ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846

A+| A| A-

A Critique of Spectrum Management in India

This article examines the recommendations of the Telecommunications Regulatory Authority of India for preventing the "winner's curse" and promoting a healthy competitive environment for sustaining the growth of this exponentially growing industry.


A Critique of Spectrum Management in India

Rohit Prasad, V Sridhar

a three-stage auction procedure and a start-up spectrum of 2 × 4.4 MHz in 1800 MHz was given to the winning bidder. In addition to the entry fees, licensees were required to pay a percentage of annual revenue as spectrum charges.

Further allocation of spectrum beyond the start-up spectrum levels was based on availability and justification and attracted additional revenue share as spectrum charges. The contractual rights of spectrum holders were incrementally established through a series of government

o rders. In 2002, the subscriber linked spectrum allotment procedure referred to as subscriber based norms (SBN) was introduced, which laid down a road map up to allotment of 2 × 12.5 MHz of spectrum per operator in each LSA (Table 1).

Table 1: Spectrum Allocation Criterion – 2002

Quantum of Minimum Subscriber Annual Spectrum Charges
Spectrum Allotted Base Required (as % of Adjusted Gross
(Crores) Revenue)
2 × 4.4 MHz - 2
2 × 6.2 MHz - 3
2 × 8 MHz 0.5 3
2 × 10 MHz 1.0 4
2 × 12.5 MHz 1.2 5

Source: Sridhar (2007).

This method of spectrum allocation was very different from the methods followed in other countries, where a sizeable spectrum block (about 2×15 MHz) was given to the operators as start-up spectrum [TRAI 2005]. The rationale for adopting a different approach was the scarcity of spectrum due to non-availability from the department of defence.

Second Stage – De-linking Licence from Spectrum (2003-06): During 2000-01, the government also liberalised the basic telecom services (BTS) market, which provided traditional landline based services. Jain and Sridhar (2003) provide details of the BTS operations and their growth in I ndia. In 2000, BTS operators approached the government with a proposal to provide local access loop at much lower cost using the alternative code division multiple a ccess (CDMA) wireless technology. After a couple of years of litigation between the BTS and GSM mobile operators, the Indian government announced unified access service (UAS) licences in November 2003 that allowed basic service licence holders to provide full mobility based services

This article examines the recommendations of the Telecommunications Regulatory Authority of India for preventing the “winner’s curse” and promoting a healthy competitive environment for sustaining the growth of this exponentially growing industry.

Rohit Prasad ( and V Sridhar ( are with the Management Development Institute, Gurgaon.

Economic & Political Weekly

september 20, 2008

pectrum is an essential resource for the provision of mobile services. The management of radio frequency spectrum is important to meet the following objectives: (a) granting of exclusive rights to spectrum; (b) ensuring its efficient use; and (c) promoting competition in services [Falch and Tadayoni 2004]. In addition, in emerging economies, spectrum management must achieve the development goals of universal telecom access at low prices. As pointed out by Falch and Tadayoni (2004), the focal point of regulation of scarce resources, namely, spectrum, has changed from pure coordination and planning to the creation of a competitive and sustainable environment for various telecom services. In this article, we critically look at the history of spectrum management in India with r espect to allocation and pricing.

Management of Spectrum

Before the introduction of mobile services, spectrum intended for commercial usage in 800 MHz, 1800 MHz and 1900 MHz was entirely in the control of the national d efence force of India. The utilisation of spectrum for commercial purposes began with the release of a limited amount of spectrum in 1995. The management of spectrum in the country can be divided into the following stages:

First Stage – Auctioning Scarce Spectrum (1995-2003): The Indian government auctioned 2 × 4.4 MHz of start-up spectrum for the Global Systems for M obile (GSM) based mobile services in 1995. Two operators were selected for each licence service area (LSA) [for details on LSAs in India, see Prasad and Sridhar 2008]. Subsequently, the third operator l icence was awarded along with 2 × 4.4 MHz of start-up spectrum in the 900 MHz band to the government operator on a pro bono basis in 2001. The fourth operator licence was issued in 2001 using


with a stipulated entry fee based on the bid price paid by the fourth operator in 2001 [DoT 2003]. The fixed fee based l icence (as opposed to auction based) theoretically allowed any number of mobile licences to be provided and implicitly delinked spectrum allocation from licensing. Though firms were awarded licences after paying the required entry fee, they were given start-up spectrum only as and when available. Following the entry of two or three CDMA based mobile operators in each LSA, one or two new firms also paid the stipulated entry fee and got licences to operate GSM services in certain LSAs.

In 2005, the Telecommunications Regulatory Authority of India (TRAI) reviewed the spectrum allocation process taking into account spectrum availability and efficient techniques for utilisation of assigned spectrum [TRAI 2005]. TRAI (2005) stated that the spectrum held by the GSM and CDMA operators was well below international averages. It was recommended that existing operators be given adequate spectrum before considering allocating spectrum to new service providers especially since “there is adequate competition in almost all service areas”. TRAI (2006) continued to maintain that there was a shortage of 2G spectrum.

The entitlement of incumbents naturally extended to future 3G spectrum since in a spectrum scarce environment, 3G could be considered as an extension of 2G.1 It also followed that the scarce resource of spectrum need not be spread too thin thereby justifying a cap on the number of operators.

Third Stage – Stricter Criterion for A llocation of Spectrum (2006-08): As technological progress took place, it began to be believed that if used maximally, spectrum held by incumbents was sufficient for meeting their near term requirements. Meanwhile, additional spectrum was also being vacated by the defence. As a result, in keeping with the principle of maximal usage of spectrum, new SBN were defined, incumbents kept out of fresh allocations, 3G treated as a separate service from 2G and the cap on the number of operators removed. As mentioned in the TRAI recommendation on 3G spectrum [TRAI 2006], “the defence services have agreed to vacate 2 × 20 MHz in the 1800 MHz band, in addition to 25 MHz in the

2.1 GHz UMTS band. The availability of additional spectrum in the 1800 MHz band provides sufficient room for growth of 2G services for the medium term.” Therefore, “the authority has recommended that the government should not treat the allocation of 3G spectrum in continuation of 2G spectrum”.

The TRAI (2007) recommendation that no cap be placed on the number of telecom access providers in the country allowed more new firms to enter the market by paying the low fixed entry fee. It even allowed CDMA operators to obtain GSM licences. The new licensees were put in queue for assignment of spectrum as and when available. TRAI (2007) further stated that the spectrum allocation criterion should take into account methodologies received for 2.1 GHz band, prorated for 2×1.25 MHz.

Critique of Spectrum Policy

The following issues need to be addressed in the policies for allocation and pricing of spectrum:

Optimal versus Maximal Use of Spectrum: The formulation of spectrum policy in India began under conditions of very limited availability of spectrum. In the initial phase the policymaker’s requirement of maximal usage of spectrum with the a ssociated SBN was justified. However, as spectral efficiency increased and additional spectrum got released, more nuanced d efinitions of efficiency needed to be applied in order to promote the growth of the industry.

The single-minded agenda on maximising the number of subscribers per unit of

such as synthesised fre- Table 2: Subscriber Base Criterion spectrum ignores the im
quency hopping and use of for Additional Spectrum Allocation 2008-Current portance of efficiently us
advanced codes for effi- Quantum of Spectrum Minimum Subscriber ing other inputs like base
ciently utilising allotted Allotted Base Required transceiver stations (BTSs).
spectrum. Following this, 2 × 4.4 MHz - The specific interpretation
the government tightened 2 × 6.2 MHz 0.5-0.8 of technical efficiency used
the subscriber base criterion for additional spectrum allocation to existing opera 2 × 7.2 MHz 2 × 8.2 MHz 2 × 9.2 MHz 2 × 10.2 MHz 1.5-3.0 1.8-4.1 2.1-5.3 2.6-6.8 by the government has resulted in high reuse of spectrum and hence, more
tors making them ineligible 2 × 11.2 MHz 3.2-6.8 BTSs and cell towers. In
for participating in the 2 × 12.2 MHz 4.0-9.0 some circles, the inter-site
new allocation of spec 2 × 13.2 MHz 4.8-7.0 distance between cell tow
trum. Table 2 gives the re 2 × 14.2 MHz 5.7-10.7 ers is less than 100 metres,
cent SBN released by the 2 × 15 MHz 6.5-11.6 which is one of the lowest

Source: WPC (2008).

government, which are much stricter than that shown in Table 1.

Fourth Stage – Policy on 3G (2008): The government announced the policy for 3G mobile services in August 2008 [DoT 2008b]. In line with TRAI’s recommendation [TRAI 2006], the government opted for a simultaneous ascending auction for allotment of a start-up spectrum of 2×5 MHz in the 2.1 GHz band with specified reserve prices for different categories of LSAs. Note that 2×5 MHz is the minimum carrier requirement for providing 3G services using wideband code division multiple access (WCDMA) technology in the 2.1 GHz band. The 3G policy also states that 2×1.25 MHz carriers will be allotted to UAS licensed CDMA operators at a price equal to the highest bid

in the world leading to ironclad structures clogging some of the cities.

The production function estimated in Prasad and Sridhar (2008) clearly indicates the presence of economies of scale in the industry. This implies that unit costs decline with an increase in the scale of

o peration, at least over the range of subscriber minutes observed in the industry to date. In view of this, the policymaker should allow accumulation of spectrum, where such accumulation leads to lower costs without increased threat of cartelisation. This would allow operators to integrate operations and aggregate spectrum holdings to take advantage of economies of scale.

Efficient Use of Mobile Infrastructure:

Underutilisation of BTSs leads to an increased demand for spectrum. There

september 20, 2008


are several new technologies to ensure

  • o ptimal yield from BTSs. These include:
  • (a) implementation of in-building solutions such as Femtocell to improve the efficiency of BTSs; (b) single antenna interference cancellation that can improve downlink bandwidth of GSM networks without changes to network configurations; and
  • (c) utilisation of smart antenna arrays that can confine channels to narrow beams thus improving capacity gain.
  • The government should monitor the
  • o ptimal use of base stations in the same way it monitors the use of spectrum. Prescribing SBN for BTSs is one possible option, though challenges of administering and monitoring would need to be addressed.
  • Further, the government must do more to promote infrastructure sharing. While it has allowed both passive (towers, rental places) as well as active (antenna, feeder cable, node B, radio access network and transmission systems) infrastructure amongst service providers [DoT 2008a] in order to effectively use the radio access network infrastructure, incentives in the form of lower regulatory levies need to be provided to improve adoption, thus i mproving the efficiency of the radio access networks.

    Spectrum Pricing: The government has chosen to provide the UAS licence for upfront fixed fees along with a revenue sharing agreement. The revenue share of government is increased by a percentage point on every fresh allocation of spectrum. SBN are used to determine eligibility for fresh spectrum.

    The fixed fee has been benchmarked to the fee paid by the fourth cellular operator in the auction of 2001. A benchmarked price is appropriate only if the object being sold (in this case, a licence) is the same and the market conditions are similar.

    The fee paid by the fourth operator in the auction process was specifically for a cellular mobile telephone service (CMTS) licence along with start-up spectrum. The UAS allows the licensee to provide access services using non-spectrum related technologies such as wireline service as well. As per the guidelines of the UAS licence [DoT 2005]:

    The unified access services cover collection, carriage, transmission and delivery of voice and/or non-voice messages over licensee’s network in the designated service area and includes provision of all types of access s ervices. The access service includes but not limited to wireline and/or wireless s ervice including full mobility, limited mobility as defined in clause 12 (c) (i) and fixed wireless access.

    The UAS licence is therefore a super set of the CMTS licence and its price cannot be equated with the price paid for a CMTS l icence. Further market conditions in 2003, and even more starkly, in 2007 were quite different from those in 2001. Let alone factoring the increased value of spectrum in the booming telecom sector, the government did not even factor inflation to arrive at the 2003 and 2007 prices.

    Inappropriate benchmarking has resulted in underpricing of spectrum. Underpricing spectrum leads to a tendency of hoarding and therefore, should be avoided.

    The confusion regarding the CMTS and UAS licences has cut both ways since UAS licensees providing CDMA services and applying for a GSM licence in 2007 had to pay exactly the same amount that they paid when they were issued the CDMA-UAS l icence. In fact, as per the above argument, only the charge for GSM spectrum should have been levied.

    Though UAS implicitly separated licensing from spectrum, licence fee needs to be separated from spectrum fee for pricing both licences and spectrum appropriately (in case a fixed fee method is chosen).

    Regulatory Certainty: Telecommunication licences should balance regulatory certainty with the flexibility necessary to address future changes in technology, m arket structure and government policy [I ntven et al 2000]. The fixed fee is paid by the UAS licensees without any firm guarantee on the date of allocation. It therefore involves a promise to allocate spectrum at an uncertain point in the future. Further, the SBN represent a very complicated f uture contract in spectrum with an additional clause of the seller having the right to renegotiate the terms by strengthening the SBN. This contract is very non-transparent and difficult to convert into monetary terms for the purpose of decision-making by the involved parties.

    Moreover, there is no injunction on the regulator for an orderly step-by-step change of SBN in response to technological changes. Thus, the government decision on strengthening of SBN by an order of six to 15 times increased the price of spectrum from nil to infinity (since the incumbents became ineligible for fresh allocation). This does not represent a very conducive environment for business decision-making.

    Number of UAS Licences Issued: The UAS licence regime de-linked licence from spectrum in the sense that it allowed the possibility of giving licences without present availability of spectrum. However, it included a promise to provide 2×4.4 MHz of start-up 2G spectrum “as and when available”.

    In 2007-08, the government failed to take into account the availability of 2G GSM spectrum while granting licences. In theory, the maximum available GSM spectrum in the 900 and 1800 MHZ bands is 100 MHz. In practice, in India, only 60-70 MHz can be used. Over 40 MHz is a lready in the possession of incumbents. The remainder is insufficient for even the start-up needs of the six to nine new entrants, let alone the incremental spectrum required thereafter.

    The government can be faulted for not restricting the total number of licences granted (or alternatively, not de-linking licence from spectrum) based on its knowledge of spectrum availability.

    Policy on 3G: The migration from 2G to 3G requires the setting up of new infrastructure by operators and the purchase of 3G compatible handsets by end users. 2G services can continue to be provided using 3G spectrum and equipment, but 3G services cannot be provided in a 2G environment. The government is currently viewing 3G as a high value service distinct from 2G and formulating allocation policies accordingly.

    The policy announcement on 3G spectrum marked the culmination of almost two years of wait following the TRAI recommendations on allocation and pricing of spectrum for 3G presented on September 27, 2006.

    The delay in the 3G policy was in a sense responsible for the rush for licences in 2007 when entrants unsure of the 3G allocation rushed to get 2G spectrum that was promised along with the UAS licence.

    Economic & Political Weekly

    september 20, 2008


    Seen independently of the policy legacy, the 3G policy has several salutary features. These include the separation of the licence from the spectrum, the choice of an a uction mechanism to reveal the current market value and the opening of doors to foreign participants.

    However, the underpricing of 2G spectrum in 2007 and the auction determined price of 3G spectrum in 2008 will result in a new entrant to the Indian market paying much higher prices for entry through the 3G route as opposed to entry through the 2G route. While we may overlook the case of global entrants who have deeper pockets than their Indian counterparts, we cannot turn a blind eye to UAS licence holders who will not be able to get the promised 2G spectrum due to non- availability and therefore must bid for 3G spectrum to be able to provide even voice services.

    Regarding the details of the auction procedure recommended, TRAI (2006) advocated that the lowest bidder in each stage of the auction be placed in the wait list for spectrum. However, it also mentioned that the size of the wait list should take spectrum availability into account. It is imperative for the government to limit the length of the wait list in order to avoid the impasse in the allocation of 2G spectrum.

    In case of 2×1.25 MHz carrier in 800 MHz, TRAI (2006) recommended a single stage auction if more UAS-licensed CDMA operators are interested. TRAI (2006) also explicitly specified that the 2G subscriber base should not be taken in to account for 3G spectrum allocation as 3G is a different service from 2G. However, DoT (2008b) policy states that 2×1.25 MHz be given to the operator having maximum 2G subscriber base. There is inconsistency in the method of allocation stated by TRAI and DoT where DoT appears to flout its stated policy of treating 2G and 3G as d istinct services.

    The policy announcement is silent on the road ahead after the present allocation. One hopes that the mistakes of 2G are not repeated in 3G. A clear road map of spectrum availability, use of the auction mechanism for allocation, and rationalisation of the subscriber based norms to serve as a low hurdle of eligibility to participate (as opposed to requiring maximal usage) would be appropriate.

    Technology Neutrality: The introduction of CDMA based mobile service during 2001-03 was clouded by the legal battles between the incumbent GSM and BTS

    o perators. Given their defensive position and the prevailing paradigm of maximal u sage, the CDMA mobile service providers were allotted a lower start-up spectrum block (2.5 MHz as compared to 4.4 MHz for GSM) as they were considered more efficient. Parity was established between the two technologies in terms of the number of subscribers each could reach with the spectrum allotted to them.

    The definition of parity is tantamount to handicapping a more efficient player by restricting access to a necessary resource. In an environment where maximal usage is no longer relevant, the correct definition of parity is “equal access to resources” to be used as inputs. Such an approach would create a truly level playing field in which the respective technologies would be able to compete. Promoting such competition would also economise on the use of spectrum. The only qualification to this conception of parity could be the threat of monopoly power. However, if anything, the strengthening of CDMA service would increase competition in an industry that is currently largely dominated by the GSM players.

    As per TRAI (2005), both types of operators should have the same amount of s tartup spectrum and be held to the same subscriber based norms.

    Level Playing Field: As is pointed out in TRAI (2007), the ratio between actual subscriber base and subscribers that should be serviced as per the SBN is much larger for private operators compared to government operators, indicating that private operators are having to stretch spectrum much more than government operators. In some LSAs, this ratio is even less than one for some government operators, i ndicating that these government operators were given spectrum despite not meeting the SBN. Even when both the government and private operators met the


    Announces K. Subrahmanyam Award for Excellence in Research on Strategic and Security Issues

    Nominations are invited for this prestigious award to be conferred each year on an Indian scholar, journalist or analyst who has made an outstanding contribution in the area of strategic and security studies. Preference will be given to those younger than 50 years of age. The awardee will be selected from among the invited nominations by a jury of eminent persons constituted by IDSA.

    The award will be conferred on 11th November 2008, the Foundation Day of IDSA. It carries a cash award of Rs. 50,000/- plus a citation.

    Nominations may be sent in confidence to Director General, IDSA latest by 10th October 2008. Nomination forms may be obtained from the office of the Director General, IDSA at the following address:

    Institute for Defence Studies and Analyses

    No.1, Development Enclave, Rao Tula Ram Marg, Delhi Cantt. New Delhi – 110 010

    (Tele: 011-26146591, 26146831)

    Or downloaded from Institute’s Website:

    september 20, 2008


    SBN criteria, government operators are given priority in the allocation of spectrum. This places them at an advantage in an environment where spectrum is scarce and SBN change rapidly. In a highly competitive mobile services industry, such practices do not indicate the presence of a level playing field. If SBN criteria are to continue, then the additional spectrum should be allotted to eligible operators on a firstcome-first-serve basis.

    harmful for the industry. A clear defined, consistent policy is the need of the hour.


    1 The 3G spectrum is designed for providing connectivity at speeds of up to 2 Mbps and hence, is capable of providing services such as mobile television, and real time video conference while 2G spectrum is used for low bandwidth services such as voice and short message service.


    Falch, M and R Tadayoni (2004): ‘Economic versus Technical Approaches to Frequency M anagement’, Telecommunications Policy, Vol 28, pp 197-211.

    Intven, H, J Oliver and E Sepulveda (2000): Telecommunications Regulation Handbook, retrieved January 15, 2001 from the worldwide web at handbook

    Jain, P and V Sridhar (2003): ‘Analysis of Competition and Market Structure of Basic Telecommunication Services in India’, Communications and Strategies, Vol 52, pp 271-93.

    Prasad, R and V Sridhar (2008): ‘Optimal Number of Mobile Service Providers in India: Trade Off between Efficiency and Competition’, International Journal of Business Data Communications and

    Networking, Vol 4, No 3, pp 69-81.

    DoT (2000): ‘Guidelines for Issue of Cellular Mobile

    Sridhar, V (2007): ‘Growth of Mobile Services across

    Telephone Service’, retrieved January 15, 2003

    Conclusions Regions of India’, Journal of Scientific and Indus

    from the worldwide web at http://www.dotin

    trial Research, Vol 66, pp 281-89.

    Spectrum policy in India while having been

    TRAI (2005): ‘Recommendations on Spectrum Related

    – (2003): ‘Guidelines for Unified Access Service

    very successful in nurturing the growth of Issues’, retrieved May 13, 2005 from worldwide

    L icence’, retrieved January 15, 2006 from the

    web at

    industry, suffers from the lack of a long-worldwide web at

    – (2006): ‘Recommendations on Spectrum Alloca

    – (2005): ‘Guidelines for Unified Access Service

    term vision and absence of a holistic per- tion and Pricing for 3G and BWA Services’,

    L icence’, retrieved January 15, 2006 from the

    r etrieved May 13, 2007 from worldwide web at

    spective that considers all the relevant worldwide web at

    – (2008a): ‘Guidelines for Infrastructure Sharing’,

    factors before making policy decisions. The – (2007): ‘Recommendations on Reviewing of License

    retrieved March 15, 2008 from the worldwide web

    Terms and Conditions and Capping of Number of

    trajectory of spectrum policy in India has at

    Access Providers’, retrieved November 13, 2007 been marked by many flip flops: on SBN, – (2008b): ‘Guidelines for Auction and Allottment from worldwide web at

    of Spectrum for 3G Telecom Services’, retrieved WPC (2008): ‘Order No J-4025/200(17)/2004-NT dated

    spectrum pricing, 3G policy and competing

    August 10, 2008 from the worldwide web at January 17, 2008’, retrieved May 9, 2008 fromtechnologies. The resulting uncertainty is worldwide web at







    Economic & Political Weekly

    september 20, 2008

    Dear Reader,

    To continue reading, become a subscriber.

    Explore our attractive subscription offers.

    Click here

    Back to Top