NBN: The Numbers Game

Since the announcement of the new NBN, there have been various news reports claiming that the charges to users will be unrealistic. For instance, this report from the Daily Telegraph claimed consumer experts had calculated that broadband monthly usage fees would have to be $200 to justify a commercial return on the NBN. Of course, the consumer expert turned out to be AAPT’s chief — hardly a prima facie model for independence in expert advice.

[DDET So let’s try and tease a few things out.]

First, there are about 8 million households in Australia and by 2016 another million will likely be added. Anyhow, for various reasons let’s suppose only 90 percent of them will spend money on telecommunications so we can use 7.2 million as a base. Second, it is not clear what the rate of return (commercial or otherwise) should be on the NBN. 10 percent is a number thrown around but that is a number after the whole network is complete. In any case, it is useful to think of yearly flows so let’s take $4.3 billion per annum as a desired target for revenues from this. Taking these two numbers together we have about $600 per annum from households on average or $50 per month. (By the way, if you desired rate of return was 15 percent, that number would rise to $75 per month. Alternatively if only half of potential users use the NBN services, that becomes around $1,200 per annum or $100 per month per subscriber).

Can we justify this on the basis of broadband alone? Current expenditure per household on broadband is $500 per annum, according to the ACMA. That said, that involves less internet penetration than I have assumed above (5.1 million subscribers) and so if currently internet penetration were to rise to 90 percent that would require a lower revenue take per household. On the other hand, the higher speed availability will increase the willingness to pay of households for the service. More interesting, is the ACMA’s calculation that improvements in competition or other outcomes in the industry are netting us over $300m per annum in additional consumer surplus or about $60 per annum per subscriber household. Those gains accumulate and so if the NBN generated additional gains of the same order, the social returns might be there.

Once we add wired communications in general, there is another $7.65 billion being spent or $765 per subscriber (of which around $350 is in access charges alone and $177 is in fixed to mobile charges). My belief is that this is where the NBN will have its biggest impact. Access charges may rise but more importantly, usage charges (other than to mobiles) will plummet — maybe to zero. If that occurs, the gain to consumer surplus is at least $238 per subscriber but possible more like $400 in total.

The issue here is that, the NBN may itself only capture revenue from say half of all potential customers. So the amount going to the NBN would be at most $500 per annum per subscriber from broadband and, conservatively, $527 from wired telecommunications. Spread over 3.6 million households that is annually $1027, which is less than the $1,200 required (a shortfall of $320m). However, add the potential consumer surplus benefits which are spread over all households and there is another $2.88 billion there from wired telecommunications alone (ignoring the broadband benefits). Thus, socially, the NBN even if taking just half the market, seems comfortably able to earn well beyond a 10 percent return.

Now I stress that these calculations are a ‘Sunday morning’ affair but I thought the exercise would be useful just to frame a bit of the debate. I am happy for others to try out their own calculations in the comments.


11 thoughts on “NBN: The Numbers Game”

  1. $50-$75 per month access charges is well and good however that is the price paid to the network wholesaler, the consumer will pay their provider a slight more once you include data, support, and administrative charges.

    Having said that that ROI for this type of network could easily be put at 15-20 years and bring down the monthly costs significantly.

    The point about usage charges is quite true and not being identified enough by the MSM.


  2. The point about usage charges is a good one. However, I think the main problem with the NBN is that it will compete with Telstra who will probably fight tooth and nail to retain its current customer base and would presumably be willing to undermine the profitability of the NBN.

    Lastly, in the assumptions about potential profitability of the NBN everyone is assuming a $43 billion price tag, I apologise for being cynical but I suspect that there will probably be major cost blowouts on a project this size.


  3. The Telstra response is an interesting issue. Usually, aggression as been to deter expansion or entry (why spend the money otherwise) but with the NBN that is an impossible task. I think there will be a retreat.

    I’m not sure about the cost blowout. Costs continue to fall and over 8 years we can expect quite a fall. It will be interesting to see where the $43b comes from.


  4. Ok, that’s a good point about costs falling, though presumably falling costs will advantage Telstra also.

    As for where the money will come from, over 8 years I guess its ‘only’ 5 or 6 billion dollars a year, so I dunno maybe another alcopops taxs or two should do it?


  5. Kevin,

    But money can’t be spent twice. If you issue a $2500 share to a person, and they sell it, the money is spent elsewhere in the economy. If most people do this (and why wouldn’t they since there is no ability to buy a fibre connection yet) then all the money will have been spent on other stuff, and there now will be no incentive for anyone to invest in the new fibre company b/c no one has any money left to pay for their installation.

    Or am I completely wrong?


  6. invig

    when someone sells their shares they will sell it for some existing money and it will be sold at a discount to the face value at which they were issued. This is why people may or may not sell their shares. The shares can only be converted to regular currency that is used by NBN. I have expanded the description a little more to try to explain the effect but it is the reason why issuing shares or in the case of Energy Rewards issuing more Rewards is not inflationary on the regular currency.

    It does not matter how many shares are issued as long as the money the economy needs the money supply to be increased. We can see if we are issuing too much if inflation of the general currency starts to rise too much – in which case we ease off on the issuing of shares but as the money supply in 2008 as measured by M3 increased by $170billion I think we can assume 43 billion over a few years is not a problem. Of course this assumes we stop fractional reserve banking as the way to create new money and only let banks lend money they have on deposit.


    Here is a presentation I am giving to the ACT government committee on Climate Change this coming Tuesday at 2:30. It advocates zero emissions by 2020 and shows how we can guarantee to achieve that target while at the same time reducing the cost of energy.


    I think I am gradually learning how to explain why this works so that it is understandable. We fund community infrastructure by the profits obtained by the community from the infrastructure AFTER the infrastructure is built not from profits we have earned in the past.


  7. “The shares can only be converted to regular currency that is used by NBN”

    So you put a limit on what can be purchased if the person cashes in their shares.

    Well, that seems very similar to the idea of a voucher system but with some flexibility in the use of the voucher (at a discount).

    Interesting. But I wonder how difficult it will be implement and monitor.

    At the moment, there are restrictions on aboriginals buying anything but food, and they then sell their debit cards at a discount for cash. Same thing, but you want to effectively control the black market in ‘voucher resale’, yes?

    Good luck with the presentation, the more new ideas out there the better. And I agree you need to explain your ideas better (to me anyway…:)


  8. Malcolm Turnbull comments on the economics of the NBN today in the Australian – http://www.theaustralian.news.com.au/story/0,25197,25329883-5015664,00.html
    Since no comments seem to be making it in, here is my comment on the Opposition leaders assumptions.
    My assumptions are online at : http://www.thejoie.com/welcome/docs/NBNcosts.xls

    Mr Turnbull makes some good commercial sense in his article. I have made some estimates on prices and takeup, as well as costing the investment with interest and depreciation, and make the following comments.
    MT assumptions:
    Operating expenses 50%, depreciation 5%
    Takeup of homes 50%, rest with competition
    ISP $100pm, NBN (Ruddnet) 70% of $100pm : gives no economic return on investment
    My assumptions:
    Spreadsheet is online at: http://www.thejoie.com/welcome/docs/NBNcosts.xls.
    Depreciation same, Interest as for depreciation
    Takeup $200pm 10%, $150 20%, $100 30%, $75 40%, $50 60%, $30 80%
    On these assumptions, even with no operating costs (and I think 50% is very high), interest and depreciation (5%) at $180m per month each, swallow the revenue of max $240M, leaving a shortfall per month of $120M plus operating costs, and excluding competition, and ISP profit.
    Other benefits accrue to businesses, who also use the network. Looking at small, medium and large, I guess small might pay $200 20% takeup, $150 30%, $100 40%, and medium $2,000pm 10%, $1,000pm 20% and $500 30%, large $20,000pm 10%, $15,000 20%, $10,000 30%.
    Even with business cost savings on Telstra prices of 50% for large, and 20% for small, and revenue expansion for large 5%, Medium 10% and Small 20%, the most extra revenue and benefit I can get to is about $80M per month.
    Still a shortfall of $40M per month, plus operating costs, plus competition, plus ISP share.
    So, there are a lot of variables here, and it is worthwhile to think through what they are, and what they are worth to us individually, as a family, and as a nation.
    If a national project is worth $40M of national fervour (per month) then maybe it is worth it. If we can find another $40M of benefit elsewhere eg international reputation, proactive problem solving, then maybe there is value in the NBN.
    Other downsides would include potential impact on other telecoms share value, to shareholders.
    Long ago, Pliny once said: fortuna fortem favet, not long after Mt. Vesuvius erupted. Sometimes life is for living, and doing the big important things to make the world a better place. Fortune favours the brave. The NBN might cost us $40M a month ($2 per person per month), but setting an example to the world – PRICELESS.


  9. Singapore’s NBN to offer wholesale 100Mbit/s access for US$14

    Singapore consumers can expect more Internet Service Providers (ISPs) to launch 100Mbit/s consumer broadband services next year at about US$50 a month or less.

    The impetus to make available more choices in high-end broadband services is driven by Singapore’s upcoming broadband network, portions of which will be up by the middle of next year.

    StarHub, awarded the tender to install the hardware for this network, will sell ISPs access to it at a wholesale rate of US$14 per 100Mbit/s for residential lines.

    This is much lower than ISPs offering consumer broadband services currently pay StarHub or SingTel for wholesale: StarHub charges US$14.37 for a 2Mbit/s line and US$23.80 for a 100Mbit/s line; SingTel’s charges range from $10 for a 512kbit/s line to US$22.67 for a 10Mbit/s line.
    from telecomtv.com 15.04.09
    For your info.


Comments are closed.