We consider the same sample of countries (eight OECD countries) as Berlemann and Marwardt (2007) with a close data base (1963-2005 or shorter period according to the available data for the monthly unemployment rate). If we consider the sample group made up by eight countries, the estimations show favorable results for the unemployment for the partisan rational theory by taking into account the degree of the electoral surprise. We show that if we make estimations by country the obtained results for the unemployment are very different according to countries: for example, for the United States, the results are favourable to the rational partisan theory but not for France notably. In comparison to Berlemann and Markwardt (2007), we also make estimations with various partisan variables (with and without taking into ac- count the degree of the electoral surprise) to test the rational partisan theory for the unemployment. The model of Heckelmann (2006) does not give favourable results either for the whole sample of eight countries or for every country.
Keywords : rational partisan political cycles, electoral surprise, electoral uncer- tainty, unemployment, OECD countries (8)
JEL Classification : C22, C23, D78, E32