Aims - Beneficiaries
The outcome of the project will benefit econometricians and statisticians inferring about non-gaussian models and empirical macroeconomists working with the theory and practice of forecasting and evaluation of uncertainties. The common theme here is the application of the newly developed tempered stable distributions to modelling and forecasting.
Econometricians and statisticians will benefit from our results on the asymptotic and finite sample properties of the parameters estimation and hypotheses testing in the context of non-gaussian distributions. The results might also stimulate other researchers to develop the methods further and apply them to solving other non-gaussian problems.
Success of the project will benefit empirical macroeconomists who will receive new methods of probabilistic forecasting, which explicitly permit the assessment of extreme events. Approximation of uncertainties through allowing for their dependencies and deriving these approximations from tempered stable distributions will open the door to new, computationally feasible, ways of constructing such forecasts.
Our methodology will allow for changing the emphasis from solving analytically complicated nonlinear models to stochastic simulation. This will result in a new method of constructing the 'fan charts' representing probabilistic forecasts. As the appropriate computer programs will be prepared and disseminated to the academic community, further inference into forecasting problems will be more straightforward.