That is an enormously good point. It is a matter of bewildering complexity. Ofcom is over-layering four different models dependent on masts, terrain, topography and thickness of walls, and the reality is, as the hon. Gentleman says, that 90% of the time for 95% of the people is probably an overestimate of what we are currently getting.
Nevertheless, Ofcom states in its consultation document that it can see no benefits from extending the coverage further. In fact, it states on page 67 that the costs would outweigh the benefits. Why? Because it is worried about losing money in the auction—nobody knows how much—and is worried that when it tries to sell the radio spectrum, which it owns, to the mobile telephone companies and asks them to increase their coverage obligation from 95% to 98% these companies might pay less in the auction. Indeed, they may. It stands to reason they would pay less, but probably not as much less as Ofcom fears.
It may indeed stand to reason, but the evidence from past auctions of the spectrum does not show bidders producing bids while in any sense respecting the cost base of the project on which they are about to embark.
That is an excellent point. The reality of auctions is not that people operate on a fully rational basis, counting the number of their masts and then bidding exactly less than that. We have all participated in auctions. They are elaborate psychological procedures that are exactly designed to extract as much money as possible.