Of the 85 million new mobile connections of all types (both GSM and CDMA) added in the country between November 2006 and November 2007, nearly 27 per cent were accounted for by Bharti Airtel as compared to a mere10 per cent by Reliance and 13 per cent by fellow CDMA mobile provider Tata Teleservices [Get Quote], underscoring Reliance's need to begin offering the more popular GSM mobile services.

Beginning GSM mobile services is also critical since the Average Revenue Per User (ARPU) is much higher for GSM firms which, on average, earn Rs 275 per month as compared to Rs 173 for all the CDMA mobile firms. So far though, Reliance has managed to show an operating profit margin equal to Bharti's due to the lower operational and capital costs of CDMA mobiles. In the second quarter of 2006-07, Bharti's turnover was 1.2 times Reliance's and this rose to 1.4 in the second quarter of 2007-08. In terms of operating profits, this difference rose 1.3-1.4 times in the same period.

What's important for Reliance's GSM foray is that it has a big advantage over the two or three newcomers who are also expected to begin operations in the GSM mobile space. Not one to waste time, Reliance has already reportedly placed orders for 80-100 million lines to be set up over a period of about three years.

The Reliance advantage is two-fold. Bharti Airtel and Vodafone-Essar's quality of service is declining with the existing spectrum over-stretched, so if Reliance gets its GSM mobile network up and running fast, it can hope to attract several of their customers. If it gets the companies' older subscribers, this will hike its ARPU even more - in general, older subscribers, generally in the metros, pay a higher ARPU than non-metro users and are often heavier users of value-added services as well.

In the thick of the fight between the GSM mobile phone firms and the telecom ministry, Telecommunications Minister A Raja had announced that number portability would be introduced from April this year - this will allow Reliance to attract customers of existing GSM mobile phone firms.

With the Telecom Regulatory Authority of India (Trai) increasing the subscriber requirements two to three times for spectrum allocation, existing GSM mobile phone firms may be in a position to improve the quality of their services. In spite of several of them soon qualifying for fresh spectrum in some circles, they may still not get it.

The new subscriber norms notified by the ministry include allocation of spectrum till just 7.5 MHz, while firms like Bharti Airtel and Vodafone-Essar already have 10 MHz and will soon qualify for 12.4 MHz!

This is the reason why Trai chief Nripendra Misra has written letters of protest to the Department of Telecommunications (DoT). Misra has said that while he had prescribed a subscriber-linked path for the allocation of up to 15 MHz of spectrum, the DoT had notified a path of just up to 7.5 MHz.

In comparison with the new GSM-mobile players who've just been given their licences/spectrum, Reliance Communications has a cost advantage of anywhere between 20 and 40 per cent. As Romal Shetty, consulting firm KPMG's executive director, points out, 60-70 per cent of Reliance's existing CDMA infrastructure (the towers) can be used for its new GSM network.

A good way to illustrate this is to look at the country's expected mobile phone subscriber base of 500 million by 2010. If Reliance is to get 20 per cent of this market, it needs to get around 65 million more subscribers. Assume all of these are to come from GSM mobile services and not from CDMA that Reliance offers at the moment.

To service such a market, a newcomer would have to set up 65,000 towers (assuming one tower can service an average of 1,000 customers) which could cost around $4bn and another $2.5bn or so for the electronics.

Reliance, however, already has 22,000 towers, according to company officials. So, while it will need to spend around $2.5bn on the electronics for GSM services, it will need to build just 43,000 new towers - this reduces its overall capital costs by around a fifth (given that Reliance has transferred all its towers to a separate business, Reliance Telecom Infrastructure Limited (RTIL), the capital costs talked of refer to RTIL - the point, however, remains since the revenues Reliance Communications will pay RTIL will depend upon the overall cost the latter incurs).

Reliance officials, however, claim that they will have 40,000 towers up and running by the end of this fiscal. Though rivals argue it is not possible to ramp up so fast, it is useful to see the impact of this on its costs. If the company has 30,000 towers, and needs to build just 35,000 more, its capital costs will fall to $4.7bn or nearly 30 per cent less. If the company has 40,000 towers, the capital costs go down by over 35 per cent. 

In the first year of operations, however, Reliance's savings will be a lot more since it can use its 22,000 existing towers first and simply instal the GSM electronics required on them - for the first 22 million or so subscribers that these towers can accommodate, Reliance's capex costs will be a mere $35-40 per subscriber, according to Shubham Majumder, associate director-research, who tracks telecom at Macquarie Capital. This is in comparison to the $100 or so that newcomers will have to invest in capex per subscriber.

Obviously, existing mobile phone firms such as Bharti Airtel and Vodafone-Essar will also be able to expand at $30-40 per subscriber since they can use their existing towers and just need to add in electronics - but if the firms don't get additional spectrum, such comparisons are meaningless. This is the crux of the current tussle between the GSM mobile firms and the ministry of telecom.

While many believe it will be difficult for Reliance to offer both CDMA mobile and GSM mobile services at the same time, there are some precedents for this. Vivo, the largest operator in Brazil and the country's only CDMA operator, moved to a GSM network three years ago and 70 per cent of its net additions came from GSM services. For now, it is definitely advantage Reliance.