Last month I lambasted Commsday and Henry Ergas in an article for a publishing a misleading report on the cost of services to the home user in which they claimed consumers would have to pay $215/month.

This was widely reported with Ergas later defending his research by quoting services from Internode at $185. Interesting that he should pick the most expensive of all Internode's services with 100Gbytes of monthly quota. Why not choose the 15Gbyte/month $99 plan?

In my opinion the real agenda behind the Ergas evaluation was to support his mates at the Liberal party - he has been engaged in the past by them to conduct tax reviews. See here. It seems his opinion may have been a little biased on this subject. Perhaps it was why Senator Minchin was so quickly onto this economic modelling. It has also more recently come up again in the Senate Select committee on the NBN.


At 02 June, 2009 09:01 Anonymous said...

Its amazing to see everyone's opinions on NBN pricing come complete without any financial analysis.

However if you actually do the math you come up with a required return in the order of 8% ...

[Assumptions: 75% debt funding, 9.5% interest rate on debt, 7.5% risk free rate, 0.65 beta, equity risk premium of 7% and a tax rate of 30%]

assume a growth rate of 5% p.a. on a $43bn build = $1.29bn p.a. in free cash. [$43bn * (8% - 5%)]

[Note: this assumes the retail costs are built into the $43bn as well....]

Next assume the NBN's CapEx is inline with Telstra's at 14% of Revenue, and the NBN plus retailer will have an EBIT margin in line with Telstra's as well: 21%.

[Note the lower the EBIT margin the higher the revenue and monthly charges have to be to generate the required return of $1.29bn free cash]

This means that the NBN (plus retail ISP) revenue must be $1.29bn / (21% - 14%) = $18.42bn. If we assume that this is spread over 10.1 million customers each customer must pay on average $1824.61 p.a. or $152.05 per month for access....

This is an understatement of the true cost of the NBN for several reasons:

1. It assumes 100% of the ULLs will subscribe to the NBN... if only half do then double the monthly costs.

2. It assumes retail ISPs require no capital expenditure.

3. It assumes the NBN will not be sold and will continue to be underwritten by the government... Note: If the Weighted Average Cost of Capital for the NBN increases to 12.25% (in line with utilities) then required revenue jumps to $44.5bn and monthly charges jump to $367.00

4. It assumes the NBN will be profitable from the outset- Note a more complex model needs to incorporate the reduced costs of capital in deferring some of the NBN builds (lowering monthly charges) but unless the NBN is profitable during these times, they result in an overall increase in the present value of the network build.

5. This assumes the NBN and ISP retailers will operate at Telstra's margins. [i.e. NBN Co will needs to be every bit as aggressive as Telstra in taking profits or it will not generate the commercial return]...

6. This assumes retail ISPs have the same cost of capital as the NBN company. But these ISPS have much higher capital costs because they are not government underwritten.

7. In any case this simple BOTE calculation justifies monthly costs in the order of $250 per month.

Note to all other pundits and NBN pricing wannabes - don't show up unless you have done the math...

At 02 June, 2009 15:47 Stephen Davies said...

I would in general agree with your annoymous claims with the exception of the following points:

1. The cost to build the network is not $43b. Several experts have already publicly stated the cost to be around the $25b mark. I suggest you technology wannabes should check the details of the $43b and the basis of this number before making your economic calculations.

2. Your cost of debit seems rather high, considering the government has said it would raise capital itself to fund the building of the network through instruments such as bonds. It would be more reason to assume the government would borrow the money at much lower interest rates than 9.5%

3. The assumption that the price will be the same for all services is unreasonable. There will be a mix of services from a range of customers with different profiles. If Telstra had to fund the rollout of their network entirely from residential revenue, we would have nothing more than a standard telephone.

4. The assumption of the NBNco having a CAPEX inline with Telstra's 14% of revenue is unreasonable. There is no old or outdated technology to support, there is no mobile network to support, billing systems are much easier.


Post a Comment