Regulators are starting to analyse (or at least they are talking about it) model assumptions. We all know that too often our assumptions are just in the model documentation. Furthermore, we have no standard language for talking about assumptions. Indeed a given assumption within an organization that applies to more than one model may be described different by different developers – so at first sight it is not evident that it is the same assumption.
Does anybody have views of standard language for assumptions (is there a catalogue). Does anyone have examples whereby the same assumption is referred to in many ways? Your views on this topic would be appreciated.