AFR Infrastructure Summit
11 June 2015
When the Suez Canal was opened in 1859, the total cost of the project was around three times that forecast the year before the project started1.
The Panama Canal was a substantially more complex project and the technologies needed more sophisticated. The first attempt by the French was abandoned in the 1880s after almost a decade of work, more than 25,000 lives lost and only a third of the canal completed2. It's estimated that more earth was deposited in the canal by landslides and storm than the French were able to dig out. It is now thought that by the time the canal was completed by the Americans in 1914, its cost overruns were between 70 and 200%.
An expansion of the Panama Canal initiated in 2009 will see vessels carrying up to 13,000 containers pass through the canal, more than double its previous capacity. The cost of the project was forecast in 2009 at $5.25 billion and has so far blown out by $1.6 billion3.
The history of megaprojects is such that no matter how sophisticated the technology, or how great our advances in accounting practices, we have essentially not gotten any better at avoiding serious cost overruns.
One of the leading experts in the field, Bent Flyvbjerg looked at 111 megaprojects conducted between 1910 and 2000 and found that cost overruns have continued in fairly consistent magnitudes throughout the period. He concluded: 'No learning seems to take place in this important and highly costly sector of private and public spending'4.
Projects dealing with public spending typically tend to perform worse. Different publicly funded projects typically compete with each other for a set pool of discretionary funds, meaning there is an incentive to overcook the benefits and undercook the costs; politicians and project managers feel the need to become 'advocates' to win public support for the project; and there is a natural tendency for 'empire building'5 without adequate cost controls.
Sometimes politicians lapse into fits of irrational exuberance and commit billions of dollars without even committing to any cost benefit analysis at all.
Which is perhaps a good segue into discussing the National Broadband Network.
You have all heard me speak before about the foundational mistakes with the nbn.
First, the Government should never have undertaken the project. In every other comparable jurisdiction the private sector rolls out telecom upgrades and the Government provides subsidies to ensure regional and remote areas get a service. The telcos have the construction and business risk which they are best able to mitigate. The Government makes a political call and writes a cheque.
In Australia we have the complete reverse. The Government has the construction and business risk and as for the cheque - well that went to the telcos.
Second, a project of this size and complexity would tax the resources and capability of a large telco - but a Government owned start up?
Third the business plan was focussed on technology as ideology - fibre to the premises as a cult. A rational business would have focussed on the customers (remember them) and used whatever technologies to deliver very fast broadband as quickly and cheaply as possible.
Fourth the previous Government hired the wrong people. In this case under Labor there was a board with not one person who had either built or run a telecom network and a CEO whose experience was working for an equipment vendor not building linear infrastructure or running a telco.
I'm sure many of you have heard me speak before about the application of Moore's Law and why it doesn't apply to digging holes; or perhaps you've seen my whiteboard exposition of the economics of last mile telecommunications.
But today I want to talk to you about four important aspects of the nbn, which are relevant to many other Government,and indeed private sector, projects.
- the reasons for the optimism baked into the nbn and its stubborn persistence throughout its early years
- the extent of the ongoing risks associated with the nbn
- the difficulties in mitigating these risks
- and finally, wider public policy lessons that should follow.
Overcoming the nbn's optimism bias
Key questions are: Why was there such an optimism in the project? And why did it persist for such a long time?
From the very first corporate plan, there were warnings from objective outsiders that the company was highly unlikely to meet its forecasts. Writing in the Telecommunications Journal of Australia, in 2011 just after the release of the plan, experienced analyst Ian Martin summarised the outlook:
'The risks are asymmetric and point the wrong way for nbn Co: ARPU may well be lower, but there is little prospect it will be higher than planned, take up may be lower but is unlikely to be higher or sooner, and at best costs may come out in line with forecasts but there is a good chance they will be higher.'6
The nbn did not have long to wait before it was mugged by reality. By the time of the election, the company had met only 17% of its 2010 fixed line rollout target.
In the meantime, the first construction tender failed; one of the four delivery partners pulled out of the project; two of the four delivery partners were demanding higher rates; and another delivery partner was not delivering at a meaningful scale7.
And yet the management of the company still boasted that it would earn an internal rate of return above 7% even declaring: 'What is remarkable is how little the financials changed over the 3 years'8.
Creating a business plan for a large scale infrastructure project is hard at the best of times. Creating a business plan involving a wholly owned Government start-up over a 30 year time horizon was a virtual impossibility. I am reminded of the words of famous Silicon Valley investor Guy Kawasaki about the trick to assessing a start-up's business plan: 'As a rule of thumb, when I see a projection, I add one year to delivery time and multiply revenues by 0.1.'9
I say this not to open old wounds or fight yesterday's battles, but just to make a larger point about our approach to this project.
As Flyvbjerg and numerous other authors show, cost overruns are frequent and widespread. This is less an indicator of the competency of those involved with the project or of the intent to deceive, but more of the inadequate systems of checks-and-balances in place to ensure that you are looking at the project with an analytical eye, that you are getting the best information that can be debated in a rational way – in short, that you are walking into this project with your eyes wide open.
One of the most noticeable improvements in the nbn project has been in the quality of advice coming both from the company and the Department. One of the complaints of the former management was that commitments about the nbn's deployment and costs were announced publicly before the project was even started, and they had little scope to challenge these assumptions10.
We have of course introduced many transparency measures and increased oversight11, with one of the chief consequences being a much more open and frank relationship between the various stakeholders12. As I have said many times to the management and my Department, don't tell me what you think I want to hear; the glee club culture – where any one with bad news is ignored – has come to an end.
The Anatomy of the nbn's Risk
The nbn may well be the biggest single infrastructure project the Commonwealth has ever undertaken. There is a strong case to be made that it is the most complex. Unfortunately for taxpayers, knowing what we know now, it may also be the riskiest project that the Commonwealth has ever embarked upon.
The nbn, as I have said many times before, is not like building a bridge or a mine. These engineering feats are difficult enough, but projects like the nbn might be considered more difficult because it is not really a single project, which can be planned for, or designed, in a single exercise.
It is rather, the culmination of 12 million distributed and distinct projects all rolled into one. Each house the nbn hooks up will carry distinctive features and challenges – different workforce mobilisation requirements, different rock types for drilling into the ground and varying quality of existing infrastructure.
Sometimes when you overcome all of these things, you turn up only to find out the homeowner has left for the day.
The nbn is a company replete with horror stories about getting customers connected to the network. Consider some of the recent examples:
- The company had to build a new street network through 300m of solid rock to connect three premises in Toowoomba, Queensland, costing in excess of $100,00013.
- It cost up to $25,000 per premise to connect acreage properties in Victoria with lead in cables greater than 150- 200m in length.
- There was the home in Wollongong where the contractors drilled through the stormwater pipe and the ground began to subside.
But the nbn is not just a construction project. It is also one of the most complex transformations of regulatory oversight and industry settings ever undertaken by the Commonwealth; it involves one of the most complex legal and commercial agreements ever completed by the Commonwealth14; and it is the single largest IT project on which the Government has ever embarked.
On this last point, I want to reiterate how big the nbn's IT project is.
By the company's forecast, it will have spent more than $2.5 billion on IT by the 2021-22 financial year. 16
It is larger than any comparable Australian Government IT projects.
|Agency||Project||Initial Estimate||Final Spend*|
|Victorian Government||Myki Smartcard||$1.17 billion>||$1.52 billion|
|Queensland Government Health Department||Payroll Overhaul||$6 million||$1.2 billion|
|VicRoad15||Registration and Licensing Program (RandL)||$173 million||$293 million|
It is so hard to manage public IT projects, the subject forms its own subculture of the megaprojects literature which shows that though the bulk of public sector projects are completed on budget, there is a 'fat tail' of cost overruns: almost one in five projects end up running off the tracks, with an average cost blowout of 130% and overruns in time of 41%.
So the nbn is really a project unlike any other. It involves a number of highly complex work-streams, which in ordinary times the Commonwealth Government – or any government, for that matter – would be reluctant to undertake on its own. And they are all interdependent. A delay to one necessarily increases the delivery risk of another, and so on.
By way of illustration: If four interdependent work streams have an 80% likelihood they will be delivered on time, then put together the likelihood of delivery is only 40%.
But apart from the complexity, the other utterly unique feature of the nbn is its sheer scale. That is why even small changes in the execution of the business model can add up to very large consequences for the project's bottom line.
The company has modelled a range of scenarios – each of them very credible – that would add $1 billion to its overall peak funding requirement. They include:
- failing to meet the company's take-up target in the fixed line footprint by around 5%
- a seven month delay in either its FTTN or HFC product launch
- only managing to connect 80% of the number of premises per day needed at the height of the project
- a $250 increase in the average cost of HFC lead-ins
- an FTTN node costing 25% more than anticipated
- and a failure to meet its target increase in the average revenue per user by $3.
Again, this is a risk I have warned about repeatedly, given that the experience in Australia and virtually everywhere else in the world has been that telco users have been trained to expect more bits and bytes for essentially the same amount of money17.
Internal Rate of Return is often how the quality of investments are measured, so it is a useful metric to measure the commercial viability of the nbn from the taxpayers' perspective. As a general rule of thumb – and I would stress the word 'general' – each additional $1 billion in peak funding results in a deterioration of the Internal Rate of Return in the order of about 10 basis points. So getting these small assumptions right is absolutely critical.
Difficulties in Mitigating the Risk
There are a limited number of levers any management of the nbn could pull to improve the economics of the nbn.
But first I want to make a general point about the economics of the project. Normally, a project like the nbn is a trade-off between time, scope and money.
For example, if you wanted to reduce costs or increase the scope of the project, you would extend the time needed to deliver it. This was certainly true early on in the project. Public policy expert Bill Scales noted that the previous Government completely reconstituted its nbn policy, after the initial FTTN Request for Tender failed, in just 11 'rushed and chaotic' weeks18; had better planning been undertaken, I suspect we would have ended up with a much more sophisticated project.
Nevertheless, the point is that the ability to trade-off time for money or scope has lessened as the project has continued. In fact, in some ways there are now no trade-offs, only increased risks. So for instance, increasing the time taken to deploy FTTN or HFC is more likely to add to the peak funding requirement, because of the risk that customers will substitute to alternative technologies such as wireless19.
Indeed, as a study released by the ACMA today will show, in the past year alone an additional 2% of households have gone wireless only in the past 12 months alone, and almost one in four Australians aged 25-34 are now have no landline for Internet or telephony services20. So the competitive threats are very real.
Strategies to mitigate the risks of the project are necessarily limited for a number of reasons.
First is the underlying policy reasons for embarking on the project, and the mandate for universal coverage. We have sought to reduce the scope of the project in terms of the technology we will use, reducing the overall peak funding requirement by around $32 billion21, so that we are able to deliver consumers they services they need – and are willing to pay for – rather than 'gold plating' the infrastructure.
But any further reductions in the cost of the project would require revisiting nbn's universal coverage mandate.
The cost-benefit analysis into the nbn showed that we could deliver around $6 billion of additional economic benefits with an unsubsidised rollout – that is, a rollout that delivered broadband upgrades only where it was commercial to do so compared to our 'multi-technology mix' approach22.
But the Government has explicitly rejected this option. Indeed if there is any reason for Government intervention in providing broadband at all it is to ensure coverage where the market by itself would not provide.
Department of Communications research has estimated that over 30 years to 2040, the fixed wireless and satellite networks will result in a net loss of around $16.8 billion23.
While we will look at delivering those subsidies most efficiently, we are absolutely committed to every Australian having access to a world-class broadband service at an affordable price.
We believe in the potential of the Satellite Service to revolutionise life in the bush, and the delivery of services such as School of the Air and remote telehealth.
In the past fortnight, I have written to the nbn Board to ensure we are maximizing the benefits of the Long Term Satellite and that we will avoid the mistakes made in managing demand of the Interim Service Satellite Service where $350 million was spent to deliver a service which delivered little better than dial-up speeds.
The second difficulty in managing the nbn risk is ensuring we get the balance right between improving the economics of the project and ensuring we are acting in the long-term interests of end-users. That is, while it is conceivable that we could constrain competition and possibly even raise prices, does not seem to be in the best interests of consumers or the wider economy.
The mantra against infrastructure-based competition to the nbn was that the network is a natural monopoly - which doesn't explain why some very unnatural anti-competitive measures were legislated to prevent competition.
The truth is that much of what Labor was doing was protecting the monopoly of the nbn. As the Expert Panel in the cost-benefit analysis found, if competition produced even small productivity gains that would outweigh the duplicative costs of competition - and that in the long term would drive cheaper prices and better service for customers – that would be a good thing24.
On the contrary, this Government is committed to fostering competition in telecommunications in all its forms – network, wholesale, retail. However, we have to move from what we've inherited to where we want to be and that will take some time and effort. We set out a roadmap for doing this in our Telecommunications Regulatory and Structural Reform package released last December. That is an environment in which nbn Co is a player, because that is the unavoidable reality, but one in which it and other players can compete fairly and effectively.
What does this mean in practice?
- New licence conditions put in place last December do not prevent competition; on the contrary, they seek to foster competition in the provision of infrastructure, providing the infrastructure provider offers open access and supports retail competition.
- nbn has been instructed to charge for infrastructure in new developments: previously it provided this infrastructure at no charge; this made it very hard for small but innovative providers to compete.
- We are working with the Communications Alliance on rules for the deployment of VDSL2 that maximise both broadband performance and competition to the greatest extent possible.
- We are also working on legislative changes that will provide further opportunities for alternative providers, while preserving the key principle that networks with bottleneck features should be open access.
- We are requiring nbn to maintain IT and accounting systems that would enable its future disaggregation into competing business units if appropriate.
Noting these constraints – and they are fairly fundamental - we are taking important steps to manage the risk of the rollout.
We have made substantial progress on mitigating the risk in the project over the past 12 months. I am substantially more confident on its prospects than I was this time last year.
In particular, the nbn has managed to:
- Begin a construction pilot on the FTTN network ahead of the product launch later in the year. That means the nbn has been able to refine its construction process with the FTTN footprint covering more than 400,000 premises ahead of the company being in a position to begin widespread migration. This has now extended to the Northern Territory, where the rollout is well advanced, Western Australia and South Australia where the rollout is currently the most behind.
- Just yesterday, the company signed new construction contracts with five delivery partners covering around four million premises in the FTTx footprint. Whereas even the previous Government had admitted its construction model was broken25, the new model gives the company flexibility to award work to companies who are delivering and will reduce the high administrative burden on those companies26.
- Complete the renegotiation of Commercial Agreements between Optus and Telstra although there are obviously still a number of conditions precedent, some of require regulatory approval.
- Implement separate operational and business support systems so that the eventual goal of disaggregation in the long term does not compromise the delivery of upgrades in the short term.
But most importantly, the company is acting like a business. nbn Co's management has been able to complete a full top down, bottom up overview of the company. For the first time since its creation, the company has a sophisticated view of its customers, its costs and its operations. The company's forecasts are increasingly believable, and I have confidence that it is actively seeking out and mitigating risk.
There are still many unknowns of course. If only nbn Co's management could invent a time machine! But I am pleased that the management have recognised that there are many things that they don't know, and are remaining flexible. What do you do in the face of so much uncertainty? You maximise for flexibility so you can adopt the best strategy for each situation. This is, of course, the heart of the Multi-Technology Mix approach.
Implications for Future Projects
So while we have learnt many lessons specific to the nbn project, I thought I would outline some of the more general points and their application to future infrastructure projects.
The first is that the architects and managers of projects need to be accountable for the things that they can deliver on and less focussed on the things they can't. The main reason the original nbn financial model seemed so impermeable to the constant stream of real-world data that suggested the plan could not be delivered was that its IRR relied on very large revenue increases and thus operating margins in the future coupled with a hefty multiple on the project's terminal value thirty years hence.
nbn Co has been criticised for producing a corporate plan in 2014 that only had a financial and operational outlook of 12 months. In fact this was a necessary compromise given the uncertainties the new management had to deal with. The upside of this approach was the focus on delivering in the near term and I am confident to say that by the end of the financial year the NBN should be able to meet most, if not all, of the targets in that plan - a rarity throughout this project. The upcoming corporate plan will include a financial and operational forecast for at least the next three years.
Secondly, risks should be assigned to those parties best equipped to deal with them. In the telco world, this is typically those parties that have the workforces and long histories in deploying networks. As the nbn found, even construction companies who have been in the industry for a long time found it incredibly difficult to quote for work and deliver on budget.
There can be no substitute for the industry having 'skin in the game'. Everywhere in the world, the basic public policy strategy was for the Government to offer subsidies for areas which were uneconomic and then passing on the delivery risk to the private sector. In Australia, the Government has taken on all the delivery risk, while those in the industry have for the most part asked for better infrastructure, better customer service, quicker delivery - and all for lower prices.
And finally, it is important to reiterate - as indeed I have been doing ever since I started as Shadow Minister five years ago - about the need for transparency and independent oversight.
One of the distinctive features of the nbn project was the lack of benchmarking against other similar projects around the world.
So clearly, you have to do the work up-front to assess a project's costs and benefits. But crucially you need systems in place to update data from the field to get a better understanding of how the project is progressing, and when it is time for intervention. One of the unique aspects of the project was that it took almost five years to get a handle on what it was really costing to connect a premises to the all-fibre network, with the actual cost being 40-50% higher than what the company had publicly been projecting.
In my experience of the nbn - keeping in mind I have spent a lot of time on the outside looking in, and now on the inside looking out - our learnings about the project have had an inverse relationship to the excitement it has elicited in the Financial Review's editorial pages. The further up the learning curve we have scaled, the further back in the paper the stories have ended.
And that's not a bad thing. The national conversation has to move on from the pits and pipes to the cool things that people are doing with high speed broadband - to the innovations, the disruptions, the transformations and above all the imagination of people taking advantage of the opportunities afforded to them by the new digital infrastructure.
1 Flyvbjerg, B., Bruzelius N., & Rothngatter, W., (2003), 'Megaprojects and Risk', p.18
3 The Economist, (2014), 'Keeping Things Afloat', available online here: http://www.economist.com/blogs/americasview/2014/01/row-over-panama-canal
4 Flyvbjerg, B., Bruzelius N., & Rothngatter, W., (2003), 'Megaprojects and Risk', p.16
5 See for instance: Flyvbjerg, B., Garbuio M., & Lovallo, D., (2013), 'Delusion and Deception in Large Infrastructure Projects: Two Models for Explaining and Preventing Executive Disaster,' California Management Review, vol. 51, no. 2, Winter 2009, p.10 onwards. The authors note specifically: 'politicians, planners or project champions deliberately and strategically overestimate benefits and underestimate costs in order to increase the likelihood that their projects, and not their competition's, gain approval and funding … However, this misrepresentation and failure can be moderated by measures that enhance transparency, provide accountability, and align incentives.'
6 Martin, I, (2011). 'A significant gap in the nbn corporate plan' in the Telecommunications Journal of Australia, pp.51.1-51.11
7 For example, see http://www.afr.com/business/telecommunications/nbn-at-war-with-contractors-20130401-j0yo8, http://www.afr.com/business/telecommunications/nbn-contractors-fail-to-deliver-20130402-j0yr1 and http://delimiter.com.au/2013/10/15/nbn-construction-model-failed-says-conroy/
8 Quigley, M., (2013), 'Reflections on the nbn', A presentation to TelSoc, available online here: http://telsoc.org/event/national/2013-12-02/mike_quigley_reflects
9 Quoted in Anthony, S., (2012), 'The Planning Fallacy and the Innovator's Dilemma' in the Harvard Business Review, available online here: https://hbr.org/2012/08/the-planning-fallacy-and-the-i
10 For instance, former CEO Mike Quigley stated: 'There were already public commitments made before nbn Co got going. The management teams of publicly funded infrastructure projects are not in control of the delivery timetable they are held to – something, it seems, nbn Co will continue to grapple with.' Source: Quigley, M., (2013), 'Reflections on the nbn', A presentation to TelSoc, available online here: http://telsoc.org/event/national/2013-12-02/mike_quigley_reflects
11 Statement of Expectations, dated 18 April 2014, http://www.communications.gov.au/__data/assets/pdf_file/0014/221162/SOE_Shareholder_Minister_letter.pdf
12 For example, the information provided to Shareholder Ministers as part of the 2015 Corporate Planning process has been of much higher quality than in previous years.
13 Information provided by the nbn
14 David Thodey famously described the Definitive Agreements between Telstra and the nbn as 'some of the most complex, perhaps, agreements ever signed in corporate Australia's history' (Telstra, 2011, 'nbn Analysts Briefing Transcript' available online here: http://www.telstra.com.au/abouttelstra/download/document/tls781-nbn-analyst-briefing-transcript.pdf). In fact, the complexity of the agreements were so severe they presented a large ongoing burden on both companies and added to the execution risk of the project. As Andy Penn noted after the new agreements were signed: 'We've also improved and simplified the agreements based on what we have learned working with the original agreements over the last three years. This includes removing some of the complexity in reporting processes and the payments are now much more simply linked to the rollout of nbn.' (Telstra, 2014, 'Telstra signs revised nbn Definitive Agreements – analyst briefing transcript', available online here: http://member.afraccess.com/media?id=CMN://3A416771&filename=20141216/TLS_01585078.pdf)
15 Victorian Auditor General, (2014), 'Impact of increased scrutiny of HVHR projects', available online here: http://www.audit.vic.gov.au/publications/20140626-HVHR-Projects/20140626-HVHR-Projects.pdf
16 Figures are approximate and include capital and operational expenditure. An earlier transcription of this speech referred to 'IT capex'. This has been corrected.
17 See for example the Government's response to the Senate Select Committee on the nbn, 30 April 2014, http://www.malcolmturnbull.com.au/media/response-to-the-senate-select-committee-on-the-nbn
18 Scales, B., (2014), 'Independent Audit: nbn Public Policy Processes', available online here: http://www.aph.gov.au/~/media/02%20Parliamentary%20Business/22%20Chamber%20Documents/223%20Tabled%20Papers/Documents%20Presented/Out%20of%20session/040814_audit_report
19 nbn estimates that a nine month delay in the rollout of HFC will reduce overall take-up by 1%, due to substitution to alternative products.
20 As reported in CommsDay: '12% Segment of Mobile-Only Australian Adults Set to Expand: ACMA', 11 June 2015.
21 nbn, (2013), Strategic Review
22 Expert Panel, (2014), 'Independent cost‐benefit analysis of broadband and review of regulation' available online here:
http://www.communications.gov.au/__data/assets/pdf_file/0003/243039/Cost-Benefit_Analysis_-_FINAL_-_For_Publication.pdf, page 11, chart 2.
24 Expert Panel, (2014), 'Market and Regulatory Report', p.18, available online here: http://www.communications.gov.au/__data/assets/pdf_file/0009/243387/NBN-Market-and-Regulatory-Report.pdf. The report stated: 'Disaggregating the technology platforms could lead to some duplication of operating and network upgrade costs. However, the panel's modelling shows that if competition sped up delivery by one year and accelerated the annual growth rate of productivity by 2.5%, its benefits would more than outweigh those costs of duplication. Experience shows competitive markets can easily produce gains of this order. Over and above these gains, a competitive structure would deliver the benefits of greater customer responsiveness, a lowered risk of monopoly pricing and a reduced burden on taxpayers.'
25 For instance, former Minister Stephen Conroy said: 'The nbn Co board and myself as shareholders have to take the responsibility for believing that these companies could deliver on their contractual obligations. And when they didn't in the Northern Territory case, we got money back off them and we sacked them and we put in place in a different operation. So ultimately - and I think I said this on Friday - I think the construction model that nbn Co put in place hasn't delivered. But it hasn't delivered based on contractual obligations not being met by these construction companies'. Source: ABC Lateline, (2013), 'Stephen Conroy Discusses nbn and Working With the Construction Industry', available online here: http://www.abc.net.au/lateline/content/2013/s3868978.htm
26 nbn, (2015), "New nbn Construction Contracts to Spur Multi-Technology Rollout", available online here: http://www.nbnco.com.au/corporate-information/media-centre/media-releases/new-nbn-construction-contracts-to-spur-multi-technology-rollout.html