Kickstart Forum Keynote Address February 2012: "The right role for Government in ICT Policy"

Feb 26 2012

 

Introduction

We are living through an extraordinary experiment in public policy in information and communications technology in Australia. 

The current Government’s policy on the National Broadband Network is way out of line with anything previously done in Australia – and in most other countries as well.

I’ve been privileged to work on public policy in this sector since the mid nineties.  Like my colleague Malcolm Turnbull, and many others in the Coalition, I am passionate about ICT and broadband.

This is a sector which attracts huge and growing attention from governments – and rightly so.

Certainly the Rudd-Gillard Government – I think that’s the best terminology in the current circumstances – has bold ambitions in ICT, across many portfolios.  As well as the NBN, there is the multibillion promise to deliver computers in schools; the $470 million program to build personally controlled electronic health records; there is the work to automate the back office functions in the massive superannuation sector.

So I think it is timely to reflect on the right role for government in making ICT policy.

Today I want to argue that over the last twenty five years, in Australia and around the world, we have learned some important lessons about effective policy in this area. 

And I want to make the case that the Rudd-Gillard Government is violating most if not all of those lessons.

Because ten is a nice round digital number, I have identified ten such lessons which I want to run through today.

1.    You can’t mandate take up

In my view the first lesson is that you can’t mandate take up.  You cannot know in advance which applications people will find useful and incorporate into their lives, and which applications will leave them cold.

There are plenty of applications which have bombed because end users just were not excited by them.  AT&T first tried to introduce video telephony in the US in the late sixties, under the brandname ‘Picturephone’ – but it was a commercial failure because customers just did not see the point in it.[1]

Video telephony seems to have had an enduring attraction for telecommunications marketers, because when third generation mobile services were launched in Australia by Hutchison 3 in 2003, the big selling point was the capacity to make two way video calls. 

The launch event featured then Premiers of Victoria and NSW, Steve Bracks and Bob Carr, conducting a video call.  But even though 3G dominates the mobile sector today[2], how often do you see people having two way video conversations on their mobile?

Just as many applications do much worse than expected, some do much better than expected. When Optus launched, it was long distance services which were expected to be the rivers of gold; the syndicate only acquired a mobile licence because it was bundled in with the fixed line licence being issued by the government. 

But as it turned out, mobile was much more successful than long distance.

If private sector operators, with a focus on sales and marketing, can get it so wrong in predicting what people might want to use, governments have an even less compelling track record.  To quote just one example, in 1982 the UK Minister for Information and Technology predicted that by 1990 multi-channel cable television would be widely used for “armchair shopping, banking, calling emergency services and many other services.”[3]

When you narrow down the focus from take up of ICT products and applications generally, to consumer take up of high speed broadband services generally, the evidence again suggests this government’s strategy is seriously flawed.

Let me offer three data points which raise very serious doubts about the strategy of building – at huge expense – a ubiquitous super high speed fibre broadband network as a means of achieving very high take up.

One: according to the most recent OECD broadband statistics, Japan is ranked only sixteenth in broadband penetration – even though it has the second best fibre network availability in the OECD after Korea.[4] 

Second, the Netherlands, which ranks highest in fixed broadband penetration at 38.5 services per 100 inhabitants, has pretty low fibre take up – only 1.3 services per 100 inhabitants are fibre (compared to Japan’s 16.4.)[5] 

Third, in the UK, some 57 per cent of homes are able to receive what that country defines as super fast broadband, that is, 24 megabits per second or above.  Yet only two per cent of all homes actually choose to take the high speed service.[6]

In summary, I think the lesson of new IT applications, whether they be launched by government or private sector, is that you cannot mandate take up – and that is a serious problem confronting the NBN.

Another initiative of this government, the Personal Electronically Controlled Health Records program - which supposedly is going to be operational by 1 July this year – faces a similar problem. 

The key gatekeepers are doctors and other health care practitioners.  Unless they are enthusiastic supporters and adopters of the new system, it will languish unused.  But there is little evidence of any serious thought being given to how take up will be encouraged; again this government seems to assume that if you put the system out there you can essentially force people to use it.

2.    Beware of the IT big bang

A second lesson from IT implementation, in the private and public sectors, small scale and large scale, is to beware of the big bang. 

In telcos, banks, airlines and other corporates, there are plenty of horror stories about huge IT projects – for a new billing system, or customer service, or other purposes – being abandoned, after spending tens or even hundreds of millions.

A good example in public policy in recent years is the British attempt to establish a universal system for electronic health records – recently abandoned after spending some 11 billion pounds.[7]

The lesson of experience in IT is: be very, very wary of the big bang.  Often an incremental approach is more prudent. 

Yet with the NBN, what the Rudd-Gillard Government is attempting to do is build three brand new networks using brand new equipment and the very latest generation technology to pass almost fourteen million premises by 2025[8]; to do so with a brand new company, established from scratch with no existing track record; to force some ten million households off their existing network onto the new network; and incidentally to underpin all of this with brand new billing systems, operational support systems, customer support systems and all of the other IT apparatus of a huge telecommunications enterprise. 

You could not possibly come up with a higher risk, more disruptive, more eggs-in-one-basket way to upgrade Australia’s broadband infrastructure.

3.    Your goal is behavioural change; IT is just the tool

My third observation is that the goal of the exercise is not to build a new IT system.  The goal of the exercise is to change behaviour.

In the case of the NBN, the goal is to get people using high speed services for a whole range of socially worthwhile purposes: doing business with each other, providing and receiving medical services, education and so on. 

In the case of PCEHR, the goal is to capture the personal health records of all Australians, and have health professionals access those records, and update those records, every time they interact with a patient.

In each case, the goal is a pretty sensible one: the big question is whether the particular hardware and software being introduced is going to be used by those we want to use it, to the degree we want. 

Unfortunately, most of what we hear from government concerns the technology – the inputs. 

But if you can’t tell us what people are going to use 100 Mbps for, you do not have a very compelling case that people are going to change their behaviour.

Last year, I participated in the House of Representatives Infrastructure and Communications Committee inquiry into the Role and Potential of the NBN. 

The evidence we received from a range of government departments was quite underwhelming. They could not tell us what applications required speeds of 100 Mbps; nor could they tell us about how they planned to use the network to deliver new services. 

Much of what they told us was highly unpersuasive.  For example, the Department of Innovation, Industry, Science and Research tried to argue that the NBN was in some way helpful for the square kilometre array – the planned super high resolution space telescope which Australia is bidding to build, with international funds, in the Western Australian Desert.[9]

One example I thought was quite telling concerned an application developed by GE and Intel for remote monitoring of elderly patients, and trialled by an organisation called Hunter Nursing in 2010.  The government has sought to use this application as an example of the need for the NBN – indeed it is cited in a fact sheet issued by NBN Co – yet the Intel-GE executive who appeared before the Committee told us that the speed required for this application to work well was 512 Kbps.[10]

My conclusion from this inquiry: the focus so far has been on the technology, not how to achieve the behavioural change the technology is supposed to bring. 

4.    Government or private sector, you can lose a lot of money betting on technology

The fourth lesson is so obvious to those familiar with ICT – that is, everyone in this room – I almost hesitate to mention it.  But I will, because it is far from clear the Rudd Gillard Government recognises the point: you can lose a lot of money betting on technology.

The economic structure of this sector is well known: you face very high fixed costs and very low marginal costs.  You have to spend the money up front – and only after you have spent the money do you know if you are going to get the take up, or the sales, that you need to generate a positive business case.

Prudent policy makers would at the very least familiarise themselves with the long history of financial disasters in the technology sector in Australia. 

Given that NBN Co has just proudly announced it is buying two satellites, let’s start with Aussat.  This government owned satellite operator racked up negative shareholders’ funds of $800 million – before bundling it into the second telecommunications licence so the accumulated losses were assumed by Optus in the early nineties.[11]

In the fibre backbone space, there was IP1, which built a fibre network between Melbourne and Perth.  It collapsed in 2003, and was bought by Telstra for $25 million, a fraction of what it cost to build.[12]

Then there was the Nextgen, which spent $850 million on a fibre backbone network running 8400 kilometres between most of the capital cities in Australia. This also collapsed in 2003, and construction giant Leightons – after writing off its own initial investment in the venture - bought out the rest of the syndicate for less than $40 million. [13]

  I could also mention One.Tel.  After building a strong brand as a reseller in the late nineties, the Packer and Murdoch backed One.Tel decided it wanted to build and operate a mobile network.  It spent $523 million in March 2000 buying spectrum – 15 MHz across the nation – but by 2001 the company collapsed.[14] 

Two other financial failures in Australia – disguised by the broader ownership of the asset – have been the Optus and Telstra HFC networks.  When SingTel acquired Optus it wrote down the value of the HFC network by 1.2 billion dollars.[15] 

If we wanted to cast the net a bit more broadly we could cite the literally dozens of spectacular IT and telecoms collapses around the world in the 2001 tech wreck.  Who could forget Worldcom?  Or Global Crossing?  Or Pets.com? 

5.    The more you leave in the hands of the private sector, the better

Let me turn to a fifth, related lesson about ICT policy: the more you leave in the hands of the private sector, the better. 

For one thing, as I have just argued, if you can do your dough so easily, then the last thing you should do with public money is make big technology bets.

For another thing, private sector companies are simply better at sales and marketing than the government.  In telecoms you can look at the way that incumbents have been outmarketed by new entrants around the world; in airlines you can look at how government owned flag carriers have been left looking like lumbering dinosaurs by nimble low cost new entrants.  

So if your goal is driving take up, why not rely on the private sector – which is much better at achieving this than government can ever be.

I believe the mobile sector in Australia is an example of good ICT policy.   Over the last twenty years it has shown extraordinary growth.

Penetration has grown strongly and steadily and is now over 125 per cent (that is, there are 1.25 mobile services in operation for every Australian)[16]; prices have dropped steadily; there has been relentless technology, product and pricing innovation; network coverage has continually improved; and the mobile phone has become an essential tool in the life of virtually every Australian. 

This was not achieved by government building and operating a mobile network.  Instead government set the ground rules, issued licenses to three competing operators, set some coverage requirements; and then essentially got out of the way. 

All of the mechanics of building the networks, establishing the distribution channels, advertising and marketing the product and delivering successive rounds of technological innovation were done by private companies.

I need hardly add that the NBN is using precisely the opposite approach – which is why I predict it will secure precisely the opposite result.

6.    What Government owns, government is conflicted about

If maximum use of the private sector is desirable, maximum ownership by government has many dangers. Lesson number six is that what government owns, government is conflicted about. 

Inevitably, if government is a shareholder in a business venture, it is sorely tempted to fiddle the rules to maximise the success of that venture. The interests of customers – that is, citizens – come off second best.

That conflict bedevilled telecommunications policy in Australia for many years, and was central to the Howard Government’s decision to privatise Telstra.

Now we see the same conflict reasserting itself with the NBN.  Let me remind you of a few examples. 

The legislation passed last year which bars new entrants from competing with NBN Co unless they, like NBN Co, operate on a wholesale only, open access basis has the effect of giving NBN Co a statutory monopoly to reinforce its economic monopoly.  It is a return to the bad old days of egg marketing boards and the like – things we thought had gone forever after the competition reforms of the nineties.

The decision to have NBN Co own and operate its own satellites -  at odds with the recommendation of the McKinsey-KPMG Implementation Study[17]- is one which serves the institutional imperatives of an empire building NBN Co, but which is not in the interests of end users.

NBN Co’s approach to wholesale pricing similarly is not in the interests of end users.  It initially proposed that the regulator should allow it to increase prices in real terms by five per cent per year, on all but the entry level service.[18] 

After a storm of protest NBN Co backed down somewhat, and now seeks to increase nominal prices by up to half the rate of inflation each year.  But as my colleague Malcolm Turnbull has pointed out, this is still a price increase – and compares very unfavourably with OECD figures showing that the nominal retail price of ADSL broadband fell by 69 per cent between 2005 and 2010 in Australia.[19]

7.    If you want the lowest price and highest take up, your first priority is getting a competitive market structure

This brings me to lesson number seven.  If your public policy objective in IT and communications is to achieve the lowest possible prices to end users, and in turn the highest possible take up – and I think this is a pretty good objective – then the best policy tool by far is to build a competitive market.

The evidence from the Australian mobile market which I cited earlier is one powerful piece of support for this claim.

The experience of fixed broadband competition by operators using the unconditioned local loop service offers further evidence. 

Using ULL, a competitor to Telstra can lease the copper wire from an exchange to a customer’s premises for a price of, currently, around $16 a month.[20]  The competitor puts its own equipment, a DSLAM, in the Telstra exchange, and supplies the customer with a DSL modem. 

Given combined telephony and broadband revenues per customer, on average, of around $80, the competitor has quite a lot of margin to play with – and a cost structure a lot closer to that of a network owner than that of a pure reseller. 

It was ULL based competitors who were the first to introduce ADSL2+ and forced Telstra to respond.  It was ULL based competitors who were the first to introduce naked DSL – so a customer could take a DSL service without also having to pay $30 a month for a voice line rental service that the customer may well not require.  In other words, it was ULL based competitors who drove prices down in broadband.[21]

In my book Wired Brown Land I looked at pricing trends in retail DSL in Australia.  I found a pretty conclusive pattern: when competitors were simply reselling Telstra’s DSL service, they had low margins and little capacity to compete or differentiate in any meaningful way. When they moved to using ULL, that changed quite a lot. 

Yet the direction we are now following, under NBN, is to move away from unconditioned local loop and back to pure resale.  From a competition point of view, that is a retrograde step – made worse by the fact that the HFC networks which provide physical competition are going to be taken out of service.

8.    If you want the best technology, your first priority is getting a competitive market structure

Competition is critical for delivering the lowest prices and highest take up – but the next policy finding, I argue, is that it is equally critical for delivering the best technology. 

The implicit bargain with the NBN is that we’ll give you the best technology but we have to cut some corners on competition.

But getting the latest technology is not a one time exercise.  Technology constantly evolves.  You need is a mechanism for constantly bringing the latest technology to the market.  The most effective such mechanism is a competitive market.

For evidence of this, you could look at market structures with limited competition where operators have been slow to bring new technology to the market. Telstra’s leisurely approach to upgrading from first generation DSL to ADSL2+ is one example. 

Another good example is the way that Telecom New Zealand stuck with the elderly AMPS standard for many years after it had been withdrawn in Australia.  Why?  Because they had a captive market and could get away with it – so why bother to spend the capex on upgrading their network?

Conversely, you could look at the relentless pace at which new technology has been introduced in the highly competitive mobile market in Australia over the last twenty years: from AMPS to GSM to CDMA to 3G to NextG and now 4G or LTE. 

In my view, the Rudd-Gillard Government has got itself into territory it should never be in: not just specifying a technology but in fact directly investing in that technology.  The likely consequence: we will be stuck with its technology choice for many years, whether it turns out to be a good choice or a bad choice. 

9.    Let information flow freely

The ninth principle of good ICT policy, in my view, is to let information flow freely. If there is one thing government could and should do, it is to be an honest broker pumping reliable information out into the marketplace.

Let me give one recent example of where inadequate information caused real problems: the OPEL network.  This was the Optus-Elders joint venture which won a competitive selection process in 2007 under the Howard Government’s Broadband Connect Program, to build a rural broadband network, with the assistance of a government grant of almost one billion dollars.

Notoriously, Stephen Conroy cancelled this contract after coming to government.  The core dispute was not over how many premises would be served by the network, it was over how many of those were ‘underserved premises’ – defined as premises unable to receive a ‘metro-equivalent broadband service.’

OPEL had prepared a database of such premises based on publicly available data sources, applying the definitions in the program guidelines.  It became clear during contractual negotiations that the Department had prepared a database which delivered different outcomes. 

I have to declare an interest here.  I was very heavily involved in the OPEL proposal, and I consider its abandonment to be a tragedy. 

But the broader point I want to make is that this tragedy could have been avoided if the database of underserved premises was transparent and available from the outset for public scrutiny – including by potential bidders. 

As it was, OPEL only saw the Department’s database a few short days before the contract was signed – and long after its bid had been lodged and chosen.

When we did see it, there were many puzzling anomalies. To give one example, maps prepared using the Department’s database contained several circular areas on the NSW North Coast completely free of underserved premises  – on the basis, apparently, that these areas already had wireless coverage and it was ‘metro-equivalent.’

I think we should be working to avoid a repeat of this public policy failure when it comes to the NBN. 

That is why the Coalition has called – including in our policy at the last election – for a comprehensive and transparent database identifying, on a premises by premises basis across Australia, the broadband speed presently available to each such premises. 

This database could underpin rational decisions to be made about where to invest money to upgrade the network – and where such investment was not required.

Incidentally, it would also allow individual citizens to compare the speed they were actually getting with what the database showed – allowing anomalies to be highlighted and corrected.

I think there is a useful analogy in the role of Geosciences Australia in providing objective, high quality data to the resources industry.  Government allocates extensive resources to obtaining and publishing this data – but then leaves it to the private sector to use the data to inform commercial decisions about where to prospect and drill.

Don’t let the rural tail wag the metro dog

I want to close with a final lesson, concerning a distinctively Australian aspect of IT and communications policy: how do you deal with the challenge of providing services to the vast rural and remote areas of our continent.

I think an error which has repeatedly bedevilled Australian IT and communications policy is to let the rural tail wag the metro dog.

To explain this point, while we have vast geographical areas in rural and remote Australia, and these present huge challenges, the vast majority of Australians live in the cities. 

As one example, during the Howard Government years I worked on policy changes to deal with Telstra’s ‘extended zones.’  These were areas where there were no local calls; every call was a timed, STD call.  The extended zones covered 80 per cent of Australia’s land mass – yet in total there were around 30,000 services provided by Telstra in that whole area.

Of course addressing rural communications needs must be a priority, and always will be under a Coalition government.

But I think it is an error to try to solve the problem of rural communications by building in cross subsidies between metropolitan and rural users.

The success of the unconditioned local loop policy came because until recently it used the opposite approach of geographic de-averaging. 

Prices were set by the ACCC in four separate geographic bands: CBD, the suburbs, regional areas, and then remote areas.  The last set of per line charges for the three inner bands were, respectively, $6.60, $16 and $31.30.[22] 

This allowed customers in the cities – the vast majority – to have much cheaper services, reflecting the true costs of serving them.

As the incumbent monopolist, Telstra always argued for a geographically averaged charge.  That is now the approach which the Rudd-Gillard Government has mandated for NBN Co.  It is a big mistake.  It entrenches much higher costs for the majority of Australians who live in the cities, and it also disguises the cost of serving rural and remote areas.

In my view, a much better approach would be to provide explicit, on budget subsidies – allocated on a competitive basis – to operators providing services in rural and remote Australia. 

Conclusion

Let me conclude with the observation that as IT journalists, you cover a fascinating sector.  It is a sector in which governments have bold ambitions and big cheque books.  There are plenty of enormous private sector players as well. 

Inevitably, there have been some hard lessons learned, in both public and private sector approaches to ICT.   Today I have sought to distill, in a personal view, some of those lessons as they apply to ICT policy in Australia. 

In my view, the approach of the Rudd-Gillard Government, on NBN Co as well as on a number of other policy initiatives, reveals a failure to learn from many of these lessons.  I fear that the outcome of the NBN experiment – and of other aspects of this government’s ICT policy – will in consequence be an unhappy one.

Let me close with this observation.  All of you have a passion for ICT and its power to transform lives.  The Coalition shares that passion.  We are strong supporters of upgrading Australia’s broadband infrastructure. 

We simply say – as I have sought to demonstrate today – that experience shows some policy approaches work well, and others do not, when it comes to delivering the benefits of ICT as broadly as possible.  That is why our policy approach is informed by experience – rather than seeking to fly in the face of it.

ENDNOTES

[1] http://www.fastcompany.com/magazine/106/next-past-perfect.html, downloaded 25/2/12

[2] For example, around two thirds of Optus’ 9.4 million mobile services are 3G. SingTel, Management Discussion And Analysis Of Financial Condition, Results Of Operations And Cash Flows For The Third Quarter And Nine Months Ended 31 December 2011, pp 44 and 45  

[3] Cable & Wireless Optus, Submission to the Productivity Commission Review of the Telecommunications Regulatory Regime, downloaded from www.optus.com.au, 25 February 2012

[4] OECD Broadband statistics [oecd.org/sti/ict/broadband], 1d (1). OECD Fixed (wired) broadband subscriptions per 100 inhabitants, by technology, June 2011, downloaded 25 February 2012.

[5] ibid

[6] Ofcom  Communications Market Report 2011,pp 246-247, http://stakeholders.ofcom.org.uk/binaries/research

/cmr/cmr11/UK_Doc_Section_5.pdf

[7] http://www.telegraph.co.uk/health/healthnews/8780566/Disastrous-11.4bn-NHS-IT-programme-to-be-abandoned.html, downloaded 25/2/12

[8] NBN Corporate Plan 2012-13, p 134

[9] House of Representatives Committee  on Infrastructure and Communications, Inquiry into the role and potential of the National Broadband Network, August 2011, Dissenting Report, p 306

[10]House of Representatives Committee  on Infrastructure and Communications, Inquiry into the role and potential of the National Broadband Network, August 2011, Dissenting Report, p 318

[11] P Fletcher, Wired Brown Land: Telstra’s Battle for Broadband, New South Books, 2009, p 34.

[12] S Rochford, ‘Leighton back for more with Nextgen buy’, Sydney Morning Herald, 26 Dec 2003, downloaded 25/2/2012, http://www.smh.com.au/articles/2003/12/25/1072308626755.html?from=storyrhs

[13] S Rochford, ‘Leighton back for more with Nextgen buy’, Sydney Morning Herald, 26 Dec 2003, downloaded 25/2/2012, http://www.smh.com.au/articles/2003/12/25/1072308626755.html?from=storyrhs

[14] J Taylor, ‘Spectrum Price Corrects One.Tel Folly’, ZDNet, Jan 4 2012, http://www.zdnet.com.au/spectrum-price-corrects-onetel-folly-339328824.htm, downloaded 25/2/2012

[15] P Fletcher, Wired Brown Land: Telstra’s Battle for Broadband, New South Books, 2009, p 38.

[16] Paul Budde, Australia - Mobile Communications - Subscriber Statistics, September 2011, http://www.budde.com.au/Research/Australia-Mobile-Communications-Subscriber-Statistics.html, downloaded 26/2/12 (Last update: 07 November 2011)

[17] McKinsey & KPMG, Implementation Study for the National Broadband Network, March 5 2010, pp 296-297

[18] NBN Co,  NBN Co Discussion Paper: Introducing NBN Co’s Special Access Undertaking, July 2011, p 24

19] M Turnbull, ‘Why NBN Prices will be Higher’, Delimiter, http://delimiter.com.au/2012/02/07/why-nbn-prices-will-be-higher-by-malcolm-turnbull/, downloaded 26/2/12

[20] ACCC, Inquiry to make final access determinations for the declared fixed line services, Final Report, July 2011, p 8

[21] P Fletcher, Wired Brown Land: Telstra’s Battle for Broadband, New South Books, 2009, pp 198-199.

[22] ACCC,  Unconditioned Local Loop Service Pricing Principles and Indicative Prices, June 2008, p 22