quinta-feira, abril 19, 2012

French forecasting models

The April issue of French Politics (ungated) is dedicated to several fun articles forecasting the 2012 French elections. Nadeau, Lewis-Beck, and Bélanger, using simultaneous equations to deal, in the one hand, with the relationship between incumbent popularity and vote for the leftist candidates in the first round and, on the other hand, the relationship between unemployment, time in office, cohabitation, and incumbent popularity, predict that the Left will get a (narrow) majority of the votes in the first round. Nadeau and Lewis-Beck (again) and Didier then take a look at second-round vote as a function of presidential approval measured 4 months before the election and the image of the candidates, and suggest that Sarkozy is in rather bad shape for the second round, albeit not so bad as polls were suggesting a few months ago. Foucault focuses on legislative elections, using local data on unemployment, national GDP data, PM's popularity, local data again on previous electoral performance of the incumbent, and a couple of other specificities, and concludes that the Left will have a narrow majority of the votes (in the last issue of PS, Foucault and Nadeau also use local data, which is then aggregated and weighted to arrive at a prediction of a - narrow - Sarkozy defeat).  Evans and Ivaldi predict Marine Le Pen with around 17%. But Jerôme and Jerôme-Speziari, using regional data, actually give a very close advantage to Sarkozy. As one of the papers notes, "a victory in 2012 would certainly qualify Sarkozy as one of the most successful campaigners in modern times." I guess at least that is safe to say.


Enviar um comentário