Unity in diversity : studies on micro and macro panel data sets

Thanks to the progress of information technology, keeping track records of a multitude of people or entities has become a matter of routine. For instance, marketeers keep files of potential customers, employers of employees, insurers of insured, investors of companies etc. The data concern a number of different but qualitatively similar units, which have been observed over some time. The topic of this dissertation is the methodology and practice of the analysis of such multi-unit or panel data. Two micro-economic and two macro-economic applications are presented and discussed; the micro-economic ones use a large sample of households or individuals, the macro-economic ones involve a small or less small number of countries, each observed for a few years. What our results show is that relationships that are often taken for granted are in reality very fragile and may be based largely on preconceptions.