www.elsevier.com/locate/enconman

Energy Conversion and Management 48 (2007) 1975–1982

Economy wide emission impacts of carbon and energy tax inelectricity supply industry: A case study on Sri Lanka

K. Siriwardena a, P.D.C. Wijayatunga a,*, W.J.L.S. Fernando a,R.M. Shrestha b, R.A. Attalage a

a Sri Lanka Energy Managers Association, 29, Fairfield Gardens, Colombo 8 Sri Lankab Asian Institute of Technology, PO Box 4, Klong Luang, Pathumthani 12120, Thailand

Received 15 October 2004; received in revised form 7 July 2006; accepted 30 January 2007Available online 27 March 2007

Abstract

This paper presents the results and analysis of a study conducted with the objective of investigating the impact on economy wide emis-sions due to carbon and energy taxes levied within the electricity generation sector of Sri Lanka. This exercise is mainly based on theinput–output table developed by the national planning department. An input–output decomposition technique is used to analyze fourtypes of effects that contribute to the overall reduction in equivalent carbon, NOx and SO2 emissions. These four effects are: fuel mixeffect (i.e. the change in emissions due to variation I fuel mix), structural effect (i.e. change in emissions due to changes in technologicalcoefficients with taxes compared to that without taxes), final demand effect (i.e. the change in emissions associated with changes in finaldemand) and joint effect (i.e. the interactive effect between or among the fuel mix, structural and final demand effects).

The polluting fuel sources and low energy efficiency generation technologies are less preferred under these tax regimes. Of the foureffects, a change in fuel mix in thermal electricity generation and a change final demand for electricity were found to be the main con-tributors in achieving economy wide emission reductions. It was found in the analysis that a minimum of US$ 50/tC tax or US$ 1.0/MBtu of energy tax is required to have a significant impact on economy wide emissions in the Sri Lankan context. This translates intoan overall increase in electricity generation cost of approximately USCts 0.9 kW�1 h�1 and USCts 0.6 kW�1 h�1 under the carbon andenergy tax regimes, respectively. The reduction in emissions is also strongly coupled with the value of the price elasticity of electricity.� 2007 Elsevier Ltd. All rights reserved.

Keywords: Energy efficient technologies; Carbon tax; Energy tax; Greenhouse gas emissions

1. Introduction

Excessive CO2 (carbon dioxide) emission is a globalproblem calling for international coordinated actions.Therefore, to achieve a certain target for CO2 emissionsrequires creation of incentives to reduce CO2 emission bycountry and sector in a cost effective way.

Carbon and energy taxes have attracted attention andare commonly recognized as an economic instrument toachieve greenhouse gas (GHG) emission mitigation. Also,the GHG mitigation potentials of such policy measures

0196-8904/$ - see front matter � 2007 Elsevier Ltd. All rights reserved.

doi:10.1016/j.enconman.2007.01.030

* Corresponding author. Tel.: +94 11 2392606; fax: +94 11 2392641.E-mail address: priyantha@pucsl.gov.lk (P.D.C. Wijayatunga).

are yet to be assessed in the case of most Asian countries.Therefore, it is imperative to assess the potential role ofcarbon and energy taxes for GHG and other harmful envi-ronmental emissions in Sri Lanka.

By levying tax on CO2 emissions and, thus, on the burn-ing of fossil fuels, consumers are motivated to utilize energymore efficiently and to substitute away from the most CO2

intensive fuels in the power sector. Also, imposition of a taxleads to a cost effective allocation of CO2 emissions. Thetaxes, in addition to reducing GHG emissions, are the addi-tional revenue generation that can be utilized to subsidizepower generation using renewable energy sources.

The electricity generation system in Sri Lanka is gradu-ally moving toward domination of thermal generation,

1976 K. Siriwardena et al. / Energy Conversion and Management 48 (2007) 1975–1982

mainly based on petroleum fuels and coal. This situationresults in a gradual rise of GHG and other harmful envi-ronmental emissions in the power sector.

The objective of this study is to investigate the impact ofa possible carbon tax or energy tax imposed on the electric-ity generation sector on the overall emission levels of thecountry.

After Leontief and Ford’s [1] introduction of the use ofthe input–output (I–O) model for computing pollutantemissions from energy consumption, Miller and Blair [2]extended the model to include multiplier effects and inter-industry activities. A few studies have used an input–output decomposition method in the energy sector. Outof those studies; Chang and Lin [3] and Marpaung [4] havedone similar work. Chang and Lin have studied the varia-tion of CO2 emissions in Taiwan from 1981 to 1991, andMarpaung’s work deals with emissions constraints andtaxes in the Indonesian power sector. Both of these studieshave found the final demand effect to be the largest contrib-utor for emission reduction. Also, Rose and Chen [5], Chenand Wu [6], Han and Lakshmanan [7] and Labandeira andLabeaga [8] provide useful examples of applying input–output decomposition methods.

2. Methodology

The basic equation used in the input–output method [2]is the following:

X ¼ AX þ Y ; AX ¼ Intermediate Demand;

where ‘X’ is the vector of output of producing sectors infinancial terms (US$) and ‘A’ is the matrix of technologicalcoefficients while ‘Y’ contains the final demand in financialterms (US$) in each of those producing sectors

Existing I-O table

Data related to I-O sectors

Modification of I-O table

GDP growth

Updating I-O table

Data on fuel consumption by each I-O sector

Fuel price

Genew

I-

Contribution of ep

) X ¼ ðI � AÞ�1Y and if L ¼ ðI � AÞ�1;

X ¼ LY ;

where ‘I’ is the identity matrix and ‘L’ is known as Leon-tief’s matrix.

If ‘C ’ is the matrix of direct fuel requirement coefficientsdefined as fuel use per unit of total output of a sector in kg/US$ then

Total fuel consumed ¼ C0LY :

If E consists of all the emission factors, the pollutant emis-sions are given by

Pollutant emissions ¼ E0C0LY :

The change in emission with (terms with subscript T)and without (terms with subscripts 0) a tax regime suchas carbon or energy tax is given by Ref. [4]

Change in emissions ¼ E0C0TLTY T � E0C00L0Y 0

¼ E0C00L0DY þ E0 C0TLT � C00L0

� �Y T

ðaÞ ðbÞ

Ddenotes changes in the associated quantities

E0 C0TLT�C00L0

� �Y T¼E0 C00þDC

� �L0þDL½ ��C00L0

� �Y 0þDY½ �

¼E0 C00DLþDCL0þDCDL� �

½Y 0þDY �¼E0C00DL Y 0þE0DC L0Y 0þE0DCDLY 0

þE0C00DLDY þE0DCL0DY þE0DCDLDY

The component (a) of the total change in emissions,E0C00L0DY , can be named final demand effect, and thisis further divided into final demand effect due to changein construction of power plants, E0C00L0DY CPP, and finaldemand effect due to change in electricity final demand,

Existing & Candidate plant data

Electricity demand

Emission factors

ration expansion plan with & ithout carbon/ energy tax

O analytical framework formulation

ach component to total change in ollution emissions

K. Siriwardena et al. / Energy Conversion and Management 48 (2007) 1975–1982 1977

E0C00L0DYEFD. The component (b) is further subdividedinto six parts as shown above.

All these components can now be identified and namedas given below [4].

Component

Estimating equation

Power plant construction –final demand

E 0C0(t) 0L0(t)DY(t)CPP

Electricity final Demand

E 0C0(t) 0L0(t)DY(t)EFD

Fuel mix effect

E 0DC(t) 0L0(t)Y0(t) Structural effect E 0C0(t) 0DL(t)Y0(t) Joint effects E 0DC(t) 0DL(t) Y0(t)

+E 0DC(t) 0L0(t) DY(t)+E 0C0(t) 0DL(t) DY(t)+E 0DC(t) 0DL(t)DY(t)

The framework of the total study using the above I–Oanalysis is shown in the following block diagram [15].

The study used MATLAB for simulation of the aboveframework in the context of the Sri Lankan economy.

3. Assumptions

Exports are treated as part of final demand and importsare ignored; this was also adopted by Gay and Proops [9]and Proops et al. [10]. Further, the change in electricitydemand resulting from all economic feedbacks is not con-sidered in this study. That is, multiplier effects are ignored.The study presented in the paper used the I–O table of 1994in the absence of a later version.

The value of fuel used per unit output is assumed con-stant except for thermal electricity generation throughoutthe planning period. Final demand effects due to thechange in demand for end use equipment were ignoreddue to the difficulty in sourcing reliable related data.The impacts of demand side management (DSM) pro-grams in the electricity sector were also not consideredfor the study.

Table 1Emission factors for different fuels used in electricity generation

Fuel C (t/kg) NOx (kg/kg) SO2 (kg/kg)

Fuel oil 0.000848 0.049162 0.070002Diesel oil 0.000875 0.011749 0.010000Kerosene 0.000877 0.002654 0.006269Gasoline 0.000847 0.028109 0.000410LPG 0.000814 0.003312 0.000000Fuel wood 0.000000 0.000000 0.000000Coal 0.000700 0.007517 0.009499Naphtha 0.000934 0.012540 0.000000LNG 0.000798 0.000000 0.000000

4. Input data

4.1. Existing generation capacity

Existing generation capacity consisted of all the gridconnected power plants in the country. They amount to[11],

Hydropower – 1185 MW (large and small grid connected)Thermal – 828 MW (diesel, steam, gas and combinedcycle)Wind – 3 MW (pilot wind farm established in 1999)

4.2. Candidate plants

All the candidate thermal and hydroplants consideredby the Ceylon Electricity Board (CEB), Long term Gener-ation Expansion Plan [11] were used. Demand forecastswere also based on the same reference [11].Thermal

• Oil, coal and fuel wood fired steam• Oil fired gas turbines• Oil and LNG fired combined cycle

Hydropower

• Four large hydro plants in four river basins

Distributed power

• Grid connected mini-hydro and wind plants

4.3. Scenarios of analysis

Scenario analysis consisted of examining the outcomebased on varied price elasticity of demand, carbon taxand energy tax levels. They are the following.

Price elasticityof electricitydemand

�0.20, �0.33,�0.40, �0.75

Carbon Tax(US$/tC)

5,10, 25, 50, 100,200

Energy Tax(US$/MBtu)

0.5, 1.0, 2.0, 5.0,10.0

The elasticity value, �0.33, is based on the study byJayatissa [12] for the Sri Lankan electricity demand. Sensi-tivity studies were conducted around this elasticity from�0.2 to �0.75. The energy tax was applied on all the gen-erating options except distributed power plants, which areembedded in the distribution network.

Table 2Economy wide reduction in emissions with carbon tax from 2006 to 2025

Priceelasticity

Carbon tax rate(US$/tC)

Total CO2

mitigation (106 kg)% CO2

mitigationTotal SO2 mitigation(103 kg)

% SO2

mitigationTotal NOx

mitigation (103 kg)% NOx

mitigation

�0.20 5 52 0.017 34 0.003 53 0.00610 135 0.045 689 0.068 698 0.07525 1034 0.346 9112 0.902 4145 0.44750 10350 3.458 55199 5.464 29302 3.163

100 43448 14.517 306697 30.359 193817 20.919200 54477 18.202 373590 36.981 247201 26.680

�0.33 5 56 0.019 36 0.004 57 0.00610 138 0.046 708 0.070 717 0.07725 1060 0.354 9338 0.924 4248 0.45850 10675 3.567 56930 5.635 30221 3.262

100 45210 15.105 319137 31.591 201679 21.767200 54514 18.214 373840 37.006 247365 26.698

�0.40 5 61 0.020 40 0.004 63 0.00710 146 0.049 748 0.074 757 0.08225 1082 0.361 9530 0.943 4335 0.46850 10856 3.627 57896 5.731 30733 3.317

100 45868 15.325 323778 32.050 204612 22.084200 54935 18.355 376730 37.292 249278 26.905

�0.75 5 71 0.024 46 0.005 72 0.00810 209 0.070 1067 0.106 1081 0.11725 1300 0.434 11453 1.134 5210 0.56250 12611 4.213 67256 6.658 35702 3.853

100 51407 17.176 362881 35.921 229323 24.751200 62551 20.899 428959 42.462 283837 30.635

1978 K. Siriwardena et al. / Energy Conversion and Management 48 (2007) 1975–1982

4.4. Emission factors

The emission factors are based on the guidelines of theIntergovernmental Panel on Climate Change (IPCC) andthe pollution inventory data of the country. They wereassumed constant for the sectors of the I–O table. Theseemission factors are given in Table 1.

Table 3Economy wide reduction in emissions with energy tax from 2006 to 2025

Priceelasticity

Energy tax rate (US$/MBtu)

Total CO2

mitigation% CO2

mitigation

�0.20 0.5 2095 0.7001.0 5877 1.9642.0 11783 3.9375.0 51301 17.140

10.0 64308 21.486

�0.33 0.5 2168 0.7251.0 5986 2.0002.0 12361 4.1305.0 51814 17.312

10.0 66543 22.233

�0.40 0.5 2187 0.7311.0 6095 2.0362.0 12385 4.1385.0 52156 17.426

10.0 66662 22.273

�0.75 0.5 2491 0.8321.0 6617 2.2112.0 13952 4.6625.0 55353 18.494

10.0 86438 28.880

4.5. Input–output table

The original Sri Lanka I–O table of 1994 consists of58 sectors; this was disaggregated to 61 sectors byseparating electricity and water and by furthersubdividing electricity into thermal generation,hydrogeneration and transmission and distribution. It

Total SO2

mitigation% SO2

mitigationTotal NOx

mitigation% NOx

mitigation

8300 0.822 5473 0.59131552 3.123 17631 1.90352136 5.161 31133 3.360

236474 23.408 139259 15.030329841 32.650 209388 22.599

8590 0.850 5663 0.61132135 3.181 17957 1.93854692 5.414 32659 3.525

238838 23.642 140651 15.180341309 33.786 216668 23.385

8662 0.857 5711 0.61632721 3.239 18284 1.97354798 5.424 32723 3.532

240418 23.799 141582 15.281341917 33.846 217054 23.427

9869 0.977 6507 0.70235524 3.517 19851 2.14361732 6.111 36863 3.979

255151 25.257 150258 16.217443351 43.887 281446 30.376

Table 4Decomposed economy wide reduction in CO2 emissions with carbon tax from 2006 to 2025

Priceelasticity

Carbon tax rate(US$/tC)

Fuel mix effect(FM)

Structuraleffect (ST)

Final demand effect (FD) Joint effects(JE)

Total mitigation(FM+ST+FD+JE)

%mitigationPower plant

constructionElectricity

�0.20 5 31 0 0 21 0 52 0.01710 76 0 0 59 0 135 0.04525 686 1 0 347 0 1034 0.34650 9296 13 1 1043 �2 10350 3.458

100 42012 76 1 1381 �22 43448 14.517200 51985 94 3 2403 �7 54477 18.202

�0.33 5 33 0 0 22 0 56 0.01910 78 0 0 60 0 138 0.04625 703 1 0 356 0 1060 0.35450 9587 14 1 1075 �2 10675 3.567

100 43716 79 1 1437 �22 45210 15.105200 52020 94 3 2404 �7 54514 18.214

�0.40 5 37 0 0 24 0 61 0.02010 83 0 0 64 0 146 0.04925 717 1 0 363 0 1082 0.36150 9750 14 1 1094 �7 10856 3.627

100 44352 80 1 1458 �23 45868 15.325200 52422 95 3 2423 �7 54935 18.355

�0.75 5 42 0 0 28 0 71 0.02410 118 0 0 91 0 209 0.07025 862 2 0 436 0 1300 0.43450 11326 16 1 1270 �2 12611 4.213

100 49708 90 1 1634 �25 51407 17.176200 59690 108 3 2759 �8 62551 20.899

Table 5Plant additions without tax from 2006 to 2025

Year Plant additions – Base case Capacity (MW)

2008 Coal – Tr. · 1 3002009 Coal – Tr. · 1 3002011 Coal – Tr. · 1 3002013 Coal – Tr. · 1 3002014 Coal – Tr. · 1 + Gas Turbine · 1 300 + 352015 Coal – Tr. · 1 3002016 Coal – Tr. · 1 3002017 Coal – Tr. · 1 + Gas Turbine · 1 300 + 352018 Gas Turbine · 1 1052019 Coal – Tr. · 1 3002020 Coal – Tr. · 1 + Gas Turbine · 1 300 + 352021 Coal – Tr. · 1 + Gas Turbine · 1 300 + 1052022 Coal – Tr. · 1 + Comb. Cycle · 1 300 + 3002023 Coal – Tr. · 1 3002024 Comb. Cycle · 1 + Comb. Cycle · 1 150 + 3002025 Gas Turbine · 2 + Coal – Tr. · 1 + Comb.

Cycle · 135 · 2 + 300 + 150

Total addition 5485

Note: Coal – Tr. – Traditional coal fired plants.

K. Siriwardena et al. / Energy Conversion and Management 48 (2007) 1975–1982 1979

was assumed that all the generation from the thermaland hydro sectors are intermediate demand of transmis-sion and distribution. The final demand was predictedover the planning period based on projected sector wisegross domestic product (GDP) growth rates obtainedfrom the annual report of the Central Bank of Sri Lanka[13], except for electricity transmission and distribution,which was based on the generation expansion plan ofthe CEB [11].

4.6. Fuel consumption

The fuel consumption by each sector of the economywas calculated using the energy balance table [14] andannual survey of industries [15]. The energy consumed bythermal electricity generation was amended, yearly, basedon the generation expansion plan [11].

5. Results and analysis

Table 2 shows that when a carbon tax regime is inplace within the Sri Lanka power sector, the economywide reduction in emission becomes significant only fora tax above US$ 50/tC in the case of the carbon taxregime. This occurs due to two main reasons. One isdue to the change in electricity price becoming significantonly above this level of carbon tax to decrease the finaldemand, causing a drop in emissions. The second reason

is that the tax is high enough to move the technologystructure of the generation sector for it to become signif-icantly cleaner (structural effect) but expensive. This isdue to the lack of suitable less expensive cleaner technol-ogies from which to select in the generation expansionplan, both in number and capacity. The emission mitiga-tion is not increased as expected from US$100/tC to

Table 7Plant additions with $5.0/MBtu energy tax from 2006 to 2025

Year Plant additions – At $5.0/ MBtu & �0.33elasticity

Capacity(MW)

2007 Wind · 3 902008 Mini-hydro · 2 302009 Supercritical · 1 4002012 IGCC · 1 3002014 Supercritical · 1 4002015 IGCC · 1 + Wind · 2 300 + 602016 Comb. Cycle · 1 3002017 Gas Turbine · 2 35 · 22018 Comb. Cycle · 1 + Gas Turbine · 1 150 + 1052019 Comb. Cycle · 1 + Gas Turbine · 1 150 + 1052020 Comb. Cycle · 1 3002021 Coal – Tr. · 1 3002022 Coal – Tr. · 1 + Comb. Cycle · 1 300 + 1502023 Coal – Tr. · 1 + Gas Turbine · 1 300 + 352024 Coal – Tr. · 1 + Gas Turbine · 2 300 + 35 · 22025 Coal – Tr. · 1 300

Total addition 4515

Note: Coal – Tr. – Traditional Coal fired plants.

Table 6Plant additions with $200/tC carbon tax from 2006 to 2025

Year Plant additions – At of $200/tC & �0.33elasticity

Capacity (MW)

2007 Wind · 3 902009 Mini-hydro · 2 302011 Supercritical · 1 4002013 Wind · 2 602014 IGCC · 1 + Gin ganga (Hydro) 300 + 492015 Supercritical · 1 4002016 Dendro · 3 10 · 32017 IGCC · 1 + Gas Turbine · 2 300 + 35 · 22018 Gas Turbine · 1 + Gas

Turbine · 2 + Dendro · 235 + 35 · 2 + 10 · 2

2019 Comb. Cycle · 1 + Broad lands (Hydro) 300 + 402020 Comb. Cycle · 1 + Dendro · 5+ Wind · 2 300 + 10 · 5 + 30 · 22021 Comb. Cycle · 1 + Gas Turbine · 1 300 + 1052022 Comb. Cycle · 1 + Gas Turbine · 1 300 + 352023 Wind · 2 + Comb. Cycle · 1 + Gas

Turbine · 230 · 2 + 300 + 35 · 2

2024 Comb. Cycle · 1 + Comb. Cycle · 1 300 + 1502025 Comb. Cycle · 1 · 1 + Uma Oya (Hydro) 300 + 150

Total addition 4639

1980 K. Siriwardena et al. / Energy Conversion and Management 48 (2007) 1975–1982

US$200/tC due to reduced demand leading to highergeneration contribution from existing plants.

The percentage NOx (nitrogen oxides) and SO2 (sul-phur dioxide) reduction is higher compared to that ofCO2 (carbon dioxide). Most of the reduction in CO2 emis-sions occurs due to switching of coal based generation toDiesel or fuel oil based generation. Although the percent-

Table 8Generation by fuel type from 2006 to 2026 with carbon tax

Price elasticity Carbon tax rate (US$/tC) Generation (GW h)

Diesel Residual oil Furn

0 50267 10848 4350�0.20 5 49492 10848 4350

10 52132 10848 435025 50650 10848 435050 57797 10848 4350

100 172602 10848 4350200 171230 10848 4350

�0.33 5 49956 10848 435010 49750 10848 435025 51648 10848 435050 65871 10848 4350

100 168391 10848 4350200 164610 11427 4482

�0.40 5 48197 10848 435010 50806 10848 435025 45291 10848 435050 62282 10848 4350

100 165772 10848 4350200 149634 10848 4595

�0.75 5 46074 10848 435010 48204 10848 435025 38253 10848 435050 55969 10848 4350

100 157433 10848 4350200 111954 10848 4350

age wise reduction of NOx and SO2 is higher than that ofCO2, it is low in absolute terms. The total reduction inemissions is not proportional to the tax rate; the reductionpercentage seems to saturate after $100/tC. As expected,the reduction in emission increases with the price elasticityof demand due to the reduction in electricity final demand(see Table 3).

Emission mitigation with energy tax shows a similar pat-tern, and it becomes significant at $1.0/MBtu.

ace oil Naphtha Coal Fuel wood Wind Hydro Total

16086 247521 0 105 103495 43267215736 247521 0 105 103671 43172317925 241497 0 105 103687 43054418384 238472 0 432 104691 42782722969 209271 215 2564 104787 41280123200 72190 469 3892 106290 39384122929 58955 1895 7896 107520 38562315960 247021 0 105 103405 43164515849 245621 0 105 103411 42993418442 235434 0 432 104691 42584523200 199224 215 2564 104787 41105923200 66718 469 3892 106370 38423823200 59237 2622 7896 107944 38141815815 247435 0 105 103405 43015518384 239472 0 215 103440 42751518126 235416 0 432 104693 41915623200 181112 215 2684 104980 38967123200 62801 469 3892 106354 37768622955 59482 2867 8120 107944 36644515187 247305 0 105 103441 42731018725 233434 0 215 103440 41921617964 225384 0 432 104693 40192422941 171050 215 2684 104980 37303722896 40276 487 3892 106354 34653621170 60473 1387 7615 107944 325741

Table 9Impact on electricity price

Tax rate Electricity price (US cents kW�1 h�1) at different priceelasticities

�0.20 �0.33 �0.40 �0.75

Carbon tax (US$/tC)

Base case 7.94 7.94 7.94 7.945 8.04 8.04 8.04 8.08

10 8.12 8.15 8.17 8.2125 8.29 8.40 8.41 8.5150 8.83 8.84 9.05 9.11

100 9.12 9.53 9.60 10.04200 10.20 10.21 10.34 10.63

Energy tax (US$/MBtu)

0.5 8.18 8.19 8.24 8.261.0 8.52 8.55 8.61 8.692.0 9.11 9.24 9.37 9.635.0 10.65 11.10 11.19 11.37

10.0 12.24 12.41 12.55 12.57

K. Siriwardena et al. / Energy Conversion and Management 48 (2007) 1975–1982 1981

Out of the five components that contribute to the totalemission change, the electricity final demand effect and fuelmix effect are predominant (Table 4). Since the change inthe generation expansion plan is small and coal is stillcheap at the low tax rates, the fuel mix effect is comparablewith the final demand effect, but the fuel mix effect is veryhigh (as much as 97% of the total reduction) at the highertax rates. The study by Chang and Lin [2] using a differentdecomposition approach showed that the final demandeffect would be the largest contributor to the total emissionreductions.

All joint effects are very small and negative (increasingemissions) except the joint effect of fuel mix-structural-finaldemand. The structural effect and final demand due topower plant construction effect are also very small (<1%).

Tables 5–8 show that with the imposition of tax, coalfired steam plants are replaced with cleaner technologieslike oil fired combined cycle, IGCC (integrated gasificationcombined cycle), supercritical coal and renewable plantssuch as wind, mini-hydro and dendro-thermal (wood fuelfired plants). Also, the total plant addition is reduced dueto the reduction in demand with increased taxes.

For Table 9, the electricity price is obtained by addingthe change in average incremental generation cost to theoriginal price. This means when a significant reduction inemissions is achieved, electricity prices will rise by at leastUSCts 0.9 per kWh with the carbon tax or USCts0.6 per kWh with the energy tax. This leads to approxi-mately 10–15% increase in average electricity prices. Fur-ther, price increments are not proportional to taxes, asthe tax reduces the demand and higher proportions ofcheaper sources can be used.

6. Conclusions

From this case study, using an I–O decomposition anal-ysis, it can be concluded that carbon taxes above $50/tC in

the electricity sector cause a visible reduction in economywide emissions at moderate values of the price elasticityof demand. This value is $1.0/MBtu in the case of energytaxes. However, it is important to note that the reductionin emissions is also strongly coupled with the value of priceelasticity. Further, it is concluded that these figures arereflected in the final consumer electricity prices by about10–15% increase in them.

The fuel mix in thermal electricity generation and thefinal demand effect are the major components contributingto the overall emission reduction. This means that the emis-sion reduction occurs mainly as a direct result of thechange in generation composition and the drop in electric-ity demand resulting from increased electricity prices ratherthan indirect effects in the economy. However, the outcomeof these tax regimes also depends on the existing electricitysupply structure and available generating options forfuture expansion.

Acknowledgements

This study was conducted jointly by the Sri Lanka En-ergy Managers Association, the University of Moratuwaand the Asian Institute of Technology, Thailand underARRPEEC III funded by the Swedish International Devel-opment Corporation Agency (Sida). The authors arethankful to all these institutions for their valuable assis-tance in different forms.

Also, the assistance extended by Mr. Sudharshana Per-era and Ms. Chethiyangani Kulatunga of the NationalPlanning Department, Sri Lanka, and Mrs. Kamani Jay-asekara of the Ceylon Electricity Board is gratefullyacknowledged.

References

[1] Leontief W, Ford D. Air pollution and the economic structure:empirical results of input–output economics. Input Output Tech1972:9–30.

[2] Miller RE, Blair PD. Input–output analysis; foundations andextensions. New Jersey, Englewood Cliffs: Prentice Hall; 1985.

[3] Chang Yih F, Lin Sue J. Structural decomposition analysis ofindustrial CO2 emission in Taiwan an input–output approach. EnergyPolicy 1998;20(1):5–12.

[4] Marpaung COP. Environmental implications of electric utilityintegrated resources planning: a case of Indonesia. AITDissertation No. ET–98–2, Asian Institute of Technology,Thailand; 1998.

[5] Rose A, Chen CY. Sources of change in energy use in the U.S.economy, 1972–1982: A structural decomposition analysis. ResourEnergy, vol. 13. North-Holland; 1999. p. 1–21.

[6] Chia-Yon Chen, Rong-Hwa Wu. Sources of change in industrialelectricity use in the Taiwan economy 1976–86. Energy Econ1994;16(2):115–20.

[7] Xioli Han, Lakshmanan TK. Structural changes and energy con-sumption in the Japanese economy 1975–1985; an input–outputanalysis. Energy J 1994;15(3):165–88.

[8] Labandeira Xavier, Labeaga Jose M. Estimation and control ofSpanish energy related CO2 emissions: an input–output approach.Energy Policy 2002;30:597–611.

1982 K. Siriwardena et al. / Energy Conversion and Management 48 (2007) 1975–1982

[9] Gay Philip W, Proops John LR. Carbon dioxide production by the UKeconomy; an input–output assessment. Appl Energy 1993;44:113–30.

[10] Proops John LR, Gay Philip W, Stefan Speck, Thomas SchrOder. Thelifetime pollution implications of various types of electricity generation:an input–output analysis. Energy Policy 2003;24(3):229–37.

[11] Long term generation expansion plan (2002), Ceylon ElectricityBoard, Generation Planning Branch, Ceylon Electricity Board, SriLanka; 2002.

[12] Jayatissa PVIP. Econometric analysis of electricity demand; the caseof Sri Lanka. AIT Thesis ET–94–18, Asian Institute of Technology,Thailand; 1994.

[13] Annual report, Central Bank of Sri Lanka; 2002.[14] Sri Lanka energy balance 1999, Energy Conservation Fund, Sri

Lanka; 2000.[15] Annual survey of industries 1995, Department of Census and

Statistics, Ministry of Finance and Planning, Sri Lanka; 1995.