On spectrum auctions, we could well be headed for the best of times or the worst of times. Even as the American 600MHz spectrum auctions ended recently with a bang and yielded nearly $20 billion to the US government, India has sounded the bugle on Trai recommendations for the next auction of spectrum in the 4G bands (700MHz, 800MHz, 1800MHz, 2300MHz and 2500MHz) and also in the futuristic 5G bands of 3400MHz to 3600MHz. There are abundant learnings for India from a study of the American and other auctions, but there are huge challenges to address that would need political intervention.
While awaiting Trai pronouncements, one cannot help but reflect whether we will apply the required corrections and achieve spectrum allocation for the broadband age, or are we headed for one more sub-optimal spectrum auction, laden as we are with over Rs 4.5 lakh crore of industry debt with revenues at an all-time low?
First, it is indisputable that the degree of success of an auction is directly dependent on the sale of most, if not all, of the goods being put up for sale. Going by this, the latest American auction was 100% successful, since all the 70MHz of spectrum offered was sold. In contrast, in the Indian 2016 mega auctions where a total of 2350MHz in seven bands were put up for auction in 22 circles, only 964MHz, or barely 41%, got sold. Even taking all the six e-auctions held since 2010 together, only about 60% has been sold. In the case of 700MHz (very close frequency to the US 600MHz) auctioned in 2016, nil quantity was sold—clearly a total failure. There is obviously much to be done to improve the effectiveness of our auctions.
Second, it is not the case that Indian operators are flush with spectrum and hence did not need to pick up more from the 2016 auctions. Even after those auctions, while India might get by for a narrowband age, it still needs an average of 20MHz per operator to meet the current global average and is ill-equipped for the broadband era. The success of an auction is determined not only by the ability to sell a large proportion of spectrum up for auction, but also by the market/clearing price being significantly above the reserve price, viz. the auction process must help discover the true market price of the spectrum. Here again, we see a remarkable difference between the Indian and American spectrum auctions.
You might also want to see this:
In the US, the 600MHz spectrum was auctioned to support growth in next-generation mobile communication services. The regulator, Federal Communications Commission (FCC), took back 84MHz of spectrum in the 600MHz band from television broadcasters through reverse bidding. It then repackaged the spectrum into 70MHz of clean spectrum and opened it to bidding by telecommunication providers. The auction generated $19.7 billion, and the premium over the reserve price was more than 100% in the US.
In contrast, in 2016 in India, while 700MHZ got no takers at all, even the other bands saw a very poor response, with the reserve price ending up as the final clearing price in most cases. The average sale price was hardly 5% above the reserve price, i.e. there was hardly any market discovered price in India.
The abject failure of the 700MHz band in the Indian auction can be partly explained by calculation errors by Trai in the reserve pricing for the band, which was set at 4X of the 1800MHz band, and 800MHz and 900MHz bands were priced at twice that of 1800MHz and 2100MHz on the basis that the radio waves of the former travelled twice as far as the latter’s. However, the price of the 700MHz band did not follow this principle, but instead was linked to the European auction price of 800MHz, and even so, there were serious arithmetical errors in the calculation which resulted in a 100% higher multiplier for 700MHz.
However, even if one had corrected for arithmetical errors, it is likely that the outcome of the 700MHz auction may not have been vastly different as the band would still have been significantly overpriced. This is clearly borne out by analysis of the US and Indian auction results, which shows that the Indian 700MHz auction reserve price was effectively 46 times more than the US auction price of 600MHz! Given that the extremely price-sensitive Indian market desperately needs the 700MHz spectrum in order to deliver cost-effective mobile broadband, this was a wasted opportunity. Other bands auctioned in the Indian auction were nine times higher than the prices discovered at the recent US auctions. It is abundantly clear that a holistic assessment of the potential of the spectrum and a true discovery of its price is urgently needed. The American comparison is not a freak case. Even in 2010, when the first e-auction was conducted in India for 3G 2100MHz spectrum, a comparison with a contemporary German auction established that Indian auction reserve prices were unreasonably high.
Spectrum auctions are a step towards market-based spectrum management and are a way to maintain a balance between public welfare and maximisation of revenues by the government. While India was one of the early adopters of spectrum auctions, its success in implementing the auctions has been quite low. Auction design can be optimised to meet public policy objectives. There is a need to set reserve prices at levels that are high enough to keep non-serious bidders at bay, but low enough to achieve vibrant price discovery. There needs to be a meaningful correlation across bands based on factors such as efficiency, coverage and the existing ecosystem.
Above all, we need to break free of the ‘command and control’ mindset and overcome the inhibitions due to a legacy overhang of the CAG pronouncements and Supreme Court judgments that are adversely affecting the rational setting of spectrum reserve prices. Our spectrum management has caused much harm to the public interest by stifling mobile broadband and data growth. To resolve these issues, what we need is tremendous political will to ensure true discovery of spectrum market price—like the one demonstrated by the AB Vajpayee government, which gave us the trailblazing New Telecom Policy 1999.