That cost-benefit analysis

by

I thought it was time to pick up a theme on the National Broadband Network that has been going around for sometime; the lack of a clear cost-benefit analysis.

First, it is never going to happen. Put simply, the political rationale for the NBN is a combination of two things. First, that a big push on broadband was not going to happen without a big push from Government given the virtual monopoly held by Telstra and the ineffectiveness of regulation to manage that. Second, there is the Yes Prime Minister Trident/Hollowmen/GFC Big Ticket/Shiny things rationale that is wonderfully captured by this piece in The Onion. You only want a comprehensive cost-benefit analysis if it is going to change your decision. The political rationale is so strong that that is not going to happen and so there is no point to attempts at quantification.

Second, even without a choice motive, there is a downside to the lack of clear analysis: that we can’t optimise the NBN and will inevitably end up causing some waste and inefficiency. I have already mentioned many times that by placing the policy sales pitch and broadband, and worse than that, on broadband speed, we leave that as the sole metric for performance. However, the NBN can potentially yield benefits of lower prices (virtually nothing in fact) for telecommunications and also a revolution in government services if basic broadband is freely available. The problem is that the Government is not being held to account for realising those benefits and this is very troublesome. That said, once the stuff is in the ground …

Third, and this is more worrying than the lack of a cost-benefit analysis, there has been no consideration whatsoever given to issues of market design. If we were serious about this, the NBN would be regarded as a platform that would allow telecommunications markets to evolve. Instead, it is regarded as a thing rather than an institution that sets the rules of the game. That is why we end up with fibre all over the place. That is why we end up with engineering criterion.

On that latter point, and with due respect to my colleague Rod Tucker, it may well be true that only fibre can deliver the fastest broadband speeds. But time and time again we find that people are willing to sacrifice engineering metrics for other things. The evidence is compelling that consumers will sacrifice speed for wires just as they sacrificed CPU power for portability (something IT people thought would never happen). For this reason, we need to be cautious in how we let technological choices be made and to provide rules to allow it to evolve flexibly. The problem is that that does not square with the one-eyed sales pitch on the NBN.

Finally, one thing we can’t quantify well in cost-benefit analyses is ‘future proofness.’ That provides a reason to push forward with the NBN but at the same time is the reason to ensure it is built in a flexible manner. You can’t simultaneously be claiming you are insuring Australia for the future without actually taking out insurance on the details.

Oh and by the way, there is no cost-benefit analysis for the Coalition’s plan. To do that would require a consideration of opportunity cost. So if the choice is between ‘do nothing’ and ‘do the NBN,’ as any first year economics student will tell you, you can’t justify ‘do nothing’ without first doing a cost-benefit analysis on the NBN.

10 Responses to "That cost-benefit analysis"
  1. Nobody (except perhaps Rod Tucker) mentions the well -known fact that once the fiber is in the ground, speeds can continue to increase by simply changing the electronics at the end-points. It is not that simple with other technologies. So the fiber is more \future proof\ than any other technology -at least as we now know it. And if wireless is nice, you can always get it better from the the end-point, rather than from some thousands of towers (which people seem to have a worry about). Just my two bytes worth.

  2. “You only want a comprehensive cost-benefit analysis if it is going to change your decision”
    That is from an economic perspective.  From a political perspective, I would say:
    “You only want a comprehensive cost-benefit analysis if it is going support your (political) decision”
    Since there has been no CBA, the inference must be that it is expected that a CBA would give the “wrong” answer.
     

  3. I think the nature of the answer depends on a few assumptions:

    1) the discount rate used (8% or so) – justified on the basis that the NBN is “infrastructure” and not a high-risk technical implementation.
    2) the optical backbone is considered a 40-year investment (?!)
    3) wholesale costs of approx. $50-60 per subscription. (What is the likely retail rate?)
    4) the take-up rate (at stated wholesale cost) by the public is high – I honestly don’t remember what was presented in the report but I do recollect that significant numbers of the public would be signing onto NBN. (If they don’t then there is a problem.)

    If you believe those asumptions hold then perhaps the CBA holds. If you vary the assumptions then the multi-billion dollar spend looks less promising.

  4. I believe that “broadband penetration” counts as an economic indicator by the OECD these days. Perhaps the NBN is a very expensive PR campaign to ensure Australia moves up (or retains) its place in OECD rankings.

  5. See my cost benefit analysis on <a href=”http://forums.whirlpool.net.au/forum-replies.cfm?t=1481577l”>Whirlpool</a>, and my blog <a href=”http://valman.blogspot.com/2010/07/my-cost-benefit-analysis-on-nbn.html”>Value Management</a>, which shows a positive NPV, if the NBN creates a 1% increase in GDP for ten years. I look forward to seeing what the voters decide. The election is effectively voting on the cost/benefit of the NBN. If voters vote Labor, they are voting for the NBN, and if voting Liberal voting against NBN. The first suggests voters value the NBN at more than $43B, and the second suggests voters value the NBN at less than $43B. Go figure…. we will all see soon.

  6. How is the election doing that Richard? Voters optimise over a range of policy and other factors and as such there is nothing to guarantee that the election result will be coincedent with public preferences over the NBN.

  7. In my view the election and voting is like electors purchasing a package of goods, which they are valuing with their vote. The goods are policies, government programs (such as the NBN), and the skills of the various individual politicians to pit those skills against future unknown and unknowable circumstances. The voters are valuing the NBN, and aggregating that value along with the other policies that they take notice of, and deem important, probably no more than five or six out of the perhaps 100 policies on the table. For me, NBN is the most important (to me). For someone else, it will likely be something else, that has greater relevance and impact to them. I don’t believe there is a lot of optimising. Rather, a lot of filtering, some balancing, and trusting. All these processes are part of the valuing process. Which ends up with a single output – a vote (though more complex in the senate). Have a good weegend…. valuing your future….

  8. The barriers to entry for wireless data networks are far lower than they are for wired networks – that’s why we already have several competing high speed wireless networks.  If it really was such a compelling technology, then people would already be abandoning ADSL and Cable for Wireless in droves – but they’re not. Wireless will never be able to deliver the aggregate bandwidth, across all consumers, that wired can.  There is simply a limited amount of radio bandwidth available – it is a much more limited resource than space for cables underground. When thinking about future upgrades of the fibre network, consider the example of the copper network.  It was built during the early 20th century for manually-switched low-bandwidth analogue voice, and these days it’s possible to get 18 Mbit/s ADSL2 over essentially the same copper wires.  The hard (expensive) part is getting the wires into the ground – once we have that, we can work magic at the endpoints for years to come.

%d bloggers like this:
PageLines