When the Rudd government announced the construction of the National Broadband Network in 2009 with an estimated cost of A$36 billion (and public sector contribution of $28 billion), it did so based on a political position, with a business case rather than cost-benefit analysis.
The now completed CBA, released this week, considers three possible scenarios for the future from here, against a base of the work that has happened to date (which of course cannot be rewound).
The first is to complete an unsubsidised roll out (i.e. provide a variety of technologies based on commercial feasibility); the second mixed technologies (MTM) including fibre to the premises, fibre to the node, and fixed wireless and satellite – consistent with the NBN strategic review; and the last fibre to the premises (FTTP) – or the NBN as it has been evolving till now.
To the cynical there is no surprise as to the core conclusion of the study - it supports the political position of the Abbott government – that the FTTP scenario is possibly the worst of all outcomes. These results formalise what is implied in Malcolm Turbull’s five minute explanation of why the MTM scenario is better.
The conclusion of a CBA is of course based on a comparison between cost estimates – which can generally be ascertained with a degree of accuracy – and estimates of benefit, which are usually far less certain, and that is the case here.
Standard approaches, but a conservative view
Assessment of the veracity of the cost estimates in the study is best left to engineers. But from an economics point of view an interesting choice in the cost estimates in the analysis is the inclusion of a deadweight loss through taxation impacts in the context of public funding. This adds 24% to the costs of publicly provided infrastructure. While this has some theoretical justification it is not commonly used in CBAs for other public infrastructures.
In addition, the report uses (and provides some justification for) what many would consider very conservative approaches to some general assumptions – such as the residual value included (depreciated cost rather than ongoing benefit) and the choice of discount rate (higher than the standard Treasury recommendation).
In terms of private benefit estimates the CBA uses a triangulation approach – the use of choice modelling to investigate willingness to pay, the use of a technical assessment in terms of what higher speeds would be used for and the value that creates, and observed historical take-up rates of NBN speed plans.
These are very standard approaches – but one has to question the applicability in the context of a product that is rapidly evolving in terms of its impact.
The private benefit estimates – all three methods – are based on current perspectives of how the internet will impact our lives and how it will be used into the future – as reflected in not only take-up rate (which reflects the marginal consumer), but value in terms of surplus (the value above what has to be paid – which can be very high for some customer groups). All of this is discussed in the report, and to some extent addressed in the sensitivity analysis.
More than a ‘video entertainment system’
But the reality is the way we all use the internet now is very different to what it was 10 years ago – and it will be very different again in another 10 years – and this is very hard to envisage and even harder to value.
Could the study have done anything different? Maybe not, but it’s important to recognise that any estimates constrained by current perspectives as to use are almost certainly going to be conservative. It should also be noted that sensitivities to the assumptions used (greater take-up rates) support the MTM scenario even further.
But to me, for all of this, a critical factor is the smallness of the public benefits concluded as being present, estimated at 6% of the total benefit - relative to private benefit at 94% - mainly linked to private use for entertainment purposes.
Again maybe this finding is not such a surprise given Tony Abbott had a number of years ago described the NBN as a white elephant, and a “video entertainment system”.
The low benefit outcome in the study is in large part due to the reviewers ignoring any social benefits that are not dependent on household or business connections (many of the identified public benefits fall within the government sphere) and further ignores benefits that will not differ between the identified scenarios.
This effectively means the public benefits of an effective high speed broadband infrastructure are largely underestimated. It should be clearly noted that there are many studies that support much higher estimates of public and social benefit. The study could also at least reference the academic literature that concludes widespread adoption of broadband has been strongly linked to economic growth and social well-being and especially so for regional communities.
A very important point made in the report (in support of the MTM scenario) is the real options value that is present in limiting the roll out in the context that technologies and consumer preferences are hard to predict. It is great to see a CBA that takes real options value into account.
In short – the CBA certainly does its job in the context of supporting government policy – and largely does so by using appropriate and defensible methodologies.
But, and in my view it is a big but, the worry is that it’s overly conservative in the estimated value of benefits of improved broadband infrastructure. While it is not proposed, it would certainly be detrimental to all our futures for the study results to be used to support the unsubsidised case as the way forward – to go this route would miss many of the private and public benefits that are conservatively estimated in the assessment.
On alternative assumptions and methodology – the net national benefits of public investment in significantly improved broadband infrastructure are very clear.