2015, Stankiewicz, Sandra

I use the adaptive elastic net in a Bayesian framework and test its forecasting performance against lasso, adaptive lasso and elastic net (all used in a Bayesian framework) in a series of simulations, as well as in an empirical exercise for macroeconomic Euro area data. The results suggest that elastic net is the best model among the four Bayesian methods considered. Adaptive lasso, on the other hand, shows the worst forecasting performance. Lasso is generally better then adaptive lasso, but worse than adaptive elastic net. The differences in the performance of these models become especially large when the number of regressors grows considerably relative to the number of available observations. The results point to the fact that the ridge regression component in the elastic net is responsible for its improvement in forecasting performance over lasso. The adaptive shrinkage in some of the models does not seem to play a major role, and may even lead to a deterioration of the performance.

2014, Brüggemann, Ralf, Glaser, Markus, Schaarschmidt, Steffen, Stankiewicz, Sandra

We investigate non-linearities in the stock return - trading volume relationship by using daily data for 16 European countries in an asymmetric vector autoregressive model. In this framework, we test for asymmetries and analyze the dynamic relationship using a simulation based procedure for computing asymmetric impulse response functions. We find that stock returns have a significant influence on trading volume, but there is no evidence for the influence of trading volume on returns. Our analysis indicates that responses of trading volume to return shocks are non-linear and the sign of the response depends on the absolute size of the shock. Thus, using linear VAR models may lead to wrong conclusions concerning the return - volume relationship. We also find that after stock markets go up (down), investors trade significantly more (less) in small and mid cap stocks, supporting evidence for the theories of overconfidence, market participation, differences of opinion, and disposition effect.

2015, Stankiewicz, Sandra

2013, Barsoum, Fady, Stankiewicz, Sandra

For modelling mixed-frequency data with business cycle pattern we introduce the Markovswitching Mixed Data Sampling model with unrestricted lag polynomial (MS-U-MIDAS). Usually models of the MIDAS-class use lag polynomials of a specific function, which impose some structure on the weights of regressors included in the model. This may deteriorate the predictive power of the model if the imposed structure differs from the data generating process. When the difference between the available data frequencies is small and there is no risk of parameter proliferation, using an unrestricted lag polynomial might not only simplify the model estimation, but also improve its forecasting performance. We allow the parameters of the MIDAS model with unrestricted lag polynomial to change according to a Markov-switching scheme in order to account for the business cycle pattern observed in many macroeconomic variables. Thus we combine the unrestricted MIDAS with a Markov-switching approach and propose a new Markov-switching MIDAS model with unrestricted lag polynomial (MS-U-MIDAS). We apply this model to a large dataset with the help of factor analysis. Monte Carlo experiments and an empirical forecasting comparison carried out for the U.S. GDP growth show that the models of the MS-UMIDAS class exhibit similar or better nowcasting and forecasting performance than their counterparts with restricted lag polynomials.

2015, Barsoum, Fady, Stankiewicz, Sandra

For modelling mixed-frequency data with a business cycle pattern, we introduce the Markov-switching Mixed Data Sampling model with unrestricted lag polynomial (MS-U-MIDAS). Usually, models of the MIDAS-class use lag polynomials of a specific function which impose some structure on the weights of the regressors included in the model. This may lead to a deterioration in the predictive power of the model if the structure imposed differs from the data generating process. When the difference between the available data frequencies is small and there is no risk of parameter proliferation, using an unrestricted lag polynomial might not only simplify the model estimation, but also improve its forecasting performance. We allow the parameters of the MIDAS model with an unrestricted lag polynomial to change according to a Markov-switching scheme in order to account for the business cycle pattern observed in many macroeconomic variables. Thus, we combine the unrestricted MIDAS with a Markov-switching approach and propose a new Markov-switching MIDAS model with unrestricted lag polynomial (MS-U-MIDAS). We apply this model to a large dataset with the help of factor analysis. Monte Carlo experiments and an empirical forecasting comparison carried out for the U.S. GDP growth show that the models of the MS-U-MIDAS class exhibit nowcasting and forecasting performances which are similar to or better than those of their counterparts with restricted lag polynomials.