Advanced Regression Methods for Macro-Economic Indicators as Explanatory Variables

This project aims at improving the insights on the effects of macro-economic indicators on economic ratios and the prediction of economic ratios based on these indicators. The advantages of using advanced regression methods to model these interrelations will be investigated. Focus will be given to penalized regression and Bayesian model averaging and these methods will be improved and adapted to suit the specific needs in applications in finance and economics. The inclusion of lag variables as possible predictors for example requires to account for correlated independent variables as well as to develop suitable methods for determining variable importance. In addition methods developed for linear regression will be extended to generalized linear regression models. In particular economic growth and firm failure rates will be investigated.