Good work! I'm getting my Ph.D. in economics, and mechanism design is one of my focuses. Not sure who's behind this, but what's the goals of the auction? There's no terribly good reason raising revenue should be one of the goals since it can be raised with other forms of taxation with less distortion; as you pointed out, a good bit of the incidence of this tax is going to be pushed onto mobile internet users in the form of higher prices, which is ultimately just another regressive tax most of us can't afford, like that on gasoline (which doesn't meaningfully reduce driving).
I would hope the goals would be to get the spectrum in the hands of the entity that will put it to its highest and best use for societal welfare. Profitability is some measure of this, and so the highest bidder may ultimately be the one able to create the most value, but making them pay as much as possible for the spectrum just ensures they will need to capture as much of that value for themselves as possible to pay for their bid.
I'm not sure about the technical details regarding spectrum and bandwidth in this case, but some kind of required-investment bidding would be a better approach. I believe that the theoretical max bandwidth is a function of type and width of spectrum (65 MHz of electromagnetic), but there's plenty of gains from capital investment to be had regarding hardware to run the networks and probably from signal optimization, depending on how they allocate their MHz across sections/channels (e.g. minimizing noise and interference, maybe some kind of dynamic/automatically re-sizing of spectrum divisions based on bandwidth needs of each section at any given time, etc). Bidding in $ of capital investment would be a decent approach since they have some motivation to make those $'s go as far as possible, but having some projected return on investment with respect to consumer experience would be nice. Unfortunately, that gives incentive to exaggerate one's efficiency, corrupting the usefulness of all such projections in the bid, and the projections themselves are hard enough to create that you'd be hard pressed to prove any misleading conduct beyond standard corporate optimism if they fall short.
Seriously, though, who the hell designed this auction, and why did they not consider any of these standard questions any auction designer worth her salt would have started with from the get go? Makes me more curious where the money from the bids will ultimately end up, since corruption is the only strong reason (I don't buy gov incompetence here).