As populist radical right parties have become electorally successful throughout Europe, mainstream parties have been adopting more restrictive immigration policies in order to win back voters, in what has been labelled a "contagion of the right". Sweden, however, has been a deviant case both in a comparative European context and in relation to influential theories of party competition. Despite the electoral threat from a growing radical right party (the Sweden Democrats) during the last decade, no other party engaged in any significant policy co-optation prior to the 2014 elections. In this paper I consider multiple explanations for this deviance derived from empirical and theoretical literature and mapped onto an analytical framework distinguishing between the party goals policy, office, and votes. The empirical material consists of survey data, statements from party representatives, and parliamentary voting patterns. I find the goal-oriented explanations to be only partially satisfactory, and go on to explore the possibility that the deviance can be explained by the institutionalisation of the Swedish cordon sanitaire – the commitment by all other parties to politically isolate the Sweden Democrats. The path dependency of the cordon sanitaire, I argue, became a behavioural constraint that effectively hindered parties from legitimately engaging in the co-optation of SD policies, until it was removed by the external shock of the 2015 "refugee crisis".