This paper identifies and quantifies the contribution of a set of covariates in affecting levels and over time changes of happiness inequality. Using a decomposition methodology based on RIF regression, we analyse the increase in happiness inequality observed in Germany between 1992 and 2007, using the German Socio-Economic Panel (GSOEP) database, deriving the following findings. First, trends in happiness inequality are mainly driven by composition effects, while coefficient effects are negligible. Second, among composition effects, education has an inequality-reducing impact, while the increase in unemployment contributes to the rise in happiness inequality. Third, the increase in average income has a reducing impact on happiness inequality, while the raise in income inequality cannot be considered as a driver of happiness inequality trends. A clear cut policy implication is that policies enhancing education and economic performance contribute to reduce happiness inequality and the potential social tensions arising from it.