Final Activity Report Summary - AUTOFIM (Automated financial modelling)
Most financial models are technically demanding and therefore a challenge to estimate, use and maintain. In particular, automated modelling is infeasible for most types of models if many variables are considered from the outset, as in "General-to-Specific" (GETS) modelling. The objectives of the project 'Automated financial modelling' (AUTOFIM) were to implement and evaluate the usefulness of automated value-at-risk modelling and automated derivative price modelling within a GETS modelling framework.
The most important scientific achievements of AUTOFIM are three. First, a very general and flexible model that enables automated financial GETS modelling has been developed and studied. Second, empirical evaluations strongly suggest that automated financial GETS modelling can be useful in derivative pricing and in value-at-risk analysis. Finally, a methodological study suggests that the notion of financial variability is indeed a feasible study object. Contrary to a widespread view among academics, it is feasible to evaluate volatility - i.e. a prediction of financial variability - in terms of their forecast precision of variability.
The most important scientific achievements of AUTOFIM are three. First, a very general and flexible model that enables automated financial GETS modelling has been developed and studied. Second, empirical evaluations strongly suggest that automated financial GETS modelling can be useful in derivative pricing and in value-at-risk analysis. Finally, a methodological study suggests that the notion of financial variability is indeed a feasible study object. Contrary to a widespread view among academics, it is feasible to evaluate volatility - i.e. a prediction of financial variability - in terms of their forecast precision of variability.