Predicting Austrian export growth via gravity equation

This thesis aims to predict export growth rates for Austria to OECD members by a gravity model that is analyzed by a fixed effects regression. The data is organized as a panel, thus it also includes time and displays not only a cross-section. The analysis of the predictive performance includes RMSFE as well as the derivation of a confidence band that shows the models quality. To support the proposed model, a Monte Carlo simulation is made to test if the statistical model is able to correctly estimate the parameters. It is then analyzed like the results of the prediction. The main finding of the thesis is that the prediction with this model is not too far off the observed data. Also, the simulation estimates the coefficient well and even with larger error terms the parameters are fairly close to the real value. The gravity model works really well and even with high variation in the error terms, the predictions can compete with other statistical models.submitted by Dunzinger Martin, BScUniversity of Innsbruck, Masterarbeit, 2018(VLID)308008