Too much emphasis is put on maximizing revenues; for govt services, emphasis must be on raising societal welfare
Given the potential benefits to be got by getting an auction structure right – to be able to get as much of the ‘consumer surplus’ as possible – it is not surprising that Paul Milgrom and Robert Wilson are the winners of this year’s Sweriges Riksbank Prize in Economic Sciences (Nobel prize) for their work in auction design. In 1994, the FCC in the US used their methodology to conduct its first auction and rake in $617 million from spectrum sales. Instead of the policy of receiving bids in a sealed cover and then allocating the resource to the highest bidder or using a lottery system to allocate spectrum, Milgrom and Wilson’s approach involved companies simultaneously bidding on the spectrum in multiple areas to determine the best value. Nor is it surprising that one of the winners – Milgrom – was one of the advisor’s for Google’s IPO. Allowing multi-stage bidding is a good way to get more value as every participant gets more time to match/outbid the previous highest bid; to that extent, it is better price discovery.
In India, however, the model has wrought havoc on the finances of telecom companies – government greed made it worse – so the model needs to be finessed if it is to work. One possibility is to use the ‘second-price’ auction theory of another economics Nobel, William Vickrey, to avoid the winner’s curse; in this case, allow the winning bidder to pay what the second-highest bidder offered. The problem with the Milgrom-Wilson model is that it is designed to maximize value for the person holding the auction; in the case of government auctions, like those for spectrum, it needs to be designed to maximize societal welfare which would mean provision of telecom services for as many people as possible. One way to do this, is to put conditions on, say, rural penetration. If a condition is put that allows a telco to participate in a bid only after it agrees to install 30% of its base stations in rural areas, auction prices will automatically come down. Auctions for water distribution, for instance, could insist that a certain proportion of residents will get water at a certain price; it could also lay down minimum number of hours of water supply. While the bidding process will ensure complete transparency, the initial conditions will maximise societal value, and also ensure bid prices remain reasonable.
What has made the auction outcome much worse in India is government greed. In 2014, when the initial 20-year period of the first private telecom licenses expired, the government refused to renew them unless telcos rebid for spectrum; with their entire revenue stream at risk if telcos did not win back their spectrum, not surprisingly, the 900MHz auction fetched very high prices. Nor was this the only instance of poor government policy. While there are different reasons for why individual auctions saw sky-high bidding, as a general rule, the telecom regulator (Trai) used the last auction price as the reserve price for the next auction, thereby ensuring that prices were hardly ever based on demand and supply; not putting enough spectrum on auction also meant most auctions were conducted in a situation of constrained supply
While some will argue that putting conditions to ensure bidding doesn’t go out of control hurts the exchequer, this is not true either. A recent letter from RJio to the government points out that just a third of the spectrum put on auction since 2012 has been sold; and there has been no auction in the last four years. Based on the reserve price fixed in the auctions, according to RJio, the government has unsold spectrum worth Rs 3.9 lakh crore on its hands compared to the Rs 2.5 lakh crore that it got in the auctions between 2012 and 2016. Even if maximization of government revenue was to be the sole criterion, it is clear the auction strategy has failed miserably, apart from the destruction of telco balance sheets.