But is there any financial justification for a charity shop paying to stock it?

There is financial justification if they can actually sell it. And of course not if they cannot sell it. We don’t get a crystal ball. It’s a judgement call to predict whether they can sell it. They have opted to be risk averse. The decision was likely made without awareness of pre-spychip characteristics and without considering whether it can be marketed as such. It would be interesting to know if a shop has actually tried that, which can confirm that pre-spychips don’t sell even when marketed as such.

One of the problems is that the shop cannot handle supporting the public on linux. Independent of hardware performance, they abandoned linux in the store and returned to only selling Windows machines. Their staff turnover was too high to keep training staff on giving linux support.

Plus the energy cost to run it if you do buy it.

Energy cost is the same. A 65w PSU that is pegged will draw 65 watts regardless of what it drives. If the hardware is not being heavily tasked and the platform is lean, then there is only a difference on responsiveness but not on energy. Furthermore, some of these pre-spy chips have a TDP of 17w, which is interesting enough that recent machines are still being produced with them:

https://cputronic.com/index.php/cpu/amd-a4-4355m

A 17w TDP means it’s efficient enough to be fanless, or to not need to spin a fan often. It’s foolish to let laptops with 17w APUs get tossed when the very same CPU is still going into modern laptops precisely because of their efficient /for the money/.

Some of the higher TDP chips from the same family are notoriously inefficient. So sellers and buyers need some awareness. In the worst case, an inefficient chip is not so inefficient in a cold region or cold season. We cannot say from 10,000 feet whether such instances would be inefficiently deployed when speaking independant of the regional climate.