Yesterday I attended the thesis defence of a student I have supervised. The dissertation is a collection of empirical essays in political economy that I hope will make their way to the publication circuits in the near future. One of the chapters was about how the memories of economic management in previous military regimes affected the support for democracy in Latin America. The main finding is that this benchmark does significantly affect perceptions of democracy, showing that reference points are important in political preferences, as we know from behavioral economics that they are in other fields. Another apparently very different chapter was about how the Catalan public TV channel had affected preferences for nationalism in the voting population. This influence is difficult to identify, as causality may go in different directions and there may not be enough variation to find statistical significance. However, by exploiting the fact that some territories started the channel before others probably for exogenous reasons, the author found that this channel, in the early periods (the only ones that can be analyzed using this technique) had an effect on preferences that was double in size than the effect in other similar studies. Finally, the third chapter was an event study about how the Catalan secession campaign of the last years had affected the stock returns of firms with interests in the region. In this case, the effects are very small and non-significant. I found very interesting that one of the very thorough members of the evaluation committee thought that the three studies were related because all of them analyzed preference formation following autocratic regimes. It was humbling to realize that many things that happen in Catalonia and Spain are the result of a country that just experienced a 40 year military dictatorship that finished just 40 years ago. Many of those who lived under that regime, in favour and against it, are still with us. The dark shadow of the Franco period still influences us in many ways, not always in the expected direction.