Sunday, August 31, 2014

And the NBN cost-benefit analysis finds...

No wonder Labor never wanted its National Broadband Network subjected to cost benefit analysis.

For all of its talk about the public benefits that would flow from fibre to the home broadband such as e-health, e-learning and working, those benefits turn out to require very little bandwidth.

The Coalition's independent cost benefit review finds that what does need high bandwidth is high-definition TV. But even that doesn't need anything like what the NBN was going to deliver.

By 2023 the typical household will need just 15 megabits per second, according to the review's bottom-up analysis. The top 5 per cent of households will need at least 43 Mbps. The top 1 per cent will need 48 Mbps. The NBN was going to deliver 100, upgrading later to 1000.

Tech-heads might think those figures are low, but the review explains that the most common household comprises just two people.

"Even if those two are each watching their own HDTV stream, each surfing the web and each having a video call all simultaneously, then (in part thanks to the impact of improving video compression) the total bandwidth in 2023 for this somewhat extreme use case for that household is just over 14 Mbps."

Bandwidth that isn't used is bandwidth that's wasted. If it costs a lot to get it it's money wasted.

The review finds the Coalition's pared-down model will cost $7.2 billion in today's dollars. Labor's fibre-to-the-premises NBN would have cost $17.6 billion. The benefit from the Coalition's model is $1 billion. The benefit from Labor's model is minus $4.7 billion. That's right, minus $4.7 billion. The minus sign is because Labor's scheme would have delayed the spread of fast broadband. The review finds that without it the market would have delivered high-speed broadband to most Australians sooner.

Australians care about speed, but not that much. The review finds we are prepared to pay an extra $1.50 per month for an extra Mbps when our speed is low, but only an extra 70¢ when it climbs to 50 Mbps and nothing when it climbs above 90 Mbps.

On the other hand, we care deeply about getting high-speed broadband quickly.

The Coalition's plan uses a mix of technologies to deliver speeds of 20 to 100 Mbps to all of Australia and does it much sooner than Labor would have delivered 20 in the bush and 100-plus in cities.

The Coalition scheme will cost a net $620 per person in today's dollars. Labor's would have cost $2200.

It's hard to fault the review panel for thoroughness. It's review goes out until 2040, and beyond that using residual value. It developed Australia's first bottom-up analysis of what Australians actually need; something Labor never did, starting instead with a solution and then assuming it was needed. It commissioned the Institute for Choice at the University of South Australia to examine what consumers were actually prepared to pay and it has submitted its work to peer review from an international panel made up of experts from the Brookings Institution and the University of Texas.

It finds that both the Coalition and Labor's NBN needlessly waste money, but that Labor's wastes more.

In The Age and Sydney Morning Herald

Cost benefit analysis shows Turnbull plan has $16b advantage

The Abbott government's pared-back broadband plan is three times more cost effective than Labor's ambitious scheme and would leave Australians $16 billion better off, according to the first independent cost-benefit analysis of the national broadband network.

In a scathing verdict on the Rudd and Gillard governments' plan to introduce fibre directly to 93 per cent of premises, the cost benefit analysis finds the policy is so expensive it would barely leave the community any better off in net terms than if broadband investment remained frozen at present levels.

The much-anticipated report finds households and businesses will benefit from quicker downloads but the much-vaunted societal benefits of fast broadband – such as improvements to health and education services – will probably be extremely limited.

The cost benefit analysis panel, led by former Victorian Treasury head Michael Vertigan, modelled the estimated costs and benefits of expanded broadband access from 2015 to 2040.

Communications Minister Malcolm Turnbull, who commissioned the analysis, will use the findings to justify his decision to pursue a "multi-technology mix" broadband network using a combination of fibre to the premises, fibre to the node, hybrid-fibre coaxial, satellite and wireless services.

Mr Turnbull regularly attacked Labor for failing to commission a cost-benefit analysis before launching the national broadband network, the biggest infrastructure project in Australia's history.

The report finds the multi-technology mix model outperforms a fibre to the premises plan in net economic benefits in 98 per cent of scenarios.

A multi-technology mix NBN would cost $24.9 billion to launch from 2015 compared with $35.3 billion for fibre to the premises (FTTP), the report finds.

A multi-technology mix would deliver download and upload speeds of 20-100 megabits a second, while FTTP would deliver speeds above 100Mbps.

The report finds the most cost-effective option would be an unsubsidised launch in which the free market delivers high-speed broadband to 93 per cent of homes. This would have a net economic benefit of $24 billion, but would leave 7 per cent of premises in regional and rural areas without fast broadband.

Providing fast broadband to the bush through wireless and satellite services – as envisaged under both Labor and the Coalition's plans – will cost nearly $5 billion but produce only $600 million in economic benefits.

This suggests an alternative model for a government committed to fast broadband for all Australians at a low cost: subsidising the introduction to the bush while leaving private providers to serve metropolitan areas.

Compared with the unsubsidised launch scenario, the multi-technology mix model has a net cost of $6 billion ($620 a household) and fibre to the premises has a net cost of $22 billion ($2220 a household).

Australians would be only $2 billion better off, in net terms, than if there was no further launch> of broadband under a fibre to the premises model. They would be $18 billion better off under a multi-technology mix.

The report finds a multi-technology mix is more "future proof" because it can be upgraded to fibre to the premises later if demand for fast broadband booms.

"The [multi-technology mix] scenario leaves open more options for the future because it avoids high up-front costs while still allowing the capture of benefits if, and when, they emerge," the report finds.

The cost benefit analysis finds most of the benefits of fast broadband – most notably video downloads – will accrue to private users within households and businesses. By contrast, hospitals and schools require relatively low bandwidth to deliver services. 

As well as Mr Vertigan, the expert panel members were economist Henry Ergas, former Australian Communications Authority chairman Tony Shaw and former eBay Australia managing director Alison Deans.

Mr Vertigan said the findings of his report show that cost benefit analyses should be mandatory before construction begins on major public infrastructure projects.

Labor communications spokesman Jason Clare said the cost-benefit analysis was "tainted" by the involvement of figures such as longstanding NBN critic Henry Ergas.

"It's hard to take the report seriously when three weeks before the last election Malcolm Turnbull said he would get this report done by the government body Infrastructure Australia and instead what he has done is got some of the most vociferous critics of the NBN, as well as former staff, to write this report," he said.

Mr Clare said the government had a "myopic" view that fast broadband was just about video games.

This is about setting us and Australia up for the future. That is why Japan, South Korea and China and New Zealand are doing it and even in Indonesia." 

In The Age and Sydney Morning Herald SUNDAY EXPLAINER

It was going to be bigger than the Snowy Mountains Scheme and it was going to revolutionise our lives. Four years ago Labor began an audacious program of connecting every house and business in the country to the internet, all over again. It was to cost $43 billion and it commenced without a cost benefit analysis. This week the Coalition released the results of a cost benefit analysis into Labor's NBN, the first. It finds the benefits would never have approached the cost. Who's right, and who's offering the best plan now? Peter Martin runs the numbers.


Every address in every city and reasonable sized town was to be offered the chance to be wired up to an optical fibre cable. Remote locations were to be connected by satellite and wireless. As of June this year NBN Co had connected 210,000 locations and had millions to go. The promised speed for the fibre links was 100 megabits per second, with the whole system upgradable later to 1000 Mbps, each far faster anything that was available at the time.


The cost benefit study finds most uses don't require anything like 100 Mbps. Ultra high definition TV comes closest. The government's advertising stressed instead the unknowable nature of its uses saying the NBN would be used "in ways we only dream of".


Labor might have felt there was no point after it had given the go ahead. But it was big on claims. The NBN would "boost our economy", it would "open up a world of interactive health care never thought possible". It might have felt these claims couldn't be tested. After all if its uses are unknowable, how can we know what it will be used for?


A cut-down version. Instead of costing $73 billion (the revised figure for Labor's scheme determined by the strategic review in December, 2013) the Coalition's scheme would cost $41 billion. Instead of connecting every house and business in every city to an optical fibre, it would use a mix of technologies. One quarter of city addresses would get fibre, 44 per cent would get fibre to the basement in apartment buildings, or fibre to a node (a shared connection housed in a street cabinet), with the last link to the residence or business by existing copper wire and 33 per cent would be serviced by Foxtel and Optus Vision cables re-engineered to deliver high speeds.  

The Coalition would deliver 43 per cent of the nation 25 Mbps or more by 2016 and 91 per cent 50 Mbps or more by 2019 (two thirds would get 100 Mbps). By then Labor would have wired up only 57 per cent of the nation.


It can certainly deliver speed more quickly than could have fibre to each home. NBN Co says fibre to the apartment test sites in Victoria have delivered download speeds in excess of 100 Mbps and upload speeds in excess of 45 Mbps. Fibre to the node test sites at Woy Woy in NSW have delivered download speeds of 95-97 Mbps and upload speeds of 28-34 Mbps.


The cost benefit review panel headed by the former head of the Victorian Treasury Michael Vertigan> finds we won't need that much.

It says by 2023 the typical household will need 15 Mbps, the top 5 per cent will need 43 Mbps, the top 1 per cent 45 Mbps and the top 0.01 per cent 48 Mbps.

It says the biggest user of bandwidth is high definition video. Even a household watching two HD video streams at once, simultaneously searching the web twice and conducting two video calls will need "just over 14 Mbps".


Too right. But high definition video is the biggest hogger of bandwidth. The review finds other uses including e-health need much less. Hospitals and associated institutions that need serious bandwidth for e-health can connect directly to the internet via fibre.


Yes, even for fridges and air conditioners. But those sort of uses don't require high speeds, and extra use of things that do require high speed doesn't necessitate even higher speed. The panel explains it this way: If a household that previously watched one hour of high definition video a day ramped up its usage to four hours a day the amount of data it needed would have quadrupled, but the speed it needed would have not.


You can connect directly to the internet by fibre, but it'll cost you a few thousand dollars. The review believes few households will bother. It finds that after speed reaches 90 Mbps our willingness to pay for even more flattens to zero.


Yes. Around 18 per cent of the NBN's first customers have signed up for the top speed. But most, three quarters, want 50 Mbps or less. One third want only 12 Mbps.


That's Labor's argument, that the future is unknowable. But it cuts both ways. One response to an unknowable future is to build something very expensive in case it will be needed, the other is to build something cheaper which can be expanded later where needed.


Converted to present day dollars the panel says the Coalition scheme will cost $7.2 billion. The Labor scheme would have cost $17.6 billion. It says the Coalition scheme will deliver a (disturbingly low) benefit of $1 billion. In contrast Labor's scheme would have delivered a benefit of minus $4.7 billion. The minus sign is because it would have delivered high speed broadband slower than would the market itself. The review found that beyond a certain point we don't much value the extra speed we would get in return for the delay.


By 2020 under the Coalition's scheme. By 2024 under Labor's.

Peter Martin is economics editor of The Age

In The Age and Sydney Morning Herald