Spectrum Floors in the UK 4G Auction: An Innovation in Regulatory Design

Abstract

This paper evaluates the regulatory approach to competition issues in the UK 4G spectrum auction which involved the innovative use of spectrum floors, the flexible reservation of portfolios of spectrum for either a new entrant or the smallest incumbent national mobile competitor (H3G). Spectrum floors have two dimensions of flexibility: different portfolios of spectrum can be reserved for different players, eg depending on their pre-auction spectrum holdings; and the choice of spectrum to be reserved from a range of portfolios, each of which is sufficient to promote competition, is decided through the auction as the floor that minimises the opportunity cost of reservation. The use of spectrum floors appropriately emphasised output efficiency through promoting downstream competition. But its implementation in the auction sought also to maximise auction efficiency (subject to the constraint of the floor) through sophisticated modifications to the already complex design of the combinatorial clock auction. This approach represented a balance between the risks of market and regulatory failure. On the one hand, H3G or a new entrant might fail in the auction to acquire the spectrum required for promotion of downstream competition because of asymmetries in intrinsic value compared to social value or strategic investment by the larger incumbents (EE, Telefonica and Vodafone) to deny spectrum to competitors. On the other hand, using the traditional approach of spectrum set-aside there might be a high opportunity cost of reservation through choosing the wrong spectrum to be reserved. In the UK 4G auction held in January to February 2013 the full potential of spectrum floors was not, in the event, tested because there was no competition between bidders eligible to obtain reserved spectrum - H3G was the only such bidder willing to pay the reserve price of the spectrum floors. However, there was still competition over the choice of spectrum floor for H3G, which resulted in an outcome that some considered surprising. H3G won the spectrum floor of 2x5MHz in the higher-value 800MHz band, ie scarce sub-1GHz spectrum, rather than 2x20MHz in the higher-frequency 2.6GHz band. By analysing the bid data, a comparison of these two floors shows that the marginal value in H3G’s package bids of £165m exceeded the overall opportunity cost to other bidders of £107m. This outcome was not a fluke or an aberration as it clearly reflected the consistent pattern of bids made in the auction. In particular it reflected H3G’s strategy of bidding marginal values for the floor packages equal to the differences in their reserve prices, which guaranteed that it would not pay more than the reserve price for its winning spectrum floor; and EE’s consistently aggressive bidding at the margin for 2x20MHz of 2.6GHz compared to 2x5MHz of 800MHz. Based on the bids made in the auction it was clearly more efficient for H3G to win the spectrum floor of 2x5MHz of 800MHz, even if this might not have been the prediction before the auction. With the important caveat that some of the bidding may have been strategic and deviated from bidders’ true intrinsic values, this suggests that one of the key elements of flexibility in the new regulatory tool of spectrum floors proved to be important the first time it was used in the UK 4G auction. This flexibility avoided the regulatory failure that could have occurred with simple spectrum set-side of the regulator reserving less efficient pre-specified spectrum in the false belief that would minimise the opportunity cost of spectrum reservation.