1 Paper presented at the 21 st annual conference of the International Association for Feminist Economics (IAFFE), 27-29 June 2012, Barcelona. What about the 99 percent? Gender, class and promotions in Belgium Nick Deschacht Assistant professor Faculty of Economics and Business Hogeschool-Universiteit Brussel (HUB) Warmoesberg 26 – 1000 Brussel – Belgium E-mail: Nick.Deschacht@hubrussel.be LinkedIn: http://be.linkedin.com/in/NickDeschacht Website: http://hubrussel.academia.edu/NickDeschacht 20 June 2012. Draft version. Comments are highly appreciated. Please do not cite. Abstract This paper quantifies the size of the gender discrimination effect in promotions. The results indicate that women near the top face fewer obstacles than women in the lower or middle career levels. The largest obstacles to promotion are found among lower-ranking white-collar workers where men are 72 percent more likely to be promoted than women with the same characteristics. To quantify obstacles to promotion we introduce a method that applies the Oaxaca-decomposition of wage gaps to promotion gaps. By using panel data of a representative sample of the Belgian labor force, we measure how employees acquire workplace authority over time and control for productivity related characteristics and sample selection bias. To intepret our results we propose a theory that explains gender discrimination as a result of statistical discrimination in a winner-take-all market for promotions. We conclude that social policy is too focused on top positions and is not likely to affect the careers of the majority of working women and men. 1. Introduction The debate surrounding the gender career gap is heavily focused on the lack of women in top positions. One European country after another is implementing quota for the composition of the boards of directors of large companies, following the example of Norway in 2003. European commissioner Viviane Reding recently applauded these policies for mobilizing female talent and proposed to adopt them on a European-wide level.