Clean development mechanism: leverage for development?

Climate Policy 1 (2001) 251–268

Clean development mechanism: leverage for development? Sandrine Mathy∗ , Jean-Charles Hourcade, Christophe de Gouvello CIRED, Campus International du Jardin, Tropical de Paris, 45 bis avenue de la, Belle Gabrielle, 94736 Nogent sur Marne Cedex, France Received 22 July 2000; received in revised form 1 February 2001; accepted 1 March 2001

Abstract The objective of this paper is to show that the investments through the clean development mechanism (CDM) can exert a leverage effect to (i) make attractive to developing countries non-binding commitments and the adoption of national policies and measures; this comes as a guarantee of non-conditionally of the mechanism to strictly environmental concerns and (ii) create a flow of additional investments and technological transfer from Annex B countries to non-Annex B countries. The Indian power sector has been chosen as an example of a sector where institutional barriers, market imperfections, and tariff distortions create a great space for Pareto improvements and leave an important potential for no-regret measures: technological transfer, air conditioned systems, transport infrastructures and removal of subsidies on consumption. This paper presents a micro-economic formalisation on (i) the evolution of profitability of current emitting technologies used in the power sector under the adoption of national policies and measures and (ii) the impact on renewable energy technologies competitiveness of emission credits in the context of CDM. This formalisation has been developed to enable quantitative simulation. A first exercise using the Markal model (used in 77 countries) on the electric sector in India enabled us to simulate the leverage effect of emission credits on additional incomes taken as a proxy for development. © 2001 Published by Elsevier Science Ltd. Keywords: Clean development mechanism; Formalisation; Quantitative simulation

1. Introduction The clean development mechanism (CDM) is the product of a last minute ‘Kyoto surprise’ (Werksman, 2000), a semantic innovation that has lead to a substantial ambiguity in the positions of Annex B and non-Annex B countries. CDM was placed on the negotiation table in order to overcome the opposition of the G77 to the use of joint implementation (JI) in developing countries. Its original sin is well described by the chairman of the Kyoto conference, Estrada (1998): ∗ Corresponding author. Tel.: +33-1-43-94-7393; fax: +33-1-43-94-7370. E-mail address: [email protected] (S. Mathy).

1469-3062/01/$ – see front matter © 2001 Published by Elsevier Science Ltd. PII: S 1 4 6 9 - 3 0 6 2 ( 0 1 ) 0 0 0 1 3 - 4


S. Mathy et al. / Climate Policy 1 (2001) 251–268

I did not like the proposal but it got wide support and I facilitated its approval. My reservation was that the CDM is considered a form of joint implementation but I don’t understand how commitment can be implemented jointly if only one of the parties involved is committed to limit emissions and the other party is free from quantitative point of view. Such disparity has been at the root of every colonisation since the time of the Greeks. COP6 demonstrated that suspicion regarding the CDM still pervades the views of most G77 countries. The implication is that the CDM will gain real support from developing countries only if it can be demonstrated that, far from being a pure flexibility mechanism, it fulfils the following objectives by descending order as defined in article 12 of the Kyoto Protocol: • to assist non-Annex I countries in achieving sustainable development; • to assist non-Annex I countries in contributing to the ultimate objective of the convention; • to assist Annex I countries in achieving compliance with their quantified emission limitation and reduction commitments. This means that the CDM must be submitted to a double additionality condition: an environmental additionality requiring that all CDM abatements are emissions cuts which would not have occurred otherwise and a development additionality. A successful response to both these prerequisites would represent an unprecedented response to the long standing environment-development challenge raised during the preparation of the Stockholm conference in 1972: environmental protection demands an active participation of developing countries and this participation is impossible if environmental policies do not work synergistically with development policies. However, if there is a large literature on how to ensure environmental additionality (Sugiyama and Michaelowa, 2000; Janssen, 1998), the issue of the development additionality of the CDM is still under-worked despite useful insights (Thorne and La Rovere, 1999; Humphreys et al., 1998). In the first section, this paper will define the economic drivers for CDM leverage on the development of host countries. In the second section we develop a simple micro-economic model of the possible impacts of the CDM on technology and on foreign investments flows, on the top of which we will formalise the net leverage of emission credits on development. In the third section, we will tentatively capture the magnitude of this effect with a numerical experiment on the electricity sector in India. 2. Harmonisation space between flexibility and development policies It is often overlooked that a CDM project can be defined purely as an abatement project in only a few cases such as sequestration in non-occupied areas. In most of the cases, these investments cannot be analysed in isolation from the investment in energy, agriculture or transportation to which it is attached and from the externalities it generates. Three types of benefits should be considered jointly: the carbon benefit due to the gap between the international value of carbon and the unit cost of the CER yielded by the project, the commercial benefit, and the social benefit due the local positive environmental externalities or to the economic spill-over which may be considered by the public authorities of the host countries. One of the most repeated arguments against the CDM is that both of these side-effects may be negative (large scale dams, for example) and that the risk of generating social costs rather than benefits is high because of the asymmetry of negotiation power between investors and most of host countries. We will come back to this argument only in conclusion and consider, for the time being, that even though a lower

S. Mathy et al. / Climate Policy 1 (2001) 251–268


access to expertise prevents the host country from getting the biggest share of the rents associated with the project, the Protocol guarantees host state sovereignty in that it may define what type of CDM project it is ready to accept and what type of project will be rejected because of negative social costs. The co-existence of these three types of benefits is critically important; should a CDM project only yield a carbon benefit, the potential for a net development benefit for the host country would be weak. Indeed there would be no direct net surplus, if the foreign investment recovers the total cost of the abatement and all the CERs accrue to the investor; the only benefit would be then an upgrading of technical know-how. To demonstrate the mechanisms potential leverage on development we will start from a conventional macro-economic representation of the income f(I) generated by a level of investment I in an economy under given technical and institutional constraints. The curve f(I) describes the income generated by an amount I of cumulated investment if the individual investments are ranked by decreasing profitability order. The slope of the curve (f  (I ) > 0, f  (I ) < 0) represents the profitability of each marginal project and the last investment made in a country is represented at the point where the slope of the curve equates the prevailing discount rate. A higher level of income can be then generated either through lowering the discount rate or through an upward move of f(I) due to access to a new production frontier (Fig. 1).

Fig. 1. Income generated by investment.


S. Mathy et al. / Climate Policy 1 (2001) 251–268

The domestic discount rate is determined by the availability of capital in the economy and its level of indebtedness. In most of non-Annex B economies, this rate ireal South is higher than the optimal social discount rate iopt South that would prevail in a perfect capital market, which would equalise the discount rates all over the world. Indeed, these countries have to confront higher interest rates if they want to mobilise capital because they are indebted (the real interest rates on the capital markets increase with the external debt) and are perceived as presenting a ‘country risk’ higher than investments in OECD countries. The resulting level of investment is Ireal South < Iopt South yielding a lower net income. The production frontier f(I) results from technical endowments and institutional constraints including domestic market imperfections and tariff distortions in the public sector. In principle a government can move f(I) upwards by adopting domestic reforms among which those yielding the co-benefits of carbon saving projects. Those of these policies that lead to lower GHGs emissions levels are typically ‘no-regret’ policies that should be adopted regardless of climate concerns (for instance reducing coal subsidies induces efficiency improvements in coal-based thermal power stations that reduce local pollution). In the presence of such ‘no-regret’ policies, some projects that were formerly unprofitable do become economically viable while the profitability of others decreases, but the balance of the two effects is by definition positive. With the same discount rate, the level of investment moves to Imeasure South > Iopt South . The inflow of foreign investment has two related impacts: (a) it relaxes the capital constraint, lowers the domestic discount rate if the interest rate demanded by the foreign investor (including a risk premium) is iN < ireal South ; the volume of investment increases until r(I ) = i N and the output until INorth > Imeasure South ; (b) by providing more efficient technologies through technology transfer, the production frontier moves up from fmeasure (I) to fmeasure+transfer (I). This results in a new level of investment at Itransfer North > INorth . Note that the improvement of the net situation compared with the initial f(Ireal South ) is unrelated to the value of carbon. Typically, there are ‘no-regret’ policies, based on the relationship between national policies and capital inflows that should be implemented regardless of climate change concerns (IPCC, 1993, 1995). Obviously, the very fact that they have not been implemented so far demonstrate the existence of transaction and political costs inhibiting the removal of market barriers and institutional bias. The obstacles to the adoption of “Pareto improving policies” pointed out by Stiglitz (1998) do operate in this field because of the asymmetric mobilisation capacities of the losers and the winners (Jaffe and Stavins, 1994). The CDM can have a leverage effect on development by triggering the exploitation of potential no-regret policies that might otherwise remain frozen. It can do so through two interrelated channels: (a) technically, the creation and sharing of a carbon benefit through the CDM inflates the profitability of investments and moves up the curve to Itransfer+credit North > Itransfer North ; (b) this additional benefit provides an incentive to public authorities for confronting the transaction costs of Pareto improving policies since, it yields revenues to compensate the losers of such policies.

3. Micro-economic and macro-economic assessment of the leverage effect In this section we will try and disentangle the main drivers of the mechanisms sketched in the previous paragraphs. We will first concentrate on how project profitability is impacted by a public policy (a tax on fossil energies) and by the revenues from emission reduction credits for foreign investors. In a second step, we will generalise the results obtained for one project to a set of projects and a portfolio of investments. In a last step, we will examine the macro-economic impacts of the new allocation of investments.

S. Mathy et al. / Climate Policy 1 (2001) 251–268


3.1. Impact of the CDM on micro-economic choices Let S(t) be the demand for an energy service for a period of N years. This demand can be satisfied either by a carbon intensive technology (coal plant), or by a carbon free technology (wind power plant for instance). Though in the real world the criteria for decision making are more complex, for simplicity we will retain here the internal rate of return ρ(IRR) to model the investment behaviour: a project with an IRR lower than the discount rate observed in the country will not be realised and the same applies to choices of technology within a project. We have to examine how this IRR will be modified by national policies on one hand, and by the credits from the CDM yielded by foreign investments on the other. 3.1.1. Impact of a tax on the profitability of carbon intensive technologies A tax on fossil energies is used here as an illustrative example of domestic policies. This tax is applied to the energy producer. The increase of taxation induced by this new carbon tax is denoted by dT, the volume of carbon intensive energy being charged Ec ; it flows immediately that the additional energy expenditure Ec dT . In theory, the first response of an investor would be to respond to the tax burden on consumers by increasing the price of energy. To keep the demonstration simple without altering the substance of the conclusion, we will assume here that the energy prices are unchanged at this stage: in general, part of the burden of a tax falls on the producer in case of elastic demand and in a large subset of developing countries public authorities regulate energy prices and are reluctant to reflect tax increases in energy prices for equity and political reasons. 1 As a result, demand can be considered unchanged. The producer, if he does not abandon his project, will have a lower IRR as follows: 1 ρ(T ) = −1 (1) (1/(ρ0 + 1)) + (Ec /I )(T − T0 ) with ρ 0 as the internal rate of return with a tax level of T0 , before the tax increase. This equation shows that for a given tax, the higher the carbon intensity of the investment (Ec /I), the greater is the importance of the decrease of the internal rate of return. For the rest of this demonstration, we will assume that faced with a decline of the internal rate of return, a producer will nevertheless respond to the anticipated level of demand and will not reduce his investment, because the tax is not high enough to incentives a switch in investment to another sector (see Appendix A). 3.1.2. Impact of the credits on the adoption of carbon saving technologies Imagine the case of an Annex B investor. We assume that he controls an Annex B technology that is a perfect substitute to the technology locally available and otherwise employed by the national investor. Let ρ  be the internal rate of return of the carbon saving project. In each period the annual amount of carbon is reduced by this technology is Ec C (where C is the decline in carbon content compared to the carbon intensive energy and Ec the amount of energy which would have been consumed by the carbon intensive technology). If the value of credits is V, each year, receipts rise by Ec CV. 1 In India, energy consumption has long been considered as a public service. Consumption tariffs are far from reflecting production costs. Residential and agricultural tariffs are very low and will, as a result of equity considerations, remain low. Produced electricity is sold at a minimum price to state entities that in their turn distribute the electricity to customers at a specific price set in order to respect equity concerns. These prices are not high enough to cover production costs; but the government does not want to increase prices.


S. Mathy et al. / Climate Policy 1 (2001) 251–268

If ρ0 is the internal rate of return of the carbon saving project without remuneration from emission credits, the value of the internal rate of return with a value V for the credits is 1 ρ  (V ) = −1 (2)  (1/(1 + ρ0 )) − (CEc /I  )V These equations show that the internal rates of return of two technologies will get closer to one another. As the tax increases, ρ decreases and as V increases, ρ  rises. They show also that the more carbon intensive the baseline technology is, the more attractive the credit remuneration will be for the northern investor. These two results, which are valid for a single project, also apply to a program of substitution projects if ρ is interpreted as a global mean rate of return of the programme (see Appendix A). 3.2. Assessment of the leverage effect 3.2.1. Components of the leverage effect The increase in tax on fossil fuels will lower the profitability of carbon intensive projects and will at the same time increase the profitability of projects using carbon saving technologies. Thus, some projects that would have been realised by domestic investors in the reference scenario will be replaced by CDM projects. This substitution is not the only one to be considered. Indeed, CDM projects are realised with a northern contribution, and will come to substitute projects that would otherwise have been realised by southern investors. On the assumption that electricity demand is fixed as a public objective of the country, the global economic effect of the substitution between domestic and foreign investors makes domestic capital available for other investment opportunities. It is well known that since energy is a highly capital-intensive sector, it crowds out investment in other industrial development priorities. This is demonstrate, where a large part of the indebtedness of certain developing countries (Brazil) is due to the investment in the energy sector. At the macro-economic level, the leverage in developing countries is produced by this higher availability of capital. One can refer here to some kind of Keynesian investment multiplier. Every period, investments increase inducing an increase in production and so an increase in consumption. Evidently, this effect cannot be considered long-term, but rather on a period-by-period basis. To evaluate its magnitude it is necessary to estimate the amount of domestic investments that have been reoriented to non-power sectors by the entry of international investment thanks to the clean development mechanism and at which marginal efficiency. In the reference scenario (no tax increase and no CDM project), the income generated by investments in the power and non-power sectors is R(T0 , V = 0) =

RnE South + R(IE South (T0 ))       non-power sector power sector


Under the policy mix (tax and CDM), the general equilibrium effect of the CDM is then given by R2 (T , V ) = RnE South + RρOS (λIE South (T0 )) + R((1 − λ)IE South (T )) + R(IE North (V ))       power sector non-power sector +TEc (T ) − T0 Ec (T0 ) +    tax recycling

LE  less expenditures

which is composed of four basic components.


S. Mathy et al. / Climate Policy 1 (2001) 251–268


• The variation of the income from the power sector is as follows: ◦ CDM investments from northern countries generate the income R(IE North (V)). ◦ (1 − λ)IE South , which is the amount of domestic investments, which remains in the energy sector in the CDM scenario. Important for the mechanics of the leverage is the fact that, because of the tax, the least profitable projects are no more funded by domestic investments and the remaining domestic investments in the power sector have a higher internal rate of return. • The variation of the income of the non-power sector: this share of domestic investment made available by the inflow of foreign investments is supposed to be realised under a uniform internal rate of return equal to the rate of interest used in the country. A part λ of initial domestic investments is transferred to other sectors and induce the income ROS (λIE South (T0 )). • Recycling of tax in the economy captures the potential for no-regret measures and legitimates the adoption of the tax for public authorities. • The co-benefits: the decline in CO2 emissions reduces local negative externalities (NOx , particles, sulphur emissions) which lower the impact on health and the need for some public expenditure to offset these externalities. The estimate of public expenditures is a very sensitive matter because they generate ancillary benefits that are not easy to quantify (benefits linked with the decrease of pollutants which affect local populations for instance). Since the evaluation of such ancillary benefits is controversial (Krupnick, 1998) these benefits will not be considered here even if they may be of big importance in countries like China or India. Calculation of the leverage effect quantifies the additional income of one unit of credits, like a multiplier effect of credits on additional income: l=

R2 (T , V ) − R1 (T0 , V = 0) Ec CV


3.2.2. Calculation of northern investments Northern investors penetrate the market after solving Eq. (2) allowing the transfer of λIE (T0 ). They satisfy the demand λS. Since their technology is supposed to be more productive than technologies available in the host countries less investment is needed for a constant demand. Let σ be the technical efficiency improvement. Then the income is given by R(IE North (V )) = σ λ 

IE South (T0 )l t t (1/(1 + ρE North (V )) )


3.2.3. Calculation of domestic investments Power sector. Only a part 1 − λ of former domestic investments in the power sector in the reference scenario remains in the same sector under the CDM scenario. Thus, the difference between the incomes from the two scenarios is R((1 − λ)IE South (T )) − R(IE South (T0 ))   1−λ 1   = IE South (T0 ) − t t t (1/(1 + ρE South (T )) ) t (1/(1 + ρ0 ) )



S. Mathy et al. / Climate Policy 1 (2001) 251–268 Transfer of investments to non-power sectors. A part λ of initial domestic investments in the power sector IE (T0 ) is reoriented towards other sectors. A uniform internal rate of return ρ OS (here 12%) is considered for these investments: IE South (T0 ) RρOS (λIE South ) = λ  t t (1/1.12 )


3.2.4. Total additional income generated by the mechanism Finally the additional income R2 − R1 = I E (T0 )Π is given by Eq. (9) as follows: 1−λ 1 − t t t (1/(1 + ρSouth (T ))) t (1/(1 + ρ0 South ))


1 σ +λ  + t t t (1/1.12) t (1/(1 + ρNorth )(V ))


This literal expression shows the factors influencing the multiplier effect of credits regardless of the tax recycling and of the lower public expenditures to compensate pre-existing local externalities: • The gap between social marginal profitability of investments before and after tax (gap between ρ South (T) and ρ 0 South ). • λ the foreign share of total investments needed to meet the demand: λ has a negative effect on the income derived from domestic investments in the power sector because, assuming a constant demand, the bigger λ is, the smaller domestic investments will be. However, this effect is compensated for by positive variation of foreign investments, and by the transfer of domestic investments to other sectors. • V the value of carbon: the higher V is, the higher the northern investor’s internal return rate ρ North (V), and accordingly the bigger the income generated. This multiplier is not constant and changes each new period, depending on the multiplier of preceding periods. 3.3. Illustration — the case of the Indian power sector In this exercise, we will try to evaluate to potential for CDM with respect to the leverage on development and the decrease in CO2 emissions in the context of national policies and measures. This has been tested for the Indian power sector using the Markal model. The Markal model is a linear optimisation model representing the energy system at the level of a country or a region. For a given period (from 1995 to 2035) and a fixed energy demand, the model chooses the less expensive combination of technologies by minimising the discounted sum of the costs for investments, operations and maintenance. The Indian power sector is characterised by institutional obstacles, as well as political and social barriers commonly found in developing countries. The Indian situation can be considered sub-optimal. There is therefore great potential for policies that could improve the global performance of the system. In co-operation with our Indian colleagues, we selected four kinds of policies and measures currently under discussion in India. This does not mean that all of these measures would be politically acceptable, but at least they are not excluded from the debate.

S. Mathy et al. / Climate Policy 1 (2001) 251–268


The following policies and measures were selected: deregulation of the electricity production, privatisation of the electricity transmission and distribution sector, introduction of a tax on carbon intensive energies, and introduction of subsidies for renewable energies. 3.3.1. Simulation of alternative national policies and measures Given the model constraints, we represent the substitution between domestic and foreign technology through lower costs (a fixed rate of 90% of the southern technological cost has been used), a better load factor and an efficiency increase (quantity of output per unit of input). Note that in Markal, the command variable is in the hands of the northern investor who evaluates the attractiveness of the credit remuneration. This decision is the driving force of the model. However, the impact of this command variable is conditional upon the nature of national policies implemented. As we have already noticed, there is a large space for implementation of no-regret policies and measures. The main methodological difficulty is that Markal is not a forecasting tool, as it minimises the technological cost to reach a social optimum and does not capture the real behaviours under condition of market imperfections. This is why it does not express negative cost policies that could be exploited. We have overcome this difficulty by expressing these policies exogenously as a release of constraints. In this way, we can evaluate the specific impact of each measure and its political valuation with reference to the theoretical or exploitable impact of no-regret measures. Deregulation of the power sector. The effect of opening up electricity production to private investors has been translated in Markal according to Khanna and Zilberman’s results (1999) into an exogenous improvement in energy efficiency and productivity which happens more quickly under the CDM scenario than in the reference scenario. The effect is also represented by an increase in the load factor (Box 1). Better transmission and distribution. We assume a progressive removal of state interference and an optimistic evolution of electricity distribution management (a better recovery of outstanding payments), we suppose that losses decrease more rapidly until 2035 and get nearer from rates observed in northern countries (between 6 and 9%). Adjustment of electricity price on production costs. We assume the introduction of a tax on carbon intensive energies. Given the model constraints, results reflect a situation in which public entities are motivated by the redistributive effect of tax recycling, and where at an aggregated level, the electricity demand elasticity is zero, since the decrease in demand in sectors with a high price elasticity is compensated by the increase in demand by populations that benefit from redistribution (Box 2). We assume the tax is applied to all kinds of domestic energy except renewable energy sources: domestic coal, lignite, crude oil, and natural gas. In the reference scenario, fossil fuel prices are constant until 2035. To include resistance to the adoption of a tax on coal in a country like India, the scenario includes a progressive increase of the tax level from 0% in 1995 up to 20% in 2035 (<0.5% per year). This tax is too weak to reach important CO2 reductions, but significant enough to reduce external social costs as it induces a readjustment between domestic coal, and imported coal which is of far better quality and therefore emits less SO2 , NOx , and particles.


S. Mathy et al. / Climate Policy 1 (2001) 251–268

Box 1. The Indian energy system With a population of 1 billion, demographic growth of 2% per year (1400 million in 2025), and a high economic growth of 6.5%, security of energy supply is a main pre-occupations in the context of deficient markets and growing demands. During the next decade (IEA, 1999), India needs to invest US$ 285 billion (75% of its GNP) in the energy sector. The situation is critical: electricity prices have long been under governmental control in the so-called administered pricing mechanism, and have never reflected real production costs (Banerjee, 1997). Agricultural and residential tariffs are highly subsidised. Agricultural subsidies amount to Rs. 138 billion (US$ 3 billion) during the 1996/1997 fiscal year. These subsidies are compensated by high tariffs in the industrial and commercial sectors. However, this has not been sufficient and the government has covered the difference. SEBs rates of return are negative, implying that the construction of new plants is impossible. Power failures are current and often last several hours a day. Power production is principally coal-based at 69% (TERI, 1998) of the total. Coal supply is assured by Coal India Ltd. which operates an accumulated loss of US$ 800 million. Imported coal is highly taxed, and is therefore not used even where it is of better quality. Natural gas is an increasingly used resource as reserves are important. Joint venture opportunities for exploitation will soon be available. India is the biggest importer of oil with 51 Mt imported in 1996 (IEA, 1999). India is the fifth biggest primary energy consumer in the world and has an energy intensity which is twice as high as the world average. In contrast, commercial energy consumption is only about 0.3 t per capita which means one fifth of the world average. India is the fifth biggest CO2 emitter with 863 Mt in 1996, but one of the smallest emitters per capita: 0.91 Mt per capita. The power sector contributes to 48% of Indian emissions. In 1991, reforms were undertaken to open this sector to private participation. However, this participation is at the time of writing still very low. With more deregulation and private participation, the overall performance of the energy sector would improve: Khanna and Zilberman (1999) showed the influence of the mode of coal plant ownership both on the volume of intermediate consumption and on the utilisation factor. Thus, with mass participation of private and commercial actors, both productivity, energy efficiency, and load factor would be improved. They also showed that it would lead to the use of a coal of better quality (imported or not), mainly with a lower ash content. Such a shift would, according to this study, improve the productivity of coal from 25.6 to 26.93%, and would thus, decrease carbon dioxide emissions (2.5%). Transmission losses in India are estimated to be about 21% of the electricity generated. Efforts have been made to open this activity to private investors. There is a great potential for renewable energy sources in India, particularly for small hydro plants, wind power, and biomass. This potential is currently not exploited, although loan facilities and minimum guaranteed prices per kWh produced are offered to investors (TEDDY, 1998; CMIE, 1998). Nuclear electricity capacity exists in India, but is quite limited; and for political reasons this capacity is not expected to be built on. Subsidies for investment in renewable energies. In Markal, these measures can be translated into an investment cost reduction that will lead to a modification in the penetration rate of each technology. This method will be applied to a set of renewable energies: biomass, photovoltaic, solar hybrid, geothermal, wind power, hydroelectricity (small and high capacity plants), co-generation.

S. Mathy et al. / Climate Policy 1 (2001) 251–268


Box 2. The Markal model Indian Markal has been developed at the Indian Institute of Management of Ahmedabad by Prof. Shukla and A. Kanudia and refers to periods of 5 years until 2035 (Loulou et al., 1997). Required raw materials (inputs) are transformed by a set of technologies at the country’s disposal into intermediates or consumer goods. Mechanisms of extraction, conversion, or consumption are quantified for each technology (for instance, how many tons of input are required to produce 1 t of output). Thus, the model is a large panorama of about 40 technologies, linked together by energy and material flows. It computes a partial economic equilibrium of the energy system at each time period of 5 years until 2035 and consists of a set of equations and inequations, collectively referred to as the constraints, representing different demands for energy services in a desegregated manner (Loulou et al., 1997): • • • •

agriculture (irrigation, traction); industries (paper, textile, chemistry, manufacture, cement and steel production); commercial and residential (lighting, cooking, household); transport (fret, two and three wheels, passengers, air).

Once, the set of demands exogeneously fixed for each period, the energy system specified and its technological components described, Markal generates 7000 equations and inequations and constraints: • • • •

flow conservation (for each energy flow, the consumption must not exceed the supply); demand satisfaction; constraint on the capacity utilisation which must not exceed the installed capacity; capacity transfer (the capacity results from initial capacity plus previous investments which are still productive); • use of a resource must not exceed the annual capacity of its source; • growth constraint; • emission constraint. The objective function is the long-term total discounted sum of the energetic system cost required to respond to the demands. It can be written as TDSC = technology cost + import cost − exports revenue where technology cost is the discounted sum of all technological investments and all operation and maintenance costs.

3.3.2. Significant results Simulation results show a sensitive decrease of the total electricity generation to satisfy demands due to an increase in energy efficiency (less losses, better quality of coal, better management of plants). A new technology mix reduces the usage of coal from 69% in the reference scenario to 59% in the CDM scenario. Coal is mainly replaced by renewable energy sources. In 2035, renewable energies produce twice as much electricity in the CDM scenario as in the reference scenario. The level of investments in renewable energies remains quite stable until 2020 (Fig. 2). After this date, these energy sources become more competitive. Even if over the whole period considered in the


S. Mathy et al. / Climate Policy 1 (2001) 251–268

Fig. 2. Investments in the power sector.

simulations, the cost-efficiency ratio is better in the CDM scenario than in the reference scenario (where the penetration rate is lower), the investments in renewable energy become more productive earlier, and renewable energies therefore have a higher penetration rate. In a nutshell, technology substitution and efficiency improvements induce significant emission reductions as they start at −3% in 2005, due to equipment change inertia, and increase up to 16% in 2035 (Fig. 3). This result derives mainly from domestic technological reallocations, and northern stakeholder participation. A second significant result is the high participation of northern investors. As showed in Fig. 4, the impact of national policies results in a shift, but does not affect the aggregated level of national

Fig. 3. CO2 emissions in the reference scenario and in the CDM scenario.

S. Mathy et al. / Climate Policy 1 (2001) 251–268


Fig. 4. Domestic and international investments in the power sector.

investments. This result is partly due to the assumption of non-price elasticity of electricity demand (see the justification in Section 2). With constant electricity demand in the reference scenario and the abatement scenario, we can see an important increase of northern investor participation that reaches 8% of total investments in the power sector. This result is even more important than the technology transfer that happens only in non-coal-based production, i.e. 30% of the reference scenario production, and 40% in the CDM scenario. This share replaces investments that would have been carried through by national agents in the absence of CDM. They may now invest in other sectors, which is the real leverage effect of CDM. The valuation of this effect is very complicated without reference to the macro-economic context. In the absence of a general equilibrium model, however, we can try to evaluate it. It can be no more than little, because even if foreign investments are important for the power sector, they remain modest compared to the fix capital formation. The calculation of the indicator Π gives us the additional income induced by foreign capital flow and by more productive investments. L1 =

R2 − R1 = ΠIE South (T0 ) = 3 Ec CV


As one can see in the literal expression of Π (Eq. (9)), the two determining parameters, here are ρ E South which describes the profitability of “dirty” power plants, and ρ OS , which is the mean internal rate of return of investment transferred to other sectors. Of course the additional income is all the more important per unit of credit if we consider a high potential for no-regret policies in the CDM scenario. The more profitable the transfer of investments towards other sectors (mean internal rate of return ρ OS South of 12%), the higher the leverage of CDM on development. We have also calculated another kind of leverage that could be very interesting for Annex 1 investors: the ratio between foreign investments and the amount of credits generated by the mechanism


S. Mathy et al. / Climate Policy 1 (2001) 251–268

following: L2 =

IE North (V ) = 35 Ec V


This means that credits represent an average 3% of an investment. This represents an additional subsidy for foreign investors. These evaluations are really fragile, but the magnitude of the multipliers appears sufficient enough to make it impossible to exclude the potential leverage for development CDM represents from discussions concerning the CDM, and go beyond a purely sectoral approach. 4. Conclusion This paper aimed to provide both theoretical and empirical materials to prevent deadlock around the requirement of ‘developmental additionality’. This constitutes a prerequisite for the acceptance of the CDM by the G77 and most Annex B countries are reluctant to take more than a declaratory stance. In the view of Annex B countries, developmental additionality will lead to constraints that undermine the capacity of the CDM to deliver cheap CERs and arbitrarily increase the overall cost of their Kyoto commitments. The bottom-line of our argument was to demonstrate how the commercial benefit and co-benefit of a CDM project together with its carbon benefit enlarges substantially the negotiation space between the host and investing country. In the absence of an explicit general equilibrium model, assuming constant end-use demand, and with no account of local co-benefits, we do not pretend that any operational conclusion can be derived from the exercise carried out on the electric sector in India. However, the order of magnitude of the multiplier effect on growth is high enough to demonstrate the utility of further numerical investigation, and of the generalisation of the approach to other contexts. It confirms the intuition that the CDM, conceals a potential for environment and development benefits high enough to reorientate the debate on many contentious issues: • The sovereign authority of host countries over CDM investments: beyond the formal capacity to refuse certain projects, there this is room for pro-active policies by which government can attract CDM projects to increase the efficiency of development policies and government-promoted programs consistent with the objective of the FCCC (rural electrification, refurbishment of public facilities, etc.). • The trade-off between the integrity of CDM credits and the monitoring costs: because reductions are measured against inherently counterfactual baselines, and because it is difficult to set up precise ex-post measurements, there is trade-off between the inherent “leakiness” of the CDM, and the costs of scrutiny which might reduce the economic viability of projects, especially smaller-scale investments in the least developed countries. It should be examined whether, in case of program-based CDM (a) the aggregate potential error from misspecifying the baseline before the fact may be reduced thanks to baselines built on the bases of average technical references employed for each individual project and (b) risks of awarding unearned credits after the fact may be reduced by sampling instead of a project-by-project scrutiny. • Harvesting low-hanging fruit and support for structural policies; this problem can be mitigated by the capacity of the host country to make more attractive investments on equipment programs with a development impact through domestic policies. This is only apparently contradictory to the objective

S. Mathy et al. / Climate Policy 1 (2001) 251–268


of the foreign investor. The objective of the latter is to maximise the total of the commercial rent and the carbon rent from the projects with reference to parameters such as search costs, transaction costs, risks of entering in a totally new activity (tree plantations for example) uncertainties, certification risks, political risks and strategic opportunities of opening new markets in its activity domain. Beyond the semantic, this paper concludes that there is a conceptual distinction between joint implementation and the CDM. While the first is a carbon cost minimising tool, the second creates a negotiation space where a foreign investor aims at maximising the sum of a commercial benefit and a share of the carbon rent, while the host country aims at maximising local economic co-benefits and the remaining share of the carbon rent (Hourcade and Toman, 1999). This perspective opens a new suite of institutional issues (unilateral initiatives, capacity of the industry of the host country to gain credits, role of the ODA, capacity building) but we are inclined to think that this is the only route to transform the “Kyoto surprise” in an effective innovation.

Acknowledgements This study was conducted in co-operation with Pr. Shukla at the Indian Institute of Management in Ahmedabad, India. The authors would also like to express their gratitude to Martin Hession of Imperial College, London, for his assistance in corrections to the English grammar in this paper.

Appendix A Let S(t) be a demand for an energy service for a period of N years. This demand may be satisfied either by a carbon intensive technology (coal plant), or by a carbon free technology (wind power plant for instance). Lifecycles of these technologies are considered to be equal to the same duration N, to simplify calculations, and mathematical expressions. 2 Let F(t) be the annual net income (the difference between receipts and expenditures) (F (t) for a carbon intensive technology). We will suppose F(t) is constant over the whole period N. Let the internal return rate of one of these projects be denoted ρ, and x = 1 + ρ. The definition of ρ (or x) is I =F


N  1 1 x N+1 − 1 = F = F (1 + ρ)t xt x N (x − 1) t=0 t=0


If Φ = Φ(x, I, F ), Eq. (A.1) can be written as Φ(x, I, F ) = x N+1 I − x N I − FxN+1 + F = 0


We can easily differentiate this equation for each of the three command variables: the internal return rate ρ, the annual net income F, the investment I. We can thus calculate the conditions of decision neutrality 2 If technology lifecycles are different, we have to consider the same demand for the two kinds of service, but calculations remain the same (Mathy and de Gouvello, 1999).


S. Mathy et al. / Climate Policy 1 (2001) 251–268

for the investor if at least one of these three parameters, according to a specific project, change.

∂Φ ∂Φ ∂Φ dI + dF + dx = 0 ∂I F,x constants ∂F I,x constants ∂x F,I constants

∂Φ = x N+1 − x N = 0 ∂I F,x constants

∂Φ = −x N+1 + 1 = 0 ∂F I,x constants

∂Φ = [(N + 1)x N − NxN−1 ]I − (N + 1)x N F = 0 ∂x F,I constants

(A.3) (A.4) (A.5) (A.6)

After simplifications, we can inject the Eqs. (A.4)–(A.6) in Eq. (A.3) and we have [x N+1 − x N ] dI + [−x N+1 + 1] dF + [x N−1 I ] dx = 0


A.1. Impact of a tax on the profitability of carbon intensive technologies This tax is applied to the energy producer. If the level of the tax T is dT on the energy Ec , expenditures will be up Ec dT. We obtain with dI = 0 and Eq. (A.7) the variations of x: dx Ec −x N+1 + 1 = dT I x N−1 Therefore, if x0 is the internal rate of return with a level of tax T0 , before the increase in tax, ρ(T ) = x(T ) − 1 =

1 −1 (1/(ρ0 + 1)) + (Ec /I )(T − T0 )



A.2. Impact of the credits on the adoption of carbon free technologies Let ρ  (ρ  = 1+y) be the internal rate of return of a carbon free project. As we wrote, y counterbalances the net present value of the project of investment I , where F is the annual net income considered as constant during the whole project. We can lead the same argumentation as before: the annual amount of carbon reduces by Ec C (where C is the decline in carbon content compared to the carbon intensive energy) each period, if this technology is used to replace the carbon intensive technology. If the value of credits rises by dV, each year receipts rise by Ec C dV. So dF  = CEc dV and dy CEc −y N+1 + 1 =  , dV I y N−1

dy CEc ≈  dV 2 dy I


If ρ0 is the internal rate of return of the project without remuneration of the emission credits (V = 0), the value of the internal rate of return with a value V of the credits is 1 ρ  (V ) = y(V ) − 1 = −1 (A.11)  (1/(1 + ρ0 )) − (CEc /I  )V

S. Mathy et al. / Climate Policy 1 (2001) 251–268


where the quantity of carbon not been emitted is equal to the quantity of energy used to respond to the same demand with the carbon intensive technology. A.3. Generalisation to a set of projects Actually, we can demonstrate that, there is a global intern rate of return ρˆ for a set of K investments projects (Ik )k=1,... ,K (they all have a constant annual net income (Fk )k=1,... ,K and an internal rate of return (ρk )k=1,... ,K ). This project is defined by a global investment of Γ , and a constant annual income over the whole period: xk = 1 + ρk , Γ =F


∀(xk )k=1,... ,K , ∃ρˆ

1 , (1 + ρ) ˆ t t=0

where Γ =

  Ik , F = Fk k



Each project consumes an annual amount of energy Eck . So, if the tax on energy increases, expenditures  increase: dF = k Eck dT = Ec dT . We can thus apply the preceding results to ρˆ if we consider a set of national projects in the power sector and if we assume a tax on fossil fuels. 1 1 Ec − (T − T0 ) = ρ(T ˆ ) ρˆ0 I


where ρˆ0 is the initial internal rate of return and T0 is the initial tax level. References Banerjee, R., 1997. Electricity pricing, policies and DSM in India. In: Shukla, P.R. (Ed.), Energy Strategies and Greenhouse Gas Mitigation. Allied Publishers Limited, New Delhi, pp. 108–122. Center for Monitoring Indian Economy (CMIE), 1998. Current Energy Scene in India, Mumbai. Estrada, R., 1998. First approaches and unanswered questions. In: Issues and Options: The Clean Development Mechanism. UNDP, 1998, pp. 23–29. Hourcade, J.C., Toman, M., 1999. Policies for the Design and Operation of the Clean Development Mechanism. In: Summary of Discussions at Multinational Workshop, Centre International de Recherche sur l’Environnement et le Développement (CIRED), Paris, 24–25 September 1999. Resources for the Future (RFF), Washington, DC. Humphreys, Sokona, Thomas, 1998. Equity in the CDM, ENDA TM, Dakar. IEA, 1999. Draft Report on India’s Energy Challenges and Recent Policy Development. IEA/NMC, p. 6. Intergovernmental Panel on Climate Change (IPCC), 1993. Second Assessment Report. Intergovernmental Panel on Climate Change (IPCC), 1995. Working Group III, Chapters 8 and 9. Jaffe, A.B., Stavins, R.N., 1994. The energy paradox and the diffusion of conservation technology. In: Resource and Energy Economics, Vol. 16, pp. 91–122. Janssen, J., 1998. Strategies for risk management of joint implementation investments. In: Riemer, P., Smith, A., Thambimuthu, K. (Eds.), Greenhouse Gas Mitigation: Technologies for Activities Implemented Jointly. Elsevier, Amsterdam, pp. 357–365. Khanna, M., Zilberman, D., 1999. Barriers to energy-efficiency in electricity generation in India. Energy J. 20 (1), 25–40. Krupnick, 1998. Climate Change, Health Risks and Economics, Weathervane Feature available on http://www.weathervane. Loulou, R., Shukla, Pr., Kanudia, A., 1997. Energy and Environment Policies for a Sustainable Future. Allied Publishers Limited, New Delhi.


S. Mathy et al. / Climate Policy 1 (2001) 251–268

Mathy, S., de Gouvello, C., 1999. Le Mécanisme de Développement Propre: Mécanisme de flexibilité pour les engagements de réduction d’émission des pays de l’Annexe1 ou levier pour le développement des pays en voie de développement. Communication pour le colloque de Strasbourg, 1–2 December 1999, les journées du PIREE. Stiglitz, J., 1998. Distinguished lecture on economics in government: the private uses of public interests: incentive and institutions. J. Econ. Persp. 12 (2), 3–22. Sugiyama, T., Michaelowa, A., 2000. Reconciling the design of CDM with inborn paradox of additionality concept. Central Research Institute of Electric Power Industry (CRIEPI) Working Paper no. Y00905, Tokyo. Tata Energy Research Institute (TERI), 1998. Tata Energy Data Directory and Yearbook (TEDDY), New Delhi. Thorne, S., La Rovere, E., 1999. Criteria and Indicators for Appraising Clean Development Mechanism (CDM) Projects. Helio International. Werksman, J., 2000. The clean development mechanism: unwrapping the Kyoto surprise. In: Chichilnisky, Heal (Eds.), Environmental Markets, Equity and Efficiency, Columbia University Press, 20 pp.