Analysis: Did NBN Co pay too much for wireless spectrum?

David Havyatt crunches the numbers - and the results aren't rosy.

There is probably a good reason Senator Conroy avoided mentioning the $120 million price NBN Co paid Austar for wireless spectrum last week in his otherwise laudatory press release. 

The simplest of analyses suggests NBN Co has paid way too much for this spectrum. 

For those needing a refresher on last week’s news, NBN Co paid $120 million to acquire five year leases on the 2.3 GHz and 3.4 GHz bands.

Usually, the price paid for spectrum is governed by a price-based allocation (usually an auction) for spectrum licenses. The 900 MHz band originally used for GSM is the major exception.

Every price based allocation has occurred in different economic climates, with different lot structures.  A” lot” is a small band of the frequency available for a specific geographic area. 

Below I have tabled some of the key prices paid for wireless spectrum in the past.

Prices vary wildly – as shown by the excessive price in the PCS2000 auction when One.Tel and Hutchison went toe-to-toe for some 1800 MHz spectrum versus the relatively lower price for the 3G spectrum. 

But our interest today is in what NBN Co paid – being that it is a wholly-Government funded entity that requires our scrutiny.

The 2.3 GHz licences were originally auctioned as 7 MHz channels for doing narrow-cast television (MMDS), and this is what Australis and Austar used them for. The advent of both satellite and HFC Pay TV meant this was no longer their best use, and the licences were converted to fifteen year spectrum licences in 2000.  As you can see in the table above, the various holders paid a total of $71.17 million to convert the licences.  The original five-year licences had already expired. 

The 3.4 GHz spectrum was auctioned in 2000.  It could be argued that the price was suppressed by the decision to exclude Telstra from the bidding.  Nevertheless the price stated ($112.50 million) is how much was paid. 

For the two bands combined, the total was just over $180 million. 

Since those auctions, a number of transactions (spectrum swaps) have resulted in Austar basically retaining these two spectrum bands for Australia's regional areas and Unwired retaining them in the cities. 

It’s important to factor this in – spectrum prices are usually higher in denser areas due to the lower capital cost to build.

Allowing for the fact that Austar’s licences cover about 40 percent or less of the population and the effect of lower values, the Austar licences should represent about a third of the value of the original price. So you could say that the spectrum Austar sold to NBN Co was worth about $60 million when allocated in 2000.

Further, the original licences were for fifteen years of which NBN Co has only bought the last five years. So a fair value for what NBN Co has bought - ten years removed - is really about $20 million. 

I would thus suggest that NBN Co has paid about six times what it should for spectrum the NBN implementation study noted (at page 274) could only provide a cell radius of 7km, compared to 14km for the 700 MHz band.

Unfortunately for NBN Co and Conroy - which obviously feel the need to kick some quick goals, the 700 MHz band isn’t yet available for auction (until all Australians switch from analogue to digital television).

NBN Co might argue that they were prepared to pay a price premium because the Government has discussed the potential for fifteen-year licences to be “renewable” if services are  “in use” on that spectrum. 

But there is no reason why NBN Co was required to buy spectrum on the market when the Government can resume spectrum licences. 

The procedure for resumption, as specified by the Radiocommunications Act, specifies that on resumption the market value will be paid where “the market value of the licence, or the part of the licence, at a particular time is the amount that would have been paid for it if it had been sold at that time by a willing but not anxious seller to a willing but not anxious buyer.”

I would argue that the value paid by NBN Co is that of an anxious buyer.

There is a risk with NBN Co that the eagerness of both the company and the supervising department to get deals done and make progress on the project will result in all parties being prepared to pay too much for assets.

The challenge for NBN Co is that its prices have to be approved by the ACCC – and the competition watchdog is more interested in the “efficient price” of assets than the price paid. Stay tuned for more on this.

What do you think? Did NBN Co pay too much?