Models and Data in Finance: les Liaisons Dangereuses

Why Finance Needs Philosophy (and Vice Versa): Some Epistemic and Methodological Issues

“…That is, a model becomes a way of intervening and not simply of describing. When this happens, a sort of 'reverse finance' emerges (see Ippoliti, 2019) and we could start from the model and then try to produce the corresponding 'reality', or better, we could start from the wanted outcome, the one predicted by the model, and then try to engineer as much as possible (with norms, rules, technologies) the hypotheses of the model needed to produce that specific outcome so that the market will behave as the model predicts. Several financial tools, for example algo-trading or high frequency trading, could be viewed in this way, that is, as a (legal) way to obtain certain outcome in this reversing fashion.…”

“…Of course, not every model can be reversed in order to align a market with its expected outcome, but this possibility increases under certain conditions (see also Svetlova, 2012;Svetlova & Dirksen, 2014 on this point). For instance, it increases by means of two factors: the time-lapse between when the model makes its prediction and when the predicted event should happen, and how much a market is run by algorithms (see Ippoliti, 2019).…”

Introduction: Philosophy for Finance

“…Financial phenomena offer interesting material for philosophical debate-for instance in ontology, methodology, epistemology, and theory-building. Just to mention two examples, finance raises interesting questions about the relationship between data and hypotheses (Ippoliti 2019) or the relationship between the three fundamental notions such as explanation, prediction and control ( §5).…”

“…This property makes theory and practice so intertwined in finance that one can operate on the theoretical product, a model, to adjust the world (Boldyrev and Ushakov 2016): there is a continual and mutual influence of practice and theory in financial systems. This interplay opens the way to a specific scenario that can be labelled 'reverse finance' (Ippoliti 2019).…”

“…In effect, the nature of the relations between models and data, reality and theoretical product-the liaisons dangereuses (Ippoliti 2019)-paves the way for possible manipulation of stock markets through an appropriate use of models. In fact, since neither data nor models, at least in certain circumstances, are simply representations or descriptions of a system, the reversal becomes possible and the boundaries between performativity and manipulation get very thin in certain circumstances, especially at micro level or timescale.…”