Home > Economics, Environment > Carbon tax and evidence-based policy

Carbon tax and evidence-based policy

June 6, 2011

I accept what I see as the mainstream science on climate change. The world has had a warming trend in recent decades, this is partly caused by humans, and I expect it to continue into the future. I am comfortable with the IPCC range of estimates of warming, which has a mid-point estimate of 2.8 degrees over pre-industrial temperatures (ie 2.1 degrees above today). I hope the skeptical scientists are right and that warming will be lower, but I find the IPCC estimates plausible and an appropriate starting point for analysis.

But before I accept a climate policy, such as a price on carbon, I want to see that the policy has more benefits than costs.

It’s great to have good intentions. We all want to do good. But poor people can’t eat good intentions, and good intentions will not cool down the globe. If we really do care about good public policy and making the world a better place, we must not introduce policies that fail a benefit-cost analysis. And the simple truth is that a carbon tax fails under any reasonable set of assumptions.

To do an analysis of a carbon price we need five inputs to the model. We need (1) an estimate for the costs of climate change; (2) an estimate for how much of those costs will be prevented with a global binding deal; (3) an estimate for how much Australia will impact the likelihood of a global binding deal; (4) the costs of the carbon price; and (5) the discount rate. For all of these issues, if we use the most dramatic and pro-tax assumptions we can find, the policy still fails. Even the Garnaut Report admitted that “the accrued benefits [from a carbon price] were almost as large as the costs”. Almost. Amazingly, this didn’t make it into the papers. Of course, Garnaut goes on to say that if you keeping adding in extra benefits (while not adding in extra costs) then eventually you can make the policy pass, but with that approach you could make anything pass.

Even then, the Garnaut report does not use a reasonable set of estimates. In all assumptions, he picked extreme “pro-tax” values. While the standard approach to benefit-cost analysis uses a 5% discount rate, Garnaut prefers 1.35%… but the policy still fails. While the IPCC predicts 2.8 degrees of warming by 2100, Garnaut uses 5.1 degrees of warming (does this mean he’s a science denialist?)… but the policy still fails. While the peer-reviewed analysis of the costs of climate change are at 1-3% of GNP by 2100, Garnaut finds a 10% cost (more than three times higher than the Stern report)… but the policy still fails. While we will likely not be able to stop most of the costs of climate change, Garnaut hopes we can stop 80% of the costs… but the policy still fails. While Treasury estimates the costs of a carbon price at 5-7% of GDP by 2050, Garnaut uses only 4.5%… but the policy still fails. Even with all of these dodgy assumptions, it is just not possible to make a carbon price into a winner.

And all of this excludes Garnaut’s biggest exaggeration. Australian politicians can only set Australian policy, not global policy. All honest commentators know that carbon mitigation can only succeed with a strong, binding, global agreement, and that Australia acting on our own is useless. Garnaut responds (correctly) that Australia has influence on the world and so our policy settings can still have a benefit. True. But we do not control the world, so our influence is less than 100%. The relevant question is: “how much more likely is a strong, binding, global agreement because of Australia’s carbon price”? An honest assessment might be 1% or 5% or 10%… but Garnaut implicitly uses 100%.

Indeed, it could be that passing an Australian carbon price reduces the incentives for a global deal, because we are promising to act irrespective of whether there is a deal or not. A better strategy for encouraging a global deal might be to pass legislation that introduces a carbon price only if there is a strong, binding, global deal, which gives other countries an incentive to pass such a deal.

But even if we use Garnaut’s assumption that Australia controls the world… the policy still fails!

There is just no way for an honest person to come up with assumptions that make a carbon price pass a benefit-cost analysis. If you use more reasonable assumptions, the policy is a massive failure. Consider these assumptions: (1) costs of climate change = 2.9%, taken from Stern report; (2) a global binding deal could prevent half of these costs; (3) Australian policy influences the likelihood of a global deal by 10%; (4) costs of a carbon price = 5%, which is the lower end of the Treasury estimate; and (5) a discount rate of 3%, which is a compromise between the mainstream 5% and Garnaut’s preferred number. This gives a benefit cost ratio of 1:100, and many would claim that the assumptions are still very generous.

Pro-tax people will need to think of some heroic assumptions to get their preferred policy to pass. If you have some ideas, feel free to tell me your assumptions and I’ll tell you the associated benefit cost ratio. Personally, my “best guess” estimates would be (1) costs of climate change = 2.5%, taken from Nordhaus; (2) a global binding deal could prevent half of these costs; (3) Australian policy influences the likelihood of a global deal by 2%; (4) costs of a carbon price = 5%; and (5) a discount rate of 5%. This gives a benefit cost ratio of 1:1000.

Global warming and carbon taxes are emotional issues, with two intrenched sides often shouting past each other. But for people interested in evidence-based policy the conclusion is clear: a carbon price is bad policy. If we want to see action taken on climate change, we need to start considering “no regret” policies. And nuclear power.


This article is an edited overview of my speech at a recent debate (3 June 2011) hosted by Property Council of Australia. The other speakers against the tax were Bob Carter and Michael Matusik, while the pro-tax side was Matthew Bell, Kellie Caught, and Kirby Anderson, with Mark Ludlow (AFR) as the adjudicator. At the end of the debate, the anti-tax side was determined the winner by audience applause.

  1. jc
    June 18, 2011 at 6:47 am

    good post dude

  2. Mitch
    June 19, 2011 at 4:37 am

    So what if a $30/tonne carbon tax replaced fuel and diesel excise?

  3. June 19, 2011 at 9:43 am

    @Mitch — An excellent question, and not one I can answer properly without more analysis. It depends on several things, including (1) the comparative size of the new tax vis-a-vis the old taxes, and (2) the relative efficiency costs of the two taxes.

    Importantly, it would also hinge on assumptions about the future of the carbon price. If it was to stay fixed at $30/tonne then the associated economic costs would be lower than the Treasury estimates (which assumed an ultimate price of between $100 and $200/tonne).

    If we can arrange a carbon tax in such a way as it is simply good tax reform, and good for the economy, then it would fall into the category I call “no regret” policy. I have no in-principle opposition to tax reform that includes a carbon tax, but it needs to be “no regret”, and that means not turning into an ETS and not increasing the rate (unless as part of further future good tax reform).

  1. No trackbacks yet.
Comments are closed.
%d bloggers like this: