Open your kimono and show us your modelling

I WAS wrong. Treasury's modelling doesn't assume the US has an emissions trading scheme in place by 2016. It merely assumes its economy operates as if it did.

Unnamed government sources have told the Fairfax press that Treasury assumes the US will "reach emission reduction targets at a cost no higher than the international price", that is, at least cost, without having to bother with a market-based mechanism.

That would not be a mere accomplishment: it would be a miracle. At least if you take Climate Change Minister Greg Combet seriously who has repeatedly said imposing a carbon tax is "essential" for "achieving emissions reductions at cheapest cost". As have Julia Gillard and Treasurer Wayne Swan.

But Treasury apparently knows better. If those government sources are to be believed, its officials have discovered a way of getting the benefit of a carbon tax without actually having one.

Yet, oddly, that miracle cure doesn't seem to have been around when Treasury wrote its modelling report. Rather, Treasury assumes that from 2016, under the auspices of a "co-ordinated international policy regime", industrial countries, including the US, would "trade, either bilaterally or through a central market" delivering a "harmonised world carbon price".

How the US will trade if it doesn't have some form of permit system is a mystery worthy of Hercule Poirot. And the mysteries don't end there. If the US can achieve least-cost abatement without an ETS in 2016, why put one in place later? Indeed, why would anyone bother with such a scheme?

Answering these questions would be easier if the government opened the kimono on the actual model. Given access to the model itself, we would know exactly what it assumes. And the implications of changing those assumptions could be tested.

It would be possible to assess the costs to Australia if we adopt a carbon tax and our major competitors don't: the scenario Treasury's report fails to detail. And it would also be possible to examine the effects of other crucial features of Treasury's modelling.

For example, the model does not provide for the mandated decommissioning of the Hazelwood and possibly Yallourn power stations in Victoria. These generators have low operating costs and even with a rising carbon price would operate until at least 2025. Replacing them sooner requires substantial investment in generating plant and transmission. That will need to be paid for. But when?

Although the government is talking of decommissioning those generators now, Treasury's modelling seems to defer the cost until at least 2025 and maybe until 2040. That conveniently reduces the estimated hit to electricity prices.

The model also assumes unlimited access to permits overseas. Those permits provide two-thirds of our mitigation to 2020, "resulting in lower economic costs". But the government has now said it will cap purchases of foreign abatement at far less than that. So here, too, the policy's costs are underestimated.

And the model also unrealistically assumes the government's policy is revenue neutral (so that other taxes don't need to be raised to finance any shortfall), with all revenues returned to taxpayers as lump sum payments, so the compensation does not distort any decisions.

But here's the best bit, tucked away in a technical annex. The modelling assumes emitters can borrow permits from the future. And borrow they do, on a scale that puts Greece to shame.

By 2050, emitters worldwide have borrowed four years' global permit allocations from the future. Using Treasury's estimate of future carbon prices, that is equivalent to a net debt of $10.7 trillion in 2011 dollars, or 10 times Australia's current national output. And the total value of those net borrowings would rise at 6 to 8 per cent a year, far exceeding the growth rate of world incomes.

Why assume debt accumulation on such a plainly unsustainable scale? Because it postpones the pain, shifting emissions reduction to beyond the modelling period. The problem, however, is no current or likely scheme allows such net borrowing, much less on the scale Treasury envisages. So that further underestimates the policy's costs, probably greatly.

But without access to the model no one can say by how much. And that suits the government. For Treasury's modelling presumably reflects assumptions determined by the government, such as that all industrial countries have carbon taxes in place by 2016 or behave as if they did.

The government also presumably determined what was not to be disclosed: most importantly, the consequences if we tax our mineral exports and our competitors don't. Treasury then modelled and explained those scenarios as best it could.

Fair enough; that is the government's prerogative and Treasury's job. But when caught out, spare us the contorted denials.

Rather, when the facts come home to roost, have the good grace to make them welcome. Or as Gillard put it: "Don't write crap. It can't be that hard."