Export in different aggregations

The work concentrates on estimating determinants of Czech export. The key novelty of the research consists in estimating export flows at three different levels of aggregation (total, machinery, automobiles) and a subsequent comparison of results. An augmented gravity model is implemented for the empirical research and estimated with the use of Ordinary least squares (including time or country dummies) and Poisson pseudo maximum likelihood estimators. Consequently, the work proposes the comparison based on the Poisson estimates due to incontestable weaknesses of the OLS estimator. Besides, the models are specified based on time series and cross-sectional structures and the different obtained results prove the importance of identifying the panel data structure. The predicted models reveal that total and machinery exports follow the same tendencies, whereas the development of Czech automobile export differs and evolves on an idiosyncratic path. The analysis hence points out the inconsistency of crucial factors determining export at different aggregation levels.