Big risks require bigger rewards

THE recently released McKinsey/KPMG study on the National Broadband Network is 534 pages long. How many times does it say the NBN is worth doing? Not once. As postmodernists say, read the silence.

What the study shows is that the NBN will earn a rate of return that fails to cover the project's risk-adjusted cost of capital.

But risk is as real a cost as any other. And a project that doesn't cover that cost makes a loss every bit as real as one that cannot pay its employees' wages. Shifting that loss on to taxpayers in no way makes it disappear. It merely hides it, undermining both fiscal honesty and economic efficiency.

The study does not gloss over the high risks the NBN, with its estimated $43 billion outlays, involves. But the best return it can find barely equals that on a deposit account. And even to come up with that gloomy forecast the study makes some brave assumptions.

First, it assumes wireless broadband will not be an effective competitor. True, guaranteed high speeds may at some point give fixed networks a decisive advantage. However, that point seems far away. Rather, it is consumers' insatiable demand for mobility, as evidenced by the iPhone and the iPad, that now drives the development of services and applications. Difficult to believe then that wireless competition will not erode the NBN's revenues.

Second, the study assumes that even without an agreement with the government, Telstra will simply hand its customers to the NBN. As a result, the scenarios it models do not include one in which Telstra competes aggressively with the NBN. But given the study's recognition that in most countries, networks such as Telstra's compete successfully with those based on fibre, that scenario cannot be ignored.

Third, and most surprising, the study assumes that NBN Co will get away with massive price increases. Specifically, the study assumes that NBN Co will charge retailers $25 to $30 a month to use its lines for basic telephony. But providing basic telephony involves other costs as well. Given those costs, NBN Co's proposed charges mean users who only want phone service will face a price increase of some $5 to $10 a month, or 20 to 40 per cent.

The study then assumes that, as a monopoly, NBN Co will be able to increase its prices by about 1 per cent a year above inflation. As a result, charges in thirty years will be 35 per cent higher in real terms than at the start of the NBN's life.

Yet telecommunications prices in Australia declined by over 50 per cent in real terms in the past 30 years. And the trend internationally is for prices to continue falling.

Will Australia really reverse course, and become the only developed country with steadily rising telecommunications prices?

But even assuming it does (and imposes a virtual prohibition on competition to boot), the study finds the new network can only expect to earn the bond rate. This is deeply problematic.

First, the bond rate provides taxpayers with no compensation for being forced to bear the high risk the NBN involves.

Second, it contradicts claims made when the NBN was announced. Communications Minister Stephen Conroy said the new network would provide taxpayers with "a utility-style rate of return". However, recent regulatory decisions show Australian regulators believe even very low-risk monopolies such as electricity distribution require rates of return twice the bond rate.

Third, allowing the NBN to set prices to such low required returns would, as the study recognises, breach the government's own Competitive Neutrality Guidelines. These require government businesses to cost their capital commercially.

Fourth, it implies a misallocation of scarce savings. As a recent Productivity Commission study of public investment criteria stresses, each dollar invested in projects such as the NBN is a dollar not invested in alternatives that could earn a fully commercial return. Each tax dollar needed to fund the NBN's losses imposes further costs by distorting decisions to work, save and invest.

How big is the loss due to the shortfall between the bond rate and the project's real cost of capital? The government has not released the study's detailed modelling (but should). However, using the Australian Competition and Consumer Commission's methodology, the study estimates Conroy's promised utility-style rate of return for the NBN at 8.8 to 12.4 per cent (compared with the government 10 year bond rate of 5.48 per cent).

This implies the project will lose some $12.5 billion in present value terms, even on the study's assumptions about costs, prices and take-up. Effectively, nearly half the project's capital has to be written off at the start. It follows that the project cannot meet the national accounting guidelines' criteria for it to be taken off-budget: that it will "recover a considerable proportion of its costs", "cover its capital and other costs" and not be dependent on grants or subsidies. Rather, the loss is a taxpayer-funded grant to the project that should have been disclosed in the budget.

Many criticisms can be made of the study. But it is a serious piece of work, and its findings deserve respect. Kevin Rudd committed the government to evidence-based policy. Well, here is the evidence.It's up to the government to show it can listen.

And it should listen in another area too: the proposed mining tax. On Tuesday, Treasury Secretary Ken Henry said the commitment to eventually repay a share of miners' costs was "identical" to giving miners a government bond. This is incorrect. The commonwealth has never defaulted on its debt; but it has made countless changes to the tax rules, particularly when strapped for cash. Promises to repay costs through the tax system are therefore high risk. Miners can no more borrow against that promise at the bond rate than they can fly to the moon. If the rate at which they're compensated does not reflect the risks, they will take their dollars for capacity expansion elsewhere. Not even government can wish risk away. Trying to force it on to miners, who have a choice about where they invest, is foolish. Forcing it, as in the NBN, on taxpayers merely makes the community poorer. Experiments show even monkeys understand that higher risk requires higher reward. Surely government can too.