The politics of modelling

Models are ruled by a politics whereby the epistemic authority borrowed from mathematics offers to modellers a considerable power to make their voice heard in the making, evaluatingand monitoring of policy. This may lead to a misuse of models, in that they are used to de-politicize issues that are political. Paradoxically, this type of political use of models is at its greatest when models can be presented as apolitical, based on pure facts and their scientific manipulation, thus veiling the normative assumptions that necessarily underpin any given modelling activity. I illustratewhat is peculiar or exceptional in the use of models with examples from various fields, including epidemiology, ecology, economics, and climate change, and reconnect this analysis with existing historical and sociological literature on quantification in general and models specifically. I also consider briefly what existing guidelines and checklists for responsiblemodelling can and cannot do, cautioning about the medium- and long-term dangers associate with model misuse.