The term ‘Westminster model’ is frequently used by political scientists and practitioners. But our recent examination (Russell and Serban 2021), published in this journal, uncovered wide variation in how it is defined – demonstrating that it is more ‘muddle’ than ‘model’. This sparked a response article from Flinders et al. (2021), which we in turn respond to in this piece. We briefly revisit our initial research questions, methodology and findings, before reflecting on their analysis. We emphasize that, notwithstanding the critics’ negative tone, we and they agree on certain fundamentals. In particular on our original central point, that the ‘Westminster model’ is an ill-defined term with a long history, which mostly makes it unsuited to positivist comparative political science research. While the ‘Westminster model’ may have some valid vestigial uses, within the UK political system, or as an object of study for interpretivist political science, it does not provide a rigorous basis for case selection and comparative political science analysis.

The term ‘Westminster model’, widely used in both the academic and practitioner literatures, is a familiar one. But detailed examination finds significant confusion about its meaning. This article follows Giovanni Sartori's advice for ‘reconstructing’ a social science term whose meaning may be unclear through review of its use in the recent literature. It finds that many authors in comparative politics use the term ‘Westminster model’ without definition, while those providing definitions associate it with a large (and sometimes conflicting) set of attributes, and a set of countries often not demonstrating those attributes. Some have sought to respect this diversity by proposing variants like ‘Washminster’ or ‘Eastminster’, while others suggest that the term should be seen as a loose ‘family resemblance’ concept. But on examination it no longer meets even the – relatively weak – requirements for family resemblance. To end the muddle, and the risk of flawed inferences and false generalization, comparative scholars should drop this term, and select cases based on more precise attributes instead.