Despite this, however, lack of progress on two fronts can derail the governments hopes of getting R40,000 crore in the January auctions. One, as this column has pointed out before, the governments inability to get more 2100 MHz spectrum cleared in a swap with the defence ministry
(Its in the data, FE, November 12, goo.gl/zjQsMH) will lower investor appetite since it is 3G spectrum that telcos are really short of.
Equally important is the Trai proposal on a flat and uniform spectrum usage charge (SUC)apart from the annual license fee, telcos also pay a SUC depending on how much spectrum they have. Since this is a finite resource, the more the spectrum they have, the more they have to pay for it. Though a spectrum auction bid for a flat amount of spectrumwhat will happen in Januaryis theoretically very different from an SUC which is an annual charge, the two are related. Think of it as a security deposit on a flat and a monthly rental, the higher the monthly rental, the lower the security deposit, and vice versa. Except, spectrum is a bit more complicated.
To begin with, there are six slabs of spectrumas firms get more spectrum, they pay a higher rate for it, 3% of their annual revenues if they have up to 4.4 MHz of spectrum, 4% if they have up to 6.2 MHz of spectrum, and so on. It gets complicated because, while there is a GSM and a CDMA type of mobile, if a CDMA firm has 2.5 MHz of spectrum along with 4.4 MHz of GSM spectrum, the two spectrum amounts are not addedso, while a pure-play GSM firm would pay 5% of revenue as SUC if it had under 8 MHz of spectrum, a dual-technology firm with 2.5 MHz of CDMA spectrum and 4.4 MHz of GSM spectrum will pay a lesser 3%. Theres more. If a firm provides only internet services, it pays a lower rate. So, both Bharti Airtel and Reliance Industries are to pay a 1% of revenue fee for their 4G or broadband wireless access services, but a lot more for the voice part of their business.
In other words, its complicated and leaves a lot of room for what in polite society is called regulatory arbitrage. A firm that has both 4G and 2G voice revenues can, for instance, declare more revenues in the 4G bucket and pay a 1% SUC on them. Indeed, this has happened many times in the past with firms showing more internet-based revenues; when different rates were charged as ADC for incoming calls, firms disguised their international calls as local ones to pay lower ADC charges. Indeed, when 3G services were first allowed, the original proposal was to charge a flat 1% charge on 3G traffic but when it was pointed out that segregating 2G and 3G traffic would
be a nightmare, the government simply raised the SUC on all 2G and 3G revenues after clubbing them together.
While this was one of the main reasons for Trai coming out with a recommendation in favour of a uniform SUC across all telecom services, there were more. One of the reasons why India has such poor quality telecom services is that firms have sub-optimal amounts of spectrum, a fraction of what they have overseas. This is why the government changed the M&A rules, this is why it is allowing spectrum trading and spectrum sharingallowing sharing only for firms in the same telecom circle takes away from the idea of increasing competition, but thats for another column/edit. The problem, however, is that as firms take up more spectrum by either buying or sharing/trading, they end up paying dramatically higher SUCs. Thats one of the reasons cited by Trai for recommending a flat SUC.
Take a Bharti Airtel for instance. With adjusted gross revenues of around R32,000 crore, every 1 percentage point cut in SUC saves it R320 crore and every hike costs it that much more. A cut in SUC to 3% as recommended by Trai saves is R960 crore since it pays a spectrum charge of 6% right now; but an M&A which takes its spectrum charges up to 8% will add R640 crore to annual chargesthats a lot on a net profit of R5,096 crore.
But it is more than higher profits for a few firms, it is about the government losing revenues from firms misclassifying revenues by declaring voice revenues as data revenues, it is about firms not going ahead with M&As and having insufficient amounts of spectrum which results in poor quality telecom services.
Theres more. The telecom bureaucracy argues that escalating SUCs encourage firms to use scarce spectrum efficiently. Nothing could be further from the truth. Take BSNL which has 340 MHz of spectrum across various circles as compared to Bhartis 233 MHz and Vodafones 217 MHzbecause BSNL has few customersits revenues were R1,830 crore in the June quarter versus Bhartis R8,337 crore and Vodafones R6,614 croreit ended up paying R0.33 crore per MHz versus Bhartis R2.12 crore and Vodafones R1.54 crore. MTNL pays just R0.36 crore of SUC per MHz (see graphic).
Given the effort the government has taken to bring telecom back on track after five wasted years beginning 2008 when A Raja starved the industry of spectrum while giving it to a few chosen firms, it is a pity the entire exercise will come to nought with the telecom bureaucracy arguing it will lose revenues with a 3% SUC. This probably wont happen since lower SUCs will be passed on as lower tariffsit will also mean less misdeclaring of revenues to the governmentbut lets assume the telecom babus are right and that revenues will fall with a 3% SUC. Well, then, just hike it to 4% but still ensure it is flat and uniform.
It is also worth keeping in mind that the current escalating spectrum charge was devised in the days firms didnt pay for spectrum but got it free based on the number of subscribers they hadan escalating charge was one way to get them to pay for the scarce resource. Since theyre now paying for spectrum at market rates, where is the justification for escalating charges And keep in mind, the higher the SUC, the lower the auction bid in Januaryremember the upfront deposit and annual rentals