'...which enables them to keep their REF positions'. This erroneously connects REF performance to the presence or prominence of quantitative work. Essex did very well last time, but Strathclyde did not. Sheffield ranked either the same or better than Essex last time, with no discernible quant presence. Quant work does not necessarily do better in the REF, an exercise which does not reflect the preferences or fashions of American political science.
If only there were systematic methods to test this. Oh wait, not in your version of reality. Essex did very well, Sheffield gamed the rankings, and Strathclyde started hiring more quant folks only after the last REF (though I seem to recall they were already in the top 10 last time). The next REF will tell if that leads to an improved position. I predict yes. Their chair is a former Essex dude.
I'm pretty sure Strathclyde were not in the top ten last time. The top ten departments were all pretty mixed in terms of their profile, which was the main point I was trying to make. There is no evidence to suggest that departments that lean more towards quant methods perform better in the REF and, if anything, Essex looks likely to slip in 2021 for various reasons that have been rehearsed here already.