AbstractWe conceptualise and measure welfare state change across 21 high‐income countries as a continuum delineated by a double movement, that is, the combined change of compensatory and employment‐oriented policies. Our double movement framework readapts Polanyi's concept into the context of welfare state change. We analyse this double movement across four decades using Principal Component Analysis and a new indicator that compares spending in 2015 to maximum and minimum spending levels since the 1980. We contribute to the literature in three ways. First, we empirically document an overall change in spending for compensatory and employment‐oriented policies, with the latter becoming more prominent over time. This change is more pronounced in the 1990s and even more so in the 2000s, and partially reduced classic regime differences. The PCAs generate a Cartesian space where each country is positioned across time (the 1980s, the 1990s, the 2000s and the 2010s) and space (within four quadrants, i.e., ‘the strong employment‐oriented space’, ‘the weak employment‐oriented space’, ‘the strong compensatory space’, and ‘the weak compensatory space’). Second, we develop a fivefold taxonomy of welfare state change characterised by: (1) retrenchment in Canada, Germany, Ireland, Netherlands, Sweden and the United Kingdom; (2) abridged adaptation in Australia, Belgium, Finland, Norway, Spain and the United States; (3) minor expansionary adaptation in Greece, Italy, Japan, Portugal and Switzerland—where spending levels were low in the 1980; (4) adaptation with an equilibrium between the two movements in New Zealand and (5) strong expansionary adaptation in Austria, Denmark and France. Overall retrenchment and abridging adaptation prevail over expansionary adaptation—this is due to cutbacks of unemployment, family allowances and active labour market programmes not being counterbalanced by the expansion of childcare. Third, we critically interpret these changes, introducing the double movement concept into comparative social policy.