How Conroy Lost His NBN Seoul

June 16, 2011
Communications & Broadband
Speeches

Labor holds up South Korea as the very model of a modern broadband society to whose technological sophistication Australia must aspire if we want to ‘keep up’.

But far from supporting Stephen Conroy’s recklessly costly and needlessly risky $50 billion national broadband network, Korea’s recent broadband experience throws into question some of the key assumptions underpinning the NBN.

Labor argues that only running fibre to 93 per cent of Australian homes and businesses can satisfy what it is certain will be greater and greater demand from consumers for bandwidth.

But the idea that bandwidth supply creates its own demand is a vast oversimplification. It completely ignores the critical role of other factors such as price, quality of service, availability of applications and competition among providers.

And it turns out it is exactly these factors that in Korea are proving more powerful than mere availability of high speeds.

The latest figures from Korea Telecom reveal that after five years of rolling out fibre-to-the-basement (a far less costly version of NBN Co’s fibre-to-the-home architecture), demand for the highest bandwidth plans it enables has gone pear-shaped.

Since January 2010, the number of Korean households subscribing to 100 megabits per second broadband has actually shrunk by 70,000, despite increased overall take-up of broadband.

At the start of last year more than 17 per cent of consumers were opting for higher speeds, but a year and a half later the proportion has dipped to less than 15 per cent. Presumably, many Korean families have discovered there are few, if any, applications for that much bandwidth. Or they have decided they would rather not spend the extra $5 a month for faster access.

What might this mean for the NBN?

While KT is struggling to get Korean customers to pay a retail price of around $31 a month for 100 megabits per second, the NBN Co assumes Australian consumers will happily pay at least $56 a month (and many argue a lot more) – and that by 2015 around 25 per cent of consumers will be opting for 100 megabits per second.

In other words, supply will create its own demand regardless of price.

NBN Co’s 2011-2013 corporate plan (which in December it was kind enough to provide in a bowdlerised form to the Australian taxpayers so generously funding its phalanx of highly-paid executives and massive construction cost) includes its version of a famous graph which shows the steady rise in internet access speeds since 1985.

The graph shows typical download speeds climbing from 10 kilobits per second in 1990 and 512 kilobits per second in 2000 to 10 megabits per second in 2010 (although in reality, average download speeds in Australia in 2011 are a fraction of that).

NBN Co extrapolates the past into the future and predicts that by 2025, download speeds will need to be 1 gigabit per second.

Since the mid-1980s, growth in the use of bandwidth has reflected an explosion of new applications using fixed-line communications networks. In 1985, most copper wire carried only voice telephony or faxes; today ADSL over copper delivers everything from email to interactive gaming and video-on-demand.

But it’s not just the technological revolution and proliferating applications of the past quarter century that have driven demand.

There has also been a revolution in economic thinking: monolithic government-owned telecommunications monopolies have been deregulated, opened up to competing carriers and technologies, and sold to private owners more focused on satisfying customers and controlling costs.

This process has not always been smooth or without controversy; in Australia, many people argue Telstra should not have been privatised as an integrated business, for example. But it has delivered immense benefits to household and business users.

Just think of the sophistication and variety of today’s broadband offerings and devices, or the steep fall in broadband costs over the past decade. According to the OECD, for instance, ADSL prices in Australia fell by 45 per cent between 2005 and 2008 alone.

So it has been a vast array of changes in the market for communications – including deregulation, increased competition, falling prices, new applications, greater focus on customers, and more varied devices and carriage technologies – rather than simple increases in feeds and speeds that has driven increased bandwidth use.

The NBN as conceived by the government will remove or neuter some of the most important forces at play in the past quarter century.

It will halt facilities-based competition between fixed-line technologies or networks. It will mark a return to monopoly public ownership. And it will reverse two decades of falling prices (if the NBN Co corporate plan, which projects real increases in the fees it collects from customers, is any indication).

Senator Conroy doesn’t care about any of that. He’s seen the future and to him it looks like Korea (even though it turns out most Koreans do not receive broadband over fibre that runs into their home, contrary to Senator Conroy’s occasional claims). In November 2009 he stated: “For other countries targeting broadband and ICT investments to stimulate economic growth, Korea provides a significant example of the capacity to succeed.”

But as we saw earlier, Korea does not seem to support the proposition that consumer demand for bandwidth will grow inexorably in line with supply. On the contrary it reminds us forcefully that price, competition and applications matter immensely.

And if we review the NBN Co’s published claims about the applications which it claims will drive usage, we find they quickly verge on the fanciful.

While high-definition video can currently be streamed over less than 10 megabits per second, and improved compression software has slashed the bandwidth needed to stream video 16-fold over the past decade, NBN Co assumes Australian households will inevitably migrate in massive numbers to broadband plans offering more than 100 megabits per second because they need a “full-scale entertainment solution (3D perspective, HD, interactive)”.

A few years later the same users will flock to 250 megabits per second so they can engage in “full 3D virtual collaboration” – I’m not sure exactly what that means, much less why it should be heavily subsidised by taxpayers.

With visions like that dancing before their eyes, no wonder the government and the NBN’s most enthusiastic supporters quickly lose patience with those, such as the Coalition, who suggest a better approach might be to conduct a real-world assessment of how much fixed-line bandwidth households will actually need at various points in time and how to most cost-effectively provide it.

That brings us to another assumption underpinning NBN Co’s plans – its conviction that as “demand for bandwidth increases it will become increasingly difficult for any non-fibre delivery technology to compete”.

This ignores the fact that the market has continually responded to demand for bandwidth up to now with more advanced techniques for getting more out of our existing infrastructure. There is simply no economic justification for junking networks that as yet show few signs of reaching their technological limits if these are in good nick.

Few people imagined in the early 1990s that Telstra’s copper network could deliver VDSL capable of 100 megabits per second over short distances, as it today can, or that network equipment suppliers such as Huawei would be testing technologies which potentially deliver bandwidth of up to 700 megabits per second over copper.

Nor would they have guessed in the mid 1990s that the HFC cables passing a third of Australian households could easily accommodate superfast broadband connections of 100 megabits per second, as they can and in some cases do (at least until their use to deliver broadband is contractually banned as part of the proposed Telstra/NBN deal).

Or that roadmaps to speeds of up to 400 megabits per second over HFC in the next few years would be an integral part of next-generation network rollouts in Europe and North America. Indeed, this week we saw the first reports of techniques claimed to achieve multi-gigabit speeds over HFC.

Although there are severe shortcomings in access to broadband in significant parts of Australia, such as blackspots affecting roughly a million premises in our cities and under-served uneconomic areas in the bush, these should be urgently addressed with targeted government subsidies, rather than used as an excuse to spend the next decade overbuilding perfectly good infrastructure across the rest of the country.

And if the past is any indication, movement by consumers up the speed ladder will have to be driven by competition and the innovation it encourages, rather than reduce the need for it.

As South Korea so clearly affirms, price, competition, service and market structure do matter and we should not set aside the hard-learned economic lessons of the past quarter century.

Recommended Posts