Most assessments still don't have one
131 impact assessments were published in 2026. Forty are post-implementation reviews, which use a different template and aren't expected to carry a scorecard. That leaves 91 forward-looking assessments that should — and 48 of them don't. The flagship transparency device of the reformed framework is absent from more than half the documents it was designed for.
The gap isn't random, and it isn't evenly spread. Roughly a third of the missing ones are Northern Ireland assessments, which run on a separate devolved template that predates the scorecard. The rest are Great Britain assessments that simply don't have one: some state explicitly that they sit outside the Better Regulation Framework, some are tax impact notes, and a long tail are older summary sheets that still print an NPSV and an EANDCB but no directional ratings at all. Nothing in the public listing distinguishes any of these from an assessment that does carry a scorecard — you have to open all 131 PDFs to find out, which is what we did.
It matters because the scorecard is the only place a department has to state, in public and in one word, which way each impact points. Where it's missing, that statement was never made.
Nothing is bad overall
Twenty-four of the 29 distinct assessments rate their own effect on total welfare as positive. Three say neutral, one says uncertain, one leaves the line blank. Not a single department, across a full year, published a scorecard saying its measure would make the country worse off. A scale on which the top box is ticked five times in six is not measuring much.
The whole cost lands in one column
Business is the only line where negatives appear at all — eleven of twenty-nine. Nine of those eleven sit directly beside a positive welfare rating. That pairing is the framework doing its job: it makes the transfer visible, and it says out loud that the case for these measures rests on gains landing somewhere other than the regulated firm.
Part B is mostly blank space
Eleven of the twenty-nine say nothing but neutral, uncertain or nothing at all across all three wider-priority lines. International trade is neutral in seventeen of twenty-nine. Where Part B does carry a signal it is almost entirely the decarbonisation line. The business environment and international questions are being answered, but they are not being used.
The same assessment is published several times
Fourteen of the 43 rows are repeat publications — one assessment attached to two or more statutory instruments. The higher education tuition fee assessment appears four times; the lifelong learning fee limits assessment three. Anyone counting published impact assessments without deduplicating is over-counting by about a third.
Departments aren't filling it in the same way
Part B has its own vocabulary — supports, may work for, may work against — and several departments answer it with Part A's positive/negative wording instead. One assessment merges two Part B lines into one; another drops the business and household rows entirely. The template is standard; the practice isn't yet.