The UK approach is highly interesting. It's assuming that a government telling people to social distance themselves is highly effective. You then direct the elderly to social distance themselves; they are mostly protected from the virus; it runs rampant through the general population; and after a two months of relatively mild, mass cases among the young, you've effectively built a herd immunity firewall around the elderly.
Three possible issues:
1) Telling people to social distance themselves has a weak effect, not a strong one. In that case, the fire burns through the young population and burns through the old population, and you end up with a pile of flaming corpses.
2) I don't know if herd immunity of the young population is enough, epidemiologically, to give herd immunity to an older population. Old people can pass it to each other, and their day to day interactions are disproportionally among the elderly. So you risk an older subgraph all catching it at once still. I don't think this is a big risk compared to 1).
3) Individuals don't develop an immunity. This would be disastrous for this strategy. Then again, it's disastrous for all strategies.
The first issue seems like the biggest issue to me, and I think it's pretty damning. If that's the case, the elderly never have the protection of herd immunity, and you're not flattening the curve but sharpening it.
Another possibility: the government knows it won't work to seriously change the number of deaths, but it also believes all measures possible in Western countries are likely to fail. In that case, you immolate the elderly population as fast as possible and then get used to a new norm instead of spending months in a futile battle. Essentially, it's possible that even with heroic efforts that the curve can only be flattened a bit, so it's just a matter of timing the deaths and how long you want to be enduring terrible economic damage. And if every other country suffers pretty much as much as you do, political impact is limited.