An Interview with PK Garg
P. K. Garg: In 1991, the teledensity of India was less than 2
per cent. The then government, as part of the economic liberalization
policy wanted the teledensity to grow fast. At that point of time the
government was supporting the telecom sector for the growth of
teledensity to the tune of about four to five billion dollars per year
through the annual budget. During the same time frame it was felt that
the government would further need to support the telecom sector to
almost double the level if teledensity was to reach five per cent and
also if every village was to have a telecom connection to make
telecommunications accessible to each and every citizen. At that time
the thrust was more on accessibility rather than having an individual
connection for every home or every citizen.
Yelena: So if a telecom connection is accessible to the whole village then all villagers can share one line instead of having an individual connection?
P. K. Garg: Yes, until that time telecom facilities were not accessible to many people. So with the accessibility at least if there is one telephone in the village, whether it is a PCO, whether it is a phone in the Panchayat, all can use it. So the thrust of the government at that time was that the telecom facilities should be made accessible to all citizens. And even for taking the teledensity to five per cent it was estimated that it would require about 25 to 30 billion dollars over a period of five years. The government made provision for almost three-fourth of this money in the Indian budgets and the planning processes. So the fiscal imbalances, etc, that is an overall economic aspect, not only related to telecom. I think you have plenty of material from other sources on what measures the government took to deal with the issue – one of which was the liberalization: entry of the private sector was decided upon and it was also decided that the government telecom department would concentrate on increasing the accessibility through line telephones. The entry of wireless was to be left to the private operators, because the government money that was available for this was limited.
Yelena: And this was prior to 1994?
P. K. Garg: This was around 1991-1992 time frame. The government decided that the wireless telephony (mobile telephony) should be left to the private sector because it would require quite a significant investment and if the government funds were diverted to the mobile wireless service provision then it would have to reduce the investments available for line telephony and increase of teledensity. Another factor at that point of time was that nobody felt that mobile telephony will grow to that extent. They were feeling that one might get a few million connections, maximum four to five millions, because it was considered to be a service for rich and elite people, business people, top bureaucrats, but not so much for the common masses.
Yelena: So where did that shift come where the government realized that mobile connections are important for all citizens?
P. K. Garg: I would say it was a combination of many factors. To understand this one has to go through the history of our wireless mobile telephony a little bit. In 1994, licenses for mobile telephony in India were granted for the four metro service areas – Delhi, Mumbai, Chennai and Calcutta. Two operators in each of these areas were selected. At that time it was not pure bidding; it was sort of a beauty contest where certain pre-conditions were made and only those who technically qualified for those conditions could bid for that. At that point of time the government had indicated a certain ceiling for the tariffs and at that point of time outgoing calls as well as incoming calls – both were charged. Then around 1995-96 the mobile telephony licenses for other areas of the country were auctioned. It was slightly different from the 1994 bids which were more on the pattern of a beauty contest. And not only the mobile telephony licenses were auctioned; it was considered then that one more operator in each area should come in the line telephony sector also as a competitor to the government operator – first, to provide competition, secondly, to add to the government efforts to increase the teledensity.
Yelena: So the government operator was BSNL right?
P. K. Garg: At that time it was not BSNL, it was called the Department of Telecom and only in Delhi-Mumbai it was MTNL. MTNL was formed in 1986, and it was managing the Delhi and Mumbai areas. The rest of the country was still under the Department of Telecom. Later on they converted into the Department of Telecom Services. And in 1996-97, one private licensee or one operator for line telephony areas was also selected through auctions. The private operators began functioning in 1995, starting their services slowly with first mobile telephony networks in the metro areas. Mobile telephony networks in the other areas started functioning around 1996-97 and the private operators for line telephony were allowed to use wireless in the last mile, the WLL wireless in local loop. So they also started rolling out their networks in 1997.
However, around the beginning of 1998 most of the operators, especially the mobile telephony operators, felt that their business case which was the basis of their bids was not right. And 1) the growth was not as expected; 2) the ARPU (average revenue per unit/ subscriber) was far below their expectations which they had assumed for their business case. So the net result was that they approached the government saying sorry, we made a mistake in our business calculations and the government has to bail us out, otherwise we will not be able to service the existing customers also. And as you can appreciate most of these funds were lent by banks and financial institutions. That was not necessarily the money of the promoters. So then the lending banks and financial institutions also supported the cause of the operators with the government. Because many of our banks who had given money to the operators were nationalized banks, some of them were private also, but most of them were nationalized banks, and if the national banks were not able to recover their loans, it means the public money is gone. So the government had set up a high level group, ministerial level group, in the last quarter of 1998 to examine this issue and the group had detailed discussions with the operators, the banks, the user groups, the economists in the country and many others. They also consulted international experts. Then finally government agreed that, yes, some relief needs to be given to these companies/ operators. And that is how the new telecom policy of 1999 came into play.
The most salient aspect of that policy was revenue share. The license fee was earlier based on their bids, and in India it was not a total payment of auction amount upfront, it was staggered. The bid itself was payable over 10 years and the operator had the choice to indicate at the time of bidding how he would like to pay – whether he would like to pay uniformly or whether he would like to pay less in the beginning, more in the end, or he would like to pay more in the beginning, less in the end – he had the choice. And as per the normal economic rules the government then calculates the net present value of that bid amount for the purpose of comparison. So the result was that instead of the fixed license fee, the government said OK, now you will not pay any prefixed amount but you will pay a percentage of the revenues which you will earn. That was a big relief because then the operator was not required to pay any fixed amount which was quite large, and he was to pay a percentage of the revenue towards the license fee.
Yelena: So it depends on how much money they make…
P. K. Garg: So if the subscriber growth is there, they get revenues, they pay that much, and if they don’t get adequate revenues, they will pay less, if they get more revenues they will pay more. And then the spectrum charges also, which were earlier based on a formula, were also changed over to the revenue share or percentage of their revenues.
Yelena: What was the formula?
P. K. Garg: Earlier the formula was that for each city, say 1MHz of paired spectrum was charged at about $10 000 per year, per city and in addition there was a fee of Rs.100 (appox $ 2) per subscriber, per year that is the wireless license charge. So instead of separate spectrum royalty and the license fee, both were combined and it was taken as the revenue share. So the burden of a fixed fee was taken away and the operators were to pay a percentage of their earnings; whatever the revenues they get, they pay a part of that. So that was the biggest consequence of the New Telecom Policy 1999. This allowed the operators to reduce their tariffs because anybody can appreciate the two barriers for the growth of mobile telephony: cost of the handset which is entry level cost and the tariff costs. During 1999-2000 onwards the cost of the handsets was already coming down. Plus the revenue share allowed the operators to reduce the tariffs, because earlier if they were charging half a dollar per minute then they were paying a fixed amount of license fee and spectrum charges out of that, but now if they were to reduce their tariff from half a dollar to 10 cents, they would be paying only the license fee and the spectrum charge as a percentage of the revenue out of the 10 cents only. So they didn’t have much of a fear to reduce their tariffs, because the payment to the government will also be less. These factors encouraged them to reduce their tariffs – one was competition and the second was that they could get more customers.
Then the next point which encouraged the fast growth was the decision of our regulator in 2003 that only the calling party will pay; the receiving party will not pay. Earlier even the receiving party on the mobile had to pay.
Yelena: In Canada, if you don’t have a good plan, you still have to pay to receive a call.
P. K. Garg: In US also. It is there still. Even if I receive a call I will have to pay. So the charges for receiving a call were a hindrance, especially with the lower strata of society, because they were hesitant to receive the call if they had to pay. But once the liability was removed from the receiving party, many people started using the phone for receiving only. For example, the self employed artisans, electricians, painters, carpenters, anybody else could give their mobile telephone numbers to the prospective customers or clients and when they need their services they could call them without any financial penalty to the receiving party. One doesn’t have to pay anything except for the normal monthly or rental charges if one uses the phone for call receiving purposes only. That gave a tremendous boost to our telecom sector growth. For example, in early 2003, until this regulation was enforced, the total number of mobile customers in India was in the region of 10 million. These 10 million subscribers came, you can say, spread over at least 4-5 years starting from 1995. But from there onwards, the growth started picking up to almost more than 1 million subscribers per month. Another factor as I mentioned is that by that time the cost of the handset had come down to about $50-$60.
Yelena: How much was it before?
P. K. Garg: Initially in the 1995 time frame it was very much in the range of about $600. And from $600 it came to $60 (one-tenth). So these factors – the entry level cost of the handsets coming down, the tariffs also becoming low and the receiving calls becoming free – they were the factors which put mobile telephony on a very fast growth. Thereafter it’s well known to everybody because we crossed 100 million mobile subscribers in the middle of 2006 and from there onward it has been picking up and the country has seen almost 15 million per month new connections. Now also it is about 10 million new connections per month but it had even crossed 15 million during some time frame around 2007-08. So that is how the growth took place.
Yelena: The unified access services license: I think it was introduced in 2003. Could you speak about some of the reasons it was introduced?
P. K. Garg: Basically it is not my area of specialization, because mine was spectrum related, but I will give you some general aspects which I know. As part of the telecom policies most of these concessions were given to the operators. Then the government got from them some concessions or levied upon them certain conditions. For the mobile operators, government said that it will bring in a third operator which was the government operator. Moreover, the government also said that because we are giving you a concession with the license fee, etc, the government will have the right to introduce more operators as and when it feels the necessity for it in the public interest. Similarly for the line telephony also, since line telephony operators were also allowed certain concessions, in 1999 the government said that it may introduce more operators in addition to the one government and one private operator. So then in the year 2000 the government asked the government operator to enter the mobile telephony as the third operator. The name of the operator was MTNL in Delhi and Mumbai and then the remaining areas the operating wing of the telecom department was renamed as BSNL. BSNL was formed in October 2000. Then BSNL was asked to enter the mobile area for the rest of the country except Delhi-Mumbai where MTNL was there. Then in 2001, the government felt that one more private mobile operator can be introduced for which bids were called. They also said that for line telephony, because the issue of spectrum was not there, anybody who wished to come for line telephony would be allowed to do so. And then one operator came almost for all over India, that was Reliance in the line telephony. TATA also came into the play for a few areas and then couple of others were there like Shyam telecom for one or two areas as well as HFCL for one or two areas for line telephony. For mobile telephony, because it was only one license to be given, no single operator could get it for the areas all over India. Then some places Airtel got, some places went to Vodafone, and some places Idea received. So like that, all of the 22 circles went to different operators.
Yelena: And prior to this, per circle there were only two wireless operators?
P. K. Garg: Initially there were two mobile operators, third one became BSNL-MTNL, and the fourth one was introduced in 2001. And then what happened was that the fixed line operators were allowed wireless in the last mile. So they adopted CDMA; Reliance chose CDMA system to provide WLL and this CDMA technology had developed by that time to a level that could provide full roaming. Not only could it provide the wireless in local loop in the last mile but it could also provide full roaming and it was sort of alleged by many people and even the wireless mobile operators that Reliance had started providing full mobile services and not restricting to WLL which was the mandate of their license. That gave rise to many legal issues; there were legal cases by the mobile operators saying that the bid amounts offered by the fourth cellular operator in 2001 and the fixed line operators were quite different. The line operators’ bids were very low and the mobile operators’ bid amounts were very high; not as high as 1996-97 time frame, but still very high as compared to the bids of the fixed line operators. So they said that this was unfair competition and this was not allowed as per policy; why are they doing it? They went to court where they made the government also as a party, claiming that the government was allowing them to do it. At that time, in 2003, the government realized that this dispute in the telecom sector should be ended. After consultations with the Telecom Regulatory Authority the government decided to no longer issue separate licenses for basic fixed line services and cellular services, and instead introduced the Unified Access Services License (UASL). The earlier fixed line operators were given a choice that if they wanted to go for full mobility using CDMA or whatever technology, they could do so by paying the bid amount which was paid for the cellular licenses in 2001.
Yelena: And dual technology licensing came about in 2007 I believe?
P. K. Garg: Yes that was in 2007. Until that time an operator could have either mobile services in a CDMA or GSM, but not both.
Yelena: Great. Thank you for this history overview of telecommunications in India.
As part of CIS’s research interests in unlicensed spectrum policies, P. K. Garg was asked to comment on international and national level policies, as well as his perspective on the matter.
Yelena: Do ITU radio regulations reserve any bands for unlicensed use?
P. K. Garg: ITU radio regulations include the international allocations of different frequency bands. These international regulations are agreed by all the member countries of ITU at the world radio conferences. So it is agreed by all the countries, all the countries have to abide by that. In the international allocations there is no band which is unlicensed. There are certain bands which are allocated for ISM (Industrial Scientific and Medical) uses. For example, some bands are earmarked for microwave ovens because that’s an industrial use. Now of course it is for home use also but microwave ovens initially were for industrial use. Similarly certain frequency bands are for operating medical devices. And there are certain other scientific requirements for other bands. So there is a category called ISM (Industrial Scientific and Medical).
Yelena: And are these the 5GHz and 2.4 GHz bands or are there more?
P. K. Garg: Yes, they are also there, but there are many other sub-bands which are allocated for ISM services. Now many of these bands have been de-licensed for public use in many countries and as you said just now, the 2.4GHz, 5.2GHz, and 5.7GHz are the bands. There are other bands also. Many social requirements like cordless phones, let’s say individual requirements of the society, were developed in many of these unlicensed ISM type bands, because it was considered impractical both for the users, the vendor as well as the regulating authority in the country to issue licenses for each and every cordless phone. That is why they were developed in these bands. Some of these bands were de-licensed first in US and few other countries just like the 1500 or 1600 MHz sub bands. Then there was a band earlier around 150MHz, subsequently there were some parts even in the 900MHz band and the 450MHz band. That is where these cordless phones were developed. The cordless phones were one of the first de-licensed usages. Prior to that there were, you might have heard of them, walkie-talkies. They operated on 400MHz and covered a range of maybe a kilometre. That was to be used when people would go in the forest or trekking or camping, etc. So they were also developed in the ISM bands which were de-licensed in few countries.
Then as the requirements of the cordless phones came subsequently for WiFi modems, all these technologies were developed in ISM bands and even in those countries where they were not de-licensed, it was felt that it was better to de-license them, because 1) it will provide benefit to the society 2) it would be impractical to regulate their use or issue a license for everybody, because, for example, if one thinks of even regulating it or issuing licenses for WiFi modems it is practically impossible. And so the spectrum management authorities in the countries, any country, have to weigh how much is the benefit to the society by de-licensing; that is number 1. Number 2 – whether they can de-license or if there are some other users already there in that band, and how to shift them if possible, because those licensed operators/users have to be protected.
So, for example – the 2.4GHz band was de-licensed in India in 2004. For Licensed and unlicensed bands as such, there are many considerations before the spectrum management authority can decide to de-license them. The government could de-license the 5GHz band only for indoor use, because there were some existing users and it was difficult to shift them away from that band. The outdoor usage by the public cannot be allowed, because it will cause interference to those existing users as well as the public will get interference from them if they use it in the outdoors. That is why it could not be de-licensed for outdoor use. Only a small, 50 MHz portion (5825 to 5875MHz) could be de-licensed for outdoor use, but in the 2.4 GHz band the existing users were able to shift out of that band so it could be de-licensed for the outdoor use also.
There is also another important aspect which you have to keep in view while de-licensing any band: though the public requirement will be there, vendors naturally, you will agree, try to force the issue because they develop some equipment and they want to sell it. Now you would have seen that spectrum is a very valuable commodity, it’s a resource, a very valuable resource and billions of dollars have been spent by the operators to get the spectrum. On the one hand one operator is spending billions of dollars, on the other hand, another operator is using de-licensed spectrum providing the commercial services. There is a big gap as you can see. So when the government de-licenses any spectrum, the idea is that the public will use it for personal use. The intention is not to de-license it for the commercial use, because commercial usages will continue to grow, continue to increase. Today if you de-license 100 MHz, tomorrow requirements will grow and they will say that they need another 100 MHz, and after another few years they will ask for another 100 – 200 MHz. Hence the de-licensing requirements will never end. So this creates quite a difficult situation with respect to other commercial operators who have paid for this spectrum.
Yelena: That’s definitely something to keep in mind.
P. K. Garg: Yes. Even now, for example – some of these Internet service providers, they are using this 2.4 GHz band. Now this 2.4GHz de-licensed band which they are using, they are using quite extensively. As a result, the availability of this band for the public gets restricted. If members of the public want to use it they will get interference and if WiFi modems find one channel busy, they will find another channel, and if they find that channel busy too, then they will go to the third channel. Thus the number of channels which are available to the public will continue to decrease.
Yelena: So by the public you mean community projects or individuals?
P. K. Garg: Yes. For example – you and me– if we are to use a WiFi modem at home and if we find because of other commercial usages around our houses that we are not able to function, that means the benefit of that unlicensed band is lost for us.
Yelena: And there is no way to regulate so that the public gets more access?
P. K. Garg: No, because once the spectrum management authority de-licenses any band it sets certain parameters – how much power, what is the bandwidth of each channel, etc. Beyond that it will not control it. Only if somebody is using a higher power than what is allowed, it can be checked and then the person can be penalized. But whether it is for commercial use, personal use, or community use is a very grey area. It is difficult for the government to control this, requires a long litigation process and other interventions.
Yelena: That is certainly understandable. However, I’m looking into how unlicensed spectrum can be used for public good. There are projects that provide cost-effective wireless communication networks in remote areas, such as the Dharamsala wireless community network project set up by AirJaldi, or networks set up by the Digital Empowerment Foundation. So, our advocacy for de-licensed spectrum is to benefit such projects.
P. K. Garg: Basically you are right. For example – if you see now for the indoor use, almost 250MHz of spectrum in 2.4GHz and 5GHz band is available unlicensed. I would say even close to 300MHz is available. And out of this 300MHz almost 150MHz is available for outdoor use also. Now if any community wireless network or a city-wide network operating on WiFi is to be created, 150MHz is more than enough if it is used judiciously. However, supposing I am an operator and I started using the unlicensed spectrum for giving services to my customers throughout the city, and their data and the total time requirements are quite extensive, then what will happen is that it will either reduce the availability of spectrum to other users or it will become totally unavailable to them. So the basic purpose of de-licensing is defeated then.
Yelena: That makes sense. During my interviews with several other experts who are really strong advocates of unlicensed spectrum, it was stated that a solution to de-licensed spectrum being clogged is to de-license more spectrum. Considering what you mentioned, why are they still advocating for this?
P. K. Garg: Well, as I was mentioning to you, unfortunately many of these people are not fully aware of the ground realities of the situation. Now no part of spectrum is completely virgin. If somebody says today it is a question of 100GHz band to be de-licensed that is a different issue. But for the 100GHz band there are no devices available today and no usage. But if we come to the band anywhere below 10GHz, all the bands are used by somebody or the other. So before de-licensing, one has to shift those users and shifting anybody is not easy. Which band will the government shift them to?
Yelena: That is definitely something to consider.
P. K. Garg: The government will have to give them alternate bands; where will they provide them from when all the bands are in use? So the proponents of de-licensing sometime are not able to appreciate the full gravity of the situation. And another thing is that the government cannot be regulating whether it is a commercial use/private use or societal use. And many of the problems in the existing unlicensed bands have come up because they are really used more and more for commercial purposes. That is the unfortunate part of it. Otherwise 2.4GHz and 5GHz bands themselves are quite adequate to serve the societal requirements. Tomorrow, the ultra wide band systems are coming. The requirements will be there, but they will be operating in higher bands and they will be confined to very small areas; even home, even your office. So they can be tackled by indoor de-licensing.
Yelena: And this is according to the 2011 National Frequency Allocation Plan?
P. K. Garg: Well, the National Frequency Allocation Plan doesn’t talk of any additional de-licensed bands but they will consider it. However that will be for indoor, very low power applications. As I mentioned to you for indoor usage it is almost like 300 or 350MHz which is available; for outdoors it is only 150MHz which is available. So for indoor it is already 200MHz of additional frequencies which are available as unlicensed. And if all these bands are put to proper use it can very easily allow even up to 100Mbps indoor usage for every house. There shouldn’t be much of a difficulty. One is that all this indoor usages are not continuous; this 100Mbps is not 24/7 because there are short spurts usages over a few milliseconds, then there is a gap and all these WiFi modems and other devices make use of the dynamic situation, so they are able to coexist. Even if I am using it here and somebody else is using it at the other end of the room in this large hall it is possible to coexist. And certainly if it is the next house or the next building the same frequencies can be used; they will be reused. So the solution lies in the greater reuse of the same spectrum; whether you call it reuse or you call it sharing, it’s the same thing. So one has to share the spectrum; reuse the spectrum.
Yelena: Great. Thank you for your time and for sharing your expertise.
[1]. Indian Telecom Sector. (2010). Government of India Department of Telecommunications. Retrieved March 11, 2012, from www.dot.gov.in/osp/Brochure/Brochure.htm.
