A comparative spreadsheet of Net Present Value (NPV) of Copper (Cu) vs Fibre Optic (FO), using Discount Rates of 0% (raw $ amounts) and 6%,7% & 8% and replacing Copper with Fibre: never, 5 years, 10 years and 15 years.
I don't attempt to model inflation, nor include Revenue. These effects cancel each other out in comparisons. This is meant to be a "first-order", not detailed calculation.
The first tab (Cu-FO) is the "never replace" scenario.
Three other scenarios of 5, 10 and 15 year replacements are included. The 10 and 15 year scenarios are after the 6-8% cross-over point, and are redundant.
Unsurprisingly, with the values constructed, not backed by evidence, by the Coalition, Copper is always cheaper in the "never replace" scenario. It's like the value were chosen to yield a desired result.
For my better estimates, I've used the $1350 construction cost from the 2012 NBN Co Corporate Plan, excluding Telstra payments as they are excluded from the Copper, but quite incredibly included in the Coalition's $3,600 figure.
Fibre maintenance I've reduced by half, not to the proportional amount ($20).
Copper and Node line maintenance I've increased by 20% from $90/year to $108/year.
Usually with I.T. systems, hardware maintenance costs are 10-20% of initial cost.
The Coalition's $90 estimate is just 10% of line install cost, with no obvious allowance for line and node maintenance.
The Coalition NBN background document provides (p14) an example NPV calculation, with the cost cost/premise of Fibre reducing by 10% in 4 years. I've created a tab calculating discounted figures for the Coalition "highball" scenario and my more realistic scenarios.
This is an absurd figure, given that it includes Telstra payments, Labour costs and machinery, fittings and fixtures costs which do not reduce with Moore's Law, but increase at least as fast as inflation.
In my realistic Fibre install, I've allowed 25% of the cost in electronics, and a very generous reduction-rate of 20% per 4 years, twice as fast as the Coalition rate.