Using spectrum for the common good

Service delivery is lost in pursuing auctions

“There is nothing either good or bad, but thinking makes it so.”  Individual profiling by marketers, for instance, epitomised by internet platforms such as Google, Amazon, Facebook, WhatsApp, Netflix and so on. But tread carefully, O Reader, because whether we think about it or not, what we do has real consequences. 

Here is a cautionary tale about auctions, more particularly, spectrum auctions.  Recently, the Nobel for economics, the Sveriges Riksbank prize, was awarded to Paul Milgrom and Robert Wilson for auction design. In the words of the award committee, “for improvements to auction theory and inventions of new auction formats”. And it is widely touted that auctions are the best way of allocating spectrum and other public resources for the common good. What remains unstated are all the assumptions necessary for good outcomes, especially service delivery.

When these auction designs for wireless spectrum or telecom licences were applied in many countries, two things happened. One, an enormous amount of money was collected by governments from auctions, as in the US (1994) and India (for licences, not spectrum), the UK (2000), and countries across Europe (2001). Two, and a devastating consequence, was that several “successful” bidders declared bankruptcy — in the US, the UK, and Europe; in India, some reneged on their bids. Those that survived were so debt-laden that they struggled to invest in building networks to use their hard-won spectrum/licences.  The welfare loss from service deprivation was incalculable.

 This combination of successful government collections and ruinous debt for bidders and industry, together with deprivation for consumers/society, was repeated in the US in 1995-96, and in India in 2010 onwards. Meanwhile, following the dotcom boom and bust in 2000, the telecommunications sector collapsed worldwide, compounded by overreaching auction bids. This was 10 times larger than the dotcom collapse (1), and the sector remained crippled by unserviceable debt in high-bid countries for nearly a decade. India was an exception because of the change to revenue-sharing in 1999. Auction mavens wrote disparagingly of countries with low bids such as Switzerland and Sweden, and countries that did not auction spectrum, such as South Korea, Japan and Finland (until 2009). Unsurprisingly, these countries achieved the best services, as they were not burdened by investments sunk in spectrum auctions. They built networks instead.  Yet until now, the Indian government, among others, is reluctant to apprehend what should be obvious from an objective analysis.  The alternative of getting stakeholders to cooperate for resolution is indeed difficult, but not having countrywide high-quality broadband is a severe impediment that we ignore at our peril.  We cannot hope to build real strengths without changing this, no matter what governments proclaim.

A number of reasons have combined to perpetuate the idea that sunk costs are irrelevant among theoreticians, policy makers, the judiciary, and lay people. There is the classical economics sunk-cost argument for ignoring past investments, premised on the assumption that future investment decisions are unaffected by past investments.  Only theoretical academics can sustain such assumptions, as also that economic decisions are entirely rational, or that there are zero transaction costs. Anyone with responsibility for costs and profits understands the reality that constrained resources affect investment capacity, and therefore investment and pricing decisions.

illustration: Binay Sinha

For years, a preponderance of research appeared to support the belief of sunk costs being irrelevant by theoreticians and especially policy-makers believing in free-markets, ignoring the collapse of the markets. The expectation that high costs affect investment capacity, and therefore must have significant consequences, was dismissed as erroneous.  Apart from occasional contradictory publications, it is only more recently that some evidence from financial and behavioural economics is being adduced to counter the notion that sunk costs do not affect future decisions, and to support the reality of how constraints on resources affect behaviour.

There have been occasional experts who disagreed with the emphasis on auctions, such as the MIT Media Labs Director Nicholas Negroponte, and publications to the contrary, such as, “Do Sunk Costs Matter?” (2), “What Really Matters in Spectrum Allocation Design” (3), and the GSMA report in 2017, “Effective Spectrum Pricing” by NERA Economic Consulting (4), on high spectrum costs holding up network investments and resulting in higher consumer prices.  For the most part, one-sided academic research, uninformed bureaucracy, auditors, and judiciary treat the amounts bid in auctions as the measure of success.

One explanation for research missing out on this aspect may be the problem of “unknown unknowns”, that is, the infeasibility of evaluating the consequences of actions not taken, because the data are not there for the paths not taken. In Donald Rumsfeld’s other inimitable phrase, the absence of evidence is not evidence of absence. Meanwhile, the reality of India’s unused spectrum juxtaposed with the state of its telecommunications reflect the lost possibilities for us, but there are no data to support what is not there.

Good, countrywide broadband in India is likely to give rise to vast benefits from greater inclusion, healthcare and education for economic activity, productivity, and better living. This requires far better connectivity, and for that, we need a much more constructive approach to the use of spectrum and networks. Spectrum is a public good that is beneficial only when used fully, and connectivity and communications are its primary uses. For providing an essential service, India’s approach is hugely flawed for a developing economy. It requires capital upfront to pay for the right to use spectrum, then using more capital to deploy networks to provide services, generate cash to run the business, as well as service the debt for spectrum and equipment. There is something very wrong when we have so much spectrum available, and considerable unmet demand, yet are unable to formulate practicable spectrum- and network-sharing or other policies to provide broadband services to everyone.  Our policies appear to be ignorant of the need for sustainable cash flows for the sector, and this obviously needs to be remedied.

shyamponappa@gmail.com

1: The Economist, 2002: https://www.economist.com/­leaders/2002/­07/18/the-great-telecoms-crash

2: R. Preston McAfee et al, 2007: https://www.mcafee.cc Papers/PDF/SunkCostFolly.pdf

3: Thomas W. Hazlett et al, 2011: https://papers.ssrn.com/ sol3/Delivery.cfm/SSRN_ID1961225_code410506.pdf

4: Richard Marsden et al, 2017: https://www.gsma.com/­spectrum­/wp-content/uploads/­2017­/02­/Effective-Spectrum-Pricing-Full-Web.pdf