In defence of spectrum auctions - TechCentral

In defence of spectrum auctions

Graham-Friend-180Spectrum is a finite, national resource which has enormous value, as has been highlighted in recent spectrum auctions around the world. Across Africa, regulators have an opportunity to realise the benefits of the 800MHz digital dividend and, following the International Telecommunication Union’s World Radiocommunication Conference later this year, they will also be able to benefit from the release of the 700MHz band. However, the extent to which African nations will be able to benefit from this spectrum will be greatly affected by the method of allocation.

While the use of auctions for allocating spectrum is commonplace in many markets and regarded as best practice, lots of countries in Africa have historically adopted administered or non-market-based approaches. Despite the popularity of these approaches, in opting not to use auctions to allocate new spectrum, regulators in Africa are unable to realise the spectrum’s full value. African regulators should use some form of auction mechanism to maximise the benefits from the allocation of any new spectrum, including the digital dividend.

The activities of national telecoms regulators are usually shaped by a set of public policy objectives determined by government. While policy objectives vary slightly from country to country, a common objective is to ensure the most economically efficient use of spectrum, which means allocating it to companies that will generate the greatest economic value from it. A well designed and competitive auction ensures that spectrum is allocated on an efficient basis, as the companies with the highest spectrum values will outbid their weaker competitors and win access to it.

The alternative to an auction is for the regulator to try and “pick the winners”, and award the spectrum to those operators it thinks will maximise the value from it. The approach is akin to the “central planning” of former communist states, and often has similarly disastrous results. The failure of regulators to identify the winners is no surprise, as they are faced with an almost insurmountable asymmetry of information regarding which operator values the spectrum most highly.

Achieving an efficient allocation of spectrum, however, may not be the only public policy objective, with governments often having a range of goals. Two common additional objectives are: 1) to provide extensive mobile voice and, increasingly, broadband coverage, and 2) to ensure downstream (post-auction) competition to maximise consumer choice and welfare.

Spectrum auction design has progressed dramatically since the early days of the notoriously inefficient first price, sealed bid auctions, which were once common. Modern spectrum auction design is now capable of addressing, simultaneously, the multiple policy objectives of coverage requirements, downstream competition and economic efficiency, meaning that by choosing to auction, spectrum regulators can ensure that they achieve their public policy objectives.

Extracting maximum revenues
If a regulator does not use some form of auction to allocate spectrum then not only does it risk failing to meet public policy objectives, but it may also fail to raise the maximum possible revenue from the sale of a valuable national resource. In the absence of an auction, the regulator will have to determine the price that the mobile operators will have to pay arbitrarily. In theory, a mobile operator will be prepared to pay up to the maximum value it places on the spectrum. Once again, the regulator faces an impossible task of trying to estimate what the maximum value is, due to an asymmetry of information. The only thing a regulator will know with certainty is that if an operator accepts the spectrum at the regulator’s arbitrarily determined price, then the price must be at, or most probably below, the operators’ spectrum value, meaning that the regulator could almost certainly have sold the spectrum for a higher price.

While in the context of a competitive auction, where maximising economic efficiency and maximising auction revenues are not necessarily aligned, it is reasonable to presume that if the spectrum is awarded to those that value it most highly then, in doing so, the regulator is also maximising government revenues. As the spectrum awards in the US and India certainly demonstrate, auctions undoubtedly have the ability to raise significant amounts of money for the public purse.

However, the benefits to the public purse from auctioning spectrum are not limited to the revenues generated at auction, as an auction ensures the spectrum is allocated to the operators that will generate the greatest value or profit from the spectrum. An auction therefore also means that the government will maximise tax receipts from the allocation of spectrum, with the result that regulators can be confident that by opting to auction, they are generating the greatest possible revenue from spectrum allocation.


Spectrum auctions are completely transparent, objective and significantly less prone to corruption, says the writer

The amount paid at auction by a mobile operator for spectrum is often referred to as a “sunk cost”, which is a cost that has already been incurred, and which is non-recoverable. Economists argue that sunk costs are irrelevant for future decision making, such as setting prices. Since the sunk cost has already been incurred, it does not matter at what level prices are set, as it will have no bearing on the money that has already been spent. Economists also contest that in a competitive market, what determines price is the marginal cost of providing the incremental unit of the item being sold. Although others argue that firms typically use accounting-based measures of total average cost to set prices, and so sunk costs do have an impact on end user prices, there is limited empirical, or even anecdotal, evidence to suggest that high spectrum prices have led to higher mobile prices. Regulators can therefore use an auction in the confidence that any revenues raised are not likely to distort consumer pricing decisions.

The same sunk cost argument can also be made for the impact of spectrum auction prices on future investment. When a mobile operator contemplates future network investment, the only consideration is the associated revenues and costs. Economists also typically invoke the assumption of perfect capital markets to further support the view that auction revenues do not impact future investment decisions by reducing the amount of capital available for future investment. The recent history of the financial markets, however, strongly challenges the efficiency of capital markets, although if a regulator is concerned about future investment, then a deferred payment scheme is entirely compatible with auctions, and can alleviate budgetary concerns. Regulators can therefore be confident that the use of auction will not damage incentives for investment.

Transparency, confidence and investment
In the past, spectrum and mobile licences were frequently awarded on the basis of beauty parades. Mobile firms would often have to employ a small army of consultants, for extended periods of time, to prepare elaborate and largely fictional “bid books” or business plans, and the operators with the “best” business plans would secure the spectrum. The scoring of the bid books and identification of the winners was extremely time consuming, highly subjective, lacking in transparency and open to corruption, often resulting in the entire award process becoming subject to legal challenge.

An auction, in contrast, is completely transparent, objective and significantly less prone to corruption. A well designed auction will increase investors’ confidence in the regulator and government and will help reduce regulatory uncertainty. By reducing uncertainty the regulator will de-risk the market, and by lowering the risk profile they will encourage inward investment. An auction will also significantly reduce the risk of legal challenge, and a well-executed auction will enhance the reputation and standing of the regulator.

A key issue with beauty parades and drawn-out administered processes is that the award procedure can be overtaken by events in the fast-moving mobile telecommunications world. A well designed auction, in contrast, can be executed quickly, which reduces risk and uncertainty, leading to potentially greater auction revenues. The process of conducting an auction can also be considerably less expensive than the use of a lengthy beauty parade. Furthermore, the benefits of the additional revenues raised are likely to more than offset any incremental costs associated with the administration of an auction.

The use of an appropriately designed and competitive auction will provide regulators in Africa with a wide range of benefits compared to non-market-based mechanisms.

With the release of the 800MHz digital dividend and World Radiocommunication Conference 2015 just around the corner, it’s therefore vital that African regulators re-evaluate their method of allocating spectrum, in order to fully realise its potential.

  • Graham Friend is an economist and MD of Coleago Consulting, a specialist telecoms strategy consulting firm that advises regulators and operators on issues relating to spectrum, regulation and network strategy


  1. William Stucke on

    Hi Graham,
    I’m broadly in agreement with you on the utility of auctions. A couple of minor points, though:

    While economists talk of “allocating” value and other things, regulators talk of “assigning” spectrum to licensees.

    Secondly, while this may change with the new spectrum policy expected soon, and especially in the light of Eskom’s current funding woes, the SA Government in the past has preferred a lower cost to communicate than high windfall income from the sale of spectrum.

    Thirdly, as you acknowledge, capital markets are not perfect, and high spectrum costs can indeed lead to high fees to the consumer.

  2. Graham,

    Overall excellent post and I’m in agreement with vast majority
    of the comments in your article. One key
    factor however that seems to get light treatment is the competitive
    situation. In my experience this seems
    to be the single biggest driver of where prices might fall out in an auction.

    The underlying assumption is that the auction is “competitive”
    when reality is always a group of players that come to the table with
    varying abilities to participate and pay.
    Incumbent players always have the most to gain and most to lose if they
    do not secure the new spectrum. The
    barriers to entry are huge in the telecom business making it impractical for
    the majority of new entrants to be truly competitive. As a result, bidding tends to stop when the smaller
    players exceed their ability to raise enough capital or debt to pay for the
    spectrum not when the larger players reach their valuation limits. An auction is ideal in a green-field
    situation but that is rarely the case today.

    If a government’s sole objective is to structure an auction
    to maximize revenue the formula is simple.
    Determine the number of incumbents or deep pocketed participants and
    offer one fewer license blocks than there are incumbents. It’s would be like auctioning off the last
    three oxygen tanks to four emphysema patients.
    The auction would simply proceed until the poorest patient could no
    longer pay. The end result may be death
    of an incumbent player and a less competitive situation long term which would
    harm consumers. We know this is true because of one key observation
    that in recent auctions the vast majority of spectrum tends to go to incumbent
    players. Very few newcomers are willing
    to place “bet the company” bids to compete against dominate players with all the
    advantages of incumbency.

    An alternative situation might be auctioning off 4 suitable
    blocks of spectrum in a market with 4 incumbent players. In this scenario the auction results are
    likely to generate much less revenue as the participants gravitate towards a
    something for everyone status quote. However,
    this type of situation would lead to a scenario that is both more competitive
    and one where the participants will have more flexibility to invest additional
    capital in a network build and still achieve the financial returns required by
    their shareholders. This scenario is likely
    a better public policy to pursue if expanding the use of telecommunications is
    important for the government.

    Auctions are excellent public policy tools however, pushed
    too far, the ultimate trade-off of the extra revenue gained by assigning
    spectrum to companies that can generate the greatest value is a less
    competitive market with higher consumer prices, fewer choices and lower network

© 2009 – 2020 NewsCentral Media