Yesterday, TRAI came out with its recommendation on 5G spectrum pricing and other related issues. TRAI did a commendable job in recommending prices that are significantly lower than it had recommended in its earlier recommendation on 1st Aug 2018. The purpose of this note is to understand what TRAI did differently this time and whether the prices it recommended will create undue biases in the operator’s preference for opting for one spectrum band over the other. If yes, then how will it impact the operator’s ability to roll out a reliable and effective 5G network.
TRAI’s Pricing Strategy
This time TRAI digressed in its pricing strategy on many accounts.
A) Unlike, in the past, (when the TRAI called a price “market-determined” when even a fraction of the offered spectrum got sold), this time TRAI has decided to call the spectrum price as such only when all the spectrum offered in an LSA got sold and nothing was left unsold. This is a huge digression from its past approach, which helped in driving the prices down significantly.
B) In the past, for all such LSAs, where a “market-determined” price was established, TRAI just scaled up this price with an appropriate index (cost of inflation) in order to determine the reserve price for a subsequent auction. Since, in the earlier auction of Mar 21, a large amount of spectrum was left unsold, hence 99.9% of the prices that emanated out of this auction cannot be pegged as “market-determined”.
C) As in the past for LSAs where market-determined prices could not be arrived at, the TRAI used a ground-up-based model to determine spectrum valuation and then peg the final reserve price as 80% of that valuation. Now the TRAI this time has further diluted this number from 80% to 70% — thereby, driving down the prices significantly.
D) For those circles where all the offered spectrum got sold, TRAI indexed the auctioned price of that LSA and compared it with 70% of the “calculated valuation” of the same LSA, and pegged the reserve prices as the higher of the two numbers.
TRAI’s Pricing Models
A) 1800 MHz
You can see here that only Orissa the TRAI was required to adopt strategy D, and for the rest of the circles, the reserve price is just a 30% discount from the average value of the valuation models.
B) 900 MHz
You can see here that only in Gujarat & Kerala the TRAI was required to adopt strategy D, and for the rest of the circles, the reserve price is just a 30% discount from the average value of the valuation models.
C) 800 MHz
Here none of the LSAs had spectrum in the offer which got fully sold in the past auction, hence reserve price is just a simple discount of 30% from the average valuation of the ground up valuation.
D) 2300 MHz
Here in 6 LSAs, the TRAI used strategy D, but fortunately in many circles where spectrum got fully sold in the past auction is totally irrelevant now as no new spectrum is available for sale in these LSAs.
E) 2500 MHz
In none of the circles, rule D was required to be applied. Hence RP was just a 30% discount from the average valuation.
F) 3500 MHz
In order to arrive at the price of this band, the TRAI used the earlier approach that is used in its recommendation dated 1st Aug 2018. RP of 3500 MHz = 30% of 1800 MHz. The reason — TRAI estimated the radio waves in the 3500 MHz traveled only 30% of what they did in the 1800 MHz band. Hence, the operators will be 70% more BTS to cover an area in the 3500 MHz compared to the 1800 MHz band.
G) 26/28 MHz
In order to arrive at the RP for this band, they simply took the ratio of the auction prices for various countries of 3500 MHz and Millimeter-Wave Bands and calculated the average. This came out to be 2.2%. Hence the RP of the 26/28 GHz band = 2.2% of 3500 MHz.
TRAI Recommended Prices
The following is the comparison of the TRAI recommended prices with those of the recent past (2021 AP/RP) at a Pan India level.
The band wise % discounts from past prices are as follows — 700 MHz (40%); 800 MHz (24%); 900 MHz (26%); 1800 MHz (36%); 2100 MHz (48%); 2300 MHz (38%); 3500 MHz (36%)
Since the 26/28 GHz band is a new band we do not have past prices available for the purpose of calculation of discounts. Also, the prices of 600 MHz have been not listed, as it is a new band its prices have been pegged with the 700 MHz band.
The Challenge
A) The prices of the new 5G band (except the millimeter-wave 26/28 GHz band) are still high. For example, the prices of 600 & 700 MHz are 8.4% higher compared with the existing 800 MHz. The reason this is a problem as the operators will need larger chunks of this band (at least 2x15 MHz each) in order to create a reliable and robust 5G network. And at the current rate, this will cost them individually Rs 60K Cr (7.8 Billion US). You can learn more by referring to my earlier article — “5G in India — The Real Challenge”.
B) The price of the 3500 MHz (which is called the Core band for 5G) is still very high. To get a sense of the magnitude, an operator will need min 100 MHz (2x50 FDD equivalent) in order to create a reasonable impact on 5G. And for this, the operator has to spend Rs 31 K Cr at the RP ($ 4.17 Billion).
The Opportunity
The price of the millimeter band looks much more reasonable compared with other bands. The SP has to spend just Rs 2796 Cr ($368 Million USD) to get a chuck of 400 MHz of this band. The other advantage is that it will drive the SP's overall SUC (spectrum usage down) significantly by expanding the denominator is a larger value when the weighted value number is calculated.
Conclusion
Though TRAI has done a fantastic job in cutting prices of spectrum using strategies that can be called revolutionary, it could not undo the burden of our past — when spectrum prices were set unreasonably high. Hence, in my view, if the prices in the key 5G bands (700 & 600, 3.5) are not revised further down then 5G in India will start incrementally and continue to ride on legacy 4G networks till the new 5G bands, especially in the low-frequency bands, become further affordable. Hence, till that time a common man will not see a huge difference, as the coverage of 5G will be limited to a few pockets in the densely populated urban areas, and voice will continue to run on legacy 4G networks.
(Views expressed are of my own and do not reflect that of my employer)
PS: Find the list of other relevant articles in the embedded link.
