////…but boils down to people who earn (significantly) more than the UBI rate simply having their UBI taxed right back,//
So not quite so “universal” as the name suggests then? //
I don't think that this would undermine the idea that it's a UBI. After all, the point is to ensue that nobody gets below that amount per month. It stands to reason that those getting (significantly) more than it might not see a literal x-hundred-a-month increase in their pay in order to make the scheme affordable. It's just an extension of how the usual welfare state "deal" works: from each according to ability, to each according to need. That would be as true for UBI.
// So tell me then, Clare, if there is to be a net reduction in the welfare bill of (just using your numbers for argument’s sake) 7% of GDP, who will bear the brunt of that reduction. State pensioners, of course will be more than happy as this proposal will double their existing pension ... //
This is presumably where details matter as to how to design the existing scheme, what level to set the UBI at, etc. In this sense, as I hope I'd made clear later in the same answer, my numbers were illustrative only rather than a fully-costed proposal: the main point was to show that the effective cost is not literally the same as (total adult population)*(UBI rate), but rather a fraction of that, and perhaps even less than -- or, at any rate, comparable to -- current spending on welfare.
// Why run a trial? Of course it will work.//
Depends, again, on how you define "work[ing]". As shown in the various links I provided earlier, you could -- and should! -- test the change in various key measurements, including but not limited to healthcare outcomes, educational levels for children in the test sample, etc., but also effects on whether these people work, what jobs they enter, if any, and its effect on the job market.
untitled has pointed out one possible source of bias in this particular scheme, and another is that this scheme is likely far too small to draw robust conclusions. For that, I'd say, you really need a city or a province, or better yet a country, and test the effect over a long period of time. Not easy at all. I said earlier that the only nationwide scheme anywhere seems to be in Iran, which introduced a flat payment to all citizens of c.$40/month -- equivalent, by using the "29% of median income" rule, to about £800/month in the UK.
My main point is that, while it's clear that details will matter and the scheme will need to be carefully-designed, the idea should be at least taken seriously, and explored on its merits, rather than written off as "lunacy".