About this project
In the light of the financial crisis it has become clear that the world economy can be hit by large shocks and that many economic relationships are varying over time. Many of the models in use today are not well suited for situations where data of this type prevails. Consequently this project aims to contribute to the development of models that are better suited for this and use them to analyse economic and financial issues.
It is not trivial to extend and modify econometric models so that they are well adapted to data of this kind. In some classes of models the problems are close to being overwhelming. One kind of models which is relatively well suited for this is Bayesian VAR models (BVAR models). At the same time they are the standard tool for macroeconomic research and has many uses in financial economics.
This project thus aims at developing and extending BVAR models so that they are better suited for analysis of time varying relationships in combination with more realistic assumptions about the distribution of the error terms. The models will then be used to analyse and forecast macroeconomic and financial phenomena.