Le-Yu Chen, Ekaterina Oparina, Nattavudh Powdthavee, Sorawoot Srisuma, 18 March 2019

Recent critiques of wellbeing research have shown that mean comparisons of reported and latent happiness across groups are valid only under strong assumptions that are usually rejected by the data. This leads to scepticism over whether econometric analysis of wellbeing data can be used to inform policy. This column suggests using the median rather than the mean, because the median ranking is stable across all increasing transformations. When focusing on the median of wellbeing data, the Easterlin Paradox still holds.

Federica Liberini, Eugenio Proto, Michela Redoano, 15 November 2013

Retrospective voting – voting for incumbents if one’s situation has improved under the politician’s watch – is a well-established pattern. This column shows that this pattern also applies when ‘improvement’ is measured by a subjective measure of well-being. Among the stark results discussed is the finding that newly widowed women are 10% less likely to be pro-incumbent than the control group.


CEPR Policy Research