A large part of the problem lies in the differential rates of licence fees the government charges for different services. While it charges nothing from pure internet services (it's 6 per cent if ISPs offer VOIP services), it charges around 12 per cent on mobile phone revenues (licence fee varies from 6 to 10 per cent, depending upon the telecom circle plus a spectrum fee of 3-5 per cent).

In the case of Reliance, the company has been giving different revenue figures in its financial statements and a different one to the Telecom Regulatory Authority of India (Trai) - while the difference was as low as 7 per cent in June last year, it was up to 23 per cent in June this year.

The difference, according to a report by UBS Investment Research last month, is that the company has not included its revenue streams like those from the internet business. Why internet revenues should have shot up so dramatically in just a year boggles the mind, but the differential licence fees could just have something to do with it.

Indeed, Trai is currently investigating a case of an integrated telecom firm where, all of a sudden, there has been a huge spurt in the proportion of international long-distance calls as well - it's too early to say if the spurt is genuine, but in this case too, there could be a licence fee angle - at 6 per cent, such licence fees are half those charged on mobile telephony.

The ministry's refusal to clarify whether telecom operators can classify non-voice revenues (which accrue on the same spectrum they use for voice operations) separately and report them under their ISP revenues has implications in even the case of the 3G spectrum that will be auctioned soon.

Since the real worth of 3G comes into play in the case of data services (downloading movies on your phone), new operators can argue a large portion of this data is part of their ISP licences. So, instead of paying a 9 per cent revenue share (8 per cent licence fee and 1 per cent spectrum fee), they could end up paying next to nothing.

The other issue likely to cause a serious problem is the different fees for 2G and 3G spectrum. Why 3G spectrum should bear a 1 per cent spectrum fee when it has vastly more capacity to carry traffic than 2G spectrum (which pays 4 per cent) is not clear. In which case, it is always possible that operators will shift as many 2G subscribers as possible on to 3G networks to lower spectrum fee outgo.

Theoretically, the auctions for 3G will take care of this since the bids will take into account both the vastly higher carrying capacity and the much lower annual spectrum fees. But if the auctions discriminate against new players as they are, the bids will be low and allow existing players to make a killing once again.

One solution proposed by the COAI to ensure firms don't misclassify revenues is to measure traffic. This cannot be done in the old way of knowing how many active subscribers there are on each network - the Visitor Location Register treats a customer who makes one call a month in the same manner as one who makes a thousand. Instead, the COAI proposes measuring actual traffic, or erlangs, on each base station, whether 2G GSM or CDMA or 3G.

This may or may not work (technical solutions have a way of not working as Reliance showed some years ago when it used technology to change the numbers of incoming international calls to reduce its ADC levies) but it is worth trying - it's also worth Mr Behura making a quick call to regulators in each country to see if such revenue separation of the type COAI is advocating actually works.

In the long run, however, a single-if-lower revenue share is a more elegant solution. Of course, till the government continues to favour firms by offering virtually free spectrum and insisting on subscriber-norms for more allotment, none of this will be possible.