28 starting a family for most, not sure the grant will stretch. Great picture of university life but think we may miss the Dr, scientists, chemist's, engineers etc.
There is a link between well educated workforce and productivity
The first part of the idea is that everybody can take up to four years of funded higher education, once they have made enough qualifying tax contributions (or whatever the equivalent is for the unemployed - separate discussion, also relevant to qualifying for pensions, benefits etc).
At 28, you might decide to have kids. Or you might decide to cash in your higher education credits. You might well want to do both - after all, you'd have been working and saving for ten years by that point. You might wait until you were forty, or until you retired.
And this is mainly for theoretical work: arts and pure sciences.
The second part of the idea is to reexamine the link between education, training, research and qualification. We know that a well educated workforce is productive. In some cases, there's a link between education (in the sense of training) and productivity. Engineers, as you say, need to learn some fundamental principles before they are let loose on bridge design.
Quite why a sociology graduate is going to be better at software sales is unclear. For the software vendor, the degree evidences a few soft skills, and if the candidate didn't have a degree, he or she would be suspect - because nice, reasonably intelligent kids are meant to have degrees these days. The well-educated workforce may be productive, but quite often it's
post hoc ergo propter hoc to argue for a causal link.
So. As you say, we clearly need doctors, chemists and so on. But medical training is already pretty much vocational; it's usually centred around a teaching hospital. It's essentially an NHS apprenticeship delivered in an academic/industry partnership. We don't notice, because the industry in this case is public sector. Why not do the same thing with pharma, civil engineering and so on? Let industry teach the basics, outsourcing some of the learning to academia if they want.
Under this model, a 30 year-old bridge designer would already have a couple of years of relevant, academically-delivered professional training under their belt. At this point, they might spend a couple of state-funded years doing a combination taught and research course in some wider aspect of civic design. Then perhaps, as a 55 year old, sick to the back teeth of bridges, but financially independent, they would spend another two years doing an intensive sociology degree.
A problem with this argument is that there's no incentive for civil engineering firms to train apprentices if there's a global supply of civil engineering graduates. So either this model applies globally, or we have quite a complex series of corporate tax breaks and incentives to make it work.
But it does seem to me that the current dynamic undervalues academic institutions by treating then as a vocational training providers or - worse - as finishing schools for upmarket service economies. My suggestion is just one of the ways in which we could do this better.