Its not just selecting a piece of the spectrum. As well as the physics highlighted above it is the practicalities.
Most of the spectrum is allocated to existing uses, some of which has to stay (military) or would take time to clear (e.g Analogue TV, FM). Any change requires sufficient notice and sufficient time for the migration.
You cannot assume always that we can do without the old or that the new addresses all the functionality of the old. Also you get issues like in the US where they have just had issues using N77 for 5G and the interference with altimeters made to earlier specifications. So due diligence is always required.
The technology will determine what is the best to use initially but then this is then influenced with what is practical for harmonisation across countries which will keep any technology/device costs down. If a country goes alone then they may be saddled with more costly adoption.
We will be getting mmWave going forward but for general mobile I think its our fault we haven't got the coverage/speeds we need. The lack of sharing between providers, implementations of 3G, 4G and now 5G at primarily on existing GSM sites rather than optimised locations has meant a patchwork of mast coverage. Added to this the monetisation of the spectrum by the UK Government released with restricted non contiguous allocations.
I do wonder what it would look like if all providers shared a single infrastructure with taller masts located centrally to cover the required cells.
My view is that I don't think it is the mast that people dislike it is the eyesore of cabinets on our road sites (which are getting bigger). They should be suitably sited in an enclosure (like electric sub stations). But of course its free to use the pavement/verge.