Key points

Type 1 (modelled median outcomes) plus Type 2 (estimates of other median outcomes) costs of climate change in the 21st century are much higher than earlier studies suggested. The Platinum Age emissions grow much faster than earlier studies contemplated.

The modelling of the 550 mitigation case shows mitigation cutting the growth rate over the next half century, and lifting it somewhat in the last decades of the century.

GNP is higher with 550 mitigation than without by the end of the century.

The loss of present value of median climate change GNP through the

century will be outweighed by Type 3 (insurance value) and Type 4 (non-market

values) benefits this century, and much larger benefits of all kinds in

later years.

Mitigation for 450 costs almost a percentage point more than

550 mitigation of the present value of GNP through the 21st century.

The stronger mitigation is justified by Type 3 (insurance value) and Type 4

(non-market values) benefits in the 21st century and much larger benefits

beyond. In this context, the costs of action are less than the costs of inaction.

Does participation in global mitigation, with Australia playing a proportionate

part, and with all the costs of that part, make sense for Australia? If so, what

extent of mitigation would give the greatest benefits over costs of mitigation for

Australians?

The costs of mitigation come early, and the benefits of mitigation through

avoided costs of climate change come later. The costs of mitigation are defined

clearly enough to be assessed through standard general equilibrium modelling.

The benefits of mitigation come in four types, only one of which is measurable

with standard modelling techniques. This chapter applies the decision-making

framework of Chapter 1 to the fundamental question before the Review. The

analysis is informed by the modelling undertaken jointly with the Australian Treasury

and independently by the Review.

11.1 The three global scenarios

This chapter analyses the three scenarios introduced in Chapter 4—the no mitigation

scenario, in which the world does not attempt to reduce greenhouse

gas emissions; and the 550 and 450 scenarios, which represent cooperative

global efforts to reduce emissions to varying degrees. To answer the question

of whether Australia should support, and play its full part in, a global mitigation

effort, the Review compared the costs and benefits of the no-mitigation and the

550 scenarios. To answer the question of how much mitigation Australia should

support the Review compared the 450 and 550 scenarios. What is compared,

through a mix of modelling and analysis, is the cost to Australia of participating in

a global agreement to mitigate climate change, and the costs of climate change

under the three scenarios.

In 2005, the atmospheric concentration of greenhouse gases was about

world, under the view of business-as-usual emissions presented in

Chapter 3, this would reach 550 ppm by 2030, 750 by 2050, 1000 by 2070, and

1600 by 2100.

The concentration of carbon dioxide (the main greenhouse gas) in this scenario

would reach 1000 ppm at 2100, compared to a band of natural variability of carbon

dioxide over many millennia of between 180 and 280 ppm, and 280 ppm in the

early years of modern economic growth in 1840.

In the 550 scenario, concentrations of greenhouse gases stop rising by around

of the level reached in the no-mitigation scenario, by the end of the century. In

the more stringent 450 scenario, given the current concentration, significant

the 22nd century.

Atmospheric concentrations of greenhouse gases are important primarily

because of their impact on global temperature. Table 11.1 shows the expected

increases in global temperature associated with each of the three scenarios, as

well as the temperature consistent with the highest climate sensitivity in the ‘likely’

range defined by the IPCC—that is, two-thirds probability of remaining within the

limits (IPCC 2007). In the absence of mitigation, in the median case, the world

century, or higher if the climate sensitivity were above its central estimate.

The 550 and 450 scenarios will limit median expectations of end-of-century

the central estimate for climate sensitivity, and stabilise global temperatures at

respectively, above 1990 levels is still within the likely range of the 550 and 450

scenarios (Table 11.1).

Table 11.1 Temperature increases above 1990 levels under the no-mitigation, 550

and 450 scenarios

2050 2100 2200

Best

estimate

Upper end

of likely

range

Best

estimate

Upper end

of likely

range

Best

estimate

Upper end

of likely

range

No

mitigation

2.3oC 2.9 oC 5.1o C 6.6 oC 8.3 o C 11.5oC

550 1.7o C 2.2 oC 2.0 oC 2.7oC 2.2 o C 3.1oC

450 1.6 oC 2.1o C 1.5 oC 2.1oC 1.1o C 1.7oC

Note: The ‘best estimate’ and ‘upper end of likely range’ temperature outcomes were calculated using

climate sensitivities of 3ºC and 4.5ºC respectively. There is a two-thirds probability of outcomes falling within

the likely range. Temperatures are derived from the MAGICC climate model (Wigley 2003). Temperature

outcomes beyond 2100 are calculated under the simplifying assumption that emissions levels reached

in each scenario in the year 2100 continue unchanged. They do not reflect an extension of the economic

analysis underlying these scenarios out to 2100, and are illustrative only. It is unlikely that emissions in the

reference case will stabilise abruptly in 2101 with no policies in place, and hence the temperatures shown

underestimate the likely warming outcomes if continued growth in emissions is assumed. 1990 temperatures

are about 0.5ºC above pre-industrial levels.

11.2 Comparing the costs of climate change

and mitigation

To understand the potential economic implications of climate change for Australia,

appropriate scientific and economic frameworks must be combined to estimate

impacts. This is not a trivial task. There is uncertainty in many aspects of climate

change science at the climate system, biophysical and impact assessment levels.

These compounding sources of uncertainty mean that quantifying the economic

impacts of both climate change and its mitigation is a difficult, and at times

speculative, task. The Stern Review (2007: 161) cautioned that ‘modelling the

overall impact of climate change is a formidable challenge, demanding caution in

interpreting results’. Moreover, modelling alone will not provide an answer to the

two questions posed at the beginning of this chapter. As explained in Chapter 1,

many of the costs of climate change cannot be modelled.

The framework set out in Chapter 1 distinguished between four types of costs

of climate change. The first type of cost (Type 1) has been measured through

a computable general equilibrium model, based on measured market impacts

of climate change in the median or ‘average’ cases suggested by the science.

That is the easiest part of the problem, but still involves the most complex long-term

modelling of the Australian economy ever undertaken. The requirement to

model changes in the structure of the Australian economy in a general equilibrium

framework to the end of the 21st century takes the models to the limits of their

capacities. For details on the combination of models used, see Box 11.1.

The Review’s innovation was to model the cost of imposing mitigation

policy alongside the benefits of the climate change avoided. This was done

as follows:

policy was developed jointly by the Review and the Australian

Treasury.

that entailed shocking the reference case to simulate a world of

unmitigated climate change.

a carbon constraint on the model, and (2) imposing ‘positive’ climate

shocks to simulate a lesser degree of climate change as a result of

successful global mitigation policy (that is, 550 ppm or 450 ppm).

This was a highly complex and technically path breaking process, which

required the Review to draw on a wide range of expertise and models

within individual sectors. The modelling of the expected impacts of climate

change by the Review included individual areas of impact, including

agriculture, human health and several aspects of infrastructure.

There is currently no single model that can capture the global,

national, regional and sectoral detail that was necessary for the Review’s

approach. As a result, the Review drew on a number of economic models to

determine the costs of climate change and the costs and benefits of climate

change mitigation for the Australian economy. The key models used were

the Monash Multi Regional Forecasting (MMRF) model, the Global Trade

and Environment Model (GTEM) and the Global Integrated Assessment

Model (GIAM). The focus of the modelling of both mitigation and climate

damage costs was on Australia. However, GTEM and GIAM—which

extends GTEM to model the interaction between the climate system and

the economy—were used to model global mitigation and climate change

damages. The outcomes of this modelling (in particular, the global carbon

price, and Australia’s emissions entitlement and trade impacts) were fed

into the MMRF model. MMRF was augmented by a series of scientific and

economic models, including for the electricity, transport, and land-use

change and forestry sectors. This allowed the determination of the costs of

unmitigated climate change and the net costs of mitigation to Australia.

GTEM was also used to derive mitigation costs for Australia, but, unlike

MMRF, without calculation of avoided climate change.

In interpreting the results of the modelling, it is important to bear

in mind that only one of the four types of costs of unmitigated climate

change, and therefore only one of the four types of benefits of avoided

climate change from mitigation, could be captured in the model.

Further details of the modelling and the climate change impact

work undertaken by the Review are available in a technical appendix

at <www.garnautreview.org.au>. The Review looks forward to

further empirical work and refinement by others of its modelling of

climate change.

The second type of cost (Type 2) involves market impacts in the median cases,

for which effects cannot be measured with sufficient precision and confidence to

feed into a computable general equilibrium model. By their nature, these costs and

benefits are not amenable to precise quantification. The Review formed judgments

about likely magnitudes, relative to the size of the impacts that were the focus

of the formal modelling. These assessments were applied in a transparent way

in adjustments to some of the model results, to remove the bias that would

otherwise be associated with the exclusion of obviously important market impacts

for which data were not available at the time of the modelling work. This is not

as good as modelling these costs within the general equilibrium framework would

have been if the data had been available, but it is clearly better than leaving them

out altogether.

The third type of cost (Type 3) is that associated with the chance that the

impacts through market processes turn out to be substantially more severe

than suggested in the median cases. Type 3 costs derive their importance from

the normal human aversion to risk in relation to severe outcomes, and from the

possibility that the bad end of the probability distribution includes outcomes that

are extremely damaging and in some cases catastrophic. Since the modelling

undertaken was not of a probabilistic nature, these ‘worst case’ impacts could not

be quantified.

The fourth type of cost (Type 4) involves services that Australians value,

but which do not derive their value from market processes. Examples include

deterioration of environmental amenity; loss of species and, more generally,

of biodiversity; and health and international development impacts that do not

necessarily have their effects through the imposition of monetary costs on

the Australian community. By definition, these costs cannot be included in the

modelling.

In contrast to three of the four types of climate change induced costs, all

mitigation costs have a market impact and so can be measured.

The other difference between mitigation costs and climate change costs is their

profile over time. The Review’s modelling of the effects of climate change ends in

2100. The long time frames and large structural shifts involved in climate change

analysis present considerable challenges for modelling the way the economy

is likely to respond. As in most economic models, the assumed behavioural

responses in models used by the Review are determined by parameters and data

that have been derived from recent history. Into the second half of this century

and beyond, the assumptions that must be made about economic parameters and

relationships become highly speculative. And yet all of the detailed assessments

of the economics of climate change indicate that the main costs of climate change,

and therefore the main benefits of mitigation, accrue in the 22nd and 23rd centuries

and beyond (Stern 2006; Nordhaus 2008; Cline 2004). Whereas the costs of

mitigation can be expected to stabilise this century, the costs of climate change

can be expected to accelerate over this century and into the next. Consideration

of the long-term costs and benefits must be a feature of any evaluation of the net

benefits of climate change mitigation policy.

Because of the importance of the non-quantified costs of climate change—

whether they are Type 3 or Type 4 costs this century, or any of the types of costs

beyond this century—a comparison of the modelled costs of mitigation and no

mitigation, or even of differing degrees of mitigation, can only contribute to any

comparison of two scenarios. What can be modelled are the gross costs of

mitigation (purely the costs of mitigation without any of the benefits of avoided

climate change) and the net costs of mitigation (the gross costs of mitigation minus

the modelled Type 1 impacts and the estimated Type 2 impacts). These costs of

mitigation (modelled at most out to 2100) need to be compared with judgments

concerning the non-quantified Type 3 and Type 4 benefits from mitigation this

century, and with the likely benefits (of all types) from climate change avoided in

the next century and subsequent centuries.

11.3 Modelling mitigation

The modelling of the two mitigation scenarios is based on costs associated with

Australia’s adherence to an emissions allocation, derived from an international

agreement commencing in 2013, to limit the concentration of greenhouse

the 550 scenario Australia’s emissions entitlement allocation in 2050 relative to

2000 emissions in absolute terms falls by 80 per cent, and in the 450 scenario

by 90 per cent. Australia’s emissions can exceed that allocation if we buy permits

from other countries at the global carbon price, which prevails across all sectors

within Australia, and across all countries around the world. The global carbon price

increases over time, along a path which ensures that emissions fall sufficiently

global terms, Australia’s emissions do not affect the global carbon price, which

is taken as a given in the domestic modelling. All revenues raised from sales of

carbon permits are distributed back to households. No payments are made to the

trade-exposed sector, as all industries around the world face the same emissions

penalty. No compensation payments are made to industry.

A critical determinant of the costs of mitigation is the assumptions made about

technologies that are or will become available to reduce emissions. Technological

development of any type is difficult to predict. When powerful incentives to

innovation are introduced to a market environment, however, human ingenuity

usually surprises on the upside. How will this ingenuity manifest itself in the face of

high emissions prices and increased public support on a global scale for research,

development and commercialisation of low-emissions technologies?

We do not know, but there are good reasons to believe that, if we get the

policy settings right over the next few years, the technological realities later in the

century will be greatly superior to those which, for good reason, are embodied in

As one alternative to the standard technology assumptions, the Review

modelled an enhanced technology future, embodying various assumptions of more

As another possibility for the future, the Review examined the implications of

the commercialisation of a backstop technology encouraged by high carbon prices,

atmosphere for recycling or permanent sequestration. In the Review’s modelling of

the backstop, deployment starts between 2050 and 2075.

The backstop technology has been introduced into the modelling in a stylised

manner. No single technology has been modelled. Rather, the backstop technology

is assumed to be available for all industries. In practice, the most likely backstop

technology will not be industry specific, but will, at a substantial cost, extract

carbon dioxide from the air for recycling or sequestration.

While the backstop and enhanced technologies are possibly complementary,

they are assumed to be alternatives in the modelling.

Which of these three visions of the technological future, or which combination

of them, or which alternative to all of them, defines the opportunities that evolve

through market processes over the years ahead will be revealed in due course.

Technological developments in response to a rising carbon price will have a large

mitigation strategies in future years.

Figure 11.1 shows the 450 and 550 global carbon prices in 2005 Australian

dollars under different assumptions about technology.

Figure 11.1 Australia’s carbon prices under different mitigation scenarios and

technological assumptions

Carbon prices, A$ (2005)

0

100

200

300

400

500

600

2013

2018

2023

2028

2033

2038

2043

2048

2053

2058

2063

2068

2073

2078

2083

2088

2093

2098

700

550 standard

550 backstop

550 enhanced

450 enhanced

450 standard

450 backstop

Note: The rising carbon price paths are derived in GTEM and implemented in MMRF, except for the prices

derived under the enhanced technology assumptions, which are implemented only in GTEM and reported

up to 2050. The 450 and 550 price paths move on to the horizontal backstop path when they reach about

$250/t CO2-e. The two arrows show the extent to which the enhanced technology assumptions reduce the

carbon price relative to the standard technology assumptions.

Taking into account the deadweight costs, negative and positive, of various

policies that are used to achieve reductions in emissions is just as difficult as

imagining the technological future. Nordhaus’s pioneering work (1994) emphasised

the reductions in deadweight costs that could come from replacing distorting forms

of taxation, such as income tax, by a carbon tax. Mitigation through a carbon tax—

no exemptions, no shielding of trade-exposed industries—had a positive economic

benefit, because the carbon tax was less economically distorting than the taxes

that it replaced. A similar result could be obtained by replacing distorting Australian

taxes with revenue from the competitive sale of permits from an emissions trading

scheme. However, a distorted Australian emissions trading scheme that diverts

management effort from commercial activities into applying pressure for political

preferment could have large negative deadweight costs.

The modelling has assumed no net transactions and other deadweight costs

of the mitigation regime. History will reveal whether this was an optimistic or

pessimistic assumption.

The costs of mitigation depend on who bears them. Generally, an increment of

money is judged to be more valuable to the poor than the rich. It follows that the

costs of mitigation are higher, and the optimal amount of mitigation effort lower, the

more the costs are carried by the poor. More mitigation is justified if compensation

for low-income Australian households is a major feature of the policy framework.

(Chapter 16 explores the distributional impacts of mitigation.) Similarly, more

global mitigation can be justified if low-income countries carry a low proportion

of the costs. Australia has a strong interest in the burden of mitigation being

borne equitably across countries and therefore disproportionately by developed

economies, as Australia’s terms of trade would be damaged most by any setback

to income growth in developing countries.

11.4 The decision to mitigate

Is global mitigation in Australia’s interests? To test the case for action, the

Review compared the no-mitigation and 550 scenarios, and compared the costs

of mitigation of climate change with the benefits of avoiding climate change (the

difference between the costs of climate change with and without mitigation).

The costs of mitigation and the benefits of avoiding some of the costs of climate

change are those associated with implementation of a 550 stabilisation strategy.

The case for mitigation rests on the large temperature increases that would

be a likely outcome—not a remote possibility—of the rapid emissions growth that

can be expected in the absence of mitigation. The updated, realistic projections of

emissions growth developed by the Review, combined with mainstream scientific

estimates of climate sensitivity, result in a best estimate of the no-mitigation

This would at best impose severe costs on the world and on Australia.

Type 1 costs: modelled expected market impacts of climate

change

The Review’s economic modelling focused on five key areas of impact (primary

production, human health, infrastructure, tropical cyclones and international trade).

In each of these areas, climate change shocks were imposed reflecting the best

(See Table 11.2.) The modelled market impacts of unmitigated climate change

relative to a world without climate change (the reference case of chapters 3 and 7)

are shown in Figure 11.2.

Figure 11.2 The modelled expected market costs (median case) for Australia of

unmitigated climate change, 2013 to 2100 (Type 1 costs only)

Costs (per cent) relative to reference case

0

1

2

3

4

5

6

7

8

9

10

2013

2018

2023

2028

2033

2038

2043

2048

2053

2058

2063

2068

2073

2078

2083

2088

2093

2098

Real wages Real GNP Real consumption Real GDP

Note: All variables in this figure and throughout this chapter are in 2005 prices.

The modelled costs of climate change rise over time. Household consumption

and GNP on the one hand, and GDP on the other, diverge through time due to

this occurs, a greater volume of exports is required to pay for the same volume of

imports. Since consumption tracks GNP closely, most of the results are reported

in terms of GNP.

Changes in labour demand are captured in large changes to wages, rather than

unemployment, as the wage rate moves to eliminate any short-run employment

effects from climate change. Unmitigated climate change causes real wages to

be around 12 per cent lower than they would otherwise have been. The fall in

real wages increases significantly in the second half of the century in response to

reduced demand for labour as a result of climate change.

Table 11.2 Assessing the market impacts of climate change

Sector Direct impact Modelled Risk Ability to adjust

or adapt

Comment Likely economic

consequence by

2100

Economy wide—

international trade

Changes to import prices Yes High – Commodity-specific shocks, but

methodology overlooks sectoral

dimensions of climate change.

High

Changes to world demand (commodity specific) Yes High –

Economy wide—

infrastructure

Impacts on commercial buildings—changes to

building codes and planning schemes

No High High Capital stock of dwellings (current

prices) in 2006–07 was around $1.3

trillion, 40% of total capital stocks

(ABS 2007a).

Medium

Accelerated degradation of buildings—maintenance

and repair costs

Yes High High Medium

Economy wide—

extreme events

(tropical cyclones,

storms/flooding,

bushfires)

Increased intensity of tropical cyclones—damage to

residential infrastructure and home contents

Yes High Medium The current average annual cost of

tropical cyclones is estimated at $266

million, a quarter of the cost of natural

disasters (BTE 2001).

Low

Increased intensity of tropical cyclones—damage to

commercial buildings and business interruption

No High Medium Low

Southward movement of tropical cyclones—

infrastructure and business interruption

No High Medium High uncertainty regarding southward

movement of tropical cyclones.

Low

Higher frequency of storm events (e.g. flooding from

non-cyclone events)—damage to infrastructure

No High Medium in short run

High in long run

Estimated average annual cost of

floods in Australia is $314 million (BTE

2001).

Low

Bushfires—infrastructure damage, crop loss,

emergency response

No High Low–medium Bushfires estimated to pose an annual

average cost of about $77 million (BTE

2001).

Low

Economy wide—

sea-level rise

Increase in sea levels of 0.59 m, impacts on coastal

settlements

Yes High High Assessment assumes there is no

significant sea-level rise this century.

Low

Economy wide—

human health

Heat-related stress, dengue fever and

gastroenteritis—impacts on productivity

Yes High High Assumes management and prevention

of health impacts.

Low

Other health impacts (productivity) No Low High Low

Agriculture Changes in dryland crop production due to changes

in temperature and CO2 concentrations

Yes High Medium All crops $19.5 billion (gross value of

commodities) (ABS 2007b).

High

Table 11.2 Assessing the market impacts of climate change (continued)

Sector Direct impact Modelled Risk Ability to adjust

or adapt

Comment Likely economic

consequence by

2100

Sheep, cattle, dairy—changes in carrying capacity

of pasture from CO2 concentrations, rainfall and

temperature

Yes High Medium High

Impacts on sheep and cattle from heat stress due to

temperature increases

No Medium Medium to high Low

Impacts on pigs and poultry from heat stress due to

temperature increases

No Low Medium Limited research on potential impacts. Low

Irrigated agriculture—reductions in water runoff Yes High Medium Adaptation through land-use change,

water conservation.

High

Fisheries Reduced yields due to changes in water

temperature

No Medium Low Potential for higher adaptive capacity

for aquaculture. In 2004–05, fisheries

contributed 0.14% of total GDP (ABS

2008a).

Low

Forestry Yields affected by water availability and CO2

concentrations

No Low Low Possibility of using other species. In

2004–05, forestry contributed 0.12%

of total GDP (ABS 2008a)

Low

Mining Slower growth in demand due to slower increase

in world income (relative to a world with no climate

change)

Yes High Low In 2006–07, the mining industry

contributed 7% of total GDP (ABS

2007a). In the reference case, this is

projected to be 10.2% in 2100.

Medium to high

Reduction in water availability No High High Low

The Garnaut Climate Change Review

256

Table 11.2 Assessing the market impacts of climate change (continued)

Sector Direct impact Modelled Risk Ability to adjust

or adapt

Comment Likely economic

consequence by

2100

Tourism International tourism affected by slower growth in

demand due to slower increase in world incomes

(relative to a world with no climate change)

Yes High Low Tourism as a share of GDP is 3.7%,

which equates to $38 billion. In

2005–06, this was 10% of exports of

goods and services (ABS 2008b).

Medium to high

(highly uncertain)

Reduction in international demand for Australian

tourism as a result of reduced natural amenity of

tourism products

No High Low Requires assumptions of changes

in preferences and relative amenity

versus absolute amenity.

High

Changes in domestic tourism as a result of reduced

amenity of tourism products

No Medium Low to medium Domestic tourism is worth $29 billion

(2.6% of GDP) (relative to international

tourism—$10 billion,0.9% of GDP).

(ABS 2008B).

Low

Government

(health)

Increased expenditure on prevention and treatment

for dengue virus, heat stress and gastroenteritis

Yes Medium High Public and private expenditure low to

moderate.

Low

Increased expenditure on prevention and treatment

for other health impacts (air pollution, mental

health etc.)

No Medium High Low

Government

(defence/aid)

Increase in defence and aid expenditure due to

geopolitical instability in neighbouring nations

No High Low Combined aid and defence budget

for interventions in Timor-Leste and

Solomon Islands, $900 million per

year. Foreign aid 2006–07, 0.3% GNI

(AusAID 2006). Defence expenditure

2006–07, $17 billion (ABS 2007c).

Medium to high

(highly uncertain)

Residential

dwellings

Building degradation and damage resulting from

temperature, rainfall, wind etc.

Changes to building codes and increased

maintenance and repair

Yes High High Costs of adaptation likely to be high. High

Impacts on buildings due to extreme events (e.g.

flooding)

No High High Costs of adaptation likely to be high. Medium to high

Transport Degradation of roads, bridges and rail due to

temperature and rainfall

No High High In 2004–05 maintaining and improving

the total road network cost $9 billion

(BITRE 2008).

Low

Table 11.2 Assessing the market impacts of climate change (continued)

Sector Direct impact Modelled Risk Ability to adjust

or adapt

Comment Likely economic

consequence by

2100

Ports Port productivity and infrastructure affected by

gradual sea-level rise and storms

Yes High High Low

Airports Impacts on infrastructure due to sea-level rise,

temperature and rainfall

No Low Medium Low

Water supply Decrease in rainfall reduces reliance on traditional

water supply for urban use and increases demand

for alternative water supply options

Yes High High High costs associated with adaptation

options.

Low to medium

Degradation of water supply infrastructure increases

maintenance costs

Yes High High Low

Electricity

transmission and

distribution

Degradation of infrastructure increases maintenance

costs

Productivity losses from blackouts due to severe

weather events

Yes High High Net capital expenditure for electricity

transmission and distribution was $6.2

billion in 2005–06 (ABS 2008e).

Low

Electricity

generation

Increased demand for electricity resulting from

greater use of air conditioners

No High High New generation costs. Net capital

expenditure for electricity generation

was $2.8 billion in 2005–06 (ABS

2008e).

Low to medium

Note: High economic consequence implies 0.5–1.5% per cent loss of annual GDP by the end of the 21st century; medium economic consequence, 0.1–0.5 per cent; and low, less

than 0.1 per cent.

The Garnaut Climate Change Review

258

The negative impacts of climate change on infrastructure have a significant

effect on Australia’s output and consumption of goods and services, and is

infrastructure impacts affect a wide range of assets, including commercial and

residential buildings, water supply and electricity infrastructure, and ports.

By the end of the century, another 40 per cent of Type 1 GNP costs of

climate change arise from the negative effect on our terms of trade. Coal demand

is significantly lower than in the reference case, primarily as a result of the

deceleration of global economic growth in response to climate change. The global

modelling (GIAM) suggests that global GDP is likely to fall by around 8 per cent

by 2100, with losses in developing countries likely to be higher than the global

average. This is important for Australia as, by 2100, developing countries in Asia

are projected to be overwhelmingly our major trading partners. The international

modelling shows that Australian terms of trade are more adversely affected than

those of any other country or region by climate change. By 2100, Australia’s terms

of trade are 3 per cent below the reference case, whereas for Japan, the European

Union and the United States they are about 1 per cent below, and for Canada

0.5 per cent above.

About 20 per cent of Type 1 GNP costs arise from the direct climate change

impacts on agriculture. The loss of agricultural productivity as a result of climate

change results in agriculture drawing more resources from the rest of the

economy in order to meet an assumed inelastic demand for agricultural products,

and to maintain production at levels determined by domestic and world demand

Agriculture is hit particularly hard by climate change. Agricultural activity is

reduced by more than 20 per cent relative to the reference case. These impacts

would be unevenly distributed. Without mitigation, the best estimate for the

Murray-Darling Basin is that by mid-century it would lose half of its annual irrigated

agricultural output (Chapter 6). By the end of the century, it would no longer be a

home to agriculture.

The other sector that is hit hard is mining. Output of this sector is projected

to decline by more than 13 per cent by 2100. This result is mainly driven by

the deceleration of global economic growth. Most coal produced in Australia is

exported. The international modelling implies that the world demand for coal falls

for coal.

The health-related impacts considered by the Review are estimated to have

relatively small market effects, though this does not take into account the intrinsic

value of the lives lost.

The economic effects of tropical cyclones, taken as a series of annualised

losses, are estimated to be small. While a single cyclone event has the potential to

create significant economic damage, particularly if it were to hit a population centre,

these events are, and are likely to remain, relatively infrequent. These results may

be underestimated, however, as it was not possible to consider either the impacts

of flooding associated with cyclones or the impacts that might be associated with

a southward shift in the genesis of tropical cyclones.

Estimating and incorporating Type 2 costs: non-modelled

expected market impacts

The second type of cost, Type 2, covering the expected market costs of the

median outcome of impacts for which data are too unreliable to feed into general

equilibrium modelling, are estimated at about 30 per cent the size of the Type 1

GNP costs (see Box 11.2). Taken together, Type 1 and Type 2 costs amount to

approximately 8 per cent of GDP, about 10 per cent of GNP and consumption, and

a higher percentage of and real wages.

Table 11.2 identifies some major expected (median case) market impacts of

climate change this century, indicating which ones were included in the

modelling and which were excluded, and provides a qualitative assessment

of their importance in terms of market effects. The following market

impacts were judged to be significant, but could not be included in the

modelling, mainly because of lack of data.

road network in Australia, the effects of climate change on building

infrastructure and roads and bridges that have not been estimated could

subtract an additional 0.8 and 0.25 percentage points respectively from

to cool buildings could also be a significant omission.

to climate variables. It is likely that climate change will also affect the

variability and predictability of the climate (especially rainfall). In the

absence of forecasts describing the level of future variability, it is difficult

to provide an estimate of the degree to which increased climate variability

would affect the economy.

has captured the impact of climate change on tourism through incomes

and relative prices, but not through the deterioration of environmental

assets. Some major environmental assets that are important for Australian

tourism are highly susceptible to climate change. These include the Great

Barrier Reef, south-western Australia (a biodiversity hotspot) and Kakadu

(see Chapter 6). International travel to Australia is projected to increase

substantially in the reference case as global incomes rise strongly over

the coming decades. This suggests that even small changes in demand

could have significant economic implications for Australia.

an increase in the capability and requirements of Australia’s defence force

and an increase in the level of Australia’s spending on emergency and

humanitarian aid abroad. Previous Australian interventions in small

neighbouring nations provide some indication of the potential size of

future defence costs that may arise from climate change. The combined

aid and defence budget for the five-year intervention in Timor-Leste has

exceeded $700 million per year. Australia’s intervention in Solomon

Islands is estimated to cost around $200 million per year (Wainwright

2005). This level of intervention is likely to continue until at least 2013.

Climate change could lead to the involvement of larger countries through

geopolitical pressures, and thus may lead to much higher spending than

would be indicated by recent history. A 10 per cent increase in defence

spending would be a cost of 0.2 per cent of GDP. Although extra defence

spending does not automatically lead to reduced GDP, the Review treats

it as a cost since it represents resources that would otherwise have been

available for productive use elsewhere.

To summarise, the omitted impacts on infrastructure and defence

alone could subtract an additional 1.2 per cent from GDP (or GNP, the

main modelling output reported) at the end of the century. The effects on

tourism, variability and predictability effects on agriculture, additional

impacts of geopolitical instability on Australia, and the range of other

possible impacts noted in Table 11.2 need to be added to this. The total

omitted market impacts could contribute an additional 1 to 2 percentage

points to the loss of GNP at the end of the century, taking the estimated

Type 2 loss to between 2.2 and 3.2 per cent. This would imply that the

modelling has captured 77 to 70 percentage points of the 2100 nomitigation

cost to GNP. Applying only the upper bound of this range

implies that Type 2 costs are about 30 per cent of the Type 1 costs of

climate change.

In comparative costings of the three scenarios analysed in this chapter,

it is assumed that this relationship holds not only for 2100 and the nomitigation

scenario, but for the entire century and for all three scenarios.

This approach is clearly based on a significant degree of judgment and

simplification. However, the inclusion of Type 2 costs is considered to

be crucial to an appropriate evaluation of the expected market effects

of climate change and the corresponding benefits from mitigation. The

Review considers these estimates to be conservative.

These combined end-of-century Type 1 and Type 2 costs are much higher than

estimates from earlier quantitative studies of the global costs of climate change

during the 21st century. Stern (2007), for example, found a reduction in global GDP

per capita as a result of climate change of only 2.9 per cent in 2100 after taking

into account all four of the categories of costs described above, two of which

(types 3 and 4) are excluded from the calculations in the preceding paragraph.

The earlier and larger costs of climate change in the Review’s study derive in

substantial part from the application of realistic, Platinum Age assessments of the

growth in emissions in the absence of mitigation (Chapter 3).

One main theme of the Review is that the accelerated growth of the

developing world, the Platinum Age, has not been factored into expectations of

emissions, concentrations or temperatures. This growth, centred on but now

extending well beyond China, is unprecedented, and likely to be sustained over a

considerable period.

The Fourth Assessment Report of the IPCC presented a range of best-estimate

scientific judgments of the IPCC, on a balance of probabilities, as a reasonable

source of scientific knowledge on climate change. But the economic analysis of

the IPCC rests on work from the 1990s, which the Review has shown to have

been overtaken by events. Chapter 3 shows that the IPCC’s SRES scenarios,

on which its projections of climate change impacts were based, systematically

underestimate the current and projected growth of emissions. Far from being

alarmist, it is simply realistic to accept the conclusion from analysis that, if the

the central range for temperature increases in the 21st century under business as

Costs of catastrophic climate change (Type 3)

The rapid growth in emissions associated with the Platinum Age has an unfortunate

consequence: in the absence of mitigation, it is making outcomes likely that were

once seen as having low probability. The economist Martin Weitzmann justifies

strong mitigation action on climate change on the basis of prevention of a possible

catastrophic outcome. A recent article (Weitzmann 2007: 18) specifies a 3 per cent

result of which will be:

a terra incognita biosphere within a hundred years whose mass species extinctions,

radical alterations of natural environments, and other extreme outdoor consequences

of a different planet will have been triggered by a geologically instantaneous

temperature change that is significantly larger than what separates us now from

past ice ages.

How much stronger, then, is the justification for mitigation when the probability

50 per cent (recall that the no-mitigation best estimate for temperature increase

a temperature increase once we move from the now-outdated SRES scenarios

on which Weitzmann bases his 3 per cent calculations, to the more recent and

realistic projections of emissions reported in Chapter 3.

The end-of-century temperature increase expected from the no-mitigation

scenario is above the estimated range of ‘tipping points’ for seven of the eight

The Garnaut Climate Change Review

262

catastrophic global events (enumerated in section 11.5) for which Lenton

et al. (2008) present such a range, and it is at the top end of the range for the

eighth.

That catastrophic events have become more likely does not make them

more amenable to modelling. Table 4.1 presented results from a survey of the

recent scientific literature. This indicated that, given the best estimate for climate

sensitivity, the triggering of a large-scale melt of the Greenland ice sheet under

temperatures expected by the end of century in a no-mitigation scenario would be a

sure thing. Given the uncertainties of when the melt would start, and when it would

translate into sea-level rises, the effect has been neither modelled nor included in

our Type 2 estimates of the costs of climate change. As the sea level rose over

a matter of centuries by 7 m, there would certainly be a large negative impact on

the world, and on Australia, through the risks of severe and possibly catastrophic

effects on non-market values and on the basis of median expectations of market

impacts. These would also be the conditions under which irreversible melting of

the west Antarctic ice sheet would be most likely to occur, so that the correlation

of risks increases the chance of severe outcomes.

Non-market (Type 4) costs

The non-market risks of climate change will be significant in a no-mitigation scenario.

As Table 4.1 shows, at the high levels of temperature increase expected under the

no-mitigation scenario, 88 per cent of species would be at risk of extinction, and

coral reefs as we know them would be destroyed. In the Australian context, under

the no-mitigation scenario, by halfway through the century the Great Barrier Reef

would be destroyed, and by the end of the century the Kakadu wetland system

would be inundated by sea water. Non-market impacts also include the greater

number of deaths due to hotter weather, the inconvenience of a greater number of

extremely hot days, and much higher bushfire risk.

While there are methods through which non-market impacts can be monetised,

the Review found it more useful simply to identify them, and to note that, as

incomes and consumption rise, as is anticipated in the no-mitigation scenario, the

relative value people assign to non-market costs and benefits will rise as well.

Costs beyond the 21st century

The lags and non-linearity of climate change impacts, even looking only at the

expected market impacts, is reflected in the Review’s modelling. The gradient of

the modelled market impacts of the unmitigated scenario (Figure 11.2) at 2100

indicates that the costs of unmitigated climate change would grow rapidly into the

22nd century. The estimated impacts in the unmitigated scenario increase threefold

from 2050 to 2075, and then threefold again from 2075 to 2100. This rate of

increase in damages far outstrips the projected rate of increase in temperatures.

It is obvious that if the analysis were continued into the 22nd century, estimated

market impacts from climate change would be dramatically higher than for the

latter decades of the 21st century.

Other studies of climate change show much higher costs in the next than in

the current century. Cline (2004) used a modified version of Nordhaus’s climate

change model going out to the year 2300. Cline’s emissions growth is lower than

in the Review’s modelling, but a scenario with a higher climate sensitivity yields

temperature outcomes close to the Review’s no-mitigation scenario at 2100,

climate change damages as a percentage of global GDP are 9 per cent by 2100,

about 25 per cent by 2200, and a remarkable 68 per cent by 2300.

The Stern Review attempted a more comprehensive assessment of global

climate change damages, including market as well as non-market impacts and a

probability distribution over a range of possible outcomes to 2200. Stern’s analysis

shows impacts on expected per capita consumption at 3 per cent at 2100, rising

to 14 per cent at 2200.

As discussed in Chapter 1, under reasonable assumptions, the present value

of a percentage point of Australian GNP in a century’s time is about as high as a

percentage point of GNP this century. The Review has not tried to model climate

change impacts beyond the 21st century. It is clear that they matter.

Summary of unmitigated climate change costs

There is a risk that temperature increases, and therefore all the impacts that

are related to temperature, will be much greater than anticipated in the standard

cases of the modelling because of positive feedback effects. These are difficult to

quantify, but they are real and potentially significant. Once temperature increases

above certain threshold points, massive carbon and methane stores on earth and

in the oceans may be destabilised, leading to much greater volumes of greenhouse

gas release from the natural sphere, and further temperature increases.

To summarise, temperature increases of the order of magnitude associated

the top of the likely band, and a smaller probability of a double-digit temperature

increase—would not lead to a marginal reduction in human welfare. Their impacts

on human civilisation and most ecosystems are likely to be catastrophic. As the

Center for Strategic and International Studies recently noted in its study of climate

change scenarios, this extent of climate change ‘would pose almost inconceivable

challenges as human society struggled to adapt… The collapse and chaos

associated with extreme climate change futures would destabilize virtually every

To point to the devastating impact of temperature increase for this century,

and of significant further increases next century, and to the possibility that such

increases would leave both global and Australian welfare at the end of this century

lower than at the start, is not to be alarmist. It is simply to recognise the reality of

rapid emissions growth, its likely continuation in the absence of climate change

mitigation, and the possibly catastrophic consequences of such large, rapid

temperature increases.

How much would it cost to greatly reduce the extensive climate change damage

outlined in the previous section? Figure 11.3 depicts the costs of mitigation up to

2050 under stabilisation at 550 ppm, as implemented in GTEM (Box 11.1). The

results are shown under both standard and enhanced assumptions concerning

After an initial modelled shock to GNP growth of around 0.8 percentage points

(a cost which in reality would be spread over several years), the gross costs

of mitigation as modelled in GTEM typically shave a bit above 0.1 per cent per

annum from GNP growth until after the halfway mark in the century under standard

technology assumptions and a bit below 0.1 per cent per annum under enhanced

technology assumptions. This can be seen as sacrifice of material consumption in

the early decades.

The gap between the standard and enhanced cases opens up in the second

half of the century. On average, annual GNP growth is 0.07 per cent faster in the

enhanced case than in the standard, and the economy actually grows marginally

faster in the enhanced case with mitigation than without. (The enhanced technology

case is further explored in chapters 21 and 23.)

Modelling the costs of mitigation in the second half of the century is more

complex, for two reasons. First, as discussed earlier, technological options become

more uncertain. It is unrealistic to expect that carbon prices will continue to rise

beyond many hundreds of dollars (as in Figure 11.1) without the development

of new technologies to offset emissions. Accordingly, long-run cost modelling is

best undertaken with the assumption that at some price a backstop technology

Figure 11.3 Change in annual Australian GNP growth (percentage points lost

or gained) due to gross mitigation costs under the 550 scenario

strategy compared to no mitigation, and under standard and enhanced

technology assumptions, 2013–50

Enhanced Standard

-0.8

-0.7

-0.6

-0.5

-0.4

-0.3

-0.2

-0.1

0.0

0.1

2010

Per cent

2015 2020 2025 2030 2035 2040 2045 2050

Note: Results are from GTEM (Box 11.1). For details on the standard and enhanced variants of GTEM, see

section 11.3.

develops, even if there is uncertainty about the price at which that technology will

develop and the precise form it will take.

Second, avoided expected climate damages become significant, so measures

of costs without them (such as modelled in GTEM) become less informative.

The net mitigation costs calculated in MMRF take into account both the gross

costs of mitigation and the benefits of avoided expected climate change market

impacts (the Type 1 costs). In addition, the net costs shown here make an

adjustment for the Type 2 costs of climate change, along the lines developed for

the no-mitigation scenario.

These net costs of mitigation as modelled and accounted for here are not meant

to represent the full benefits of mitigation, as they do not seek to capture the

Type 3 and Type 4 and post–21st century benefits of mitigation. They do, however,

provide an indication of the amount Australia would need to pay to have access to

the additional benefits of climate change mitigation.

Figure 11.4 shows the net cost of mitigation (including expected market costs

as well as benefits) for the 550 scenario, using the MMRF model with an extension

implemented by the Review to allow for a backstop technology to emerge

post 2050.

Figure 11.4 shows that, in the second half of this century, mitigation towards

the 550 reduction target adds to the growth rate of the economy, as, at the margin,

more new climate change damages are avoided than new mitigation costs added.

In fact, by the end of the century, GNP is higher than it would have been without

mitigation, even when all the costs and only the expected market benefits (avoided

costs types 1 and 2, but not types 3 and 4) of mitigation are taken into account.

Figure 11.4 Change in annual Australian GNP growth (percentage points lost or

gained) due to net mitigation costs under the 550 scenario compared to

no mitigation, 2013–2100

-0.8

-0.6

-0.4

-0.2

0.0

0.2

0.4

2010 2030 2050 2070 2090

net benefits to

GNP growth

net costs to

GNP growth

Per cent

Note: The figures reflect results modelled in MMRF (Box 11.1) adjusted to incorporate Type 2 costs using

the method described in Box 11.2.

Figure 11.5 tells this story by sector. At the aggregated industry level, the big

winner from mitigation is the agriculture and forestry sector, which in the second

half of this century is growing about 0.3 per cent faster on average every year,

as forestry expands and less climate change damage is imposed on agricultural

productivity. Mining does worse with mitigation in the first half of the century, but

better in the second as the terms of trade improve. Growth in manufacturing and

services is little affected by mitigation.

The terms of trade effects of mitigation are ambiguous. On the one hand,

reduced climate change damage improves the terms of trade; on the other, a

swing away from fossil fuels associated with global mitigation harms them. Under

standard technology assumptions, the net impact is a further worsening of the

terms of trade (from 3 per cent below the reference case in the no-mitigation

scenario to 4 per cent below in the 550 scenario in 2100). But under enhanced

technology assumptions, where clean coal technology is more competitive, the

terms of trade in 2100 are 2 per cent above the reference case. Likewise, the

backstop technology improves terms of trade to 1 per cent above the no-mitigation

scenario.

Figure 11.5 Change in Australian sectoral growth rates (percentage points lost or

gained) due to net mitigation costs under the 550 scenario compared to

no mitigation, 2013–2100

-0.4

Per cent

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

2010 2020 2030 2040 2050 2060 2070 2080 2090

Agriculture and forestry Mining Manufactures Services

Note: Sectoral growth rates are the growth in sectoral activities (value added), as modelled in MMRF. Only

Type 1 costs are shown here. Ten-year forward moving averages are used to smooth annual variability.

Growth rates are presented for aggregated sectors under the backstop technology case. Uncertainty

regarding the backstop technology was recognised by applying costs evenly across all sectors, and reducing

emissions from all sources. Any actual backstop technology would likely have a differential impact on

sectoral growth. It is therefore not possible to project sectoral changes with a high degree of certainty in the

second half of the century.

Despite the boost to growth in the second half of the century, the sacrifice

in the first half is substantial, even though the loss to the annual level of GNP is

fully recovered with a margin by the end of the century. Of course, stabilisation at

550 ppm does not eliminate all costs associated with climate change. Temperatures

associated risks. Nevertheless, the benefits that are purchased by the cost of the

550 strategy are substantial, and take several forms.

One is insurance against the effects of severe and possibly catastrophic

outcomes on material consumption during this century. Another is increased

protection against loss of non-market services this century. Yet another is

avoidance of all of the rapidly increasing costs throughout the 21st and into the

22nd century and beyond: the rapidly increasing negative impact on material

consumption under median outcomes (types 1 and 2); the risk of outcomes much

worse than the median expectations from the applied science (although throughout

and beyond the 21st century the median outcomes are more severe and possibly

catastrophic) (Type 3); and the impacts on non-market values (Type 4).

Figure 11.6 compares expected market damages from climate change under

temperatures that would be expected for a 550 scenario with those damages

associated with temperatures expected under a no-mitigation scenario (see

Figure 11.2). Climate damages under the 450 scenario are also shown for later

reference. This figure makes for an incomplete comparison for all the reasons

Figure 11.6 A comparison of the modelled expected market costs for Australia

of unmitigated and mitigated climate change up to 2100 (Type 1

costs only)

0

1

2

3

4

5

6

7

8

2010 2020 2030 2040 2050 2060 2070 2080 2090 2100

Per cent GNP lost, relative to the reference case

No mitigation 550 scenario 450 scenario

Notes: The graphs show the cost as a percentage of GNP of expected market damages from the level

of climate change associated with the three scenarios. These estimates are achieved by ‘shocking’ the

reference case with the differing levels of impact associated with the temperatures expected from the three

scenarios.

already cited, but it is a telling one. Market damages under the 550 scenario

flatten, stabilise and begin to decline relative to GNP by the end of the century, just

as market damages under the no-mitigation scenario start to accelerate markedly.

The choice is between a future with bounded expected climate change costs, but

still significant risks, and one with unbounded expected costs, a high probability

of severe outcomes and some chance of outcomes that most Australians would

consider to be catastrophic.

The rapid growth in global emissions is increasing the costs both of mitigation

(Gurria 2008) and of no mitigation. The costs of well-designed mitigation, substantial

as they are, would not end economic growth in Australia, its developing country

neighbours, or the global economy. Unmitigated climate change probably would.

11.5 How much mitigation?

Is it in Australia’s interest to support a global goal of limiting the concentration of

portion of the Review’s modelling went into weighing the relative benefits of

Australia’s participation in a 450 ppm and 550 ppm global climate change mitigation

agreement.

Figures 11.7 and 11.8 present the same comparison in terms of growth rates

as in Figures 11.3 and 11.4 above, but this time comparing 450 and 550 scenarios

rather than no-mitigation and 550.

Figure 11.7 Change in annual Australian GNP growth (percentage points lost

or gained) due to gross mitigation costs under the 450 compared

to the 550 scenario and under standard and enhanced technology

assumptions, 2013–50

-0.8

-0.7

-0.6

-0.5

-0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

2010 2015 2020 2025 2030 2035 2040 2045 2050

Per cent

Enhanced Standard

Note: Results are from GTEM (Box 11.1). For details on the standard and enhanced variants of GTEM, see

section 11.3.

In GTEM, there is an additional growth penalty when the 450 regime is

introduced (though again, this would be spread over several years), but after that

there is no real growth differential between the 450 and 550 scenarios, whichever

set of technological assumptions is made.

Net mitigation costs modelled in MMRF (Figure 11.8), with the backstop

technology as before, show greater volatility in the growth differential between the

450 and 550 mitigation scenarios over the full century than GTEM does for gross

mitigation costs in the first half of the century. Overall, however, the story is a

similar one. After the initial shock, there is, on average, no difference in the GNP

growth rates under the two scenarios over the course of the century. Sectoral

differences in growth rates under the two mitigation scenarios are relatively minor

under backstop technology assumptions. Gains to agriculture and manufacturing

are offset by losses to mining. Sectoral differences under different technology

assumptions are explored in chapters 20 to 22.

Figure 11.8 Change in annual Australian GNP growth (percentage points lost or

gained) due to net mitigation costs under the 450 compared to the 550

scenario, 2013–2100

-0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

2010 2030 2050 2070 2090

Per cent

Note: The figures reflect results modelled in MMRF (Box 11.1) adjusted to incorporate Type 2 costs as

per the method described in Box 11.2. Since annual differences in the 550 and 450 growth rates show

considerable volatility, they are smoothed in this graph by using a three-year forward moving average.

Table 11.3 summarises the cost differences between these two scenarios.

It presents results using the GTEM model for gross mitigation costs out to 2050

(where avoided climate damages are small), and the MMRF model (with the post-

2050 backstop) for net mitigation costs (costs net of Type 1 and Type 2 benefits).

It calculates the net present value of these costs using two discount rates

the Review considers appropriate, namely 1.4 and 2.7 per cent (as derived in

Chapter 1).

At the bottom of the range of discount rates considered (1.4 per cent), and

for both models, and both time periods and methods, the net present value

of the excess cost of the 450 over the 550 scenario (the ‘450 premium’) is

0.7 to 0.8 per cent of discounted GNP. At the top end of the range of discount

rates (2.7 per cent), the premium is in the range of 0.7 to 0.9 per cent of

discounted GNP.

Table 11.3 Net present cost of the 450 ppm and 550 ppm scenarios (in terms of

no-mitigation GNP) and the ‘450 premium’ to 2050 and 2100

Discount rate equals 1.4% 550 450 450 premium

Gross mitigation cost to 2050 (per cent)

GTEM standard 2.6 3.3 0.7

GTEM enhanced 1.9 2.6 0.7

Net mitigation cost to 2100 (per cent)

MMRF 3.2 4.0 0.8

Discount rate equals 2.7% 550 450 450 premium

Gross mitigation cost to 2050 (per cent)

GTEM standard 2.4 3.2 0.7

GTEM enhanced 1.8 2.5 0.7

Net mitigation cost to 2100 (per cent)

MMRF 3.3 4.2 0.9

Note: The figures give the discounted costs as a percentage of discounted GNP. The ‘450 premium’ is the

excess of the 450 ppm cost over the 550 ppm cost. Costs in GTEM are gross costs of mitigation; costs in

MMRF are net costs (gross costs net of Type 1 and Type 2 benefits). MMRF modelled results are adjusted

to incorporate Type 2 costs using the method described in Box 11.2.

Figure 11.6 helps to explain why the 450 scenario is always the more expensive

one in terms of modelled results. Expected climate change damages are less in

the 450 scenario than in the 550 scenario, but only by half a per cent of GNP. The

small expected market gain from the 450 scenario to 2100 is not in itself adequate

to justify the additional mitigation costs associated with it. Rather, the large

difference between 450 and 550 scenarios is in terms of additional Type 3 and

Type 4 avoided costs. What are the non-market (Type 4) and insurance (Type 3)

benefits of a 450 relative to a 550 outcome?

Differential avoidance of non-market climate change impacts

Neither strategy will lead to the complete avoidance of non-market climate

result in a 50 per cent decrease in the area of rainforests in North Queensland.

Nevertheless, the analysis suggests that the difference between a 450 and a 550

outcome could be of major significance for a range of environmental impacts.

For example:

• A 550 outcome would be expected to lead to the destruction of the

Great Barrier Reef and other coral reefs as we recognise them today. The

450 outcome would be expected to damage but not destroy these coral reefs.

Under a 550 scenario, the three-dimensional structure of the corals would be

largely gone and the system would instead be dominated by fleshy seaweed

and soft corals. At 450 ppm, the reef would still suffer—mass bleaching would

be twice as common as it is today—but its disappearance would be much less

likely (Table 6.2).

• The 550 ppm outcome would lead to a greater incidence of species extinction.

Under the expected temperature outcome from the 550 ppm scenario,

12 per cent of species are predicted to be at risk of extinction. This percentage

is reduced to almost 7 per cent under the 450 scenario (Table 4.1).

Differential insurance value of the 450 ppm and 550 ppm

scenarios

As important as these differential non-market impacts are, perhaps the decisive

advantage of the 450 scenario over the 550 is its insurance value. While neither

scenario would eliminate climate change risks, the 550 scenario would leave the

world, and Australia, open to larger risks of exceeding threshold temperature

values, even if these tipping points cannot be known in advance with certainty.

Lenton et al. (2008: 1786) identified nine important ‘tipping elements’ (‘subsystems

of the Earth system that are at least subcontinental in scale and can be switched—

under certain circumstances—into a qualitatively different state by small

perturbations’), and conducted a survey of experts to estimate the associated

temperature tipping points. For one of the nine (disruption of the Indian summer

monsoon), the tipping point could not be identified. For two—melting of the Arctic

summer sea ice and the Greenland ice sheet—the tipping point range was put at

sheet, disruption of the Atlantic thermohaline circulation, disruption of the El Niño

– Southern Oscillation, disruption of the Sahara/Sahel and West African monsoon,

dieback of the Amazon rainforest, and dieback of boreal forest.

define for climate sensitivity, but, as reported in Chapter 2, the IPCC notes that

‘[v]alues substantially higher than 4.5°C [which is at the upper end of the likely

range] cannot be excluded’ (IPCC 2007: 12). This means there would be a smaller

but still significant (say 10 per cent) probability that the 550 scenario could produce

century. This could also happen under the 450 scenario, but even at the top end of

To give just one comparison, according to the estimates in Table 4.1 the

temperature increase expected from the 550 scenario would give a 26 per cent

probability of initiating large-scale melt of the Greenland ice sheet. The temperature

increase expected from the 450 scenario would give a 10 per cent probability.

The large temperature changes associated with the higher end of the 550

scenario probability distribution, and the tipping points that this might breach, could

have far-reaching consequences. A Center for Strategic and International Studies

study of climate change scenarios (Campbell et al. 2007), referred to earlier in this

chapter, included a scenario of ‘severe climate change’, within which temperatures

by the Review, this is not far above the top end of the likely range by 2050 (see

Table 11.1). The study found that, under this scenario:

nations around the world will be overwhelmed by the scale of change and pernicious

challenges, such as pandemic disease. The internal cohesion of nations will be

under great stress … both as a result of a dramatic rise in migration and changes

in agricultural patterns and water availability. [There will be] flooding of coastal

communities around the world ...

Is it worth paying less than 1 per cent of GNP more through the 21st century

for the insurance value and the avoided market and non-market impacts of the

450 scenario?

This is a matter of judgment. Judgment will be affected greatly by the success of

mitigation regimes and progress in research, development and commercialisation

of low-emissions technologies over the years ahead. The Review thinks it likely

that, with a significant and rising carbon price and support for emergence of lowemissions

technologies, and confidence that the new policies are permanent

features of the economic environment, there will be technological progress in areas

not currently anticipated. Such developments would greatly favour a 450 outcome

over a 550 outcome.

Given the benefits after 2200 of stronger mitigation and the greater risks of

catastrophic consequences to the natural environment under the 550 scenario, the

Review judges that it is worth paying less than an additional 1 per cent of GNP as

a premium in order to achieve a 450 result.

on its own. Chapter 9 concluded that a credible agreement to secure the

450 scenario looked difficult for the international community as a whole in the

year or two immediately ahead. Chapter 12 discusses how Australia can most

effectively pursue support for a 450 global mitigation strategy in these inauspicious

circumstances.

Notes

1 The global emissions path is determined within the global modelling (GTEM) using a Hotellingstyle

carbon price function. The start price is fixed and then increases over time at the

prescribed interest rate of 4 per cent. The real interest rate is assumed to be 2 per cent. This

is adjusted upwards by a 2 per cent risk premium. This approach provides a proxy for banking

and borrowing, and imitates an efficient intertemporal distribution of abatement effort. The

resulting emissions pathway is then used for international trading simulations. The concept

of a resource price rising with the interest rate comes from resource economics. Hotelling

(1931) demonstrated that profit from the optimal extraction of a finite mineral resource will

increase over time at the rate of interest. Since only a finite amount of greenhouse gases

can be released into the atmosphere prior to stabilisation, the optimal release of greenhouse

gases into the atmosphere over time is a problem similar to the optimal extraction of a finite

resource (Peck & Wan 1996).

2 The standard technology assumptions represent a best estimate of the cost, availability

and performance of technologies based on historical experience, current knowledge and

expected future trends. The standard scenario includes some technological cost reductions

through learning by doing and improvements in existing technologies and the emergence and

wide-scale deployment of some currently unproven technologies such as carbon capture and

storage, hot rocks (geothermal) and hydrogen cars. It does not, however, include a backstop

technology in any sector.

3 S pecifically, the enhanced scenario implemented in GTEM included the following

assumptions:

• Faster energy efficiency improvements of an extra 1 per cent annually from 2013 to

2030, an extra 0.5 per cent from 2031 to 2040 and no extra improvements thereafter.

• More effective carbon capture and storage in response to higher carbon prices. The

share of combustion CO2 captured increases from 90 per cent to 99 per cent as the

permit price rises from zero to $140/t CO2-e.

• Faster learning by doing for electricity and transport technologies by increasing the

parameter for the learning functions by 50 per cent relative to the standard assumptions

over the whole simulation period.

• Non-combustion agricultural emissions are eliminated when the carbon price exceeds

$250/t CO2-e.

4 For technical reasons, it was necessary for the Review to use different global average

temperature changes for the assessment of domestic impacts than for the assessment of

international climate change impacts. For the Australian impact analysis, median rainfall and

local temperature outcomes are assumed in response to an average global temperature

change of 4.5oC degrees by 2100 (above 1990 levels). This temperature change is based

on the A1F1 SRES (see Chapter 3). This temperature change differs from the temperature

estimated based on the Garnaut–Treasury global emissions profile used in the economic

modelling. This emissions profile gives an increase in global average temperature of 5.1oC

degrees by 2100 (above 1990 levels). These differences could not be avoided in the time

frames available for the Review.

5 The terms of trade describe the ratio of export to import prices.

6 This decomposition is obtained by running each of the five shocks separately. Due to

interactive general-equilibrium effects, the decomposition is not exact.

7 The extent to which imports replace domestic food production is limited. Two factors are

influential here. First, growth in developing countries, combined with land constraints,

exacerbated by climate change, is likely to result in increases in the cost of food produced

overseas over the next 100 years. Second, while the Review has not undertaken detailed

modelling to estimate the impacts that climate change may have on the cost of food

production in the rest of the world, any change is presumed to influence the availability of

food exports to Australia.

8 Caution needs to be exercised in interpreting changes to world demand. A 23 per cent

decline implies that, with prices fixed, exports will decline by 23 per cent. However, prices

are not fixed in MMRF. With a typical export price elasticity of around 5, small changes to

prices will change the export results.

9 An increase in the cost of constructing and maintaining buildings is equivalent to a

productivity loss since more capital inputs per unit of output would be required. If buildings

make up around 40 per cent of capital stocks, and capital incomes make up approximately

40 per cent of total income, then a 5 per cent reduction in productivity of the building stock

would be expected to reduce GDP by approximately 0.8 per cent. In 2004–05, the cost

of maintaining and improving the road network was $9 billion (Table 11.1). If the cost of

maintaining the road network were to increase by 25 per cent, GDP might be reduced by

around 0.25 per cent.

10 The report was prepared by, among others, former CIA Director James Woolsey, former

Chief of Staff of the President John Podesta, former National Security Advisor to the Vice

President Leon Fuerth, Pew Center Senior Scientist Jay Gulledge, and former Deputy

Assistant Secretary of Defence for Asia and the Pacific Kurt Campbell.

11 Due to the modelling procedures followed, and the different models employed by the Review,

the emissions entitlement allocations for Australia modelled in GTEM were slightly different

to those presented in Chapter 9 and modelled in MMRF: over the century, a 1 percentage

point greater reduction from the reference case for the 550 scenario (82 compared to 81

per cent), and a 2 percentage point greater reduction for the 450 scenario (88 compared to

86 per cent). The Chapter 9 (and MMRF) allocations are more generous early on, and less

generous later than the ones in GTEM. Simulations suggested very little difference in cost

over time. A partial equilibrium adjustment to account for the differential purchase of emission

permits made no discernible difference in the growth rates over time. There is, however,

a difference in the first year, where the GTEM allocation profile exaggerates the negative

shock on GNP. See the technical appendix on the modelling at <www.garnautreview.org.au>

for further discussion.

References

ABS (Australian Bureau of Statistics) 2007a, Australian System of National Accounts, 2006–07,

cat. no. 5204.0, ABS, Canberra.

ABS 2007b, Australian Farming in Brief, 2007, cat. no. 7106.0, ABS, Canberra.

ABS 2007c, Government Finance Statistics, cat. no. 5512.0, ABS, Canberra.

ABS 2008a, Australian National Accounts: Input – Output, 2004–05 (preliminary), cat. no.

5209.0.55.001, ABS, Canberra.

ABS 2008b, Australian National Accounts: Tourism Satellite Account, 2006–07, cat. no. 5249.0,

ABS, Canberra.

ABS 2008c, Electricity, Gas, Water and Waste Services, Australia, 2006–07, cat. no. 8226.0,

ABS, Canberra.

AusAID 2006, Australia’s Overseas Aid Program Budget 2006–07, Commonwealth of Australia,

Canberra.

BITRE (Bureau of Infrastructure, Transport and Regional Economics) 2008, ‘Information Sheet

27: Public road-related expenditure and revenue in Australia (2008 update)’, Commonwealth

of Australia, Canberra.

BTE (Bureau of Transport Economics) 2001, Economic Costs of Natural Disasters in Australia,

Report 103, BTE, Canberra.

Campbell, K., Gulledge, J., McNeill, J., Podesta, J., Ogden, P., Fuerth, L., Woolsey, J., Lennon,

A., Smith, J., Weitz, R. & Mix, D. 2007, The Age of Consequences: The foreign policy

and national security implications of climate change, Center for Strategic and International

Studies, <www.csis.org/media/csis/pubs/071105_ageofconsequences.pdf>.

Cline, W.R. 2004, ‘Climate Change’, in B. Lomborg (ed.), Global Crises, Global Solutions,

Cambridge University Press, pp. 13–43.

Dasgupta, P. 2007. ‘Commentary: the Stern Review’s economics of climate change’, National

Institute Economic Review 199: 4–7.

Gurría, A. 2008, ‘Energy, environment, climate change: unlocking the potential for innovation’,

keynote speech by the OECD Secretary-General, during the World Energy Council: Energy

Leaders Summit, London, 16 September, <www.oecd.org/document/18/0,3343,en_2649_

33717_41329298_1_1_1_1,00.html>.

Hotelling, H. 1931, ‘The economics of exhaustible resources’, Journal of Political Economy

39(2): 137–75.

IPCC (Intergovernmental Panel on Climate Change) 2007, Climate Change 2007: The physical

science basis. Contribution of Working Group I to the Fourth Assessment Report of the

Intergovernmental Panel on Climate Change, S. Solomon, D. Qin, M. Manning, Z. Chen, M.

Marquis, K.B. Averyt, M. Tignor & H.L. Miller (eds), Cambridge University Press, Cambridge

and New York.

Lenton, T.M., Held, H., Kriegler, E., Hall, J.W., Lucht, W., Rahmstorf, S. & Schellnhuber, H.J.

2008, ‘Tipping elements in the Earth’s climate system’, Proceedings of the National Academy

of Sciences of the USA 105(6): 1786–93.

Nordhaus, W. 1994, Managing the Global Commons: The economics of climate change, MIT

Press, Boston.

Nordhaus, W. 2008, The Challenge of Global Warming: Economic models and environmental

policy, Yale University, New Haven, Connecticut.

Peck, S. & Wan, Y. 1996, ‘Analytic Solutions of Simple greenhouse Gas Emission Models’,

Chapter 6 in E.C. Van Ierland & K. Gorka (eds), Economics of Atmospheric Pollution,

Springer Verlag, New York.

Stern, N. 2007, The Economics of Climate Change: The Stern Review, Cambridge University

Press, Cambridge.

Wainwright, E. 2005, ‘How is RAMSI faring? Progress, challenges, and lessons learned’,

Australian Strategic Policy Institute Insight 14.

Weitzmann, M. 2007 ‘The Stern Review of the Economics of Climate Change’, book review,

Journal of Economic Literature, <www.economics.harvard.edu/faculty/weitzman/files/

JELSternReport.pdf>.

Wigley, T.M.L. 2003, MAGICC/SCENGEN 4.1: Technical Manual, National Center for

Atmospheric Research, Colorado.