Global vs. Partial Cooperative Climate Agreement

A Global Cooperative Climate Agreement (Scenario-1)

This section focuses on the comparison of the results obtained from TIAM used in a standalone manner (with elastic demands), and the results obtained from the coupled models TIAM-GEMINI-E3, (TIAM without elastic demands) in the case where the radiative forcing is limited to 3.5 W/m2 and all countries cooperate.

The global radiative forcing is limited to 3.5 W/m2. All countries and all emissions are covered by the climate agreement. In other words, this scenario represents a situation of perfect long-term cooperation between all countries, and the identified decisions are equivalent to the most efficient solution available to the World to limit the radiative forcing (First-best solution).

The iteration process for the coupling

The convergence of the climate scenario is obtained after 6 iterations (Figure 1). In the case of the reference case, the convergence criteria stays at a very low level (< 0.05%) what confirms that the two models were well harmonized before the coupling.

Figure 1: Evolution of the convergence criteria


The emissions obtained with TIAM-Elastic and the Coupled models remain very close over the time horizon (Figure 2). Emissions are very slightly smaller in TIAM-Elastic during the intermediate horizon, and very slightly higher at the very end (2050). The resulting radiative forcing follows a very close trajectory in both scenarios (Figure 3).

Figure 2: Comparison of emissions obtained with TIAM-Elastic and Coupled TIAM-GEMINI-E3 (case with a maximal radiative forcing of 3.5 W/m2)
Figure 3: Comparison of the radiative forcing obtained with TIAM-Elastic and Coupled TIAM-GEMINI-E3 (case with a maximal radiative forcing of 3.5 W/m2)

Demands for energy services

All sectors: The better representation of the factors influencing the demands for energy services is one of the expected added value of the coupling to TIAM and GEMINI-E3. It appears that the demands for energy services differ quite drastically between TIAM-Elastic and the Coupled models for some sectors (Figure 4).

Figure 4: Variations of the World demands for energy services (represented by an aggregated indicator)

First, agriculture, commercial, residential and road transport behave similarly in both approaches, with slightly higher reductions in the Coupled models in residential and non-road sectors, and slightly smaller reduction in commercial and agriculture. Second, the demands for non-road transport (aviation, navigation) are more dras tically reduced in TIAM-Elast (the even increase in the intermediate horizon in the Coupled models). Third, the industry sectors are detailed in Figure 5. While all industry demands decrease in TIAM-Elast, the dynamics of the variations obtained with the Coupled models appear to be more complex:

  • stronger reductions of the demands for chemicals products and iron & steel in the long term, with smaller reductions of the demands in the short term;
  • smaller reduction of the long-term demands for pulp & paper, non-ferrous products and other industry, with even an increase in the first part of the horizon;
  • same reduction of the long-term demands for non-metal, but with an increase in the first part of the horizon

Iron and Steel: The annual World demand for Iron & Steel decreases by until 14% in the Coupled models in 2050 against 8% in TIAM-Elast in 2050. The countries with the highest absolute and relative reductions of iron & steel production are China, India, which also are the highest producers. Some countries increase their production of Iron & Steel in the Coupled models: Australia, Eastern Europe, Japan, Other Developing Asia, South Korea, USA and Western Europe, what is not observed in TIAM-Elast where all regions decrease their production (Figures 6 and 7).

Figure 5: Variations of the World industrial demands
Figure 6: Variations of Iron & Steel in the different regions during the whole time horizon
Figure 7: Variations of the demands for Iron & Steel in 2050 on a regional basis

The changes in regional production as modeled in GEMINI-E3 result from:

  • changes of domestic consumption,
  • changes of export/imports.

In other words, under the requirement to reduce their emissions, countries can: either adapt their mode of production of Iron & Steel so that it becomes less GHG-emitting, or increase the imports of Iron & Steel from countries able to produce it in a low emitting mode, or decrease the domestic consumption, or decrease the exports (Figure 8). It appears that all regions decrease their domestic consumption of iron & steel, what could be related to the general result observed with TIAM-Elastic. The increase of production observed in several regions, as identified above, is motivated by the increase of their exports: those regions, which also reduce their domestic consumption, produce more with the aim of exporting and compensating for the decrease of production of other regions, mainly China and India. Indeed, the latter reduce their domestic consumption and their exports and increase their imports, all resulting in a decrease of their domestic production as observed above.

Figure 8: Variation of Iron & Steel consumptions and trade flows in 2050 (GEMINI-E3)

The analysis of energy dynamics helps understand these decisions. First, at the World level and under the climate constraint, the Iron & Steel sector is globally characterized by a substitution of coal by natural gas (Figure 9). A slight decrease of biomass consumption is also observed, compared with the Reference case, because the use of biomass for electricity production (with possibility of carbon capture, what means negative emissions since the biomass is considered as neutral in terms of GHG emissions on its life-cycle) is preferred. The decrease of the total energy consumed by the subsector is explained by the decrease of the total production of Iron & Steel, and, in a lesser extend, by the better effciency of gas technologies than coal technologies. The World energy efficiency of the production of Iron & Steel increases by 10% at the end of the horizon. Second, as regards China and India, it appears that given their high contribution to the World emissions (respectively 47% and 11% in 2050 in the reference case), any possible reduction options is implemented, including high reductions of domestic demands such as iron & steel. But why the increase of iron & steel production (and exports to China for example) in some other countries rather than the local production in China of iron & steel with clean energy and processes? One factor is that the remaining clean technical opportunities for China are quite reduced: the electricity is almost fully zero emissions, and the biomass potentials are fully used, contrary to some other countries where some biomass potentials, for example, remain. Therefore, other countries might be able to produce iron & steel in a cleaner way than China (more electricity, itself based for example on biomass plants with carbon capture).

Macroeconomic cost

Figure 10 shows the macro-economic costs computed on the basis of GEMINI-E3 outputs. We have computed various measures of this cost:

  • the sum of actualized variation of household consumption on the whole period of the simulation;
  • the sum of actualized surplus divided by the actualized household consumption of the baseline;
  • the sum of actualized variation of GDP.

Surplus is measured by the Compensating Variation of Income (CVI) or the Equivalent Variation of Income (EVI) that don't differ significantly from each other in the case of the climate change scenarios. As it is shown in Figure 10, the results are not modified noticeably between the different measures. We observe that:

  • the cost is important for developing countries especially for Latin America, India and China;
  • for developed countries with high energy intensity and which are energy importing countries, the cost is small, this is the case of European Union and Japan;
  • In contrary, for US, Canada, Australia the cost is significant but less important than the cost of developing countries.
  • These results are in line with previous studies and show that the implementation of a world carbon tax without tradable permits and initial allocation rule of burden sharing would not be acceptable for developing countries which bear an important amount of the worldwide cost of the climate policy. These results justify the implementation of a tradable permits scheme.
Figure 9: Energy consumption by the Iron and Steel sector
Figure 10: Cost variations between S1 and Reference in GEMINI-E3

Partial Cooperative Climate Agreement (Scenarios 2 and 2B)

On top of the Global Cooperative Agreement (previous section), two scenarios are modelled to represent other kinds of climate agreement and their underlying R&D agreements:

Scenario 2: Climate Agreement Limited to the Energy Intensive Industries: Using the same target as in Scenario 1 (3.5 W/m2), all sectors of the OECD countries are covered by the climate agreement while in Non-OECD countries, only energy intensive industries (including electricity generation and upstream) are covered. The aim is to avoid penalizing too much the households (residential and transport energy) but also to limit the loss of competitiveness of developed countries.

Scenario 2B: Climate Agreement Limited to the Electricty generation: All sectors of the OECD countries are covered by the climate agreement while in Non-OECD countries, only the electricity generation is covered. The modeling of scenario 2B with the same target as scenario 1 (3.5 W/m2) was infeasible. In other words, the participation of developing countries to the climate mitigation can not be limited to their electricity generation sector if the wanted radiative forcing is 3.5 W/m2, since the contribution of the other sectors of these countries to the global emissions is too high. Therefore, the target used for this scenario has been fixed to 4.0 W/m2. Our analysis shows that the smallest feasible radiative forcing would be 3.8 W/m2. S2B can hardly be compared to the others since the climate target is different.

The trade of permits between all regions is possible in these scenarios. The sectors not covered by the Climate agreement in Scenarios 2 and 2B still might indirectly react to the Climate constraint: either because of changes in energy prices and macro-economic factors, or because of the competition between different energy services for the use of energy.

CO2 price and total system costs - Scenarios 1, 2 and 2B

Not surprising, the CO2 price of S2, where only a part of the economy of non-OECD countries participate in the climate agreement, is 1.5 times higher than the price observed in S1: 421$/tCO2 compared to 286$/tCO2 in 2050 (Figure 11). Lower price of S2B is due to the less strict climate target.

Figure 11: CO2 price

As regards the total discounted cost of the system as provided by TIAM (Tables 1 and 2), the World cost is multiplied by more than 1.5 in S2 compared to S1: the cost of OECD countries is multiplied by 1.8 while the cost of Non-OECD countries is multiplied by 1.3. In other words, all regions, including the Non-OECD countries, face a higher total cost when only the Intensive Energy sectors of the Non-OECD countries participate in the climate agreement: the effort supported by the covered sectors is higher, resulting in more expensive strategies. The same dynamics are observed for China and India, compared to Western Europe: the cost increase between S2 and S1 is higher in Western Europe than in China and India, but the latter still bears a higher cost in S2B compared to S1.

Table 1: Discounted costs of OECD and Non-OECD countries from ETSAP-TIAM(see footnote 1)
Table 2: Discounted costs of World, China, India and Western Europe from ETSAP-TIAM (see footnote 1)

Macro-economic costs - Scenarios 1, 2 and 2B

The aim of scenario 2 is to avoid penalizing too much the households (mainly residential and transport sectors), in this case the relevant measure is the surplus. In Figure 12, one compares the surplus in the scenarios 1, 2 and 2B. As it can be expected, S2 is more costly for developed countries than S1. Indeed, in S2, the CO2 price increases by a factor 1.5 and this applies without exemption in all energy consumption, therefore the cost is much more important. In contrary, developing countries are welfare increasing with respect to S1: households are exempted from carbon taxation and benefit from a double dividend coming from the decrease of fossil fuel prices compared to the Reference. The gain coming from S2 is important especially for China and India (Figure 12). This result is different from those obtained from the TIAM model where the costs for developing countries also increase: TIAM does not reflect some macro-economic impacts modeled in GEMINI-E3.

Figure 12: Macro economic cost in S1, S2, S2b in GEMINI-E3 (surplus in % of household consumption)

At the World level, S2 is less efficient: the worldwide cost to reach the same emission target increase by 60%. As regards the variations of consumptions and trade, if we use the same example of Iron & Steel as analyzed in previous section, we observe that S2 follows the same dynamics as S1 (Figure 13). In contrary, S2B results in completely different variations since industry is not covered by the Climate agreement: developing and emerging countries, including China and India, reduce their imports and increase their exports compared to the Reference, while the opposite occur in OECD countries: there is delocalization of the production, as indicated previously.

Energy and technology decisions in the covered and non-covered sectors: Scenario 2

For the electricity sector, the comparison of S2 with S1 shows the following differences:

  • Higher electricity production at the World level (+27%), with a slightly higher increase in OECD (+34%) than in non-OECD (+25%) regions (Figure 14).
  • The increase of electricity production is satisfied by both renewable electricity generation and plants with CCS, with a slightly higher increase of renewable power plants, especially in OECD countries (Table 5).
  • In contrary, for US, Canada, Australia the cost is significant but less important than the cost of developing countries.
  • The generation by nuclear plants does not increase in S2 because it already reaches in S1 the maximal level allowed in the model as a way to represent socio-political limits in the penetration of nuclear power plants; in other words, nuclear is a competitive option to reduce GHG emissions given the data included in the database of course, it might be interesting to explore the consequences of an unlimited generation of nuclear power plants; however, such an unlimited nuclear generation is generally considered as unrealistic.

In OECD, the additional electricity is consumed mainly by the industry (substitution of gas) and the residential sector (substitution of oil) (Table 3). In Non-OECD, the additional electricity is consumed mainly by the industry sector, where electricity substitutes natural gas (Table 4).

Analysis of Scenario 2B

The analysis of the Scenario 2B is not as detailed as for Scenario 2 since it can not be directly compared to the other scenarios, given the different climate target used. In fact, one important result, as already explained above, is that the limitation of the covered sectors of Non-OECD countries to the electricity sector makes infeasible the limitation of the World radiative forcing to 3.5 W/m2. Our analysis shows that the smallest feasible radiative forcing would be 3.8 W/m2. A value of 4 W/m2 has been used to represent the Scenario 2B. The Non-OECD countries don't increase their electricity consumption compared to the Reference case, but of course, the structure of the electricity generation is modified in favor of low-emitting power plants, first of all CCS, much more than renewables (Figure 16).

Figure 13: Variation of Iron & Steel consumptions and trade flows in 2050 wrt Reference (GEMINI-E3)

As regards the macroeconomic cost, in S2B, even if the CO2 tax is slightly higher than in S1, the cost supported by developed countries is comparable to the cost computed in S1 (see Figure 12). For developing countries, we find that the cost decreases with respect to S1, but slightly increases for some countries with respect to S2: this is the case for China, India and rest of Asia, probably explained by higher oil prices in S2B than in S2 (the climate constraint is less strict, thus a higher consumption of oil in S2B compared to S2).

On the contrary, the cost supported by energy exporting countries falls: they are less negatively affected by S2B than by S2. S2B is certainly more acceptable by all parties but the climate target is less strict. S2B induces also pernicious impacts consisting in the delocalization of GHG emitting industries in developing countries as indicated in Figures 13 and 18.

Figure 14: Electricity generation
Table 3: Final energy consumption in OECD regions
Table 4: Final energy consumption in Non-OECD regions
Table 5: Share of electricity generation by type of plants in S1 and S2 in 2050
Table 6: Variations of the industrial productions between Reference and S2B
Figure 15: Emissions in end-use sectors
Figure 16: Electricity generation in S2B
Figure 17: Natural gas production, including associated gas
Figure 18: Steel production in scenario 1, 2 and 2b

1. In the Coupled approach, the system cost provided by TIAM does not include any costs related to the reduction of demands since TIAM is not used in the elastic mode, and the demands come from GEMINI-E3.


Contraction and Convergence [C&C] is the name of the global GHG emissions-management-model introduced by GCI to the UNFCCC negotiation in 1996. Some currect information about the origins, meaning and application of C&C are at these links: