Your browser is not supported. Please update it.

9 January 2020

Malaysia rejects 5G auction, but Taiwan’s 3.5 GHz sale breaks records

The first regulatory developments of the year highlight the contrast between authorities which are thinking out of the box about efficient spectrum allocation, and those which are clinging to traditional auctions in which operators end up paying huge sums for relatively small numbers of airwaves.

Malaysia’s government has published complicated but interesting plans to allocate spectrum in a way that would minimize the infrastructure investment required for 5G, while ensuring return on existing 4G investments. It is not going as far as some European regulators, like those in Germany and The Netherlands, in opening up bands for non-operators, but it believes its proposals would help to improve the business case for Malaysia’s telcos to support diverse industrial use cases, and to deliver benefits to different verticals.

By contrast, Taiwan’s operators have already paid over NT$100bn (US$3.3bn) for just 270 MHz of spectrum in the 3.5 GHz band, and the auction is still going on. The prices will far outstrip those in Italy, whose 3.5 GHz sale was deemed the most expensive to date, on a per-MHz/pop basis, in a recent study by the GSA. That puts the island nation in danger of being bracketed with India as clinging to old views of spectrum as a revenue source for government – in India, the operators are threatening to boycott this year’s auctions altogether if new rules are not introduced that would reduce their costs.

The Malaysian Communications and Multimedia Commission (MCMC) has outlined a strategy which shows good intentions and creative thinking, though it does carry the risks of over-complexity. The plans relate to four 5G bands – 700 MHz, 3.5 GHz, 26 GHz and 28 GHz.

In the lower two, MCMC wants to avoid the trap of forcing operators to compete for relatively scarce capacity (in 5G terms at least), and so ending up paying very high sums, with a poor knock-on effect on the ability to deploy rich networks. There are only 2×30 MHz of spectrum available in Malaysia in 700 MHz and 100 MHz (unpaired) in 3.5 GHz, so if that had to be shared between multiple MNOs, not only would prices skyrocket, but no operator would be likely to come away with sufficient capacity to support the differentiated services which industries such as manufacturing, as well as consumers, are hoping for in 5G.

So the licences in these bands will be assigned to a consortium of all the MNOs – Celcom, Digi, Maxis and U Mobile – allowing them to share a network and reduce their individual costs, while encouraging them to continue to invest in expanding their 4G networks, to complement the 5G build-out and address challenges such as rural coverage for citizens and industries. The plans do not appear to apply to 2.3 GHz and 2.6 GHz, which were also part of the 5G consultation and could still be awarded individually to supplement existing 4G holdings or support 4G/5G deployments in future.

Malaysia is certainly not the only country which is considering enabling, or even mandating, a single shared network to improve the cost and ROI picture for 5G operators. Mexico already has such a network for 4G, while Poland, Russia and others are considering this for 5G.

And MCMC’s plan does not go as far as the one proposed, during last year’s consultation period, by wireline incumbent Telekom Malaysia (TM). TM wanted to be given the rights to be the exclusive deployer of 5G in sub-6 GHz bands and petitioned the regulator for nationwide spectrum to build a single wholesale 5G network. It claimed “a drastic approach” was required to make Malaysia’s 5G roll-out cost-effective and to accelerate progress, in order to support its ambitious national digital and industrial objectives.

“The basic supply-side economics of individual mobile roll-outs and multiple overlapping networks can no longer work,” it wrote, and the expense of such an approach would limit the ability to launch critical services like health and public safety, or to extend mobile broadband to underserved rural areas.

“Splitting the spectrum bands to many operators is neither efficient nor economical for the country,” TM added. “A single national InfraCo will avoid duplication of infrastructure and networks, thus reducing the total cost of ownership for the industry.”

It seems that the regulator has taken that argument on board, without necessarily agreeing to the single deployer idea (it is not even clear whether TM will be able to join the consortium). It is unclear how its shared network approach would work in practice, and conflicts of interest between operators have been the chief reason why such schemes failed in the 4G era, in countries from Kenya to Russia.

So it is not impossible that Malaysia will end up with a variant of TM’s idea, with one operator shouldering most of the build-out burden and guaranteeing fair access to the others (the incumbent argued it was best-placed because it has the most extensive fiber coverage, with more than 540,000 kilometers in deployed fiber and core capacity).

In the lower millimeter wave band, 24.9 GHz to 26.5 GHz, licences will be assigned by beauty contest, which can be a very effective way to reward innovative business models, provided a clear framework and transparent process is put in place (some of the Nordic countries are probably the gold standards).

And licences in the higher mmWave band, 26.5 GHz to 28.1 GHz, will be assigned on a first-come first-served basis. These are open to any organization, not just telcos, and will be allocated for highly localized areas, not nationally like 26 GHz. This shows an attempt by MCMC to open up airwaves for industrial or other providers. These may find a better commercial model than the operators to support some localized or specialized user groups and applications, whether for enterprises, cities, rural areas or other sectors with demanding requirements.

The high band spectrum is seen by regulators as a good option for such policies because there is plenty to go around; it would therefore not attract huge prices in competitive auctions, so the Treasury is not losing out; and its limited propagation makes it better suited to very local networks than to wide area coverage. Ofcom, the UK regulator, for instance, published proposals last year to open up some mmWave spectrum for non-traditional operators with a view to encouraging small and medium enterprises and local services for rural, campus or industrial environments.

Over in Taiwan, competition in the 3.5 GHz auction – run on very conventional lines – has been so fierce that bidding broke the NT$100bn (US$3.3bn) mark last week and the projected total proceeds, when it ends next week, is around NT$114.9bn (US$3.8bn). By way of comparison, Germany made US$7.4bn from its 3.5 GHz auction, but has a population 3.5 times the size and a far larger coverage territory.

Prices have been driven up by the relatively small amount of spectrum (270 MHz) to be shared between five bidders – Chunghwa Telecom, FarEasTone, Taiwan Mobile, Asia-Pacific Telecom and Taiwan Star. There is a cap of 100 MHz per MNO, so in theory two operators could get the maximum amount each – also the amount many believe is the minimum required for sophisticated 5G services of the kind envisaged by Taiwanese users. But that would leave little or nothing for the other three. Alternatively, they could all get similar amounts, but none gain enough to launch highly differentiated services.

“The train has stalled and the five passengers on board have been injured,” said Rachel Liu, deputy secretary-general of the Taiwan Telecommunications Industry Development Association (TTIDA).

The bidding so far has gone well beyond the NT$44bn ($1.47bn) target set by the regulator, or the NT$60bn ($2bn) forecast by analysts. There has also been some auction of 1.8 GHz and 28 GHz spectrum, but 98% of the bidding relates to the anchor 5G band of 3.5 GHz. Indeed, there have been no bids for 1.8 GHz, even though the regulator has presented this as a key 5G band, especially for uplink – in terms of the global ecosystem, however, this remains primarily a 4G band and will not be extensively refarmed for many years in most regions.

The TTIDA’s concern is that most of the value of 5G, for operators and for the economy at large, lies in enterprise and industrial services, but inflated spectrum prices will distort this business model. To redress the balance, Liu has called for 5G to be exempt from spectrum usage charges, and for the government to offer incentives for operators which build up vertical industry partnerships and services.