Today’s news includes reports that to be compensated for the carbon tax, electricity generators will be required to commit not to reduce production. Que? Isn’t the whole point of a carbon tax to force them to do so? And how can one reconcile such commitments with the projected reductions in electricity sector emissions in Treasury’s carbon change modeling? So will the revised modeling Wayne Swan has promised now show generation with these new commitments in place?
Or is it that the promised reductions in emissions will still occur, but – perish the thought! – by means of ‘direct action’, including the mooted plant closures and the requirement that is to be imposed on the generators (also linked to the compensation) to “commit to increasing efficiency”?
To which one can only say “increased ‘efficiency’ as determined by who?”. Despite all the hot air about relying on markets, are we now assuming the Gosplan will set sensible efficiency targets for generators? And if it can do that, why have a carbon price to begin with? Why not simply regulate electricity generation? Indeed, why not go back to owning it, and then the government could simply do as it pleases, NBN-style?
I see it now: the National Generating Network – lean, green and coming to your marginal electorate any day now. Actually, why not secure an even bigger saving and merge the NGN and the NBN? Given that there will be a subsidized roll-out of smart meters, won’t combining the two allow terrific synergies? That way, you can pay more for both with all the convenience of a single bill.
Well, thank God the meters are smart. Someone needs to be. And they face no risk of competition on that front, at least from those responsible for the government’s carbon policies.
How marginal was my abatement?
Still on carbon prices, in today’s Australian I discuss the marginal abatement cost (MAC) curves in Treasury’s carbon tax model. It is simply not possible to get into technical details in an op-ed for a general audience. But some aspects are worth clarifying for those with an interest in the specifics. So, taking what I put in the op-ed as read, here goes.
Broadly, Treasury seems to use different specifications for the MAC in its models for Australia (MRRF) and for the rest of the world (for which it uses GTEM). In the GTEM modeling, abatement secured by the MAC seems completely costless. MRRF, in contrast, allows for some cost, but it is not at all clear what that cost is in the Treasury specification.
In the MRRF documentation, the costs were proxied by the carbon price. Obviously, the developer of a technology that allowed an firm to save one unit of emissions could not charge more for that technology than the cost to the firm of that unit of emissions; so in that sense, using the carbon price as the proxy of user cost makes some sense. (The user saves one emission, i.e. avoids having to pay one carbon price unit; the cost of doing so is capped at one carbon price unit.)
However, one would expect the supply curve of emissions-reducing innovations to be steeply upward sloping. As a result, what should happen is that as one reduces emissions levels, the cost of eliciting further emissions-reducing innovations should escalate rapidly.
But that will not be captured if the cost simply increases linearly with the carbon price (so that when one increases the carbon price from $100 to $101, the estimated cost of eliciting further emissions-reducing innovations increases by $1, even though by then emissions would presumably already be at low levels). To that extent, the MAC will result in an ever greater under-estimate of costs as the emissions price rises.
This is all the more the case as the Treasury MAC uses unrealistically bullish technological change parameters, for instance for cement. It assumes, in other words, that it is significantly easier to develop ways of reducing emissions than it realistically could be. How those parameters have been set and tested is left unexplained.
In short, Treasury’s MAC assumes it is far cheaper to develop innovations that reduce emissions than it could reasonably be thought to be. Indeed, it seems to assume that for everyone but us, doing so is free.
Of course, the problems with the MAC don’t end there. Two points are worth making, above and beyond those discussed in my op-ed.
First, if you believe in induced innovation caused by changes in relative prices, why only look at emissions? Compared to the base case, other prices also change as the carbon price rises, and those changes too should trigger induced innovation. A priori, there is no reason to rule out the possibility that some of those other innovations will be emissions-creating, rather than emissions-saving.
Second, the model seems to assume full diffusion in three years of the emissions-reducing innovations generated by the MAC. That seems almost unbelievably rapid, especially for highly capital intensive industries, which usually have long asset lives and high retro-fitting costs. And it far exceeds the rates of diffusion of industrial innovations found by Mansfield, Rogers and the other classics in the field. Yet no explanation is given for that assumption.
A final point. In this op-ed and previous ones, I have often had to use phrases such as “the Treasury model seems to assume”. The problem lies in the documentation, which is far from being clear and comprehensive. It would obviously have been far better if the actual model had been released along with proper user documentation. After all, taxpayers paid for it – why shouldn’t they see it? And that way, the impact of assumptions made about the MAC – and all the others – could be properly tested.
PS: Aficionados of this sort of thing will have tweaked to an intriguing question ignored in the Treasury model’s documentation. How could the same innovations be free in (say) the United States but not here? Is the model assuming restrictions on the import into Australia of new capital goods and other ways of diffusing technology? What is the nature of those restrictions and where are their costs modelled? Or is there global price discrimination? If so, how can it be that the only country with a non-zero price is Australia? How could that outcome be consistent with the standard assumption of profit maximisation (given that if the firms that control those innovations have market power, they wouldn’t only price discriminate against Australia)? And in such a world, why would there only be price discrimination on innovations generated by the MAC? In short, how can that difference in assumptions between GTEM and MRRF make any sense in an integrated world economy?