The lesson learned over and over in the West, is that any add-hock development of the spectrum results in huge future inefficiencies with large blocks of spectrum locked out of re-use due to poor planning in the early days.

The temptation is to let companies use new frequencies wherever they want, but in the longer term this results in chaos, and the only solution is to rip it all up and start again (at huge cost and disruption to service).

The only workable approach is to design an interlocking network of spacial and frequency allocations and then stick with it, and make sure that each block is fully utilised before moving on to a new block.

With all of the above considerations, the cost of the actual network hardware is negligible, especially if the consequence is that poor villages will have to abandon their total investment in equipment at some arbitrary point in the future.