global financial crisis/GFC). Yet the question remains debated as to whether these crises have induced 'genuine' changes in economic theories, institutions and policies or as to whether the latter have remained 'broadly' unchanged. Such a question, however, requires a definition of the concept of change itself. Indeed, only large and visible shifts à la Kuhn may be qualified as changes. In contrast, in a Hirschmanian perspective, transformations may occur via small departures and innovations within a consensus, via 'partial, limited, and pragmatic responses' that may be erratic adjustments in institutions and policies: here, mainstream theories have evolved since the GFC towards more openness to non-mainstream concepts (e.g., instability). In addition, the observation of change depends on the time spans considered (long-term historical approaches perceiving types of inertia or change that differ from short-term ones). In this context, the paper argues that a consequentialist approach of the concept of change has a greater explanatory power: observable changes in outcomes and causal mechanisms 'prove' changes in theoretical paradigms, institutions and policies. The example of the mainstream theories that have been used in the analysis of Sub-Saharan African economies and the policy reforms applied to them shows that changes have been limited. Theoretical paradigms, financial governance institutions and the associated policies claim to have changed since the GFC. Behind 'innovations' in theories and policies, however, outcomes remain poor, and their underlying causal mechanisms are stable since the 1980s-90s 'lost decades': i.e. theories relying on irrefutability mechanisms and absorbing other theories that could induce changes, and cumulative constraints generated by 'externalisation' processes (commodity-based structures; conditional lending to policies that are identical across time and space). More than a Hirschmanian transformation through small changes, a consequentialist perspectiveand, interestingly, consequentialism may precisely be viewed as a typical mainstream or IFI stance-suggests processes of lock-in and cumulative causation.< Réduire