NBN Sensitivity Analysis: Cost Benefit Part II

I reworked the figures (spreadsheet) in my costs benefit analysis spreadsheet with new assumptions. Four new scenarios show impact of changing assumptions. From low to high takeup adds $10 billion to NPV. But from low to high GDP impact adds $30 billion to NPV. More than 1% growth in GDP then NBN is a no brainer. Less than 0.5% then NBN is too expensive for the benefit it provides. Takeup has much less impact than GDP effect.

See spreadsheet here. Originally posted on Whirlpool NBN forum. Original Cost Benefit Analysis: July 2010 | April 2009.

Assumptions:
1a.Takeup high: Business 90%, Consumer 70%
1b.Takeup low: Business 65%, Consumer 50%
2a.GDP high: 1% impact (to year 15, pro rata to % built)
2b.GDP low: 0.5% impact

Scenarios:
1.High takeup, High GDP: NPV $10 billion
2.Low takeup, High GDP: NPV $nil
3.High takeup, Low GDP: NPV negative $21 billion
4.Low takeup, Low GDP: NPV negative $32 billion

With low takeup but high GDP impact, the NPV of the NBN is around $zero. Therefore, the NBN is positive, when business takeup exceeds 65%, AND consumer takeup exceeds 50% and GDP impact is 1% (per annum pro rata to year 15 from build start).

With high takeup, and 1% GDP growth (high GDP impact), NPV is positive $10.66 billion.

With low takeup (business 65%, consumer 50%) and 0.5% growth (low GDP impact), NPV is negative $32.5 billion.

If GDP growth only reaches 0.5% per annum (to year 15) from NBN, then NPV is negative $21 billion. This is with high takeup (business 90%, consumer 70%).

Therefore impact on GDP has the greater impact on NPV of NBN. From low to high takeup adds $10 billion to NPV. But from low to high GDP impact adds $30 billion to NPV. Sounds like we need to be arguing about how much impact the NBN will have on GDP. More than 1% growth in GDP then NBN is a no brainer. Less than 0.5% then NBN is too expensive for the benefit it provides. Takeup has much less impact than GDP effect.

Research on broadband impact on GDP.

1. Access Economics in their recent report for IBM, suggest 1.5% increase in GDP "within a few years" (p.37) for Transport, E-Health, Water, Electricity

However this is for FTTN (for $10-20 billion investment). Access Economics, The Economic Benefits of Intelligent Technologies, May 2009. Would be interesting to try to back out the costs. Though I don't think the 1.5% includes the cost of the network (p.ii), but depending on the degree of spare capacity in the economy (??jobs impact of building network??).



2. The World Bank considers there is a expectation of 1.2% increase in GDP for every 10% increase in broadband (Qiang 2009 pdf). This cites an earlier Qiang working paper of 2008. See picture above from Qiang 2009.

Australia in contrast has much high speed broadband. But many Australians do not use broadband. The ABS 8153.0 (June 2010) cites 9.6M internet services (3.2M over 8mbps; 8.2M over 500kbps). Of these 9.6M, 4.2M are DSL, 3.45M are wireless 0.8M dialup. Therefore there are at least 1.3M with dialup or slow broadband (less than 500kbps). These numbers relate to accounts rather than persons, so in our house one connection is used by me and my wife. Other houses may share a single connection with a family of five. Work and home connections might count twice. Still with 9M households and 2M businesses, there is still some scope for probably another 20% penetration of broadband from nothing, let alone those households with no computer. Malcolm Turnbull cited this week only 43% of homes with income $40,000 or less have internet.

Overall we have 9M households and 7.0M household broadband connections. But with only 4.3M DSL connections, there must be at least 2.7M wireless broadband connections, maybe iPhones and other 3G phones. At least 2M households don't have wireline broadband. Possibly as many as 4M households. Thus there is plenty of scope on the World Bank numbers for GDP growth in excess of 1%. However, the Cost benefit model needs a 1% GDP rise (about $7B per year) to run for about 8 years (from the end of the NBN build until the model ends at year 15).

The community needs to have a conversation about whether we think 1% growth in GDP is viable under the NBN. The NBN if spent in one year is about 3.5% of GDP. A 1% return increasing GDP is then around 30% return on investment. In these terms, the return seems quite high. Yet I am optimistic. Your comments are as always appreciated.