The economics of stabilisation (página 3)

There are a number of possible emissions trajectories that can achieve any given stabilisation goal. For example, emissions can peak early and decline gradually, or peak later and decline more rapidly. This is demonstrated in Figure 8.2, which shows illustrative pathways to stabilisation at 550 ppm CO2e.

Figure 8.2 Illustrative emissions paths to stabilise at 550 ppm CO2e.

The figure below shows six illustrative paths to stabilisation at 550 ppm CO2e. The rates of emissions cuts are given in the legend and are the maximum 10-year average rate (see Table 8.2). The figure shows that delaying emissions cuts (shifting the peak to the right) means that emissions must be reduced more rapidly to achieve the same stabilisation goal. The rate of emissions cuts is also very sensitive to the height of the peak. For example, if emissions peak at 48 GtCO2rather than 52 GtCO2 in 2020, the rate of cuts is reduced from 2.5%/yr to 1.5%/yr.

70

60

50

40

30

2015 High Peak – 1.0%/yr 2020 High Peak – 2.5%/yr 2030 High Peak – 4.0%/yr 2040 High Peak – 4.5%/yr (overshoot) 2020 Low Peak – 1.5%/yr 2030 Low Peak – 2.5%/yr 2040 Low Peak – 3.0%/yr Source: Generated with the SiMCaP EQW model (Meinshausen et al. 2006) 10 For example, Meinshausen (2006) and US CCSP (2006) STERN REVIEW: The Economics of Climate Change 199

Table 8.2 PART III: The economics of stabilisation

Illustrative Emissions Paths to Stabilisation The table below explores the sensitivity of rates of emissions reductions to the stabilisation level and timing and size of the peak in global emissions. These results were generated using the SiMCaP EQW model, as used in Meinshausen et al. (2006), and should be treated as indicative of the scale of emissions reductions required.

The table covers three stabilisation levels and a range of peak emissions dates from 2010 to 2040. The centre column shows the implied rate of global emissions reductions. The value shown is the maximum 10-year average rate. As shown in Figure 8.2, the rate of emissions reductions accelerates after the peak and then slows in the second half of the century. The maximum 10-year average rate is typically required in the 5 – 10 years following the peak in global emissions. The range of rates shown in each cell is important: the lower bound illustrates the rate for a low peak in global emissions (that is, action is taken to slow the rate of emissions growth prior to the peak) – in this example, these trajectories peak at not more than 10% above current levels; the upper bound assumes no substantial action prior to the peak (note that emissions in this case are still below IEA projections – see Figure 8.4).

The paths use the assumption of a maximum 10%/yr reduction rate. A symbol “-“ indicates that stabilisation is not possible given this assumption. Grey italic figures indicate overshooting. The overshoots are numbered in brackets ‘[ ]’ and details given below the table. Notes: overshoots: [1] to 520 ppm, [2] to 550 ppm, [3] to 600 ppm. 2005 emissions taken as 45 GtCO2e/yr. Source: Generated with the SiMCaP EQW model and averaged over multiple scenarios (Meinshausen et al. 2006)

The height of the peak is also crucial. If early action is taken to substantially slow the growth in emissions prior to the peak, this will significantly reduce the required rate of reductions following the peak. For example, in Figure 8.2, if action is taken to ensure that emissions peak at only 7% higher than current levels, rather than 15% higher in 2020 to achieve stabilisation at 550 ppm CO2e, the rate of reductions required after 2020 is almost halved.

If the required rate of emissions cuts is not achieved, the stock of greenhouse gases will overshoot the target level. Depending on the size of the overshoot, it could take at least a century to reduce concentrations back to a target level (discussed later in Box 8.2).

Table 8.2 gives examples of implied reduction rates for stabilisation levels between 550 ppm and 450 ppm CO2e. A higher stabilisation level would require weaker cuts. For example, to stabilise at 650 ppm CO2e, emissions could be around 20% above current levels by 2050, and 35% below current levels by 2100. As described in section 8.2, this higher stabilisation level would mean a much greater chance of exceeding high levels of warming and therefore, a higher risk of more adverse and unacceptable outcomes. The paths shown in Table 8.2 are based on one model and should be treated as indicative. Despite this, they provide a crucial STERN REVIEW: The Economics of Climate Change 200

PART III: The economics of stabilisation

illustration of the scale of the challenge. Further research is required to explore the uncertainties and inform more detailed strategies on future emissions paths.

To stabilise at 550 ppm CO2e, global emissions would need to peak in the next 10 – 20 years and then fall by around 1 – 3% per year. Depending on the exact trajectory taken, global emissions would need to be around 25% lower than current levels by 2050, or around 30-35 GtCO2.

If global emissions peak by 2015, then a reduction rate of 1% per year should be sufficient to achieve stabilisation at 550 ppm CO2e (Table 8.2). This would mean immediate, substantial and global action to prepare for this transition. Given the current trajectory of emissions and inertia in the global economy, such an early peak in emissions looks very difficult. But the longer the peak is delayed, the faster emissions will have to fall afterwards. For a delay of 15 years in the peak, the rate of reduction must more than double, from 1% to between 2.5% and 4.0% per year, where the lower value assumes a lower peak in emissions (see Figure 8.2). Given that it is likely to be difficult to reduce emissions faster than around 3% per year (discussed in the following section), this emphasises the importance of urgent action now to slow the growth of global emissions, and therefore lower the peak.

A further 10-year delay would mean a reduction rate of at least 3% per year, assuming that action is taken to substantially slow emissions growth; if emissions growth is not slowed significantly, stabilisation at 550 ppm CO2e may become unattainable without overshooting.

Stabilising at 450 ppm CO2e or below, without overshooting, is likely to be very costly because it would require around 7% per year emission reductions.

Table 8.2 illustrates that even if emissions peaked in 2010, they would have to fall by around 7% per year to stabilise at 450 ppm CO2e without overshooting11. This would take annual emissions to 70% below current levels, or around 13 GtCO2 by 2050. This is an extremely rapid rate, which is likely to be very costly. For example, 13GtCO2 is roughly equivalent to the annual emissions from agriculture and transport alone today.

Achieving this could mean, for example, a rapid and complete decarbonisation of non- transport energy emissions, halting deforestation and substantial intensification of sequestration activities. The achievability of stabilisation levels is discussed in more detail in the following sections and in chapter 9.

Allowing the stock to peak at 500 ppm CO2e before stabilising at 450 ppm (an ‘overshooting’ path to stabilisation, Box 8.2) would decrease the required annual reduction rate from around 7% to 3%, if emissions were to peak in 2010. However, overshooting paths, in general, involve greater risks.

An overshooting path to any stabilisation level would lead to greater impacts, as the world would experience a century or more of temperatures close to those expected for the peak level (discussed later in Figure 8.3). Given the large number of unknowns in the climate system, for example, threshold points and irreversible changes, overshooting is potentially high risk. In addition, if natural carbon absorption were to weaken as projected, it might be impossible to reduce concentrations on timescales less than a few centuries.

Given the extreme rates of emissions cuts required to stabilise at 450 ppm CO2e, in this case overshooting may be unavoidable. The risks involved in overshooting can be reduced through minimising the size of the overshoot by taking substantial, early action to cut emissions. 11 An atmospheric greenhouse gas level of 450 ppm is less than 10 years away, given that concentrations are rising at 2.5 ppm per year (chapter 3). However, in the scenarios outlined in Table 8.1, aerosol cooling temporarily offsets some of the increase in greenhouse gases, giving more time to stabilise. This effect is illustrated in Box 8.2. STERN REVIEW: The Economics of Climate Change 201

Greenhouse Gas Level (CO2 equivalent ppm) PART III: The economics of stabilisation

Box 8.2 Overshooting paths to stabilisation

The figure below illustrates an overshooting path to stabilisation at 450 ppm CO2e (or 400 ppm CO2 only) – this is characterised by greenhouse gas levels peaking above the stabilisation goal and then reducing over a period of at least a century.

The light blue line shows the level of all Kyoto greenhouse gases in CO2e (the Review definition) and the red line shows the level of carbon dioxide alone. The dark blue line shows a third measure of greenhouse gas level that includes aerosols and tropospheric ozone. This is the measure used in the Meinshausen et al. trajectories shown in this chapter. The gap between the two blue lines in the early period is mainly due to the cooling effect of aerosols. Critically, by 2050 the lines converge as it is assumed that aerosol emissions diminish.

550

500

450

400

350 1990 2010 2030 2050 2070 2090 2110 2130 2150 2170 2190 Source: Generated with the SiMCaP EQW model (Meinshausen et al. 2006) 8.6 Timing of Emissions Reductions Pathways involving a late peak in emissions may effectively rule out lower stabilisation trajectories and give less margin for error, making the world more vulnerable to unforeseen changes in the Earth’s system.

Early abatement paths offer the option to switch to a lower emissions path if at a later date the world decides this is desirable. This might occur for example, if natural carbon absorption weakened considerably (section 8.3) or the damages associated with a stabilisation goal were found to be greater than originally thought. Similarly, aiming for a lower stabilisation trajectory may be a sensible hedging strategy, as it is easier to adjust upwards to a higher trajectory than downwards to a lower one.

Late abatement trajectories carry higher risks in terms of climate impacts; overshooting stabilisation paths incur particularly high risks.

The impacts of climate change are not only dependent on the final stabilisation level, but also the path to stabilisation. Figure 8.3 shows that if emissions are accumulated more rapidly, this will lead to a more rapid rise in global temperatures. Figure 8.3 demonstrates the point made in the last section, that overshooting paths lead to particularly high risks, as temperatures rise more rapidly and to a higher level than if the target were approached from below.

Early abatement may imply lower long-term costs through limiting the accumulation of carbon-intensive capital stock in the short term.

Delaying action risks getting ‘locked into’ long-lived high carbon technologies. It is crucial to invest early in low carbon technologies. Technology policies are discussed in chapter 15. STERN REVIEW: The Economics of Climate Change 202

Stock of Greenhouse Gases (CO2e) Temperature Change (deg C) 450

400

350 600

550

500 PART III: The economics of stabilisation

Figure 8.3 Implications of Early versus Late Abatement

The figure below is an illustrative example of the rate of change in (a) the stock of greenhouse gases and (b) global mean temperatures, for a set of slow (SC, black), rapid (RC, blue) and overshooting (OS, red) paths to stabilisation at 500 ppm CO2e.

On the slow paths, emissions cuts begin early and progress at a gradual pace, leading to a gradual increase in greenhouse gas concentrations and therefore, temperatures. On the rapid paths, reductions are delayed, requiring stronger emissions cuts later on. This leads to a more rapid increase in temperature as emissions are accumulated more rapidly early on. The overshooting path has even later action, causing concentrations and temperatures to rise rapidly, as well as peaking at a higher level before falling to the stabilisation level.

The higher rate of temperature rise associated with the delayed action paths (RC and OS) would increase the risk of more severe impacts. Temperatures associated with the overshooting path rise at more than twice the rate of the slow path (more than 0.2°C/decade) for around 80 years and rise to a level around 0.5°C higher. Many systems are sensitive to the rate of temperature increase, most notably ecosystems, which may be unable to adapt to such high rates of temperature change.

650 2000 2050 2100 2150 2200 Year OS500 RC500 SC500 (a) 1 0.8 0.6 0.4 0.2 0 2 1.8 1.6 1.4 1.2 2000 2050 2100 2150 2200 2250 2300 OS500

RC500 SC500 (b) Year Source: redrawn from O’Neill and Oppenheimer (2004). The temperature calculations assume a climate sensitivity of 2.5°C (see chapter 1), giving an eventual warming of 2.1°C relative to pre-industrial.

Paths requiring very rapid emissions cuts are unlikely to be economically viable.

To meet any given stabilisation level, a late peak in emissions implies relatively rapid cuts in annual emissions over a sustained period thereafter. However, there is likely to be a maximum practical rate at which global emissions can be reduced. At the national level, there are examples of sustained emissions cuts of up to 1% per year associated with structural change in energy systems (Box 8.3). One is the UK ‘dash for gas’; a second is France, which, STERN REVIEW: The Economics of Climate Change 203

PART III: The economics of stabilisation

by switching to a nuclear power-based economy, saw energy-related emissions fall by almost 1% per year between 1977 and 2003, whilst maintaining strong economic growth.

However, cuts in emissions greater than this have historically been associated only with economic recession or upheaval, for example, the emissions reduction of 5.2% per year for a decade associated with the economic transition and strong reduction in output in the former Soviet Union. These magnitudes of cuts suggest it is likely to be very challenging to reduce emissions by more than a few percent per year while maintaining strong economic growth. Box 8.3 Historical reductions in national emissions Experience suggests it is difficult to secure emission cuts faster than about 1% per year except in instances of recession. Even when countries have adopted significant emission saving measures, national emissions often rose over the same period. •

•

•

•

• Nuclear power in France: In the late 1970s, France invested heavily in nuclear power. Nuclear generation capacity increased 40-fold between 1977 and 2003 and emissions from the electricity and heat sector fell by 6% per year, against a background 125% increase in electricity demand. The reduction in total fossil fuel related emissions over the same period was less significant (0.6% per year) because of growth in other sectors.

Brazil’s biofuels: Brazil scaled up the share of biofuels in total road transport fuel from 1% to 25% from 1975 to 2002. This had the effect of slowing, but not reversing, the growth of road transport emissions, which rose by 2.8% per year with biofuels, but would otherwise have risen at around 3.6% per year. Total fossil fuel related emissions from Brazil rose by 3.1% pa over the same period.

Forest restoration in China: China embarked on a series of measures to reduce deforestation and increase reforestation from the 1980s, with the aim of restoring forests and the environmental benefits they entail. Between 1990 and 2000 forested land increased by 18m hectares from 16% to 18% of total land area12. Despite cuts in land use emissions of 29% per year between 1990 and 200013, total GHG emissions rose by 2.2% over the same period.

UK ‘Dash for Gas’: An increase in coal prices in the 1990s relative to gas encouraged a switch away from coal towards gas in power generation. Total GHG emissions fell by an average of 1% per year between 1990 and 2000.

Recession in Former USSR: The economic transition and the associated downturn during the period 1989 to 1998 saw fossil fuel related emissions fall by an average of 5.2% per year.

Source for emission figures: WRI (2006) and IEA (2006).

The key reason for the difficulty in sustaining a rapid rate of annual emissions cuts is inertia in the economy. This has three main sources: •

•

12 13 First, capital stock lasts a number of years and for the duration it is in place, it locks the economy into a particular emissions pathway, as early capital stock retirement is likely to be costly. The extent and impact of this is illustrated in Box 8.3.

Second, developing new lower emissions technology tends to be a slow process, because it takes time to learn about and develop new technologies. This is discussed in more detail in Chapter 9.

Zhu, Taylor, Feng (2004) Chapter 25 notes that some of this gain was offset by increased timber imports from outside China. STERN REVIEW: The Economics of Climate Change 204

• PART III: The economics of stabilisation

Third, it takes time to change habits, preferences and institutional structures in favour of low-carbon alternatives. Chapter 15 discusses the importance of policy in shifting these.

These limits to the economically feasible speed of adjustment constrain the range of feasible stabilisation trajectories. Box 8.4 The implications for mitigation policy of long-lived capital stock Power generation infrastructure typically has a very long lifespan, as does much energy-using capital stock. Examples are given below. Source: World Business Council for Sustainable Development (2004) and IPCC (1999).

This means that once an investment is made, it can last for decades. A high-carbon or low- efficiency piece of capital stock will tend to lock the economy into a high emissions pathway. The only options are then early retirement of capital stock, which is usually uneconomic; or “retrofitting” cleaner technologies, which is invariably more expensive than building them in from the start. This highlights the need for policy to recognise the importance of capital stock replacement cycles, particularly at key moments, such as the next two decades when a large volume of the world’s energy generation infrastructure is being built or replaced. Missing these opportunities will make future mitigation efforts much more difficult and expensive. 8.7 The Scale of the Challenge Stabilisation at 550 ppm CO2e requires emissions to peak in the next 10-20 years, and to decline at a substantial rate thereafter. Stabilisation at 450 ppm CO2e requires even more urgent and strong action. But global emissions are currently on a rapidly rising trajectory, and under “business as usual” (BAU) will continue to rise for decades to come. The “mitigation gap” describes the difference between these divergent pathways.

To achieve stabilisation between 450 and 550 ppm CO2e, the mitigation gap between BAU and the emissions path ranges from around 50 – 70 GtCO2e per year by 2050.

Figure 8.4 plots expected trends in BAU emissions14 against emission pathways for stabilisation levels in the range 450 to 550 ppm CO2e. The exact size of the mitigation gap depends on assumptions on BAU trajectories, and the stabilisation level chosen. In this example, it ranges from around 50 to 70 GtCO2e in 2050 to stabilise at 450 – 550 ppm CO2e. For comparison, total global emissions are currently around 45 GtCO2e per year.

Another way to express the scale of the challenge is to look at how the relationship needs to change between emissions and the GDP and population (two of the key drivers of emissions). To meet a 550 ppm CO2e stabilisation pathway, global average emissions per capita need to fall to half of current levels, and emissions per unit of GDP need to fall to one quarter of current levels by 2050. These are structural shifts on a major scale.

Stabilising greenhouse gas concentrations in the range 450 – 550 ppm CO2e will require substantial action from both developed and developing regions. 14 Business as usual (BAU) used in this chapter is described in chapter 7. STERN REVIEW: The Economics of Climate Change 205

Global Emissions (GtCO2e) PART III: The economics of stabilisation

Even if emissions from developed regions (defined in terms of Annex I countries15) could be reduced to zero in 2050, the rest of the world would still need to cut emissions by 40% from BAU to stabilise at 550 ppm CO2e. For 450 ppm CO2e, this rises to almost 80%. Emissions reductions in developed and developing countries are discussed further in Part VI.

Figure 8.4 BAU emissions and stabilisation trajectories for 450 – 550 ppm CO2e

The figure below shows illustrative pathways to stabilise greenhouse gas levels between 450 ppm and 550 ppm CO2e. The blue line shows a business as usual (BAU) trajectory. The size of the mitigation gap is demonstrated for 2050. To stabilise at 450 ppm CO2e (without overshooting) emissions must be more than 85% below BAU by 2050. Stabilisation at 550 ppm CO2e would require emissions to be reduced by 60 – 65% below BAU. Table 8.2 gives the reductions relative to 2005 levels.

100

90

80

70

60

50

40

30

20

10

0 2000 2010 2020 2030 2040 2050 2060 2070 2080 2090 2100 Stabilisation at 550 ppm CO2e or below is achievable, even with currently available technological options, and is consistent with economic growth.

An illustration of the extent and nature of technological change needed to make the transition to a low-carbon economy is provided by Socolow and Pacala (2004). They identify a ‘menu’ of options, each of which can deliver a distinct ‘wedge’ of savings of 3.7 GtCO2e (1 GtC) in 2055, or a cumulative saving of just over 90 GtCO2e (25 GtC) between 2005 and 2055. Each option involves technologies already commercially deployed somewhere in the world and no major technological breakthroughs are required. Some technologies are capable of delivering several wedges.

In their analysis, Socolow and Pacala only consider what effort is required to maintain carbon dioxide levels below 550 ppm (roughly equivalent to 610 – 690 ppm CO2e when other gases are included) by implementing seven of their wedges. This is demonstrated in Figure 8.5.

While the Socolow and Pacala analysis does not explicitly explore how to stabilise at between 450 and 550 ppm CO2e, it does provide a powerful illustration of the scale of action that would be required. It demonstrates that substantial emissions savings are achievable with currently available technologies and the importance of utilising a mix of options across several sectors. These conclusions are supported by many other studies undertaken by industry, governments and the scientific and engineering research community. 15 Annex I includes OECD, Russian Federation and Eastern European countries. This is discussed further in Part IV. STERN REVIEW: The Economics of Climate Change 206

Carbon Dioxide Emissions from Fossil Fuel Combustion and Cement Manufacture (GtCO2/yr) PART III: The economics of stabilisation

Figure 8.5 Socolow and Pacala’s “wedges”

Socolow and Pacala compare a simple mitigation path for 30

20

10

0 fossil fuel emissions with a projected BAU path. In the BAU path, fossil fuel CO2 emissions grow to around 50 GtCO2e in 2055. In the mitigation path, fossil fuel CO2 emissions remain constant at 25 GtCO2 until 2055. This mitigation trajectory should maintain carbon dioxide concentrations at around 550 ppm. The difference between BAU and the stabilisation trajectory is the stabilisation triangle. To demonstrate how these emissions savings can be achieved, this triangle is split into 7 equal wedges, each of which delivers 3.7 GtCO2e (1 GtC) saving in 2055. Socolow and Pacala give a menu of fifteen measures that could achieve one wedge using currently available technologies. However, some wedges cannot be used together as they would double count emission savings. The panel to the right gives four of these suggested measures.

60

50

40 2000 2010 2020 2030 2040 2050 Stabilisation Triangle

Continued Emissions from Fossil Fuel Combustion and Cement Manufacture Four abatement measures that could each deliver one ‘wedge’ (3.7 GtCO2e) in 2055.

1. Replace coal power with an extra 2 million 1-MW- peak windmills (50 times the current capacity) occupying 30*106 ha, on land or off shore.

2.Increase fuel economy for all cars from 30 to 60 mpg in 2055.

3. Cut carbon emissions by one-fourth in buildings and appliances in 2055. 4. Replace coal power with 700GW of nuclear (twice the current capacity). Source: Pacala and Socolow (2004)

To meet a stabilisation level of 550 ppm CO2e or below, a broad portfolio of measures would be required, with non-energy emissions being a very important part of the story.

Fossil fuel related emissions from the energy sector in total would need to be reduced to below the current 26 GtCO2 level, implying a very large cut from the BAU trajectory, which sees emissions more than doubling. This implies: •

•

• A reduction in demand for emissions-intensive goods and services, with both net reductions in demand, and efficiency improvements in key sectors including transport, industry, buildings, fossil fuel power generation. The electricity sector would have to be largely decarbonised by 2050, through a mixture of renewables, CCS and nuclear. The transport sector is still likely to be largely oil based by 2050, but efficiency gains will be needed to keep down growth; biofuels, and possibly some hydrogen or electric vehicles could have some impact. Aviation is unlikely to see technology breakthroughs, but there is potential for efficiency savings. A portfolio of technologies will be required to achieve this. Different studies make different assumptions on what the mix might be. This is discussed further in chapter 9.

Emissions from deforestation are large, but are expected to fall gradually over the next fifty years as forest resources are exhausted (Annex 7.F). With the right policies and enforcement mechanisms in place, the rate of deforestation could be reduced and substantial emissions cuts achieved. Together with policies on afforestation and reforestation, net emissions from land-use changes could be reduced to less than zero – that is, land-use change could strengthen natural carbon dioxide absorption. STERN REVIEW: The Economics of Climate Change 207

PART III: The economics of stabilisation

Emissions from agriculture will rise due to rising population and income, and by 2020 could be almost one third higher than their current levels of 5.7 GtCO2e. The implementation of measures to reduce agricultural emissions is difficult, but there is potential to slow the growth in emissions.

In practice the policy choices involved are complex; some actions are much more expensive than others, and there are also associated environmental and social impacts and constraints.

The following chapters discuss how to achieve cost-effective emissions cuts over the next few decades. These activities must be continued and intensified to maintain stabilisation in the long run. Over the next few centuries, section 8.3 showed that emissions would need to be brought down to approximately the level of agriculture alone today. Given that preliminary analyses indicate that it would be difficult to cut agricultural emissions (Chapter 9 and Annex 7.F), this means that, in the long term, net emissions (which includes sequestration from activities such as planting forests) from all other sectors would need to fall to zero. 8.8 Conclusions Stabilising the stock of greenhouse gases in the range 450 – 550 ppm CO2e requires urgent, substantial action to reduce emissions, firstly to ensure that emissions peak in the next few decades and secondly, to make the rate of decline in emissions as low as possible. If insufficient action is taken now to reduce emissions, stabilisation will become more difficult in the longer term, in terms of the speed of the transition required and the consequent costs of mitigation.

Stabilising greenhouse gas emissions is achievable through utilising a portfolio options, both technological and otherwise, across multiple sectors. The cost-effectiveness of these measures is discussed in detail in the following chapters. STERN REVIEW: The Economics of Climate Change 208

PART III: The economics of stabilisation

References

The analyses of emissions trajectories presented in this chapter are based on those presented in den Elzen and Meinshausen (2005), using the same model (Meinshausen et al. 2006). These papers provide a clear and concise overview of the key issues associated with stabilisation paths. Pacala and Socolow (2004) discuss ways of filling the ‘carbon gap’ with currently available technologies.

Den Elzen, M. and M. Meinshausen (2005): 'Multi-gas emission pathways for meeting the EU 2°C climate target', in Schellnhuber et al. (eds.), Avoiding Dangerous Climate Change, Cambridge: Cambridge University Press.

Intergovernmental Panel on Climate Change (1999): 'Aviation and the global atmosphere', A Special Report of IPCC Working Groups I and III in collaboration with the Scientific Assessment Panel to the Montreal Protocol on Substances that Deplete the Ozone Layer [J.E. Penner, D.H. Lister, D.J. Griggs et al. (eds.)], Cambridge: Cambridge University Press.

International Energy Agency (2006): 'CO2 emissions from fossil fuel combustion on-line database', version 2005/06, Paris: International Energy Agency, available from http://data.iea.org/stats/eng/main.html

Jacobs, M. (1991): ‘The green economy: Environment, sustainable development and the politics of the future, London: Pluto Press.

Jones, C.D., P.M. Cox and C. Huntingford (2006): 'Impact of climate-carbon feedbacks on emissions scenarios to achieve stabilisation', in Avoiding Dangerous Climate Change, Schellnhuber et al. (eds.), Cambridge: Cambridge University Press.

Meinshausen, M. (2006): 'What does a 2°C target mean for greenhouse gas concentrations? A brief analysis based on multi-gas emission pathways and several climate sensitivity uncertainty estimates', pp.265 – 280 in Avoiding dangerous climate change, H.J. Schellnhuber et al. (eds.), Cambridge: Cambridge University Press.

Meinshausen, M., D. van Vuuren, T.M.L. Wigley, et al. (2006): 'Multi-gas emission pathways to meet climate targets', Climatic Change, 75: 151-194

Murphy, J.M., D.M.H. Sexton, D.N. Barnett, et al. (2004): 'Quantification of modelling uncertainties in a large ensemble of climate change simulations', Nature 430: 768 – 772

O’Neill, B. and M. Oppenheimer (2004): 'Climate change impacts are sensitive to the concentration stabilization path'. Proceedings of the National Academy of Sciences 101(47): 16411-16416

Pacala, S. and R. Socolow (2004): 'Stabilization wedges: Solving the climate problem for the next 50 years with current technologies', Science, 305: 968-972

Prentice, I.C., G.D. Farquhar, M.J.R. Fasham, et al. (2001): 'The carbon cycle and atmospheric carbon dioxide', in Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change [J. Houghton et al. (eds.)]. Cambridge: Cambridge University Press.

Ramaswamy, V. et al. (2001): 'Radiative forcing of climate change', in Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change, [J. Houghton et al. (eds.)], Cambridge: Cambridge University Press.

US Climate Change Science Program (2006): ‘Scenarios of greenhouse gas emissions and atmospheric concentrations and review of integrated scenario development and application’, Public Review Draft of Synthesis and Assessment Product 2.1 STERN REVIEW: The Economics of Climate Change 209

PART III: The economics of stabilisation

Wigley, T.M.L. and S.C.B. Raper (2001): 'Interpretation of high projections for global-mean warming', Science 293: 451-454

World Business Council for Sustainable Development (2004): 'Facts and trends to 2050 – energy and climate change', World Business Council for Sustainable Development, Stevenage: Earthprint Ltd., available from www.wbcsd.org

World Resources Institute (2006): Climate Analysis Indicators Tool (CAIT) on-line database version 3.0, Washington, DC: World Resources Institute, 2006, available from http://cait.wri.org

Zhu, C., R. Taylor and G. Feng (2004): 'China’s wood market, trade and the environment', available from: http://assets.panda.org/downloads/chinawoodmarkettradeenvironment.pdf STERN REVIEW: The Economics of Climate Change 210

9 PART III: The Economics of Stabilisation

Identifying the Costs of Mitigation Key Messages

Slowly reducing emissions of greenhouse gasses that cause climate change is likely to entail some costs. Costs include the expense of developing and deploying low-emission and high-efficiency technologies and the cost to consumers of switching spending from emissions-intensive to low-emission goods and services.

Fossil fuel emissions can be cut in several ways: reducing demand for carbon-intensive products, increasing energy efficiency, and switching to low-carbon technologies. Non-fossil fuel emissions are also an important source of emission savings. Costs will differ considerably depending on which methods and techniques are used where. •

•

•

• Reducing demand for emissions-intensive goods and services is part of the solution. If prices start to reflect the full costs of production, including the greenhouse gas externality, consumers and firms will react by shifting to relatively cheaper low- carbon products. Increasing awareness of climate change is also likely to influence demand. But demand-side factors alone are unlikely to achieve all the emissions reductions required.

Efficiency gains offer opportunities both to save money and to reduce emissions, but require the removal of barriers to the uptake of more efficient technologies and methods.

A range of low-carbon technologies is already available, although many are currently more expensive than fossil-fuel equivalents. Cleaner and more efficient power, heat and transport technologies are needed to make radical emission cuts in the medium to long term. Their future costs are uncertain, but experience with other technologies has helped to develop an understanding of the key risks. The evidence indicates that efficiency is likely to increase and average costs to fall with scale and experience.

Reducing non-fossil fuel emissions will also yield important emission savings. The cost of reducing emissions from deforestation, in particular, may be relatively low, if appropriate institutional and incentive structures are put in place and the countries facing this challenge receive adequate assistance. Emissions cuts will be more challenging to achieve in agriculture, the other main non-energy source.

A portfolio of technologies will be needed. Greenhouse gases are produced by a wide range of activities in many sectors, so it is highly unlikely that any single technology will deliver all the necessary emission savings. It is also uncertain which technologies will turn out to be cheapest, so a portfolio will be required for low-cost abatement.

An estimate of resource costs suggests that the annual cost of cutting total GHG to about three quarters of current levels by 2050, consistent with a 550ppm CO2e stabilisation level, will be in the range –1.0 to +3.5% of GDP, with an average estimate of approximately 1%. This depends on steady reductions in the cost of low-carbon technologies, relative to the cost of the technologies currently deployed, and improvements in energy efficiency. The range is wide because of the uncertainties as to future rates of innovation and fossil-fuel extraction costs. The better the policy, the lower the cost.

Mitigation costs will vary according to how and when emissions are cut. Without early, well-planned action, the costs of mitigating emissions will be greater. STERN REVIEW: The Economics of Climate Change 211

PART III: The Economics of Stabilisation 9.1 Introduction Vigorous action is urgently needed to slow down, halt and reverse the growth in greenhouse-gas (GHG) emissions, as the previous chapters have shown. This chapter considers the types of action necessary and the costs that are likely to be incurred.

This chapter outlines a conceptual framework for understanding the costs of reducing GHG emissions, and presents some upper estimates of costs to the global economy of reducing total emissions to three quarters of today’s levels by 2050 (consistent with a 550ppm CO2e stabilisation trajectory, described in Chapter 8). The costs are worked out by looking at costs of individual emission saving technologies and measures. Chapter 10 looks at what macroeconomic models can say about how much it would cost to reduce emissions by a similar extent, and reaches similar conclusions. Chapter 10 also shows why a 450ppm CO2e target is likely to be unobtainable at reasonable cost.

Section 9.2 explains the nature of the costs involved in reducing emissions. Estimating the resource cost of achieving given reductions by adopting new de-carbonising technologies alone provides a good first approximation of the true cost. The costs of achieving reductions can be brought down, however, by sensible policies that encourage the use of a range of methods, including demand-switching and greater energy efficiency, so this approach to estimation is likely to exaggerate the true costs of mitigation.

Section 9.3 sets out the range of costs associated with different technologies and methods. The following four sections look at the potential and cost of tackling non-fossil fuel emissions (mainly from land-use change) and cutting fossil fuel related emissions (either by reducing demand, raising energy efficiency, or employing low-carbon technologies).

The overall costs to the global economy are estimated in Sections 9.7 and 9.8, using the resource-cost method. They are found to be in the region of –1.0 to 3.5% of GDP, with a central estimate of approximately 1% for mitigation consistent with a 550ppm CO2e stabilisation level. Different modelling approaches to calculating the cost of abatement generate estimates that span a wide range, as Chapter 10 will show. But they do not obscure the central conclusion that climate-change mitigation is technically and economically feasible at a cost of around 1% of GDP.

While these costs are not small, they are also not high enough seriously to compromise the world’s future standard of living. A 1% cost increase is like a one-off 1% increase in the price index with nominal income unaffected (see Chapter 10). While that is not insignificant, most would regard it as manageable, and it is consistent with the ambitions of both developed and developing countries for economic growth. On the other hand, climate change, if left unchecked, could pose much greater threats to growth, as demonstrated by Part II of this Review. 9.2 Calculating the costs of cutting GHG emissions Any costs to the economy of cutting GHG emissions, like other costs, will ultimately be borne by households.

Emission-intensive products will either become more expensive or impossible to buy. The costs of adjusting industrial structures will be reflected in pay and profits – with opportunities for new activities and challenges for old. The costs of adjusting industrial structures will be reflected in pay and profits – with opportunities for new activities and challenges for old. More resources will be used, at least for a while, in making currently emissions-intensive products in new ways, so fewer will be available for creating other goods and services. In considering how much mitigation to undertake, these costs should be compared with the future benefits of a better climate, together with the potential co-benefits of mitigation policies, such as greater energy efficiency and less local pollution, discussed in Chapter 12. The comparison is taken further in Chapter 13, where the costs of adaptation and mitigation are weighed up. STERN REVIEW: The Economics of Climate Change 212

PART III: The Economics of Stabilisation

A simple first approximation to the cost of reducing emissions can be obtained by considering the probable cost of a simple set of technological and output changes that are likely to achieve those reductions.

One can measure the extra resources required to meet projected energy demand with known low-carbon technologies and assess a measures of the opportunity costs, for example, from forgone agricultural output in reducing deforestation. This is the approach taken below in Sections 9.7 and 9.8. If the costs were less than the benefits that the emissions reductions bring, it would be better to take the set of mitigation measures considered than do nothing. But there may be still better measures available1.

The formal economics of marginal policy changes or reforms has been studied in a general equilibrium framework that includes market imperfections2. A reform, such as reducing GHG emissions by using extra resources, can be assessed in terms of the direct benefits of a marginal reform on consumers (the emission reduction and the reduced spending on fossil fuels), less the cost at shadow prices3 of the extra resources.

The formal economics draws attention to two issues that are important in the case of climate- change policies. First, the policies need to bring about a large, or non-marginal, change. The marginal abatement cost (MAC) – the cost of reducing emissions by one unit – is an appropriate measuring device only in the case of small changes. For big changes, the marginal cost may change substantially with increased scale. Using the MAC that initially applies, when new technologies are first being deployed, would lead to an under-estimate of costs where marginal costs rise rapidly with the scale of emissions. This could happen, for example, if initially cheap supplies of raw materials start to run short. But it may over-estimate costs where abatement leads to reductions in marginal costs – for example, through induced technological improvements4. These issues will be discussed in more detail below, in the context of empirical estimates, where average and total costs of mitigation are examined as well as marginal costs.

It is important to keep the distinction between marginal and average costs in mind throughout, because they are likely to diverge over time. On the one hand, the marginal abatement cost should rise over time to remain equal to the social cost of carbon, which itself rises with the stock of greenhouse gases in the atmosphere (see Chapter 13). On the other hand, the average cost of abatement will be influenced not only by the increasing size of emissions reductions, but also by the pace at which technological progress brings down the total costs of any given level of abatement (see Box 9.6).

Second, as formal economics has shown, shadow prices and the market prices faced by producers are equal in a fairly broad range of circumstances, so market prices can generally be used in the calculations in this chapter. But an important example where they diverge is in the case of fossil fuels. Hydrocarbons are exhaustible natural resources, the supply of which is also affected by the market power of some of their owners, such as OPEC. As a result, the market prices of fossil fuels reflect not only the marginal costs of extracting the fuels from the ground but also elements of scarcity and monopoly rents, which are income transfers, not resource costs to the world as a whole. When calculating the offset to the global costs of climate-change policy from lower spending on fossil fuels, these rents should not be included5. 1 2 3 A full comparison of the cost estimates used in the Review is given in Annex 9A on www.sternreview.org.uk. See Drèze and Stern (1987 and 1990), Ahmad and Stern (1991) and Atkinson and Stern (1974). Expressed informally, shadow prices are opportunity costs: they can often be determined by ‘correcting’ market prices for market imperfections. For a formal definition, see Drèze and Stern (1987 and 1990). In the models used there, the extra resources for emissions reductions represent a tightening of the general equilibrium constraint and the shadow prices times the quantities involved represent a summary of the overall general equilibrium repercussions. 4 although in the non-marginal changes, the distributions of costs and benefits can be of special importance. 5 world as a whole, the rents can be included. STERN REVIEW: The Economics of Climate Change 213

PART III: The Economics of Stabilisation

If there are cheaper ways of reducing carbon emissions than the illustrative set of measures examined in this chapter, and there generally will be cheaper methods than any one particular set chosen by assumption, then the illustration gives an upper bound to total costs.

An illustration of how emissions can be reduced, and at what cost, by one particular simple set of actions should provide an over-estimate of the costs that will actually be involved in reducing emissions – as long as policies set the right incentives for the most cost-effective methods of mitigation to be used. Policy-makers cannot predict in detail the cheapest ways to achieve emission reductions, but they can encourage individual households and firms to find them. Thus the costs of mitigation will depend on the effectiveness of the policy tools chosen to deliver a reduction in GHG emissions. Possible tools include emission taxes, carbon taxation and tradable carbon quotas. Carbon pricing by means of any of these methods is likely to persuade consumers to reduce their spending on currently emissions-intensive products, a helpful channel of climate-change policy that is ignored in simple technology- based cost illustrations. Induced changes in the pattern of demand can help to bring down the total costs of mitigation, but consumers still suffer some loss of real income. Regulations requiring the use of certain technologies and/or imposing physical limits on emissions constitute another possible tool.

In assessing the impact of possible instruments, key issues include the structure of taxes and associated deadweight losses6, the distribution of costs and benefits and whether or not they disrupt or enhance competitive processes. Some of these issues are tackled in simple ways by the model-based approaches to estimating costs of mitigation considered in Chapter 10. Chapter 14 considers the merits and demerits of different methods in further detail. That discussion also examines the notion of a ‘double dividend’ from raising taxes on ‘public bads’. Chapter 11 uses UK input-output data to illustrate how extra costs proportional to carbon emissions would be distributed through the economy. If, for example, extra costs amounted to around $30/tCO2 (£70/tC), it would result in an overall increase in UK consumer prices of around 1%. The analysis shows how this additional cost would be distributed in different ways across different sectors.

In examining whether mitigation by any particular method should be increased at the margin, and whether policies are cost-effective, the concept of marginal abatement cost (MAC) is central. There are many possible ways to reduce emissions, and many policy tools that could be used to do so. The costs of reductions will depend on the method chosen. One key test of the cost effectiveness of a possible plan of action is whether the MAC for each method is the same, as it should be if total costs are to be kept to a minimum. Otherwise, a saving could be made by switching at the margin from an option with a higher MAC to one with a lower MAC. This principle should be borne in mind in the discussion of different abatement opportunities below. 9.3 The range of abatement opportunities The previous section set out a conceptual framework for thinking about the costs of reducing GHG emissions. The following sections look in more detail at estimates of the costs of different methods of achieving reductions.

This section sets out four main ways in which greenhouse-gas emissions can be reduced. The first is concerned with abating non-fossil-fuel emissions, and the latter three are about cutting fossil-fuel (energy-related) emissions. These are: •

• •

6 To reduce non-fossil fuel emissions, particularly land use, agriculture and fugitive emissions To reduce demand for emission-intensive goods and services To improve energy efficiency, by getting the same outputs from fewer inputs

The deadweight loss to a tax on a good that raises $1 of revenue arises as follows. Suppose the government has raised $1 in tax revenue, and the consumer has paid this $1 in tax. But, in addition, the individual has reduced consumption in response to changes in prices and the firms producing the goods have lost profits. In the jargon of economics, the sum of the loss of consumer surplus and the loss of producer surplus exceeds the tax revenue. STERN REVIEW: The Economics of Climate Change 214

• PART III: The Economics of Stabilisation

To switch to technologies which produce fewer emissions and lower the carbon intensity of production

Annexes 7.B to 7.G7 include some more detail on which technologies can be used to cut emissions in each sector, and the associated costs.

The array of abatement opportunities can be assessed in terms of their cost per unit of GHG reduction ($/tCO2e), both at present and through time. In theory, abatement opportunities can be ranked along a continuum of the kind shown in Figure 9.1. This shows that some measures (such as improving energy efficiency and reducing deforestation) can be very cheap, and may even save money. Other measures, such as introducing hydrogen vehicles, may be a very expensive way to achieve emission reductions in the near term, until experience brings costs down.

The precise ranking of measures differs by country and sector. It may also change over time (represented in Figure 9.1 by arrows going from right to left), for example, research and development of hydrogen technology may bring the costs down in future (illustrated by the downward shift in the abatement curve over time).

Figure 9.1 Illustrative marginal abatement option cost curve For any single technology, marginal costs are likely to increase with the extent of abatement in the short term, as the types of land, labour and capital most suitable for the specific technology become scarcer. The rate of increase is likely to differ across regions, according to the constraints faced locally.

For these reasons, flexibility in the type, timing and location of emissions reduction is crucial in keeping costs down. The implications for total costs of restricting this flexibility are discussed in more detail in Chapter 10. A test of whether there is enough flexibility is to consider whether the marginal costs of abatement are broadly the same in all sectors and countries; if not, the same amount of reductions could be made at lower cost by doing more where the marginal cost is low, and less where it is high. 7 See www.sternreview.org.uk STERN REVIEW: The Economics of Climate Change 215

9.4 PART III: The Economics of Stabilisation

Cutting non-fossil-fuel related emissions Two-fifths of global emissions are from non-fossil fuel sources; there are opportunities here for low-cost emissions reductions, particularly in avoiding deforestation.

Non-fossil fuel emissions account for 40% of current global greenhouse-gas emissions, and are an important area of potential emissions savings. Emissions are mainly from non-energy sources, such as land use, agriculture and waste. Chapter 7 contains a full analysis of emission sources.

Almost 20% (8 GtCO2/year) of total greenhouse-gas emissions are currently from deforestation. A study commissioned by the Review looking at 8 countries responsible for 70% of emissions found that, based upon the opportunity costs of the use of the land which would no longer be available for agriculture if deforestation were avoided, emission savings from avoided deforestation could yield reductions in CO2 emissions for under $5/tCO2, possibly for as little as $1/tCO2 (see Box 9.1). In addition, large-scale reductions would require spending on administration and enforcement, as well as institutional and social changes. The transition would need to be carefully managed if it is to be effective.

Planting new forests (afforestation and reforestation) could save at least an additional 1 GtCO2/yr, at a cost estimated at around $5/tCO2 – $15/tCO28. The full technical potential of forestry related measures would go beyond this. An IPCC report in 2000 estimated a technical potential of 4 – 6 GtCO2/year from the planting of new forests alone between 1995 and 2050, 70% of which would come from tropical countries9. Revised estimates are expected from the Fourth Assessment Report of IPCC.

Changes to agricultural land management, such as changes to tilling practices10, could save a further 1 GtCO2/year at a cost of around $27/tCO2e in 202011. More recent analysis suggested savings could be as much as 1.8 GtCO2 at $20/tCO2 in 203012. The production of bioenergy crops would add further savings. In this chapter, this is discussed in the context of its application to emissions savings in other sectors (see Box 9.5). Biogas from animal wastes could also yield further savings. 8 9 10 carbon to microbial activity and hence, conserve soil carbon stocks (IPCC (2001)). 11 12 STERN REVIEW: The Economics of Climate Change 216

Box 9.1 PART III: The Economics of Stabilisation

The costs of reducing emissions by avoiding further deforestation A substantial body of evidence suggests that action to prevent further deforestation would be relatively cheap compared with other types of mitigation.

Three types of costs arise from curbing deforestation. These are the opportunity cost foregone from preserving forest, the cost of administering and enforcing effective action, and the cost of managing the transition.

The opportunity cost to those who use the land directly can be estimated from the potential revenue per hectare of alternative land uses. These potential returns vary between uses. Oil palm and soya produce much higher returns than pastoral use, with net present values of up to $2000 per hectare compared with as little as $2 per hectare13. Timber is often harvested, particularly in South East Asia, where there is easy access to nearby markets and timber yields higher prices. Timber sales can offset the cost of clearing and converting land.

A study carried out for this Review14 estimated opportunity costs on this basis for eight countries15 that collectively are responsible for 70% of land-use emissions (responsible for 4.9 GtCO2 today and 3.5 GtCO2 in 2050 under BAU conditions). If all deforestation in these countries were to cease, the opportunity cost would amount to around $5-10 billion annually (approximately $1-2/tCO2 on average). On the one hand, the opportunity cost in terms of national GDP could be higher than this, as the country would also forego added value from related activities, including processing agricultural products and timber. The size of the opportunity cost would then depend on how easily factors of production could be re-allocated to other activities. On the other hand, these estimates may overstate the true opportunity cost, as sustainable forest management could also yield timber and corresponding revenues. Furthermore, reducing emissions arising from accidental fires or unintended damage from logging may be lower than the opportunity costs suggest.

Other studies have estimated the cost of action using different methods, such as land-value studies assuming that the price of a piece of land approximates to the market expectation of the net present value of income from it, and econometric studies that estimate an assumed supply curve. In econometric studies16, marginal costs have been projected as high as $30t/CO2 to eliminate all deforestation. High marginal values for the last pieces of forestland preserved are not inconsistent with a bottom-up approach based on average returns across large areas. These studies also suggest that costs are low for early action on a significant scale.

Action to address deforestation would also incur administrative, monitoring and enforcement costs for the government. But there would be significant economies of scale if action were to take place at a country level rather than on a project basis. Examination of such schemes suggests that the possible costs are likely to be small: perhaps $12m to $93m a year for these eight countries.

The policy challenges involved with avoiding further deforestation are discussed in Chapter 25.

The other main further sources of non-energy-related emissions, with estimates of economic potential for emissions reductions, are: •

13 Livestock, fertiliser and rice produce methane and nitrous oxide emissions. The IPCC (2001) suggested that around 1 GtCO2e/year could be saved at a cost of up to $27/tCO2e17 in 2020. However more recent analysis suggests that just 0.2

These figures are calculated from income over 30 years, using a discount rate of 10%, except for Indonesia, which uses 20%. 14 15 16 17 STERN REVIEW: The Economics of Climate Change 217

•

•

• PART III: The Economics of Stabilisation

GtCO2e/year might be saved at $20/tCO2e in 203018. It is important to investigate ways of cutting this growing source of emissions.

Wastage in the production of fossil fuels (so-called fugitive emissions) and other energy-related non-CO2 emissions currently amount to around 2 GtCO2e/year19. If fugitive emissions of non-CO2 and CO2 gases could be constrained to current levels, then savings could amount to 2.3 GtCO2e/year and 0.2 GtCO2/year respectively in 2050 on baseline levels20.

Waste is currently responsible for 1.4 GtCO2e/year21, of which over half is from landfill sites and most of the remainder from wastewater treatment. Reusing and recycling lead to less resources being required to produce new goods and a reduction in associated emissions. Technologies such as energy-recovering incinerators also help to reduce emissions. The IPCC estimate that 0.7 GtCO2e/year could be saved in 2020, of which three quarters could be achieved at negative cost and one quarter at a cost of $5/tCO2e22.

Industrial processes used to make products such as adipic and nitric acid produce non-CO2 emissions; the IPCC estimate that 0.4 GtCO2e/year could be reduced from these sources in 2020 at a cost of less than $3/tCO2e23. The production of products such as aluminium and cement also involve a chemical process that release CO2. Assuming that emissions from this source could be reduced by a similar proportion, savings could amount to 0.5 GtCO2e in 205024.

Table 9.1 summarises the possible cost-effective non-fossil fuel CO2 emission savings for 2050 described above. These figures are very uncertain but the estimates for waste and industrial processes arguably represent a lower-end estimate because they come from IPCC studies looking at possible emission savings in 2020, and savings by 2050 could be higher. Some of these savings cost $5/tCO2e or less, and it is possible that more could be saved at a slightly higher cost, with the technical potential for land-use changes being particularly significant. Achieving these emission savings would mean non-fossil fuel emissions in 2050 would be almost 11 GtCO2e lower in 2050 than in the baseline case. energy-efficiency measures. 18 19 20 levels or below by 2050, as in the work by Dennis Anderson described later in this chapter, and the IEA (2006) analysis discussed in Section 9.9. 21 22 23 24 STERN REVIEW: The Economics of Climate Change 218

PART III: The Economics of Stabilisation 9.5 Reducing the demand for carbon-intensive goods and services One way of reducing emissions is to reduce the demand for greenhouse-gas-intensive goods and services like energy. Policies to reduce the amount of energy-intensive activity should include creating price signals that reflect the damage that the production of particular goods and services does to the atmosphere. These signals will encourage firms and households to switch their spending towards other, less emissions-intensive, goods and services.

Regulations, the provision of better information and changing consumer preferences can also help. If people’s preferences evolve as a result of greater sensitivity to energy use, for instance to favour smaller, more fuel-efficient vehicles, they may perceive the burden from ‘trading down’ from a larger vehicle as small or even negative (see Chapter 17). Efforts to reduce the demand for emissions-intensive activities include reducing over-heating of buildings, reducing the use of energy-hungry appliances, and the development and use of more environmentally friendly forms of transport.

In some cases, there may be ‘win-win’ opportunities (for example, congestion charging may lead to a reduction in GHG emissions and also reduce journey time for motorists and bus users). But some demand-reduction measures may conflict with other policy objectives. For example, raising the cost of private transport could lead to social exclusion, especially in rural areas. Chapter 12 discusses in more detail how climate change policy may fit with other policy objectives. Part IV of the Review includes discussion of how policy can be designed to ensure that the climate change damage associated with emission-intensive goods and services is better reflected in their prices. 9.6 Improving energy efficiency Improving efficiency and avoiding waste offer opportunities to save both emissions and resources, though there may be obstacles to the adoption of these opportunities.

Energy efficiency refers to the proportion of energy within a fuel that is converted into a given final output. Improving efficiency means, for example, using less electricity to heat buildings to a given temperature, or using less petrol to drive a kilometre. The opportunities for reducing carbon emissions through the uptake of low-carbon energy sources, ‘fuel switching’, are not considered in this section.

The technical potential for efficiency improvements to reduce emissions and costs is substantial. Over the past century, efficiency in energy supply improved ten-fold or more in 25 For explanation of how BAU emissions were calculated, see Chapter 7. STERN REVIEW: The Economics of Climate Change 219

PART III: The Economics of Stabilisation

the industrial countries. Hannah’s historical study26 of the UK electricity industry, for example, reports that the consumption of coal was 10-25 lbs/kWh in 1891, 5 lbs/kWh in the first decade of the 20th century and 1.5 lbs/kWh by 1947; today it is about 0.7 lbs/kWh27, a roughly 10-fold increase over the century in the efficiency of power generation alone.

There have also been impressive gains in the efficiency with which energy is utilised for heating, lighting, refrigeration and motive power for industry and transport, with the invention of the fluorescent light bulb, the substitution of gas for coal for heat, the invention of double glazing, the use of ‘natural’ systems for lighting, heating and cooling, the development of heat pumps, the use of loft and cavity-wall insulation, and many other innovations.

Furthermore, the possibilities for further gains are far from being exhausted, and are now much sought after by industry and commerce, particularly those engaged in energy-intensive processes. Many of these opportunities are yet to be incorporated fully into the capital stock. For example, the full hybrid car (which may also pave a path for electric and fuel-cell vehicles) offers the prospect of a step change in the fuel efficiency of vehicles, while new diode-based technologies have the potential to deliver marked reductions in the intensity of lighting.

However, the rate of uptake of efficiency measures is often slow, largely because of the existence of market barriers and failures. These include hidden and transaction costs such as the cost of the time needed to plan new investments; a lack of information about the available options; capital constraints; misaligned incentives; together with behavioural and organisational factors affecting economic rationality in decision-making. These are discussed in more detail in Chapter 17.

There is much debate about how big a reduction in emissions efficiency measures could in practice yield. The IEA studies summarised in Section 9.9 find that efficiency in the use of fossil fuels is likely to be the single largest source of fossil fuel-related emission savings in 2050, capable of reducing carbon emissions by up to 16 GtCO2e per year by 2050. While estimates vary between studies, there is general agreement that the possibilities for further gains in efficiency are appreciable at each stage of energy conversion, across all sectors, end uses and economies.

Figure 9.2 provides a graphical representation of the estimated costs and abatement potential by 2020 for a selected sample of energy efficiency technologies across different sectors. 26 27 See Hannah (1979) Assuming 40% thermal efficiency and a c.v. of coal of 8,000kWh/tonne. Pounds (lbs) are a unit of weight: 1 lbs = 0.454 kg.

STERN REVIEW: The Economics of Climate Change 220

Abatement Cost – £/tC PART III: The Economics of Stabilisation

Figure 9.2 Aggregate carbon abatement cost curve for the UK – annual carbon savings by 202028

Marginal Carbon Abatement Cost Curve – £/tC Selected Technologies 2020 Abatement – MtC 300

200

100

0

-100

-200

-300

-400

-500

Source: See notes 9.7 Low-carbon technologies Options for low-emission energy technologies are developing rapidly, though many remain more expensive than conventional technologies.

This section examines the options for emissions reductions in the energy sector, their costs and how they are likely to move over time. The next section illustrates the costs of a set of policies in electricity and transport that could reduce emissions to levels consistent with a stabilisation path at 550ppm CO2e. A range of options is currently available for decarbonising energy use in electricity generation, transport and industry, all of which are amenable to significant further development. These include:- • • • •

• •

•

•

28 On and offshore wind. Wave and tidal energy projects. Solar energy (thermal and photovoltaic). Carbon capture and storage for electricity generation (provided the risk of leakage is minimised) – Box 9.2 sets out the state of this relatively new technology, and what is known about costs. The production of hydrogen for heat and transport fuels. Nuclear power, if the waste disposal and proliferation issues are dealt with. A new generation of reactors is being built in India, Russia and East Asia. Reactors have either been commissioned or are close to being commissioned in France, Finland and the USA. Hydroelectric power, though environmental issues need to be considered and new sites will become increasingly scarce. The power output/storage ratio will also need to increase, to reduce the typical area inundated and increase the capacity of schemes to meet peak loads. Expansion of bioenergy for use in the power, transport, buildings and industry sectors from afforestation, crops, and organic wastes.

This is intended to provide an indicative representation of average technology costs only (costs of individual technologies will, or course, vary). It draws together work on recent sectoral estimates undertaken by Enviros as part of the Energy Efficiency and Innovation Review (see www.defra.gov.uk/environment/energy/eeir/pdf/enviros- report.pdf) and drawing on data from the BRE and Enusim databases on the service sectors respectively, as well as Defra internal estimates for the domestic sector. The cost information presented here is based on a 3.5% social discount rate. STERN REVIEW: The Economics of Climate Change 221

•

•

•

Box 9.2 PART III: The Economics of Stabilisation

Decentralised power generation, including micro-generation, combined heat and power (dCHP) using natural gas or biomass in the first instance, and hydrogen derived from low-carbon sources in the long term. Fuel cells with hydrogen as a fuel for transport (with hydrogen produced by a low- carbon method). Hybrid- and electric-vehicle technology (with electricity generated by a low-carbon method).

Carbon capture and storage (CCS) No single technology or process will deliver the emission reductions needed to keep climate change within the targeted limits. But much attention is focused on the potential of Carbon Capture and Storage (CCS). This is the process of removing and storing carbon emissions from the exhaust gases of power stations and other large-scale emitters. If it proved effective, CCS could help reduce emissions from the flood of new coal-fired power stations planned over the next decades, especially in India and China29.

CCS technologies have the significant advantage that their large-scale deployment could reconcile the continued use of fossil fuels over the medium to long term with the need for deep cuts in emissions. Nearly 70% of energy production will still come from fossil fuels by 2050 in the IEA’s ACT MAP scenario30. In their base case, energy production doubles by 2050 with fossil fuels accounting for 85% of energy. The growth of coal use in OECD countries, India and China is a particular issue – the IEA forecast that without action a third of energy emissions will come from coal in 2030. Even with strong action to encourage the uptake of renewables and other low-carbon technologies, fossil fuels may still make up to half of all energy supply by 2050. Successfully stabilising emissions without CCS technology would require dramatic growth in other low-carbon technologies.

Once captured, the exhaust gases can be either processed and compressed into liquefied CO2 or chemically changed into solid, inorganic carbonates. Captured CO2 can be transported either through pipelines or by ship. The liquid or solid CO2 can be stored in various ways. As a pressurised liquid, CO2 can also be injected into oil fields to raise well pressure and increase flow rates from depleted wells. Norway’s Statoil, for example, captures emissions from on-shore power stations and re-injects the captured CO2 for such ‘enhanced oil recovery’ from its off-shore Sleipner oil field.

In most cases, the captured gas will be injected and stored in suitable, non-porous underground rock foundations such as depleted oil and gas wells, deep saline formations and old coalmines. Other theoretically possible but as yet largely untested ways of storing the CO2 are to dissolve it deep within the ocean, store as an inorganic carbonate or use the CO2 to produce hydrogen or various carbon-rich chemicals. Careful site evaluation is needed to ensure safe, long-term storage. Estimates of the potential geological storage capacity range from 1,700 to 11,100 GtCO2 equivalent31, or from to 70 to 450 years of the 2003 level of fossil-fuel-related emissions (24.5 GtCO232/year).

It is technically possible to capture emissions from virtually any source, but the economics of CCS favours capturing emissions from large sources producing concentrated CO2 emissions (such as power stations, cement and petrochemical plants), to capture scale economies, and where it is possible to store the CO2 close to the emission and capture point, to reduce transportation costs.

There are several obstacles to the deployment of CCS, including technological and cost 29 Read (2006) discusses how if CCS technologies were to capture emissions from the use of biofuels this could create negative emissions, that is, sequestering carbon dioxide from the atmosphere. 30 in 2050 following a technology ‘push’ for low-carbon technologies. 31 32 STERN REVIEW: The Economics of Climate Change 222

At a cost of $0.9 trillion around $23 per tonne. Sachs and Lackner, 2005 $280 to $840 billion at $19 – $49/tCO2. Page 61 IEA, 2006 PART III: The Economics of Stabilisation

barriers, particularly the need to improve energy efficiency in power stations adopting CCS. Others include regulatory and legal33 barriers, such as the legal issues around the ownership of the CO2 over long periods of time, the lack of safety standards and emission-recording guidelines. There are also environmental concerns that the CO2 might leak or that building the necessary infrastructure might damage the local environment. Public opinion needs to be won over.

Employing CCS technology adds to the overall costs of power generation. But there is a wide range of estimates, partly reflecting the relatively untried nature of the technology and variety of possible methods and emission sources. The IPCC quotes a full range from zero to $270 per tonne of CO2. A range of central estimates from the IPCC and other sources34 show the costs of coal-based CCS employment ranging from $19 to $49 per tonne of CO2, with a range from $22 to $40 per tonne if lower-carbon gas is used. Some studies provide current estimates and some medium-term costs. A range of technologies is also considered, with and without CCS, and some with more basic generation technologies as the baseline35. The assumptions set have an important impact on cost estimates. The range of cost estimates will narrow when CCS technologies have been demonstrated but, until this occurs, the estimates remain speculative.

The IPCC special report on CCS suggested that it could provide between 15% and 55% of the cumulative mitigation effort until 2100. The IEA’s Energy Technology Perspectives uses a scenario that keeps emissions to near current levels by 2050, with 14 – 16.2% of electricity generated from coal-fired power stations using CCS. This would deliver from 24.7 – 27.6% of emission reductions36. Sachs and Lackner37 calculate that, if all projected fossil-fuel plants were CCS, it could save 17 GtCO2 annually at a cost of 0.1% to 0.3% of GDP38, and reduce global emissions by 2050 from their 554ppm BAU to 508ppm CO2.

IEA modelling shows that, without CCS, marginal abatement costs would rise from $25 to $43 per tonne in Europe, and from $25 to $40 per tonne in China, while global emissions are10% to 14% higher. This highlights the crucial role CCS is expected to play39. For more on international action and policies to encourage the demonstration and adoption of CCS technologies, see Section 24.3 and Box 24.8.

Most low-carbon technologies are currently more expensive than using fossil fuels.

Estimates of the costs per unit of energy of substituting low-carbon-emitting energy sources for fossil fuels over the next 10-20 years are presented in Box 9.3; the technologies shown cover electricity supply, the gas markets (mainly for heat) and transport. The costs are expressed as a central estimate, with a range. 33 34 35 At present sub-sea storage of CO2 without enhanced oil recovery would be illegal. Sources include MIT, SPRU, UK CCS, IPCC, UK Energy Review, Sachs and Lackner. Some compare CCGT, IGCC and supercritical/basic pulverised coal with and without CCS while others compare IGCC with CCS to pulverised coal without or an alternative fossil-fuel mix. 36 37 38 39

STERN REVIEW: The Economics of Climate Change 223

Box 9.3 PART III: The Economics of Stabilisation

Costs of low-carbon technologies relative to fossil-fuel technologies -100 0 100 200 300 400 500 600 replaced

This figure shows estimates by Anderson40 of costs of technologies in 2015, 2025 and 2050 used to constrain fossil fuel emissions in 2050 at today’s levels41. For most technologies, the unit cost as a proportion of the fossil-fuel alternative is expected to fall over time, largely because of learning effects (discussed below). But, as a technology comes up against increasing constraints and extends beyond its minimum efficient scale of production, the fall in unit costs may begin to reverse. The ranges quoted reflect judgements about the likely probability distribution for unit costs and allow for the variability of fossil-fuel prices (see text below and Section 9.8 for a further discussion of the treatment of uncertainties). The 0% line indicates that costs are the same as the corresponding fossil-fuel option.

Unit costs of energy technologies expressed as a percentage of the fossil-fuel alternative (in 2015, 2025, 2050)

cost as % of fossil fuel option Electricity from gas with CCS Electricity from coal with CCS Nuclear power Electricity from energy crops Electricity from organic wastes Onshore wind Offshore wind Solar thermal (v. sunny regions) PV (sunny regions) dCHP using H from NG or coal with CCS Hydrogen from NG or coal (CCS) – industry Hydrogen from NG or coal (CCS) – distributed Electrolytic hydrogen – industry Electrolytic hydrogen – distributed Biomass for heat – distributed Bioethanol Biodiesel Hydrogen ICE vehicle – fossil H (+CCS) FC Hydrogen vehicle – fossil H (+CCS) FC Hydrogen vehicle – electrolytic H Cost in 2025 Cost in 2015 Cost in 2050 Even in the near to medium term, the uncertainties are very large. The costs of technologies vary with their stage of development, and on specific regional situations and resource endowments, including the costs and availability of specific types of fossil fuels, the availability of land for bioenergy or sites for wind and nuclear power. Other factors include climatic suitability in the case of solar ‘insolation’ (incident solar energy) and concentrated emission sources (in the case of CCS). In recent years, oil prices have swung over a range of more than $50 per barrel and industrial gas from $4 to $9/GJ; such swings alone can shift the relative costs of the alternatives to fossil fuels by factors or two or three or more. In principle, 40 Paper by Dennis Anderson, ”Costs and Finance of Carbon Abatement in the Energy Sector”, published on the Stern Review web site. 41 generation capacity), but exclude transmission and distribution. The costs of the latter are, however, included in the estimates for decentralised generation. The average costs of energy from the fossil-fuel technologies are 2.5p/kWh for central generation, 8p/kWh for decentralised generation, £4/GJ for industrial gas, $6/GJ for domestic gas, and 30p/litre (exclusive of excise taxes) for vehicle fuels; all are subject to the range of uncertainties noted in the text. STERN REVIEW: The Economics of Climate Change 224

PART III: The Economics of Stabilisation

estimates of global costs should be based on the extraction costs of fossil fuels, not their market prices, which include a significant but uncertain proportion of rents (see Section 9.2).

The cost of technologies tends to fall over time, because of learning and economies of scale.

Historical experience shows that technological development does not stand still in the energy or other sectors. There have been major advances in the efficiency of fossil-fuel use; similar progress can also be expected for low-carbon technologies as the state of knowledge progresses.

Box 9.4 shows cost trends for selected low-carbon technologies. Economists have fitted ‘learning curves’ to such data to estimate how much costs might decline with investment and operating experience, as measured by cumulative investment. ‘Learning’ is of course an important contributor to cost reductions, but should be seen as one aspect of several factors at work. These include: •

•

•

• The development of new generations of materials and design concepts through R&D and the insights gained from investment and operating experience—for example, from current efforts to develop thin-film and organic solar cells, or in new materials and catalysts for fuel cells and hydrogen production and use; Opportunities for batch production arising from the modularity of some emerging technologies, such as solar PV. This leads to scale economies in production; to associated technical developments in manufacture; to the reduction of lead times for investments, often to a few months, as compared with three to six years or longer for conventional plant; and to the more rapid feedback of experience; R&D to seek further improvements and solve problems encountered with investments in place; Opportunities for scale economies in the provision of supporting services in installation and use of new technologies, the costs of which are appreciable when markets are small. For example, if specialised barges are required to install and service off-shore wind turbines, the equipment is much more efficiently utilised in a farm of 100 turbines than in one with just ten, and of course if there are many offshore wind farms in the project pipeline. STERN REVIEW: The Economics of Climate Change 225

Box 9.4 PART III: The Economics of Stabilisation

Evidence on learning rates in energy technologies A number of key energy technologies in use today have experienced cost reductions consistent with the theories of learning and scale economies. The diagram below shows historical learning rates for a number of technologies. The number in brackets gives an indication of the speed of learning: 97%, for instance, means that unit costs are 97% of their previous level after each doubling of installed capacity (3% cheaper).

Cost evolution and learning rates for selected technologies Source: IEA (2000) pp21

After early applications in manufacturing and production (1930s) and business management, strategy and organisation studies, the past decade has seen the application of learning curves as an analytical tool for energy technologies (see IEA, 2000). The majority of published learning-rate estimates relevant to climate change relate to electricity-generation technologies. In Figure 9.5 above, estimates of learning rates from different technologies42 span a wide range, from around 3% to over 35% cost reductions associated with a doubling of output capacity.

Using evidence on learning to project likely technology-cost changes suffers from selection bias, as technologies that fail to experience cost reductions drop out of the market and are then not included in studies. In order to correct for this, the learning and experience curves used to guide the cost exercise in this chapter take account of the high risks associated with new technologies. Moreover, the projected cost reductions are based on a far broader range of factors than just ‘learning’, as discussed in the main text.

The effects of the likely fall in costs with R&D and investment are reflected in the estimates for medium-term costs shown in Box 9.3. There is a general shift down in the expected costs of the alternatives to fossil fuels, in some cases to the point where they overlap under combinations of higher fossil-fuel prices and higher rates of technical progress.

In addition, the rankings of the technologies change, with some that are currently more expensive becoming cheaper with investment and innovation. Examples are solar energy in sunny regions and decentralised sources of combined heat and power (see Chapter 25). Nevertheless, most unit energy costs seem likely to remain higher than fossil fuels, and policies over the next 25 years should be based on this assumption. These are, of course, in the main costs borne in the first place by the private sector, although the public power sector is large in many countries. It will be the role o