|Economic downturns give rise to unexpected employment shocks that can reshape the distribution of population income, and hence produce a “middle-class squeeze”. However, there is limited empirical evidence testing the latter. This paper aims at testing the ‘middle-class squeeze’ hypothesis drawing from unique data from the Luxembourg Income Study (LIS) for several years, including the period of the Great Recession, and the Integrated Values Study (IVS) obtained by merging data from the World Value Survey (WVS) and the European Values Study (EVS). We examine the association between changes in unemployment in a recession drawing upon a heterogeneous set of both income and middle-class definitions as well as an extensive list of controls and different recession periods. Our findings suggest no robust evidence that recessions produce a middle-class squeeze, though they increase the share of the population regarding itself as ‘middle class’. The effect is heterogeneous to the baseline unemployment at the time of a recession, country spending on social protection and middle-class measurements and definitions. However, when we restrict our analysis to the recent Great Recession, we do find some evidence of a ‘middle-class squeeze’.