Economic modelling may overplay the costs of Australia's 2030 climate target

The debate about Australia’s 2030 climate target, and how much it will cost Australia’s economy, continues.

Last month, Australia announced a target to reduce greenhouse gas emissions 26-28% below 2005 levels by 2030. Modelling by Professor Warwick McKibbin, economist at Australian National University, found that this would come at a cost of 0.6% of gross domestic product (GDP). McKibbin has argued that this puts Australia at the “high end” of global action in terms of economic costs.

The Climate Change Authority, which has recommended a higher target of 40-60% below 2000 levels by 2030, has argued that Australia’s target is inadequate based on the size of emissions cuts.

So is the modelling a solid foundation for Australia’s new climate target?

Making assumptions

Economic models are often the framework for making policy decisions. They start with “back-of-envelope” calculations and end with computer models requiring heavy computation of interaction between the major variables.

All economic models make assumptions about the world. For instance, the price of gas. However unless we know what assumptions a model makes, we can’t test or scrutinise it.

McKibbin’s modelling shows that action to reduce Australia’s emissions by 26-28% by 2030 is an affordable exercise. Reductions in economic growth are relatively small but in order to defend standards of living, the Australian government argues that action taken should be conservative.

He modelled a range of scenarios, including a 45% reduction target that would come at a cost of 1% of GDP.

The assumptions behind McKibbin’s modelling are not easy to unpack. The model focuses on energy related emissions and represents a highly aggregated framework of the economy. Emission reductions are based on two key levers: the transition to low-carbon energy fuels and improvements in energy efficiency.

Both of these levers involve many assumptions, evidence of which are not available in the reports thus far released.

What’s in the mix?

The most significant of the missing assumptions is what Australia’s electricity mix looks like with and without a climate target.

Australia’s fleet of electricity generators is currently composed of ageing coal-fired power stations plus a few gas-fired power stations. The government’s Renewable Energy Target indicates 23% of electricity will come from renewable sources by 2020. There is a widely held view that Australia doesn’t need any more capacity to produce electricity before 2023-24.

So even without the new climate target there will be changes to Australia’s electricity mix, such as closure of old coal-fired power plants and investment in new plants, but it’s not clear from the modelling what these changes will be.

There is also no detail about the electricity mix under new climate targets. It is essential to understand the relative contributions of gas and renewables in any target scenarios.

Mature renewable technologies, such as wind and solar photovoltaics, are cheaper than emerging technologies such as solar thermal, geothermal, wave, storage technologies. So the type of renewable energy in the mix has a large impact on how much reducing emissions will cost.

Mature renewable technologies will be a cheaper source of energy by 2025. As South Australia has found, investment in renewables for South Australia has put downward pressure on wholesale electricity prices, not increased it. Independent modelling, both commissioned for and submitted to the recent Renewable Energy Target review, which can be seen here and here arrived at similar conclusions.

So, the costs of renewable forms of electricity are likely to reduce the costs of electricity, and yet the assumptions in the modelling point to an increase in cost for pursuing renewable forms of energy not a decrease.

Gas: cheaper or more expensive?

Any transition to a low-carbon energy future will be reliant on the affordability or otherwise of natural gas. Gas, per unit, produces less CO₂ than coal, so increasing the proportion of gas in Australia’s energy mix will reduce overall emissions.

Predictions about the future price of gas are many and varied but none of them is explicitly referenced in the modelling conducted. There is no way to know whether the assumptions used included a high cost of gas or a low cost of gas, and thus what the relative contributions of renewables and gas generations are to the cost of meeting the climate target.

McKibbin readily acknowledges that his assumed costs for renewables are at the upper end of commonly accepted range of estimates. A conservative assumption would be high gas prices in the future.

In our view, the economic impact of action is likely to be lower than the headline results that were reported in the modelling when based upon more likely technology costs.

The cost of climate

The effects of climate change are expected to include significant levels of damage from unpredictable climate events. The worst of these events is thought to be a few decades into the future, but the costs should be taken into account now rather than loaded into the future.

We do this today with simple cost-shifting exercises such as depreciation and superannuation. It would be logical to start this process with economic models too.

McKibbin acknowledges that his modelling includes no costs associated with climate change damage, but in our view, excluding these costs reduces the value of the modelling exercise. Ignoring these costs is also likely to present an artificial picture of the costs of pursuing carbon abatement policies, by both exaggerating the benefits and downplaying the costs of no action.

Forecasting big changes

Furthermore, there is a question about the applicability of this model to predict the level of disruptive change likely to be experienced in the energy industry over the next couple of decades.

As seen with recent trends in wind and solar photovoltaics, production and process innovation together with scale economies often drive unexpectedly large reductions in costs, as acknowledged by McKibbin.

If we are to see levels of disruption like that experienced in the IT and telecommunications industries over relatively short periods of time, the tools used by economic modellers must be able to take into account the innovation that is gathering force in the energy industry. The tools used to represent innovation and complex policy mechanisms in the model are, in our view, inadequate to make useful predictions with respect to targets for 2030.

We welcome modelling that helps politicians make informed decisions with respect to climate targets, but the assumptions made should be transparent and defensible and that models used are appropriate for the task.

Our concern with the modelling conducted thus far, is that it is not sufficiently robust to serve as Australia’s primary methodology for setting targets at the Climate Change conference in Paris in November.

John Foster has received funding from the Australian Research Council for modelling the impacts of renewable energy on the electricity sector and the wider economy

Liam Wagner receives funding from the CSIRO to model the impacts of renewable energy on the electricity market and the wider economy.

Phillip Wild has worked on projects receiving funding from the Australian Research Council and Education Investment Fund (EIF) Research Infrastructure Program for modelling the impacts of renewable energy on the electricity sector and the wider economy.

Lynette Molyneaux does not work for, consult, own shares in or receive funding from any company or organization that would benefit from this article, and has disclosed no relevant affiliations beyond the academic appointment above.