Licensed backhaul using existing radios is already far cheaper than engineering new radios and antennas for 5G. 6,11,18ghz 1gbps radios can be had for <$20k from Exalt (for a complete link), and I am sure Ubiquiti will have a 5-10k link out in the next year or two.It could/will be used for backhaul connections for towers theoretically, lowering the cost *significantly* of new towers, which currently require expensive fiber connections, which would "trickle down" higher speeds to both congested areas (too many people per tower) as well as rural areas (too much area per tower).
If you could theoretically triple the number of cell towers in both rural and congested urban areas, overall service would greatly improve, while not costing more to the provider. Backhaul is one of the biggest costs insofar as setting up a new cell site. (or so I've read)
I believe the problem is self interference or over utilization of existing spectrum, not necessarily tower density. That is why refining the existing spectrum would be important (or moving forward with LTE-U and doing it that way). I see wifi offloading or soho routers with 5G microcells as a possibility though.