Global and aggregate impacts

Three types of aggregate impacts are commonly reported. In the first, impacts are computed as a percent of gross domestic product (GDP) for a specified rise in global mean temperature. In the second, impacts are aggregated over time and discounted back to the present day along specified emissions scenarios such as those documented in Nakicenovic and Swart (2000) under specified assumptions about economic development, changes in technology and adaptive capacity. Some of these estimates are made at the global level, but others aggregate a series of local or regional impacts to obtain a global total. A third type of estimate has recently attracted the most attention. Called the social cost of carbon (SCC), it is an estimate of the economic value of the extra (or marginal) impact caused by the emission of one more tonne of carbon (in the form of carbon dioxide) at any point in time; it can, as well, be interpreted as the marginal benefit of reducing carbon emissions by one tonne. Researchers calculate SCC by summing the extra impacts for as long as the extra tonne remains in the atmosphere - a process which requires a model of atmospheric residence time and a means of discounting economic values back to the year of emission.

This section provides a brief discussion of the historical and current status of efforts to produce aggregate estimates of the impacts of climate change. The first sub-section focuses attention on economic estimates and the second begins to expand the discussion by reporting estimates calibrated in alternative metrics. It is in this expansion that the implications of spatial and temporal diversity in systems' exposures and sensitivities to climate change begin to emerge.

20.6.1 History and present state of aggregate impact estimates

Most of the aggregate impacts reported in IPCC (1996) were of the first type; they monetised the likely damage that would be caused by a doubling of CO2 concentrations. For developed countries, estimated damages were of the order of 1% of GDP. Developing countries were expected to suffer larger percentage damages, so mean global losses of 1.5 to 3.5% of world GDP were therefore reported. IPCC (2001a) reported essentially the same range because more modest estimates of market damages were balanced by other factors such as higher non-market impacts and improved coverage of a wide range of uncertainties. Most recently, Stern (2007) took account of a full range of both impacts and possible outcomes (i.e., it employed the basic economics of risk premiums) to suggest that the economic effects of unmitigated climate change could reduce welfare by an amount equivalent to a persistent average reduction in global per capita consumption of at least 5%. Including direct impacts on the environment and human health (i.e., 'non-market' impacts) increased their estimate of the total (average) cost of climate change to 11% GDP; including evidence which indicates that the climate system may be more responsive to greenhouse-gas emissions than previously thought increased their estimates to 14% GDP. Using equity weights to reflect the expectation that a disproportionate share of the climate-change burden will fall on poor regions of the world increased their estimated reduction in equivalent consumption per head to 20%.

Figure 20.3 compares the Stern (2007) relationship between global impacts and increases in global mean temperature with estimates drawn from earlier studies that were assessed in IPCC (2001b). The Stern (2007) trajectories all show negative impacts for all temperatures; they reflect the simple assumptions of the underlying PAGE2002 model and a focus on risks associated with higher temperatures. The Mendelsohn et al. (1998) estimates aggregate regional monetary damages (both positive and negative) without equity weighting. The two Nordhaus and Boyer (2000) trajectories track aggregated regional monetary estimates of damages with and without population-based equity weighting; they do include a 'willingness to pay (to avoid)' reflection of the costs of abrupt change. The two Tol (2002) trajectories track aggregated regional monetary estimates of damages with and without utility-based equity weighting. The various relationships depicted in Figure 20.3 therefore differ in their treatment of equity weighting, in their efforts to capture the potential of beneficial climate change (in, for example,

— Mendelsohn, output

— Nordhaus, output

— Mendelsohn, output

— Nordhaus, output

Global mean temperature (°C) (above pre-industrial) 1 2 3 4 5 6 7 8

Global mean temperature (°C) (above pre-industrial) 1 2 3 4 5 6 7 8

Baseline climate, market impacts and risk of catastrophe High climate, market impacts and risk of catastrophe Baseline climate, market impacts and risk of catastrophe and non-market impacts High climate, market impacts and risk of catastrophe and non-market impacts

Figure 20.3. (a) Damage estimates, as a percent of global GDP, as correlated with increases in global mean temperature. Source: IPCC (2001b). (b) Damage estimates, as a percent of global GDP, are correlated with increases in global mean temperature. Source: Stern (2007).

agriculture for small increases in temperature; see Chapter 5, Section 5.4.7) and in their treatment of the risks of catastrophe for large increases in temperature.

Early calculations of the SCC (IPCC (1996) estimates ranged from US$5 to $125 per tonne of carbon in 1990 dollars) stimulated recurring interest, as part of wider post-Kyoto considerations, in the economic benefits of climate-change policy (Watkiss et al., 2005). After surveying the literature, Clarkson and Deyes (2002) proposed a central value of US$105 per tonne of carbon (in year 2000 prices) for the SCC, with upper and lower values of US$50 and $210 per tonne. Pearce (2003) argued that 3% is a reasonable representation of a social discount rate so the probable range of the SCC in 2003 should have been in the region of US$4 to 9 per tonne of carbon. Tol (2005) gathered over 100 estimates of the SCC from 28 published studies and combined them to form a probability density function; it displayed a median of US$14 per tonne of carbon, a mean of US$93 per tonne and a 95th percentile estimate equal to US$350 per tonne. Peer-reviewed studies generally reported lower estimates and smaller uncertainties than those which were not; their mean was US$43 per tonne of carbon with a standard deviation of US$83. The survey showed that 10% of the estimates were negative; to support these estimates, the climate sensitivity was assumed to be low and small increases in global mean temperature brought benefits (as suggested by the Tol (2002) trajectories in Figure 20.3).

Notwithstanding the differences in damage sensitivity to temperature reflected in Figure 20.3, the effect of the discount rate (see glossary) on estimates of SCC is most striking. The 90th percentile SCC, for instance, is US$62/tC for a 3% pure rate of time preference, $165/tC for 1% and $1,610/tC for 0%. Stern (2007) calculated, on the basis of damage calculations described above, a mean estimate of the SCC in 2006 of US$85 per tonne of CO2 (US$310 per tonne of carbon). Had it been included in the Tol (2005) survey, it would have fallen well above the 95th percentile, in large measure because of their adoption of a low 0.1% pure rate of time preference. Other estimates of the SCC run from less than US$1 per tonne to over US$1,500 per tonne of carbon. Downing et al. (2005) argued that this range reflects uncertainties in climate and impacts, coverage of sectors and extremes, and choices of decision variables. Tol (2005) concluded, using standard assumptions about discounting and aggregation, that the SCC is unlikely to exceed US$50/tC. In contrast, Downing et al. (2005) concluded that a lower benchmark of US$50/tC is reasonable for a global decision context committed to reducing the threat of dangerous climate change and including a modest level of aversion to extreme risks, relatively low discount rates and equity weighting.

Climate change is not caused by carbon dioxide alone, and integrated assessment models can calculate the social cost of each greenhouse gas under consistent assumptions. For instance, the mean estimate from the PAGE2002 model for the social cost of methane is US$105 per tonne emitted in 2001, in year 2000 dollars, with a 5 to 95% uncertainty range of US$25 to $250 per tonne. The estimate for the social cost of SF6 is US$200,000 per tonne emitted in 2001 with a 5 to 95% range of US$45,000 to $450,000 per tonne. These are all higher than the corresponding US$19 per tonne estimate for SCC that is surrounded by a 5 to 95% range of US$4 to $50 per tonne (Hope, 2006b). It has been known since IPCC (1996) that the SCC will increase over time; current knowledge suggests a 2.4% per year rate of growth. The social cost of methane will grow 50% faster because of its shorter atmospheric lifetime. Unlike later emissions, any extra methane emitted today will have disappeared before the most severe climate-change impacts occur (Watkiss et al., 2005).

Tol (2005) finds that much of the uncertainty in the estimates of the SCC can be traced to two assumptions: one on the discount rate and the other on the equity weights that are used to aggregate monetised impacts over countries. In most other policy areas, the rich do not reveal as much concern for the poor as is implied by the equity weights used in many models. Downing et al. (2005) state that the extreme tails of the estimates of the SCC depend as much on decision values (such as discounting and equity weighting) as on the climate forcing and uncertainty in the underlying impact models. Integrated models are always simplified representations of reality. To be comprehensive, other social and cultural values need to be given comparable weights to economic values, and there are prototype integrated assessment models to demonstrate this (Rotmans and deVries, 1997).

Table 20.2 shows the six major influences calculated by PAGE2002 and reported in Hope (2005). That the list can be divided into two scientific and four socio-economic parameters is another strong argument for the building of integrated assessment models (IAMs); models that are exclusively scientific, or exclusively economic, would omit parts of the climate-change problem which still contain profound uncertainties. The two top influences are the climate sensitivity and the pure rate of time preference. Climate sensitivity is positively correlated with the SCC, but the pure time preference rate is negatively correlated with the SCC. Non-economic impact ranks third and economic impact ranks sixth (Hope, 2005).

A few models have existed for long enough to trace the changes in their estimates of the SCC over time. Table 20.3 shows how the results from three integrated assessment models have evolved over the last 15 years. The DICE and PAGE estimates have not changed greatly over the years, but this gives a misleading impression of stability. The values from PAGE have changed little because several quite significant changes have approximately cancelled each other out. In the later studies, lower estimates for market-sector impacts in developed countries are offset by higher non-market impacts, equity weights and inclusion of estimates of the possible impacts of large-scale discontinuities (Tol, 2005).

Hitz and Smith (2004) found that the relationships between global mean temperature and impacts of the sort displayed in Figure 20.3 are not consistent across sectors for modest amounts of warming. Beyond an approximate 3 to 4°C increase in global mean temperature above pre-industrial levels, all sectors (except possibly forestry) show increasingly adverse impacts. Tol (2005) found that few studies cover non-market damages, the risk of potential extreme weather, socially contingent effects, or the potential for longer-term catastrophic events. Therefore, uncertainty in the value of the SCC is derived not only from the

'true' value of impacts that are covered by the models, but also from impacts that have not yet been quantified and valued. As argued in Watkiss et al. (2005) and displayed in Figure 20.4, existing estimates of SCC are products of work that spans only a sub-set of impacts for which complete estimates might be calculated. Nonetheless, current estimates do provide enough information to support meaningful discussions about reducing the emissions of CO2, methane and other greenhouse gases, and the appropriate trade-off between gases.

Nonetheless, estimates of SCC offer a consistent way to internalise current knowledge about the impacts of climate change into development, mitigation and/or adaptation decisions that the private and public sector will be making over the near term (Morimoto and Hope, 2004). According to economic theory, if the social cost calculations were complete and markets were perfect, then efforts to cut back the emissions of greenhouse gases would continue as long as the marginal cost of the cutbacks were lower than the social cost of the impacts they cause. If taxes were used, then they should be set equal to the SCC. If tradable permits were used, then their price should be the same as the SCC. If their price turns out to be lower than the social cost, then the total allocation of permits would have been too large and vice versa. In any comparison between greenhouse gases, according to Pearce (2003), the SCC is the correct figure to use. For reference, spot prices for permits in the European Carbon Trading Scheme since its inception early in 2005 started out towards the bottom end of the range of the SCC, but they rose quickly to around US$100 per tonne of carbon before falling by about 50% in the early summer of 2006 amid concerns that the carbon allowances allocated by the European Commission at the start of the scheme had been too generous. In the real world, markets are not perfect, calculations of the SCC are far from complete, and both mask significant differences between regions and types of impacts.

Table 20.2. Major factors causing uncertainty in the social cost of carbon. Relative importance is measured by the magnitude of the partial rank correlation coefficient between the parameter and the SCC, with the most important indexed to 100. A + sign shows that an increase in this parameter leads to an increase in the SCC and vice versa. Source: Hope (2005).

Table 20.2. Major factors causing uncertainty in the social cost of carbon. Relative importance is measured by the magnitude of the partial rank correlation coefficient between the parameter and the SCC, with the most important indexed to 100. A + sign shows that an increase in this parameter leads to an increase in the SCC and vice versa. Source: Hope (2005).

Parameter

Definition

Sign

Range

Importance

Climate sensitivity

Equilibrium temperature rise for a doubling of CO2 concentration

+

1.5 to 5°C

0 0

Post a comment