Market behaviour and the to-trade-or-not-to-trade dilemma in ‘tradable white certificate’...

Market behaviour and the to-trade-or-not-to-trade dilemmain ‘tradable white certificate’ schemes

Luis Mundaca & Lena Neij & Nicola Labanca &

Bruno Duplessis & Lorenzo Pagliano

Received: 11 October 2007 /Accepted: 7 March 2008 / Published online: 12 August 2008# Springer Science + Business Media B.V. 2008

Abstract This paper provides an empirical analysis ofmarket behaviour under ‘Tradable White Certificate’(TWC) schemes. It focuses on the entire set of ‘flexi-bilities’ granted to obliged parties to meet a mandatoryenergy-saving target cost-effectively, i.e. range eligiblemeasures, eligible end-use sectors, banking provision,market engagement of non-obliged parties, and trading assuch. We found that market behaviour responds to theunique design and context in which TWC schemes areimplemented. Contrary to expectations, limited trading isobserved so the ‘to-trade-or-not-to-trade’ dilemma isfurther analysed. A real TWC market has emerged onlyin Italy, where obliged parties (i.e. energy distributors)show preference towards ‘to-trade’. In Great Britain andFrance, an autarky compliance approach is identified,with obliged parties (i.e. energy suppliers) showing pre-ference towards ‘not-to-trade’ driven by, among manyfactors, commercial benefits of non-trading (e.g.increased competitiveness). At the same time, results

show clearer indications of cost-effectiveness for GreatBritain than for Italy. In general, high energy-saving ef-fectiveness is observed, but low ambitious saving targetsand pitfalls in the regulatory framework need to beconsidered to further develop TWC markets. Initialmarket and institutional conditions strongly suggest thattrading might not be an immediate outcome. Ambitiousenergy targets can trigger a more dynamic usage of allflexibilities by eligible parties and thus active behaviourin TWC markets.

Keywords Tradable white certificate schemes .

Market behaviour . Commercial benefits of non-trading . Ex-post policy evaluation

Introduction

In theory, the creation of tradable white certificate(TWC) markets allows obliged parties—hereafter the‘parties’—to meet a mandatory energy-saving target atlowest possible costs.1 The Government, usually in co-operation with stakeholders, sets the target. Supportingthe expanding willingness to experiment with market-based approaches in Europe (OECD 1999), TWC mar-kets are created to take full advantages of market forcesand to work in favour of increased energy efficiency; i.e.it is up to the parties to decide how to meet their giventarget cost-effectively. Parties are also given the optionto trade certified energy savings to meet their individual

Energy Efficiency (2008) 1:323–347DOI 10.1007/s12053-008-9026-6

L. Mundaca (*) : L. NeijInternational Institute for Industrial Environmental Economics(IIIEE) at Lund University,Lund, Swedene-mail: Luis.Mundaca@iiiee.lu.se

N. Labanca : L. Paglianoend-use Efficiency Research Group (eERG) atDipartamento di Energia,Politecnico di Milano,Milan, Italy

B. DuplessisCentre for Energy Processes at Ecole des Mines de Paris,Paris, France

1 For a detailed description of TWC schemes, see Bertoldi andRezessy in this special issue.

targets. For them, the market strategy depends on themarket price of TWCs compared to the costs of yieldingtheir own energy savings (i.e. credits). Whereas partiesfacing low compliance costs are likely to supply TWCson the market, parties with higher compliance costs thanthe market price of TWCs are likely to demand TWCs. Ingeneral, some parties will benefit from tradable certificateschemes, and others will be made worse-off (Ellermanet al. 2000; Harrington et al. 2004; OECD 1999).

Due to the relevance of the trading component,most of the attention has been concentrated on theexpected trading activity under TWC schemes(Capozza et al. 2006). In fact, the limited tradingactivity that has occurred so far has prompted thegeneral opinion that TWC markets show little dyna-mism, if any. However, one has to bear in mind thatmore flexibilities—described in more detail in “‘Flex-ibilities’ in TWC markets: an analytical framework”—are present in these markets: (a) the set of eligiblemeasures that parties can use, (b) the number of eli-gible end-use sectors that can yield energy savings, (c)banking provision for surplus of TWCs, and (d) marketengagement of non-obliged parties. As in any tradablecertificate/permit scheme, it is the whole set offlexibilities that deserves analytical attention (Ellermanet al. 2000; Nordhaus and Danish 2003; OECD 2002;Tietenberg 2006). In fact, to what extent a TWC mar-ket achieves energy savings cost-effectively dependson how obliged parties take advantage of all thegranted flexibilities to reduce compliance costs. How-ever, a detailed TWC market analysis has not yet beendone because of a lack of empirical evidence, whichthis paper attempts to do.

The objective of this paper is to provide a compre-hensive analysis of market behaviour in TWC schemes.Elaborating upon cost-effectiveness2, the study focuseson whether policy design, existing energy marketconditions, and corporate aspects inhibit or encourageparties to take full advantage of given flexibilities. Thispaper builds mostly upon research work developedwithin the EU EuroWhiteCert project3 and seeksanswers to the following questions:

& How can the overall market behaviour underTWC schemes be characterised?

& What are the drivers behind the to-trade-or-not-to-trade dilemma?4

We perform the analysis taking the early experienceof the Italian TWC scheme and the Energy EfficiencyCommitment (EEC; mostly first phase) in Great Britainas case studies.5 The latter is usually regarded as aTWC scheme because it allows trading of energysaving and obligationsbut it is not a certificate-basedscheme yet. In order to yield more robust conclusions,in particular for non-trading patterns, our analysis iscomplemented with early trends from the French TWCmarket and findings obtained through a simulated TWCmarketusing game theory and interviews with keystakeholders (i.e. obliged parties, authorities, and policymakers). A definitive answer on whether a TWC schemeis an adequate policy choice is beyond the scope of thisstudy.

The structure of this paper is as follows. “‘Flexibilities’in TWC markets: an analytical framework” describesthe theoretical cost-effective element embedded in TWCschemes and the set of flexibilities—including potentialtrade-offs. Specific research questions that our analysisseeks to answer are set, and altogether, these aspectsrepresent our analytical framework. In “Early evidenceof market behaviour from the Italian and British TWCmarkets”, we analyse empirical evidence of marketbehaviour coming from the Italian and British TWCschemes. Due to the fact that non-trading is observed,“Commercial benefits of non-trading behaviour forobliged parties” deepens the commercial drivers of thisscenario based on findings from Great Britain, which arethen analysed in the light of early indications fromFrance and our simulation outcomes. “Discussion onnon-trading aspects under TWC schemes” discussessome underlying policy aspects of non-trading patterns.Finally, conclusions are drawn in “Conclusions”.

‘Flexibilities’ in TWC markets: an analyticalframework

Whereas there is a tendency to focus on trading as thecrucial flexibility in TWC schemes, many more flex-

2 Defined as whether a mandatory energy saving target isachieved at lowest possible costs.3 For further information, visit http://www.eurowhitecert.org.

4 To be taken as different from the policy dilemma of whether aTWC scheme is or not the right policy choice.5 Notice that the acronyms EEC1, EEC2, and EEC3 are used todenote the first (2002–2005), the second (2005–2008), and thethird phase (2008–2011) of the EEC, respectively.

324 Energy Efficiency (2008) 1:323–347

ibilities are embedded and granted to parties to fulfiltheir targets cost-effectively. Indeed, the central argumentfor least-cost compliance relies on how obliged partiestake advantage of all the given flexibilities to reducecompliance costs. We identify the following ones.

Range of eligible measures For the first type of flex-ibility, the number of cost-effective measures andtechnology development is crucial. Policy makers defineeligible measures that qualified to yield TWCs. Thus, thepolicy commitment is expressed through technology-based rules—like per output basis for CO2-eq abatementcommitments (cf. OECD 1999). The broader the set ofeligible measures, the more flexibility is given to partiesto yield cost-effective energy-saving potentials. How-ever trade-offs exist. For instance, a larger set of eligiblemeasures can entail a heavy administrative burden forthe authorities in relation to measurement and verifica-tion (M&V) activities, increasing the total costs of theprogramme. Within this flexibility, additionality isrelevant because TWC markets are supposed toencourage measures that would not have implementedwithout the TWC financing (i.e. as depicted by thebaseline). So far, experience shows that it is very case-specific how additionality is determined and/or applied(e.g. standards above existing building regulation). Ingeneral though, one can argue that measures that areadditional are likely to be eligible. Aspects to beaddressed in our analysis are, among others, what arethe drivers behind dominant eligible measures?What arethe saving costs per measure?

Number of eligible end-use sectors For the secondtype of flexibility, the larger the group of eligible end-use sectors in which eligible measures can be imple-mented, the more options parties have for meeting theirobligation cost-effectively. Nevertheless, trade-offs ex-ist because, whereas a larger sectoral coverage may bedesired, a growing number of sources of transactioncosts (e.g. persuasion of customers, search for tradingpartner, etc.) could be expected with market actorsshouldering a higher burden—besides a heavier admin-istrative burden for the authorities. Whilst the Britishscheme allows savings only in the household sector, theItalian and French schemes allow energy savings in allend-use sectors. For the latter, savings can be realised inall end-use sectors not yet covered by the EuropeanEmission Trading scheme (EU-ETS)—including thetransport sector. When evaluating the market behaviour

for this particular flexibility, questions we attempt toanswer are, for instance, what are the dominant endsectors in which savings are realised? Does a singleeligible end-use sector hamper the efficiency (i.e.maximisation of net societal benefits) of TWCschemes?

Banking provision for surplus of TWCs For the thirdtype of flexibility, ‘banking’ is allowed so parties thatover-comply with their individual targets can save thesurplus of TWCs for future commitment periods (e.g. asin Great Britain). The banking option can prevent trading,as it allows obliged parties carrying over unredeemedsavings from one compliance period to subsequent ones.In fact, banking is usually denoted as ‘inter-temporaltrading’ (cf. Rubin 1996). In order words, it is given asan inter-temporal flexibility for saving credits in order tomitigate the costs of over-investment (Ellerman et al.2000; OECD 1999).6 When evaluating the marketbehaviour for this particular flexibility, we address thefollowing questions: How are obliged parties using thebanking option? What are the market and regulatoryaspects driving its usage?

Market engagement of non-obliged parties The fourthtype of flexibility refers to the participation of othereligible parties that do not bear any obligation, but theycan participate in the TWC market. These parties areentitled to implement eligible measures, gain TWCs fordoing so, and also trade TWCs on the market—as long asthey fulfil all the regulatory requirements. For a TWCscheme to be cost-effective, market liquidity7 is of primeimportance. In its full extension, high liquidity is affectedby a number of critical conditions: a large number ofbuyers and sellers, market information readily available,high trading volume, low transaction costs, andminimumregulatory barriers to trade (Mundaca and Neij 2006;Voogt et al. 2006).8 Within this context, a large numberof eligible actors (i.e. obliged and non-obliged) are

6 Although not yet implemented in current TWC schemes, a‘borrowing’ option for non-compliance can also be introduced.This means that a party that does not comply with its targetcommits itself to a greater target for the next compliance period.7 The term ‘liquidity’ is used to refer to the characteristic ofTWCs whereby they can quickly be converted into cash at areasonably predictable price.8 Notice that throughout the paper, we sometimes refer toliquidity to address only one critical condition: the number ofeligible parties.

Energy Efficiency (2008) 1:323–347 325

desired because they are likely to face different marginalcosts at similar levels of achieved energy savings.However, more participants can also add transactioncosts. When evaluating the market behaviour for thisparticular flexibility, we approach the analysis byanswering the following questions: To what extent doesthe participation of eligible parties reduce or increase thecompliance costs borne by obliged parties? Do eligibleparties create price volatility on the market?

Trading option to equalise marginal compliance costsThe fifth type of flexibility in a TWC scheme is quitestraightforward: trading as such. The central point isthat, as long as an allocation of individual saving targets isnot cost-effective—very likely due to asymmetric infor-mation between regulators and obliged parties—there isalways an incentive to trade. This has to be done with dueconsideration to transaction costs (e.g. search for infor-mation, legal advice, negotiation with partners, etc.), non-economic barriers (e.g. lack of awareness among endusers to increase energy efficiency) and commercial ofnon-trading (e.g. customer loyalty). Critical indicators foranalysing the trading component of TWC markets arevolume of trades, number of buyers and sellers, and pricedispersion. When evaluating the market behaviour, weseek to answer the following questions: How can thetrading activity be characterised? What are the keyinstitutional and market conditions affecting the tradingactivity? Is it possible to identify a clear preference as faras the ‘to-trade-or-not-trade’ dilemma is concerned?

Early evidence of market behaviourfrom the Italian and British TWC markets

We argue that a comprehensive approach to portray theoverall market behaviour in TWC markets is tosystematically analyse the activity level of the ‘flex-ibilities’ they involve. Table 1 summarises our findings,which are further detailed in the following sections.

Eligible measures used to realise energy savings

Before addressing the type of measures used to meetthe saving target, it is relevant to cast light on theactual amount of certified energy savings.9 According

to the Italian Authority for Electricity and Gas (AEEG2006) 286,837 toe or TWCs (approximately 3.33 TWh)were certified between January 2005 and the end ofMay 2006.10 Table 2 summarises the number of cer-tificates issued by the Electricity Market Operator(GME).11 The issuance of total certificates correspondsto 184% of the overall saving target to be achieved by2005 (i.e. 155,911).12 Under the Italian TWC, differenttypologies of TWCs exist according to the energycarrier that is saved. Whereas TWCs related to elect-ricity savings (type I) exceeded the minimum expectedamount by a factor four, TWCs related to gas savings(type II) were almost two times the minimum amountrequested to comply with the obligation in 2005. TWCrelated to other fuels (type III) had a marginal con-tribution. The higher number of TWC-I issued indicatesa higher number of cost-effective savings available in theelectricity segment. Figures recently released by AEEG(2007) show a clearer dominance of TWC-I (78%),followed by TWC-II (18%) and TWC-III (4%; as fromJanuary 2005 to December 2006).

The distribution of eligible measures that have beenused to realise energy savings in Italy is shown inFig. 1.13 Values are valid for the end of the first com-pliance year and are shown according to the eligiblesectors in which they were implemented. More than60% of the total savings were achieved, realising savingpotentials through micro-scale size and ‘low-hangingfruits’ in the commercial and household sectors. Almost20% of total savings were achieved by district heating,either through the implementation of new grids or theextension of existing ones. Solar heating panel installa-

9 To complement the findings about the Italian TWC scheme,see other papers in this special issue.

12 It has to be mentioned that the target for 2005 was actually200,000 toe. However, the target to be achieved by obligedparties was approximately 155,000 toe. This is because approx.50,000 toe could not be apportioned amongst companies thatdid not meet the apportionment criterion of having a minimumof 100,000 customers. At the time of research, apportionmentcriteria for these companies still remained to be developed.

11 GME is in charge of the issuance of TWC management ofthe market. For further information, visit http://www.mercatoelettrico.org/En/Default.aspx.

10 Under the Italian TWC scheme, notice that one TWC isequal to 1 tonne of oil equivalent (toe) or 11.63 MWh. Also,notice that obliged actors are allowed to comply with theirsaving obligation for a given year until May 31 of the followingyear.

13 No further details about the specific measures implementedin industrial sector were disclosed by the Italian authority incharge of administering and enforcing the TWC scheme (i.e.AEEG)

tion, in particular, was responsible for 8% of the totalsavings achieved.14

The trends in Italy are unlikely to reflect market pre-ferences towards any particular eligible measure. First,one has to consider that reaching the industrial sectortypically requires ex-post M&V approaches, which aremuch more case-specific than those for the household/commercial sector.15 After 2 years of operation, the

deemed or ex-ante approach has been mainly adopted,accounting for 70% of the total certified savings. Thus,the easiest approach has been to claim ‘low-hangingfruits’ in the household and commercial sectors. IfM&V approaches for industrial applications are sooncompleted/developed, eligible measures in the indus-trial sector will soon get off the ground, and a differentscenario may be observed. Second, Italian parties can

Table 2 TWC issued in the Italian TWC scheme (January2005 to May 2006)

Energy carrier Certificates issued Share (%)

Electricity (type I) 214,244 75Gas (type II) 62,826 22Others (type III) 9,767 3Total 286,837 100

Data source: AEEG (2006)

Table 1 Summary of market behaviour in TWC schemes

Flexibility Italy Great Britain

Key implemented eligiblemeasure(s)

Appliances, lighting and micro-generation (e.g. PV)but no clear trends are observed as the ‘retroactive’saving option (allows to claim savings realisedprior the implementation of the scheme) dominatedduring the first compliance year

From a portfolio of approximately 20measures, clear dominance of insulationmeasures (e.g. cavity wall) followed byCFLs

Key eligible end-usesector in which savingsare realised

Household and commercial sectors dominated thefirst compliance year, but implementation inindustrial sector may soon take off

Restricted to household sector. Requirementto yield at least 50% savings in the ‘prioritygroup’a has been met

Banking of TWCs Unlimited. Around 42% of the secondcompliance year potentially achievable withcredits generated during the first compliance year

Unlimited, largely used and driven by changesin accreditation methodology of measures.Savings carried over from EEC1 representca. 27% (or 35 TWh) of EEC2 target(130 TWh).

Market engagement ofnon-obliged parties

Great number of ESCOs have participated, butlargely explained by a less stringent definitiongiven by the regulatory framework. Combinedwith the ‘retroactive saving option’, the growingnumber of ESCOs creates conditions for thefree-riding effect

Insulation contractors have played a crucialrole. Furthermore, obliged parties havestrategically partnered with a variety of actorsto realise savings and/or deliver measurescost-effectively

Trading activity (Whatabout the to-trade-or-not-to-trade dilemma?)

Market emerged and traded TWC volume(approximately 50% of total TWCs issued) hasincreased. Bilateral and spot trading accountedfor 83% and 17% respectively during the firstcompliance year. A handful of actors participatedon spot trading market and rent-seeking behaviouris identified. Prices have fluctuated but slowlyfalling. Preference towards ‘to-trade’ is observed

Market has not really emerged, but two tradesof obligations took place and six obligedparties bought credits retroactively duringEEC1. Several market and regulatory aspectsinfluenced this pattern, including potentialcommercial benefits of non-trading. Preferencetowards ‘not-to-trade’ is observed

a Under the British scheme, at least 50% of savings must be realised in the so-called ‘priority group’, which is defined as “householdsthat received certain income-related benefits or tax credits” (OFGEM 2005, p. 4).

15 In Italy, three different M&Vapproaches exist (see AEEG 2006;Adnot et al. 2006a): (a) ex-ante approach or ‘deemed savings’, (b)engineering estimates that require some on-field measurement andrelies on simplified energy saving calculation, and (c) an energymonitoring plan (i.e. an ex-post approach based on directmeasurement of energy consumption before and after the projectis implemented).

14 More recently, the Italian authority released some aggregatedfigures covering a more extensive period, from January 2005 toDecember 2006 (AEEG 2007). See Pavan in this special issuefor more details.

retroactively claim energy savings implemented from2001 and onwards to fulfil their targets. In fact, asubstantial share of measures (approximately 60%)was actually not triggered by the TWC market as suchbut prior to 2005. This ‘early action’ feature, incombination with the ex-anteM&Vapproach, can largelyexplain the dominance of the household and commercialsectors.16 Although this regulatory aspect has givenmore flexibility to the parties, it also gives indications offree-riding, and provides little ground to identify eligiblemeasures that were actually triggered and implemented(above or beyond certain level) once the scheme startedto operate. Despite the fact that additional energysavings are defined in Italy as those “that are over andabove spontaneous market trends and/or legislativerequirements”, the retroactive provision questions itsachievement in terms of energy-saving effectiveness.17

From the end users’ standpoint, the Italian schemeseems to be minimising the costs for meeting a giventarget—with due consideration to potential free-ridingeffect. Taking TWCs prices as proxies of actual marginalsaving costs, and energy prices paid by Italian house-holds in 2006 as benchmarks, net financial savings for

end users are estimated to be 5 and 6 Euro cents/kWh ofelectricity and natural gas saved, respectively.18

Looking now at the British scheme (first phase) andin terms of energy-saving effectiveness, the scheme hasperformed well with a minor shortfall during its firstphase. Energy savings counting against the EEC1 target(62 TWh) reached almost 61 TWh19 (OFGEM 2005,p. 8), and the usage of ‘deadweight’ factors has counter-acted the free-riding effect.20 According to OFGEM(2005), the slight deficit (approximately 2%) duringEEC1 was due to two parties that went out of business.However, the total amount of savings accumulatedduring the EEC1 reached 140% of the EEC1 target or86.8 TWh (i.e. savings accounted/redeemed againstEEC1 target plus the amount of savings banked for

17 Defined as whether obliged parties meet or not a mandatoryenergy saving target.

16 This ‘early action’ provision granted to eligible parties waslikely the result that the scheme was supposed to be implementedin 2002. However, several aspects challenged its politicalfeasibility and the scheme was finally implemented in 2005.

Appliances and lightingin household and

commercial sectors33%

Eligible measures in industrial sectors

Lighting in public system27%

CHP, PV panels and districit heating in

household and commercial sectors

Heating and insulation inhousehold and

commercial sectors14%

Fig. 1 Distribution ofimplemented eligiblemeasures in relation toenergy savings (286, 837toe) under the Italianscheme. From January2005 until May 2006. Datasource: AEEG 2006, p. 20

18 This is assuming average TWC prices of €71 and 94 in 2006for electricity and gas, respectively (see “Trading activity”).Furthermore, we consider national tariffs of electricity andnatural gas of approximately 5.6 and 6.8 Euro cents/kWh,respectively. Notice that 1 toe=11,630 kWh.19 Energy savings in the British scheme are expressed in fuelstandardised lifetime discounted terms. This means that toreflect the different levels of primary energy input to a kilowattof electricity (including the carbon content), a differential isapplied to energy savings. For instance, the number of kilowatt-hours of saving is multiplied by 0.56 for coal savings, 0.8electricity savings, and by 0.35 for gas savings. For furtherdetails see DEFRA (2004b:10, 46).20 ‘Deadweight’ refers to the level of investment activitycarried out by parties under ‘business-as-usual’. A deadweightfactor is applied when energy savings from eligible measuresare estimated (including costs).

EEC2 target; OFGEM 2005, p. 7).21 Details are shownin Fig. 2. Parties scaled up their energy saving activitythroughout the EEC1 period. Whilst 20% of the totalsavings were achieved during the first year of the EEC1,35% and 45% of the savings weremet during the secondand third year of EEC1 (OFGEM 2005, p. 11). Bothpublic authorities and obliged parties concur that thislevel of energy savings would have not occurred in theabsence of the scheme.

In Great Britain, due to large cost-effective potentialsin the household sector, the dominance of insulationmeasures was clear during its first phase, in particularcavity wall and loft insulation. Due to large cost-effective potentials in the household sector, thesemeasures contributed to 56% of the total achievedsavings or nearly 38% of the savings redeemed (seeTable 3). Almost all measures can be labelled as matureand commercially available.22 Depending on theeligible measure, parties worked together with differentactors to implement eligible measures (see “Marketengagement of non-obliged parties” for further details).According to Lees (2006), the impacts of markettransformation have been visible for white goodappliances (i.e. washing, dishwashing and refrigeration

equipments). For instance, the market transformationshows high penetration of A-rated cold appliances ofup to 60% by 2006, with multiple policy programmesalso influencing such a trend (e.g. EU energy labellingprogramme, consumer advice from specialised centres,etc.; Lees 2006, pp. 33–34). For heating measures, asteady growth of condensing boiler sales was observed—actively subsidised by obliged parties—reaching a marketpenetration rate of up 80% (Lees 2006, [. 37). It wasfound that, in this case, the building regulation enactedin 2005, which now defines B-rated boilers as theminimum standard sold in Great Britain, has played acritical role to transform the market (see Lees 2006). Infact, a variety of policy instruments (i.e. informative,command-and-control and economic ones) have alsosupported the deployment of eligible technologies andthus the performance of the British scheme as a whole(cf. DEFRA 2004b; Rohr 2004).

When it comes to the second phase of the scheme(i.e. EEC2), similar market trends are identified con-cerning eligible measures. First, a growing policyambition is observed leading to a mandatory savingtarget of 130 TWh.23 After two completed years ofEEC2, approximately 88 TWh of savings have beenrealised. Again, the low energy level of the housingstock has allowed parties to yield cost-effective savingsmainly from insulation measures. They continue to haveclear dominance with a share of 85%, followed by light-ing (8%), heating and micro-combined heat and power(CHP; 5%) and appliances (2%), respectively (OFGEM2007). This activity level reflects the vast saving po-tential that the British dwelling stock still offers. If oneconsiders also banking activity under EEC1 (see “Bankingprovision for surplus of TWCs”), the level of energy-saving effectiveness of EEC2 reaches 93% (or approxi-mately 120 TWh) after 2 years of operation; see Fig. 3.

The focus on only one end-use sector has naturallylimited the number of eligible measures of the Britishscheme. However, several measures have been used tomeet the target, and the dominance of insulationmeasures is rather evident. Public authorities remainopen to enlarge the list of eligible measures and developappropriate baselines and methodologies to estimate ex-

21 The resulting environmental effectiveness of the scheme can besummarised as follows. Target achievement under the EEC1 wasexpected to save around 0.5 MtC/year in 2010 (DEFRA 2006c:(1) Notice that this year was chosen as a benchmark because it isthe mid-point of the first Kyoto Protocol commitment period.The 0.5 MtC figure was based on assumptions regarding stand-ards of comfort and electricity generation, and heating fuel mixesprojected by the authority (see DEFRA 2004b). Once imple-mented energy efficiency measures counting against target EEC1are considered (not including banking to EEC2), emissionsreductions equate to 0.4 MtC/year in 2010 (including a deductionfor estimated deadweight). Considering total UK CO2 emissionsto be approximately 150 MtC/year, of which householdemissions contribute 40 MtC/year, the achieved emissionreductions represent approximately 1%.22 In Great Britain, responsible parties must demonstrate thateligible measures would have not been implemented in theabsence of the scheme. Additionality can then be justified infinancial terms. For instance, landlords of social housingprogrammes can support obliged parties by providing writtenevidence of additionality. To implement eligible energyefficiency measures in low-income households, British obligedparties need to receive a written statement from the landlordstating that the measures would not have been implementedoutside the programme. However, authorities recognise that it isdifficult to draw the line between additional and non-additionalmeasures when local energy efficiency programmes are inplace.

23 Saving targets for EEC1 and EEC2 are not directlycomparable. This is because several assumptions in themethodology for estimating the EEC2 target are different fromthe ones used to determine the EEC1 target (e.g. timedependency of savings, discount rate, etc.). For further detailssee DEFRA (2004a).

ante savings. Whereas authorities are keen to promoteinnovative measures, including micro-CHP and micro-generation from renewable sources, it is likely that thesemeasures will have a marginal contribution in the shortterm because they are less cost-effective (cf. DEFRA

2006b). Moreover, it has been already indicated thatdue to large potentials, eligible measures will still becavity wall and loft insulation during the next phase ofthe scheme (DEFRA 2006b, 2007). For instance, halfof the UK housing stock will have loft insulation

Table 3 Eligible measures installed and total realised energy savings under the British scheme—EEC1

Category/type of eligible measure Number of measures installed Achieved savings (GWh)

Insulation Cavity wall insulation 791,524 25,069.27Loft insulation (top up) 528,496 4,138.78Loft insulation (virgin) 226,245 9,696.90Do-it-yourself insulation (m2) 15,979,367 8,101.49Draught stripping 22,743 38.56Tank insulation 195,832 433.50Radiators panels (m2) 38,878 13.39Solid-wall insulation 23,730 972.59Other insulation 2,625 21.14

Appliances Energy-efficient cold appliances 2,956,084 7,381.18Energy-efficient wet appliances 3,551,737 2,260.32Other appliances 93,837 42.49

Heating A and B rated boilers 278,991 2,361.90A and B rated boilers + heating controls 87,497 1,233.47Heating controls upgrade 2,366,128 1,220.49Fuel switching 41,077 2,763.32CHP/communal heatinga 615 39.03Other heating 202 4.66

Lighting CFLs 39,737,570 20,976.79Total 66,923,178 86,769.27

Data source: OFGEM (2005: pp. 47–66)a Number of household benefiting

Total energy savings achieved (86.8 TWh)

(including over-comply)

Total energy savings redeemed (61 TWh)

(excluding banking)

Insulation Heating Appliances Lighting

Fig. 2 Energy savingsachieved by category ofeligible measures under thefirst phase (2005–2005) ofBritish scheme—EEC1.Data source: OFGEM 2005,pp. 11–66

inferior to what is required by law in 2008 (NIA 2006,p. 4). Therefore, a rational option for parties is to con-tinue relying on insulation measures to meet theirobligation.

The British scheme grants energy savings basedentirely on an ex-ante approach. The experienceshows that an ex-ante approach is a reasonable optionwhen a number of conditions exist. This is largelyexplained by the fact that the size, type and per-formance of the measures are well known and relatedestimates use best available data.24 Thus, it would beinefficient to use an ex-post M&V approach; whichwould make eligible measures non-cost-effectiveunder this scheme.25 As a result, the burden ofM&V for the regulator has been largely reduced (cf.Adnot et al. 2006a; Capozza et al. 2006; Mundaca2007a). However, there are risks associated with purelyex-ante schemes, like partial realisation of savings orpoor additionality, for instance.

Taking into account implemented measures, severalindications of cost-effectiveness for EEC1 were found.Energy savings costs have been estimated to be around0.8∼1.4 Euro cents/kWh for lighting measures andapproximately 0.7∼1.3 Euro cents/kWh for insulationmeasures, respectively (Lees 2006, p. 27; Mundaca,2007a, p. 4349).26 These figures are lower than theestimated average cost savings of 2.5 Euro cents/kWhgenerated by the most likely alternative policy option(see OFGEM and EST 2003, p. 17)—and certainlymuch lower than energy prices paid by householdsduring EEC1, approximately 2.3 and 9.4 Euro cents/kWh for gas and electricity in 2004, respectively.27

Furthermore, it is estimated that energy savings costswere 20% lower than predicted by DEFRA (Lees 2006,p. 30). Finally, and using energy prices paid byhouseholds in 2004 as benchmarks, net financialbenefits for end-users are estimated to be 8∼8.6 Eurocents/kWh for electricity savings and 1∼1.6 Euro cents/

24 Usually, the development and usage of M&V approachesdepend on the complexities of eligible measures, number ofeligible end-use sectors, relevant data available (e.g. baseline,climatic conditions, lifetime and performance of measures,etc.), and the consensus and negotiation between policy makers,regulator(s) and obliged parties.25 Personal communication with Iris Rooney (DEFRA) andCharles Hargreaves (OFGEM), October 2005.

26 Notice that estimates given by Mundaca (2007a; i.e. 0.8 and1.3 Euro cents/kWh) include transaction costs borne by obligedparties. 1 British pound = approximately 1.41 euro as onDecember 2004.

Banking EEC2 Q1 EEC2 Q2 EEC2 Q3 EEC2 Q4 EEC2 Q5 EEC2 Q6 EEC2 Q7 EEC2 Q8

Insulation Heating Appliances Lighting

Fig. 3 Cumulative energysavings counting towardsEEC2 target (130 TWh) af-ter 2 years (out of three)under the British scheme.Notice the substantialamount of savings (mostlyfrom insulation measures)carried over from EEC1.Data source: Personal com-munication with Priya Patel,OFGEM

27 At 6.68 p/kWh (approx. 9.4 Euro cents/kWh) for electricityand 1.65 p/kWh (approx. 2.3 Euro cents/kWh) for gas. See DTI(2005). The authority has mentioned that the year 2004 can betaken as the “centre of gravity” of EEC1.

KWh for gas savings (cf. Less 2006:27; Mundaca2007a:4349).

Eligible end-use sectors used to realise energy savings

Under the Italian TWC scheme, savings can be realisedin any end-use sector. As mentioned before, Italianmarket actors have largely focused on the householdand commercial sectors. For the first compliance year,both sectors captured almost 70% of the realisedsavings, most of them retroactively though. Again, this‘early action’ or retroactive option can largely explainthe dominance of these sectors because complex M&Vissues might have prevented eligible actors to claimsavings in the industrial sector. However, this trendcould also change if one thinks about the implementa-tion of lengthy measures in the industrial sector—likelyto be a large and more cost-effective supplier of TWCs.The development of M&Vapproaches for the industrialsector is slow, but larger measures will likely be im-plemented once these methodologies are ready. In turn,this may increase also the burden for public authoritieswhen an actual larger sectoral coverage needs to bemonitored and enforced.28

As for the British scheme, which covers only thehousehold sector, the number of measures is limitedcompared to the Italian and French schemes, whichcovers multiple end-use eligible sectors. However, thelevel of ambition—as reflected in the energy savingtargets29—is much higher in the British case. Thisrelatively higher level of ambition than the Italian casealso reflects the high energy-saving potential that thehousehold sector offers in Great Britain. Importantly,British parties have faced during EEC1 and EEC2 therequirement that at least 50% of savings must berealised in the so-called priority group. During theEEC1, parties met this requirement by achieving42 TWh of savings in the priority group—against the32 TWh required (OFGEM 2005, p. 9). Available datashow that parties could maintain a balanced portfolio

of priority and non-priority eligible customers (seeOFGEM 2005, p. 10). After 2 years of EEC2, figuresshow that only 10 TWh (or 5%) of savings remains tobe achieved to fulfil the 50% requirement (OFGEM2007, p. 5).

No political signs can be identified as far as theenlargement of the coverage is concerned. This can beexplained, inter alia, by the existing trade-off between alarge coverage to boost cost-effectiveness and a higheradministrative burden and transaction costs resultingfrom a larger coverage. In any case, one could think thatthe limitation of having only one eligible sector couldhamper the cost-effectiveness of the scheme. Evidenceshows that this has not been the case under the Britishscheme, as high-saving cost-effective potentials in theinsulation segment allowed the EEC1 to yield netfinancial benefits for the end users (as noted in theprevious section).

Banking provision for surplus of TWCs

When analysing the banking of TWCs in the ItalianTWC market, one has to bear in mind that unlike GreatBritain, TWCs in Italy are issued year-by-year30 so thescheme does not use lump sum discounted savings.31

Having mentioned this, indications can be drawn aboutItalian parties using TWCs for subsequent complianceperiods. We estimated that the number of bankableTWCs for 2006 is approximately 130,926 TWCs32.This rough estimation corresponds to the total numberof TWCs issued in 2005 (i.e. 286,837) minus the numberof certificates needed to achieve the saving target for2005 (i.e. 155,911). This figure may represent anunderestimation of TWCs potentially bankable because

30 Note that issuances of TWCs is for 5 years in Italy, howeverfor heating and air conditioning the issuance period goes up to8 years.

28 At the time of research, there had been no study ontransaction costs under the Italian TWC scheme, thus it remainsto be seen whether the coverage of the scheme involvesnumerous sources and a heavy burden of transaction costs forparties.29 Mandatory energy saving targets are as follows: 130 TWh inGreat Britain (period 2005–2008), 67 TWh in Italy (period2005–2009), and 54 TWh in France (period 2006–2008).

31 Note that under the Italian TWC scheme, 1 TWC = tonne ofoil equivalent.32 This imply, for instance, that if a given saving measureimplemented in 2005 generates 1,000 certificates during the5 years of its lifetime, 200 certificates are to be issued in 2005,other 200 certificates are to be issued in 2006 and so on until2009 unless, for some reason, the measure is not capable ofgenerating the same 200 certificates in the years following2005. In other words, saving measures employed to complywith the 2005 obligation continue generating the same amountof certificates during the following years depending on the yearin which they were implemented and on the lifetimes of eligiblemeasures.

not all parties achieved their target obligation by 2005. Ifthe number of obliged parties does not increase, theobligation for 2006 is equal to approximately312,000 TWCs. Considering that the amount of bankingis unlimited, this would imply that about 42% of the 2006target could be achieved by using the 130,926 TWCs notredeemed in 2005.

Regarding the banking option under the Britishscheme, many aspects were found. First, the EEC1 wasbuilt upon the Energy Efficiency Standards of Perfor-mance (EESOP) that ran from 1994 until 2002.33 Underthe EEC1, parties were allowed to carry over savingsgenerated under the EESOP that accounted for up to10% of each obliged party’s target. At that time, nearly3 TWh in savings were banked for EEC1 compliance,representing almost 5% of the target (OFGEM 2005,p. 52). Second, savings from cavity wall insulationheavily dominated the banking activity (as shown inFig. 3). Third, parties that did not bank savings werebasically new entrants by the end of EEC1. Parties thatused the banking option were large energy suppliers.For instance, one of them achieved more than half ofits EEC2 target with savings that were carried overfrom EEC1 (OFGEM 2005, p. 67). Fourth, partieswere allowed to bank an unlimited surplus of energysavings from EEC1 to EEC2. Under EEC1, six partiesbanked savings approximately 25 TWh. Fifth, bankingactivity during EEC2 has achieved already 10 TWh,and it is likely to increase as the second phase comes toan end.

Regulatory aspects affecting the banking activity inGreat Britain have heavily drivenmost of the trends listedabove. Changes in the methodology for accreditingenergy savings (e.g. lower discount rate, different heatingpatterns, etc.) have been critical to determine the level ofbanking. Adjusted in EEC2 terms, the above-mentioned25 TWh from EEC1 are equivalent to approximately35 TWh savings under EEC2.34 Thus, parties redeemedsavings for their EEC1 target that were less attractive forcompliance under the EEC2 accreditation methodology(e.g. saving from lighting and condensing boilers). Inother words, savings with “higher” values (e.g. cavitywall insulation) were not used to meet the EEC1 targetbut carried over to EEC2. This pattern can be observed

by comparing the two bars in Fig. 2. In turn, the amountof savings carried over from EEC1 accounted for morethan 25% of the EEC2 target (130 TWh). The surplus ofsavings leading to high banking activity in Great Britainhas been also driven by the fact that suppliers wanted toavoid as much as possible the financial penalty in caseof non-compliance (i.e. up to 10% of supplier’sturnover).

Market engagement of non-obliged parties

Besides obliged parties (approx. 34), another 573 gasand/or electricity distribution companies were entitledto participate in the Italian TWC scheme as eligibleparties (see AEEG 2006).35 Furthermore, the authorityaccredited more than 550 ESCOs between November2004 and May 2006. Table 4 shows the level ofactivity of all these market actors as far as issuance ofTWCs is concerned. As observed, the activity level ofthese eligible parties, in particular ESCOs, has beenmuch higher than obliged parties. Some designfeatures can explain this high activity level. Guidelinesdeveloped by the authority in 2003 established a broador less stringent definition of ESCOs, which has driventhe large number of these eligible parties participatingin the scheme.36 However, most of these ESCOs couldbe merely classified as providers or installers of energyefficiency equipment; which also questions the achieve-ment in terms of energy-saving effectiveness. The shareof ESCOs already active is not negligible. This is becausemany of them are still waiting to participate and othershave not reached the minimum threshold of realised

34 For further information, see DEFRA (2004a, 2005).

33 For further information about EESOP, see OFGEM and EST(2003).

36 Companies submitting an energy efficiency project forcertification to the AEEG can be considered “ESCOs” if theirobject includes the provision of integrated services for therealisation and the subsequent possible management of energy-saving measures.

35 In the created Italian TWC market, the criterion to define anobliged party addresses gas and electricity distributors servingmore than 100,000 customers. This led the AEEG to identify 30obliged companies (with 10 electricity and 20 gas distributioncompanies). A trade-off between the number of obliged partiesand the complexity of the system to be administered isidentified. For 2005, the total saving target apportionmentcriteria adopted in Italy implied that 22% of the total savingtarget was not apportioned. Lowering such threshold on theamount of clients served could have increased the total amountof savings apportioned. Nevertheless, such an action could haveincreased the administration burden for the authority to enforceand monitor the TWC scheme.

energy savings required.37 Other accredited ESCOs havenever operated on the spot market as such.

Whilst the typology of actors in Italy is limited, itsnumber is substantially higher than in Great Britain. Thehigh number of other non-obliged parties, including“ESCOs”, determines a large market size for TWCs.The broad definition of ESCOs used in Italy might beencouraging the liquidity of the market (at least in termsof market players) but also stimulating free-ridingeffects. One could argue that the so-called “ESCOs”would have sold or implemented the same level ofenergy efficiency measures even in the absence of thescheme. In addition, the potential free-riding effectcould be magnified by the option to claim savingsretroactively, which also contributed had lead to anexcess of TWCs. This could also have a negative impacton future compliance periods because of the unlimitedamount of banking that is allowed.

The regulatory framework in Great Britain considersobliged parties as all energy suppliers of gas andelectricity, serving more than 15,000 customers. Theyare the only ones allowed to trade savings and obliga-tions. This feature of the scheme has resulted in only eightobliged parties acting on the market, reducing the numberof eligible parties as such (cf. Mundaca 2007a; NERA2006). The British authority has discussed the feasibilityto modify the EEC so it becomes a formal TWC tradingscheme—so the programme moves away from the

current measure-based approach. In doing so, theauthority has already pointed out, for instance, trade-offs concerning market liquidity and increased adminis-trative burden. Although a larger number ofmarket actorswould increase competition and reduce compliance costs,administrative costs might increase because of monitor-ing, reporting and verification activities (cf. DEFRA2006a).

In Great Britain, the legal framework does not haveany definition of ESCOs as such. The British schemepromotes instead ‘energy service actions’ in households(i.e. where an energy audit for a whole house is carriedout and at least two measures are installed). Toencourage energy service actions, the regulatory frame-work grants obliged parties an additional credit of 50%of the savings realised. Albeit not defined, ESCOs are inprinciple eligible to realise savings under the Britishscheme, but they are not allowed to trade them. ESCOsthat want to participate in the British scheme aredependent on the demand of the few obliged partiesand are restricted to projects in the household sector.These actions must be carried out under the fulfilment ofseveral legal clauses between the obliged party and thehouseholder (e.g. eligible measures intended to achieveimprovements in energy efficiency at the householdpremises by at least 13%; obliged party is required toundertake an energy efficiency audit, etc.; Gaudioso etal. 2007). It has been argued that some legal aspects (e.g. the ‘28-day rule’ that allows a customer to terminatean energy supply contract on 28 days’ notice) mighthamper the provision of energy services under theBritish scheme (DEFRA 2004b).38 In any case, thisdoes not prevent that energy service actions on behalfof households can be realised (Gaudioso et al. 2007).

Focusing on one critical condition for high liquidity(i.e. large number of actors), the British scheme hasembraced a greater typology of market actors than inItaly. Obliged parties have worked together with anumber of market actors to deliver and implementmeasures and thus reduce compliance costs. In fact, it isfound that parties have met their obligations partneringstrategically with multiple actors. Key players support-

37 For ESCOs, the minimum amount of energy savings to berealised is 25 toe/year in case the energy efficiency measuresare evaluated via ex-ante approaches, 50 toe/year in casemeasures are evaluated via the so-called engineering approach,and 100 toe/year in case of measures are evaluated throughmonitoring plans (i.e. ex-post M&V).

Table 4 Activity level of eligible parties during the firstcompliance year under the Italian TWC scheme

Eligible party Issued TWCs (%)

Obliged electricity distribution gridcompanies

Obliged gas distribution grid companies 23.8Non-obliged energy distribution gridcompanies

ESCOs 64.6Total 100

Data source: AEEG (2006)

38 The regulator has stressed several times that it remains opento discuss modifications to the 28-day rule with actual evidenceon whether the rule has restricted or undermined any benefitsresulting from the provision of energy services.

ing the activity level of obliged parties are as follows(Mundaca 2007a; NERA 2006; OFGEM 2005):

& Insulation contractors: To a large extent, obligedparties have relied on subcontracting insulationcompanies to realise savings. Basically 100% of allinsulation was sub-contracted under EEC1. In turn,this increased opportunities for long-term partner-ships—vital for parties to meet their obligation—andcredibility for insulation products within the house-hold sector.

& Managing agents: Obliged parties have also relied,to a lesser extent though, on subcontractingmanagingagents or ‘middlemen’ to plan and implement eligiblemeasures. They have also supported parties toidentify and bundle end users willing to implementmeasures. Due to the fact that dealing with managingagents has been considered risky, obliged partiesstarted to reduce their level of outsourcing by creatingspecialised energy service departments. These unitsmaintain a closed control over agents and contractorsbecause parties still bear the responsibility in casemanaging agents or contractors fail to deliver.

& Social housing programmes (SHP) and charityorganisations: Parties have partnered with these act-ors in order to identify customers and deliver eligiblemeasures, in particular among the priority group.Through these partnerships, parties delivered insula-tion, heating and lightingmeasures.Working togetherwith contractors, a large proportion of insulationmeasures (60%) were delivered through SHPs. Inaddition, 16 (out of 24) million of CFLs weredistributed via SHP, charities and community groups.

& Housing developers: A number of obliged partieshave also partnered with housing developers to im-plement measures. This co-operation has taken theform of parties providing necessary funding toimplement insulation measures in new houses toexceed current building requirements.

& Retailers and manufacturers: Parties set up partner-ships with these actors in order to increase thepenetration of efficient appliances. Parties havegiven financial incentives to retailers to stock moreefficient appliances in return for aggressive market-ing efforts. In turn, obliged parties used sales data toclaim energy savings. Parties have also subsidisedmarket prices and negotiated with retailers/manufac-turers rebates for customers. Working together withretailers andmanufacturers, more than 6.5millionA-

rated appliances were delivered via a partnership ofthis type (see Table 3).

It is interesting to notice that, even though the numberof eligible parties is restricted under the British scheme,numerous sources of transaction costs have been identi-fied. This includes search for information, persuasion ofcustomers, due approval of proposed measures, negotia-tion of agreements/contracts with third parties, randomquality checks, contract/agreement negotiation and lia-bility risks when trading, and due accreditation ofsavings.39 Research shows that the search for informa-tion and persuasion of customers are relevant sources ofTCs upstream in the scheme (Mundaca 2007a). Never-theless, even though the estimated scale might appearburdensome (i.e. 10% to 30% of the total direct inv-estment costs for lighting and insulation, respectively),the scheme has yielded cost-effective savings aspreviously mentioned. Furthermore, despite the fact thatthe number of identified players supporting obliged par-ties has triggered more sources of transaction costs, inparticular during the planning and implementation phasesof measures, these actors have also reduced the burden oftransaction costs for obliged parties (e.g. identification ofcustomer willing to implement measures).

Trading activity

The early experience of the Italian TWC shows intensivetrading activity, at least compared to the British case. Inthe first compliance year, 145,796 TWCs were traded—17% on the spot market and 83% through bilateralcontracts (i.e. company-to-company, not through brokersor exchanges).40 The total traded volume representedroughly 50% of the total TWCs issued (286,837) forthat period (see Fig. 4). Indications for market actorpreference towards bilateral trade can be made. Until theend of the first compliance period, it could be partlyexplained by the fact that the organised market startedon March 2006, only 3 months before the end of thefirst obligation period. It could also be explained bystrategic partnerships between obliged parties and

39 See Mundaca (2007a) for an analysis on the nature and scaleof transaction costs under the British scheme.40 Interestingly, notice that the bilateral market under the EU-ETS is estimated to be substantial. Point Carbon (2007)estimates that the direct bilateral market doubled in size, from100 Mt in 2005 to 200 Mt in 2006.

ESCOs, which were set up prior to the establishment ofthe spot market. In fact, obliged parties with the aim tofulfil their obligation explicitly created some ESCOs,encouraging an ‘intra-obliged party’ trading. In addi-tion, this could reflect the tendency in electricity andnatural gas wholesale markets in which longer, lastingand bigger volume deals on the bilateral exchange aremade and conditions can be kept privately (Philipsonand Willis 1999). Prices for bilateral trade in Italy areunknown; however, higher values than on the spotmarket could be another incentive for ESCOs and othernon-obliged parties to aim for bilateral trading.

In Italy, trading has emerged showing a clearertendency towards ‘to-trade’—contrary to the Britishtrend (discussed below). For the spot market, the GMElaunched the registry for trading on February 2006.41 Asshown in Table 5, several market actors have registeredsince then.42 With both the registry and electronictrading platform in operation, trading on the spot market

(i.e. organised by GME and as opposed to bilateralcontracts) got off the ground on March 2006. During thefirst year, TWC type-I dominated spot trades (15,253),followed by TWC type-II (10,086). TWC type III wasmarginally traded. This can be explained by the fact thatthe cost-recovery mechanism (100 € per each TWCredeemed by obliged parties) established by the author-ity does not apply to savings related to TWC type III.Overall, the volume of TWCs traded has increasedsubstantially during the second compliance year (seeFig. 4). Whereas nearly 25,000 TWCs were traded

1 3 5 7 9

Trading session

Type I (electricity) Type II (gas) Type III (other carrier)

End of 1st

compliance year

(May 31st 2006)

End of 2nd

compliance year

(May 31st 2007)

Fig. 4 Volume of TWCs traded on the Italian spot market. Data source: Gestore del Mercato Elettrico SpA 2007. Note that TWCtype-III was marginally traded (only under session # 4)

Table 5 Registered market actors during first compliance yearunder the Italian TWC scheme (until May 2006)

Market actor Number

Electricity and gas grid distribution companies 32ESCOs 79Tradersa 7Total 118

Data source: GEM (2006)a A ‘trader’ means any registered user that is not an ESCO or anelectricity/gas grid distribution company. Although TWCs areissued only for eligible measures submitted to the AEEG byESCOs and electricity/gas grid companies for certification, anymarket agent is allowed to buy and/or sell certificates on themarket.

41 Registered market actor pays an annual fee of €300 plus 20Euro cents per each TWC exchanged either through bilateralcontracts or on the market organised by GME.42 From the total number of actors, 76 obtained also thequalification of market operators by GME. Within this group,21 are grid distribution companies, 49 are ESCOs and six aretraders.

during the first compliance year, a bit more than100,000 TWC were traded during the second year. Asshown in Fig. 5, prices on the spot market have slightlyfluctuated by slowly falling, in particular for electricitysavings during the second year. This trend can be largelyexplained by the fact that the amount of certificatestraded on the spot market increased by approximately300%. TWC-II has been sold at a higher value thanTWC-I. This has been driven by the higher volume ofTWC type I, which has wider price dispersion thanTWC type II (see Table 6). Despite the large number ofnon-obliged actors, no sign of price volatility isobserved.43

Using a rather narrow definition of cost-effectiveness,pre-conditions such as common price and tradingbetween parties facing different costs are identified inItaly.44 However, the potential free-riding effect andpossible market power exercised by some obligedparties (see below) add uncertainties to any indicationof cost-effectiveness in Italy. In addition, no counter-factual or baseline was available to estimate costsavings resulting from trading.45

We found various market and regulatory aspectspotentially driving the Italian trading activity. Thealready-mentioned option to claim unlimited savingsretroactively from ‘early action’, the provision to bankan unlimited amount of TWCs and the less stringentdefinition of ESCOs may largely explain an excess ofTWCs. Furthermore, the mandatory saving target for2005 represents only a fraction of final energy use inItaly (0.3%). Thus, one can argue that the demand levelfor TWCs—scarcity determined by the mandatorysaving target—is rather low relative to the (excessive)supply of TWC and final energy consumption of thecovered eligible end-use sectors. In addition, marketactors may interpretate that the TWC scheme is only aone off burst policy effort because of the lack of long-term policy goals and political commitment post 2009.

Another factor likely to be driving the Italian tradingactivity relates to the potential market power exercise bysome obliged parties.46 For instance one obliged partyheld 90% of the apportioned electricity saving obliga-tion in 2005. Facing an excess of TWCs on the market,this party might be creating monopsonistic marketconditions (i.e. only one buyer), driving lower TWCprices. In fact, only handful number of actorsconstituted demand and supply of TWCs.47 Moreover,taking into account the excess of TWCs, rent-seekingbehaviour is also likely to be encouraged because ofthe cost recovery rate set by the authorities. Obligedparties obtain €100 for every TWC redeemed so theypocket the difference, i.e. windfall profits.48 Toconfirm this rent-seeking behaviour, notice that arather marginal number of TWC-III, which has noassociated recovery rate, has been traded since thebeginning of the spot market (see Figs. 4 and 5). Therent-seeking behaviour raises distributional equityconcerns, as obliged parties could be getting windfallprofits at the expense of taxpayers.

Concerning trading activity in the British scheme,trading has occurred but to a much lesser extent than inItaly. Despite speculations that no trading at all hadoccurred during EEC1, evidence proves the contrary.For the trading of obligations, two trades were iden-tified (cf. NERA 2006; Mundaca 2007a). First, oneparty (EDF Energy) took over the entire obligation—nearly 100 GWh—of another party (Dee Valley;OFGEM 2005). Second, ‘Opus Energy’ and ‘TelecomPlus’ decided to meet their obligations jointly. In terms

43 High volatility is understood as sudden up and downmovements of spot market prices over very short time periods.

44 Ellerman (2003) uses a similar conceptual approach to addresseconomic efficiency for the SO2 cap-and-trade programme in theUS. The author looks at the emergence of the market and volumeof permits traded as primary evidence for economic efficiency—welfare effects (e.g. social benefits and costs) are not taking intoaccount.

45 Notice that ex-post evidence of administrative and transac-tion costs has not yet been produced.

46 Market power is herein understood as how an obliged partyunder a TWC scheme can manipulate the market to its ownadvantage.

47 During the period January 2005 to May 2006, the number ofbuyers and sellers was distributed as follows (AEEG 2006): (a)On the demand side, 16 companies bought TWCs via bilateralcontracts only, 13 companies bought TWCs on the open marketonly and seven companies used both options. (b) On the supplyside, 10 companies sold TWCs via bilateral contracts only, sixcompanies sold TWCs on the open market only, six companiesused both options, and (c) five companies (two of which wereobliged actors and three ESCOs) both bought and sold TWCs.With the exception of the five companies that both bought andsold TWCs, the AEEG did not specify whether the companiesabove-mentioned were obliged actors or ESCOs, or eligible gasand electricity distributors.48 Note that according to Bohi (1994), if the regulatordetermines favourable cost-recovery mechanisms, parties areunlikely to engage in trading even when the cost of certificatesare lower— a situation that could be seen in Italy.

of trading of energy savings, six obliged partiespurchased energy savings retroactively but generatedunder other government programmes (e.g. Warm Frontprogramme49). A retroactive ‘trading’ option was givenonly during EEC1. According to OFGEM (2005, p. 56),these retroactive trades contributed to 16% of all cavitywall insulations implemented under the EEC1. Infinancial terms, one of the managing agents of theWarm Front programmeEAGAreported that around £10million in energy savings were sold to parties during2003–2004 (House of Commons 2004, p. 25). All tradeswere reported to OFGEM, which provides written ap-proval to parties involved in trading. However, partiesare not required to submit related financial data.

A number of interconnected drivers affecting the lowEEC1 trading activity were found. Among others,potential commercial benefits of non-trading (see “Ap-proach of British obliged parties” for further details),excess of individual supply of savings (driven also by ahigh cost-effective potential in the insulation segmentand certainty about the penalty in case of non-compliance), limited number of eligible parties (raisinghigh potential for market power) and perceivedtransaction costs (related to contract negotiation and

liability risks) were identified as critical factors deter-ring trading (cf. Mundaca 2007a; NERA 2007). Thebanking option also influenced a non-trading behaviour,as parties have seen the scheme as a rolling programmeand increased their own activity level to use savings forfurther commitment periods (also driven changes in theaccreditation of savings—as described in “Bankingprovision for surplus of TWCs”). Furthermore, thenumber of eligible parties has been restricted becauseonly obliged parties are allowed to trade, reducingliquidity. This is crucial because large differences insaving costs among parties are also necessary to triggertrading. Obliged parties also thought that tradingactivity could embrace high negotiation costs, asstrategically sensitive information could be disclosed

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64

Trading session

TWC-I (electricity) TWC-II (gas) TWC-III (other fuel)

Fig. 5 TWC prices on the Italian spot market (Euro per TWC), March 2006 to June 2007. Data source: Gestore del Mercato Elettrico

Table 6 Price statistics for TWC types I and II on the Italianspot market

Type I (electricity) Type II (gas)

Mean 56.3 86.5Median 54 86.7Standard deviation 14.5 7.3Minimum value 30 60Maximum value 80 98

N=64 trading sessions (March 2006 until June 2007). Values inEuros (nominal)

49 For further information, visit http://www.defra.gov.uk/environment/climatechange/uk/household/fuelpoverty/.

to a buyer/seller. In all, these factors have favouredmarket behaviour towards ‘not-to-trade’. In addition,limited trading may have been also affected by a highconcentration of obligation in certain actors and theneed of approval from the authority.

Firm evidence of equalisation of marginal costs inGreat Britain does not yet exist; however, with a handfulnumbers of actors dealing with similar contractors andthird parties, saving costs were likely to be very similarso trading could add only marginal financial gains.Importantly, interviewees indicated that trading was notnecessary because compliance costs were alreadyequated during the competitive bidding process ofsubcontracting insulation measures—which have heavi-ly dominated realised savings.50

Regarding trading activity during EEC2, no avail-able data have yet been reported. However, due to thefact that cost-effective energy efficiency/Great Britainimprovements in may soon be used up, trading islikely to arise as one concrete ‘flexibility’ for partiesto meet their targets cost-effectively. In fact, Britishauthorities have considered whether the EEC couldformally operate as a TWC scheme from 2011onwards (see DEFRA 2006a). However, it is unclearwhether a more formal TWC approach would yieldadditional benefits to the ones delivered already bythe scheme (DEFRA 2006a; NERA 2006).

Potential commercial benefits of non-tradingbehaviour for obliged parties

The lack of trading activity in TWC markets (asdescribed for Great Britain in “Trading activity”) hastriggered concerns that obliged parties are not takingadvantage of the cost savings that trading can generate.Indeed, stakeholders have mentioned that if a non-trading pattern exists, TWC schemes are significantlyless cost-effective than expected—unless the costs ofachieving the saving target are negative or zero for allobliged parties. The non-trading behaviour has reviveddiscussions among scholars and policy makers aboutthe to-trade-or-not-to-trade dilemma (i.e. preference totrade TWC or not). Besides several market andregulatory drivers identified, very little attention has

been given to commercial benefits of non-trading, anddrivers behind this pattern have not been fullyaddressed.

Approach of British obliged parties

As found, a low level of trading activity characterisedthe EEC1. Among them, interviews with obliged partiesrevealed two important commercial benefits of in-creased energy efficiency that provided little incentiveto engage in trading.

The first benefit relates to attaining strategic knowl-edge about energy efficiency. Although the EEC wasbuilt upon the EESOP, energy efficiency was still a newactivity for obliged parties. Thus, instead of relying on acompetitor for meeting their obligations, partiesembarked to autonomously meet their target. As theEEC becomes a rolling programme, a strategic learningprocess has been crucial for obliged parties to gain thenecessary operational knowledge of meeting their obli-gation in the long run. In turn, this knowledge hasbecome a key building block for some parties that havetaken energy efficiency, encouraged by the EEC, as abusiness opportunity (see below). They started to hire andtrain staff, and create energy service units to handle theirobligation. It is found that all current obliged partiesprovide advice and support their customers to increaseenergy efficiency.

The second benefit relates to increased competitive-ness. Some obliged parties have considered energyefficiency as a business opportunity for enlarging theirproduct and customer portfolio. Interviews with someparties revealed that EEC1 triggered a change in theirbusiness paradigm, leading to corporate efforts for en-hancing customer loyalty and branding (Mundaca2007a).51 Driven by climate change policies, energycompanies have started moving away from the traditionalenergy supply business. More importantly, customermobility appears to be the main driving force to useenergy efficiency as a strategy to increase client loyalty.In fact, the already mentioned ‘28-day rule’ for domesticenergy service contract—that prevents households frombeing locked into long-term energy contracts—has alsoencouraged obliged parties to use the EEC as an in-vestment mechanism to ensure long-term commercialrelationships with their clients. A concluding argument in

51 Ibid.

50 Interviews with British obliged parties were carried inOctober 2005 and March 2006. Some results were partlypublished in Mundaca (2007a).

this aspect also came from the fact that, for some parties,buying energy savings from another company wasunderstood as strategically ‘non sense’, as it would implydirect financial support to the brand of a competitor(Mundaca 2007a). One interviewee pointed out thattrading was unacceptable for executive board members.

Supplementary and more explicit indications abouthow the EEC is being used to boost the competitivenessof obliged parties are observed when reviewing recentannual reports. With an exclusive corporate focus onenergy efficiency, EDF UK has delivered a commercialreport fully dedicated to the EEC “calling for greateraction to be taken on energy efficiency across Europe”(EDF 2006, p. 1). It is claimed that the EEC hasbrought benefits not only to households but also toEDF business. Furthermore, the EEC is seen as a valuedriver also for increasing staff awareness and reducingcosts (e.g. because of competitive bidding process).Centrica—who owns British Gas, another obligedparty—claims “our commitment to meet the growingconsumer demand for energy services will be at theheart of our endeavours” (Centrica 2006, p. 1). In itscorporate responsibility policy report, Scottish andSouthern Energy (2007) mentions that is committedto promote energy efficiency amongst its customers. Ithas an energy service business unit that provides arange of products and ‘beyond-the-metre’ services toall end-use sectors.

Now, why have these benefit aspects of non-tradingnot been identified in Italy? A possible explanation liesin the design of the scheme. Obliged parties aredistributors of energy, not suppliers as in Great Britainor France. For gas and electricity distributors, commer-cial benefits of increased energy efficiencymight be lessappealing or difficult to capitalise in the distributionbusinesses, as they do not have direct access to endusers. Italian obliged parties do not have to deal withincreasing customer mobility, as their direct clients areenergy suppliers instead.

Early indications from a relevant Frenchobliged party52

Early trends in the French TWC scheme also indicatethat commercial benefits of non-trading might drive a

non-trading pattern. We focus on EDF as this is thelargest obliged party that bears 55% (30 TWh) of thetotal obligation (i.e. 54 TWh)—among 2,400 energysuppliers. Because of its high market obligation, it islikely that its strategic approach will influence much ofthe characterisation of the French TWC market.

To begin with, the business model of this partyfocuses on the changing role of energy suppliers to-wards creating value out of energy efficiency. Drivenalso by climate change policy aspects, this companywants to position itself as a provider of energy efficientand low-carbon solutions downstream on the energymarket. Increasing the provision of energy services andintegrated micro-scale renewable energy technologysystems might do this. EDF claims that a TWC schemerepresents an opportunity to strengthen and thus ma-terialise this new business model. The company views itas a clever regulatory approach to encourage marketchange through energy suppliers.

Important to our study is the fact that the company hasstated that it aims to increase its competitiveness bymeeting its obligation autonomously. The TWC schemeis perceived as an opportunity to strengthen its businessmodel. It is claimed that by integrating energy efficiencyinto its core business, increased competitiveness becomesa relevant benefit resulting from the (intended) decisionto dismiss the trading option a priori. The most likelyinter-organisational learning and strategic process be-tween EDF France and EDF UK could also explain thisapproach. Based on its business model, indications of theintended autarky approach derived from:

& Introducing energy efficiency in all market seg-ments, with specific targets for marketing and salesstaff.

& Increasing capacity building for the supply of‘energy services’,53 partnering with ESCOs, man-ufacturers, retailers and contractors of efficientmeasures.

& Increasing demand of energy efficiency by launch-ing customer awareness raising campaigns andadvice centres, including the provision of softloans.

53 The term energy service refers to the delivered benefits ofuseful energy consumption such as heating, refrigeration,lightning, cooking, transportation, etc., as opposed to thesimple provision of units of energy as such (see Blok 2006;Johansson and Goldemberg 2002).52 For details, see Urvoas (2007) and Urvoas et al. (2007).

& Developing R&D programmes for integrated sol-utions and customer behaviour.

& Purchasing TWCs to third parties only when themarket price is substantially lower than its ownenergy-saving costs.

It is unknown whether other obliged parties will adopta similar business strategy; nevertheless, one has toconsider that due to the high concentration of theobligation in only one party, market power is likely toarise in France—at least during the ‘training period’, asthe French authority has described the scheme so far.EDF could influence the performance of the TWCmarketby setting TWC prices, which would create monopolistic(seller) or monopsonistic (buyer) conditions. Further-more, given its high market share, the firm can exploiteconomies of scale in the implementation of measures. Inany case, one has to bear in mind that the supply anddemand of TWCs could be very sensitive to pricechanges and affect the EDF autarky position. This isrelevant because high price sensitivity of TWCs supplycould deter an obliged party with market power to benefitof its high market/obligation share (cf. Hahn 1984).

‘Food for thought’ from a simulated EU-wideTWC market

Based on game theory, we undertook a simulation studyto explore the feasibility and desirability of tradingunder an EU-wide TWC scheme. In this section, wefocus exclusively on aspects relevant to non-tradingmarket behaviour. For details about the simulationexercise and its outcomes, see Adnot et al. (2006a,2006b, 2007) and Duplessis et al. (2007). Projectresults and interviews with key stakeholders were usedto design and simulate features of different TWCschemes. Appointed as obliged parties, a certain numberof players were recruited among national stakeholdersand project partners (from Austria, Bulgaria, Finland,France, Italy and the UK) to represent TWC demand. Tosimulate the TWC supply, we built a mathematical modelbased on technical and financial data of existing energy-saving potentials existing in the participating countries.54

Three different schemes were simulated and tradingsessions took place via internet. During each of the threesimulation rounds, a national saving target was estab-lished for each of the participating countries: 3% of

domestic annual electricity demand in the end-use sectorsconsidered for the supply of TWCs. Such a target wasapportioned among the players of the same country(representing electricity suppliers) who could choosewhether achieving their target by (a) implementing end-use energy-saving actions within their customer portfolio,(b) buying TWCs during the 12 market sessions(simulating 12 quarters of a 3-year compliance period)and/or (c) paying penalties at the end of such a period. Anover-the-counter international market consisting of bilat-eral trades was designed for the first simulation round.55

For the second and third round, simulated national TWCschemes integrated into an EU trading platform wereanalysed.56 This entailed completely anonymous TWCsmade of national surpluses that could be tradedinternationally either in one market session at the endof the simulated 3-year obligation period (second round)or during each market session across the obligationperiod. All simulation rounds allowed testing marketaccess, auctioning rules (how bids were handled),market transparency and how variations of these featuresaffected the size and distribution of trading flows as wellas price development.

Findings suggest scarce interest towards cross EU-trading activity. For instance, during the third simulationround, only 9% of the issued TWCs were exchangedamongst obliged parties. At the same time, the observedTWC price remained almost constant during the first 10trading sessions simulated, and such a price was sig-nificantly higher than its own energy saving costs. Thedominant strategy for players was to regularly implementeligible measures within their portfolio of customers tomeet their obligation cost-effectively. What seemed alsoobservable from the bidding behaviour is that someparticipants were primarily driven by target fulfilment—

54 Such computational model automatically determines thenumber of TWCs available at a given price.

56 It has to be pointed out that the games carried out were infact auctions instead of bilateral trade or trade via an exchange.This is because players were not informed about TWC spotprice (at the opening of each market session, the ‘auctionmaster’ gave only indications about TWC average price inprevious sessions); the consequence of this was that there wasonly one way trade with the auction master according to pre-defined auctioning rules (how the TWCs exactly came fromdifferent saving potentials in different countries remainedlargely in the dark for the players).

55 Interviewees indicated that bilateral trading could be apreferred alternative compared to an open trading platform.This hypothesis is confirmed by the experience in Italy inwhich bilateral trades have dominated the trading activity sofar.

triggering over compliance—and others by cost mini-misation.57 In turn, an excess of TWCs was observedand partly explained by the penalty of €50/TWC in caseof non-compliance.

The observed non-trading behaviour might be partlybut not totally explained by the design of the simulatedTWC scheme. Indeed, such design attempted to capturethe potential loss of gross income of energy suppliersdue to increased efficiency and its possible compensa-tion by an increase in customers’ loyalty. This featurewas simulated by decreasing the obliged party energy-saving costs by 20% when TWCs came from their owncustomer portfolio. This incentive turned out to berelevant for driving a self-sufficiency complianceapproach; however, this bonus was accounted at theend of the simulation game, so it was not explicitlyrepresented when trading actually took place. Otherreasons for the observed non-trading behaviour could befound in the misinterpretation of the rules by someplayers and also by the discontinuous participation inthe trading sessions by some others.58 Nevertheless,these issues were registered with very few players anddo not seem to explain the observed trend.

Interviews performed with stakeholders whendesigning the exercise prior to the simulation gameseem to support the thesis of scarce interest towardscross-EU trading. This may be partly explained by theawareness among players of the benefits that energy-saving measures implemented with their own cus-tomers might produce. At that time, intervieweesrepresenting energy suppliers in particular argued thatimplementing eligible projects would have a negativeimpact on energy sales but a positive effect oncustomer loyalty. For the latter, they mentioned thatthe implementation of measures within their portfolioof customers would be a rational strategy to secure orincrease the number of customers and eventually alsoboost a hypothetical “green business image”. Inter-viewees mentioned that increased energy efficiency,

triggered by a TWC scheme, should be seen as anopportunity to enhance competitiveness and differen-tiate from competitors. In fact, stakeholders fromcountries familiar with energy efficiency programmesstressed the view that the provision of energy servicescould be taken as a strategy to secure customerloyalty—moreover, in competitive energy markets inwhich fierce competition for clients exist. In all,interviews revealed that game players were consciousabout ancillary benefits of increased energy efficien-cy. However, although the existing awareness of suchbenefits might partly explain the non-trading behav-iour during the simulations, it has to be mentionedthat such awareness was observed in particular inplayers from countries in which real TWC schemesare implemented. Therefore, it cannot be excludedthat a different group of players from differentcountries could have yielded a different tradingbehaviour.59

Discussion on non-trading aspectsunder TWC schemes

Ellerman et al. (2000) argue that competitive tradingmarkets will develop when policy design and imple-mentation are encouraging. Although trading activityhas shown opposite patterns in current TWC schemes,a number of exogenous and endogenous factors in-fluencing trading can be identified. Among them, fiercecustomer competition, market liquidity, early/retroactivesaving action, existence (or not) of a trading platform,transaction costs, cost-effective saving potentials, addi-tionality, banking and familiarity with trading markets. Itis clear that TWC markets are quite new so obligedparties do not have much of trading experience orbusiness models to cope with this new policy instru-ment.60 Still, we identify that policy makers can takeseveral measures to encourage trading in TWC markets.Certainly, an ambitious energy saving target is a key

58 For instance, one player decided to retain a surplus of TWCat the end of the obligation period, although no banking optionwas envisaged by the simulation rules established.

57 For example, it is remarkable that minimum and maximumprice levels did not converge during the first simulation round;this fact in conjunction with the development of the averageprice level from Q1 to Q12 indicates that several participantseither did not learn much (from an economic viewpoint) orwere—at least in some quarters—guided by non-economiccriteria. Personal communication with Adriaan Perrels (VATT),June 2006.

59 To advance research in this area, future simulations shouldinclude a control group of players and explicitly representobliged party’s decision-making behaviour as far as commercialbenefits of energy efficiency are concerned. Behavioural andexperimental economics can greatly contribute to this task.60 Evidence shows that many obliged parties were unfamiliarwith trading during the beginning of the SO2 cap-and-tradeprogramme in the US. This aspect motivated parties to exercisean autarky compliance policy (see Bohi 1994).

pre-requisite. Furthermore, the implementation of aclearinghouse to provide information about marketprices, volumes and parties (both spot and bilateral)should be considered, thereby increasing transparency.A clearinghouse can keep market actors updated andwell informed about the dynamics of the TWC marketand its regulatory framework. In addition, an electronictrading platform can reduce transaction costs (e.g.search costs) by setting the place where buyers andsellers can meet up regularly, allowing bids and bilateraltrading as well. The development of standardisedcontracts (or at least key contractual provisions) canreduce transaction costs related to legal services andperceived liability risks when trading. In Great Britain,the need of written approval for trading from theauthority could be dismissed.

A non-trading trend also reminds us that trading isrelevant but not an objective per se in TWC schemes—as in any tradable certificate scheme. We must remem-ber that the trading component aims at enhancing thescheme’s cost-effectiveness to meet a mandatory energysaving target at lowest cost. In line with some critics inthe context of a cap-and-trade scheme for greenhousegases (cf. Greenspan Bell 2005), we concur with thefact that what really matters in TWC schemes is the‘target’ as such. Target compliance depends on manyfactors, among them, a functioning and enforceable re-gulatory framework. However, a crucial pre-conditionto determine the demand level for TWCs is the es-tablishment of mandatory energy saving/efficiencytargets. If increased energy efficiency improvementsare left to market forces alone, ‘business-as-usual’trends are likely to be expected. Currently, the level ofambition of the mandatory saving targets of allimplemented TWC schemes can be considered low asfar as final energy consumption (on an annual basis) ofthe eligible sectors is concerned. For France, the savingtarget level equates to a reduction on energy consump-tion of 0.14% per year; for Italy, 0.3%; and for GreatBritain, approximately 0.6% (Mundaca 2008). Un-doubtedly, more ambitious targets can influence thetrading activity and, thus, the scarcity of TWCs in thelong term, driving a more dynamic market behaviour.

The findings on non-trading behaviour stress ancil-lary benefits with increased energy efficiency. Whereasregulators and observers seem to be mainly concernedwith the cost savings that can be accomplished throughtrading, obliged parties—at least in Great Britain andFrance—also seem interested in commercial benefits of

increased efficiency. At first, it seems that parties mightnot be taking full advantages of TWC schemes,hampering the underlying cost-effectiveness rationale.However, we found that a non-trading behaviour is alsoexplained by corporate and commercial strategies aimedat capitalising several ancillary effects of increasedenergy efficiency. As indicated by the British experi-ence, the intended strategy of the main French obligedparty and the market behaviour in the simulationexercise, policy makers and interested observers shouldnot hold their breath for trading activity to take off. Atleast in the short run, it is likely that this is not animmediate outcome of TWC schemes. Although severalaspects affected the trading activity in Great Britain, ourfindings suggest that commercial benefits associatedwith increased ‘branding’ and ‘customer loyalty’ couldyield higher financial gains for parties than the costsavings resulting from trading. However, it remains tobe seen whether this approach is an optimal choice forparties. At the EU level, one can also argue that anational autarky approach may be adopted to guaranteethat ancillary benefits (private and social) of increasedenergy efficiency are captured nationally (e.g. reducedlocal negative externalities from power production,increased security of energy supply, etc.).61 As sug-gested by the simulation game, distributional effectsmight trigger scarce interest towards internationaltrading. Altogether, findings seem to support thehypothesis that distributional effects might deterinternational trading if an EU-wide TWC scheme wereimplemented (Mundaca 2007b). Non-trading aspectsremind us that efficiency can be gained at the expenseof equity—a complex trade-off and political issue.

Finally, one should consider that a TWC scheme couldstill be efficient even if limited trading and liquidity isobserved. In fact, it is argued that a trading scheme inwhich no trading activity takes place is still likely to yieldlower cost savings than a command-and-controlapproach. This is because obliged parties still havemore flexibility in choosing their technological options(Stavins 1995; Tietenberg 2006). As shown, a high cost-effective potential in the insulation segment has allowedthe British scheme to yield net financial benefits for endusers. Indications of economic efficiency (i.e. max-imisation of net benefits for the society) exist. It isestimated that for every euro saved in electricity costs,

61 See Mundaca (2007b) for a discussion on potential distribu-tional socio-economic effects of an EU-wide TWC scheme.

another 56 to 92 Euro cents/kWh were obtained becauseof social and environmental benefits from the reductionof atmospheric pollution of power generation (Mundaca,2007a, p. 4349). These figures do not change signifi-cantly if one considers administrative and enforcementcosts borne by the regulator. Public expenditure wasapproximately 1.4 million euros (£1 million; seeOFGEM 2005, p. 4), a marginal figure compared tothe overall level of private investment costs, approxi-mately 970 million euros (£690 million; see Lees 2006,p. 27). A recent study on the economic efficiency of thefirst phase of the British scheme estimates net presentvalues for the society within the range of approximately€1,660 to 1,830 Million (see Mundaca 2008).62

Conclusions

The objective of this study was to provide empiricalevidence of market behaviour under TWC schemes.We focused our analysis on the full set of existingflexibilities present in the British and Italian TWCmarkets. We argue that the analysis of all flexibilitiesand potential ancillary effects (not captured by TWCtrading) is crucial to have a holistic understanding ofmarket behaviour under TWC markets. A non-tradingpattern was identified so the ‘to-trade-or-not-trade’dilemma was further explored. We have focused oncost-effectiveness as an evaluation criterion only. Ourresults are rather case-, policy- and country-specific;therefore, the analysed market behaviour and identi-fied performance respond to the unique design, policyand market conditions in which the studied TWCschemes are implemented. Findings suggest that theoverall market activity is dynamic and slowly emerging.However, further research should aim at ‘de-linking’ theeffects (impacts and outcomes) of TWC schemes fromother policy instrumentsa complex evaluation challenge.

In terms of cost-effectiveness, clearer indicationscan be drawn for Great Britain than for Italy.Concerning the former, indications that the schemehas met the target cost-effectively come from severalangles: lower saving costs than an alternative policyoption, equalisation of marginal costs during thebidding process of insulation measures, energy savingcost lower than expected and lower than energy pricespaid by households. Even though trading activity wasrather limited, trading did occur during the first phaseof the British scheme. However, little can be said

about savings resulting from trading because partiesare not required to disclose any related financialinformation. Regarding Italy, pre-conditions such ascommon price and trading between parties facingdifferent costs are observed; however, the identifica-tion of an alternative policy instrument and develop-ment of a credible counterfactual situation remain as achallenge for a thorough evaluation under thiscriterion. For instance, we were not in the positionto ascertain whether the substantial volume of trading(approximately 50% of the total TWCs issued) had animportant cost reduction. Furthermore, the potentialfree-riding effect and the eventual market powerexercised by some companies add complexities anduncertainties to this element. Regardless the identifieddesign drawbacks, it can be concluded that obligedparties are using, to some or to a large extent, all ofthe flexibilities granted to comply cost-effectively.Our study shows that high liquidity is crucial forparties to take full advantage of all given flexibilities.

Concerning energy-saving effectiveness, the Italianscheme achieved the imposed targets for the first year;however, the potential free-riding effect created by theretroactive option to claim savings, combined with theless stringent definition of ESCOs and the unlimitedamount of banking, prevent serious assertions in thisregard. Concerns about the free-riding effect in relationto the early action provision are valid because it isunclear which proportion of eligible technologies wereactually implemented in anticipation of the scheme or asa result of business-as-usual market trends. In addition,the low ambitious savings target questions its earlyperformance. One can argue that a high energy-savingeffectiveness is observed but at the expense of softtargets and pitfalls in the regulatory framework. Targetcompliance under the British schemewas nearly perfect,with a minor shortfall during its first phase. Partiesrelied on large cost-effective potentials in the insulationsegment to meet their targets independently. Until now,mostly commercially available eligible measures havebeen implemented so technical change does not seem tobe encouraged. However, this trend could change in thenear term when the most cost-effective potentials areexhausted. In sum, our study shows that the integrityand effectiveness of a mandatory energy-saving targetrelies critically on how ambitious the target is, effectivenon-compliance rules, due enforcement, stringent andenforceable definition of additionality, and reduced free-riding effect. Furthermore, it also depends on energy

efficiency potentials in eligible end-use sectors, and thusthe variety and related costs of current and new eligiblemeasures that can yield those potentials.We suggest thatthe energy-saving effectiveness of TWC schemesshould be weighted against the level of ambition thatthe saving targets involve. This level appears to be lowwhen measured as a proportion of final energyconsumption of eligible sectors—under 1% in all cases.Ambitious energy saving targets will trigger a moredynamic usage of all flexibilities and, thus, an activeTWC market behaviour.

Initial market and institutional conditions stronglysuggest that trading might not be an immediate outcomeof TWC schemes. On one hand, a real TWC market hasemerged only in Italy, in which obliged parties areenergy distributors. Trading volume has increased,which shows a clearer preference towards ‘to-trade’than in Great Britain. On the other hand, British parties(i.e. energy suppliers) have shown a clearer inclinationtowards ‘not-to-trade’ and seem to be more interested inincreased competitiveness resulting from energy effi-ciency project activities. Combined with strategiccommercial aspects, lowmarket liquidity, a high penaltyfee in the case of non-compliance, a banking option,among others factors, trading did take place but to amuch lesser extent than in Italy. Obliged parties havebeen actively using the banking provision—at leastunder the British scheme. In Great Britain, the patternwas predictable as regulatory certainty encouragedparties to rationally and cost-effectively use the bankingoption. As the British regulatory framework evolvese-ventually towards a certificate-based marketand cost-effective potentials become gradually used up, trading islikely to emerge in the long-term. Nonetheless, policymakers still have room to implement a number of policymeasures to actively encourage trading. In all, findingsstrongly indicate that a simple but effective institutionalframework is crucial for parties to take full advantage ofgiven flexibilities. A secured long-term policy horizonis relevant to reduce regulatory uncertainties so obligedparties can factor the costs and benefits of increasedenergy efficiency into their business plan. Politicalcommitment is critical in ensuring confidence inemerging TWC markets.

Finally an interesting, perhaps unexpected, outcomeof TWC schemes seems to be their ability to influencethe traditional business paradigm of energy supplierstowards increased energy efficiency. At first, it appearedthat companies are not taking full advantage of trading.

Despite numerous markets and regulatory driverspreventing trading, corporate business strategies aimedat increased market competitiveness also drive a non-trading behaviour in Great Britain and France. With duelimitations, the simulation game also revealed a nationalautarky approach under a hypothetical EU-wide TWCscheme. From a policy perspective, it still remains to beseen whether benefits of non-trading and positiveexternalities of increased energy efficiency—which alsoseem to influence a national autarky approach—add ornot to the overall efficiency of TWC schemes. Thesefindings seem to counterbalance the complex politicalfactor for policy makers to force energy suppliers toactively save energy; boosting the political feasibility ofTWC schemes. Further research on industrial compet-itiveness and positive externalities of increased energyefficiency needs to be done—posing a serious challengefor a thorough assessment. Aspects related to non-trading patterns strongly suggest that the performance ofTWC schemes should not be evaluated exclusively onthe basis of trading.

Acknowledgement We thank the guest editors of this specialissue and anonymous referees for proving useful comments. Thepaper builds mostly upon the work carried out under theEuroWhiteCert project (www.eurowhitecert.org). Financial sup-port from the Intelligent Energy for Europe (EIE) Programme ofthe European Community is gratefully acknowledged (contractno. EIE/04/123/S07.38640). The usual disclaimers apply.

References

Adnot, J., Duplessis, B., Almeida, A., Fonseca, P., Moura, P.,Ferreira, C., et al. (2006a). Supply side: Measurement andverification of energy efficiency projects. Task report workpackage 4.1. EuroWhiteCert Project.

Adnot, J., Duplessis, B., Perrels, A., Hacheni, M., Barroso, L.M.,Marca, E., et al. (2006b). Creation of a database of supply ofcertified projects—Modelling the supply side. Task reportwork package 4.2. EuroWhiteCert Project.

Adnot, J., Duplessis, B., Labanca, N., & Rezessy, S. (2007).Demand side: identifying alternative market participanttypes and structuring the demand side. Task report workpackage 4.3. EuroWhiteCert Project.

Autorità per l’energia elettrica e il gas (AEEG) (2006). Primorapporto annuale sul meccanismo dei titoli di efficienzaenergetica [First annual report on the mechanism of energyefficiency titles]. Retrieved June 2006 from http://www.autorita.energia.it/ee/rapporto_06.pdf

Autorità per l’energia elettrica e il gas (AEEG) (2007). Energia:risparmiate 900 mila tonnellate equivalenti di petrolio con i‘certificati bianchi’. Press release. Retrieved September2007, from http://www.autorita.energia.it/com_stampa/index.htm.

Blok, K. (2006). Introduction to energy analysis. Amsterdam:Techne Press.

Bohi, D. (1994). Utilities and state regulators are failing to takeadvantage of emission trading. The Electricity Journal, 7(2),20–27.

Capozza, A., Devine, M., Enge, A., Grattieri, W., Monjon, S.,Mundaca, L., et al. (2006). Market Mechanisms for whitecertificate trading. IEA-DSM task XIV final report: Based onnational and international studies and experiences. IEA-DSM. Retrieved July 2006, from http://dsm.iea.org.

Centrica (2006). Annual review and summary financial statements2006. Review: Securing our customers’ energy needs.Retrieved July 2007, from http://www.centrica.com/files/reports/2006ar/files/2006_annual_review.pdf

DEFRA (2004a) Comparing the Energy Efficiency Commitment(EEC) 2002–05 with 2005–08. Retrieved November 2004,from http://www.defra.gov.uk/environment/climatechange/uk/household/eec/index.htm.

DEFRA (2004b). The Energy Efficiency Commitment from April2005. Consultation proposal. Department for Environment,Food and Rural Affairs. UK: DEFRA Publications.

DEFRA (2005) Energy Efficiency Commitment 2005–08. Back-ground information on the illustrative mix. Retrieved March2005, from http://www.defra.gov.uk/environment/climatechange/uk/household/eec/pdf/measuremix.pdf.

DEFRA (2006a). The Energy Efficiency Commitment April 2008to March 2011. Initial consultation. Retrieved November2006, from http://www.defra.gov.uk/corporate/consult/eec3/consultation.pdf.

DEFRA (2006b). The first illustrative mix of measures for theEnergy Efficiency Commitment 2008–11. Retrieved De-cember 2006, from http://www.defra.gov.uk/environment/climatechange/uk/household/eec/pdf/illustrative-mix.pdf.

DEFRA (2006c). Assessment of EEC 2002–2005 carbon, energyand cost savings. Retrieved August 2006, from http://www.defra.gov.uk/environment/climatechange/uk/household/eec/pdf/eec-assessment.pdf.

DEFRA (2007). Energy, cost and carbon savings calculations for thedraft EEC 2008–11 illustrative mix. Energy Efficiency Com-mitment 2008–11. Retrieved March 2007, from http://www.defra.gov.uk/corporate/consult/eec3-2007/eec-illustrativemix.pdf.

DTI, 2005. Quarterly energy prices. Department of Trade andIndustry. Retrieved May 2006, from http://www.dti.gov.uk/files/file26972.pdf.

Duplessis, B., Adnot, J.Moura, P., Labanca, N. (2007). Simulatinga European-wide white certificate schemes: Design issuesand main lessons. Paper presented at the Summer Study ofthe European Council for an Energy Efficient Economy, Coted’Azur, France.

EDF (2006). Energy for everyone—Energy Efficiency Commit-ment. Annual report 2005–2006. Retrieved March 2007, fromhttp://www.edfenergy.com/edf-energy/downloads/EEC%20Annual%20Report%2005-06.pdf.

Ellerman, A. D. (2003) Ex post evaluation of tradable permits: TheU.S. SO2 cap-and-trade program. Center for energy andenvironmental policy research at MIT. Retrieved March2004, from http://web.mit.edu/ceepr/www/publications/workingpapers/2003-003.pdf.

Ellerman, A. D., Joskow, P. L., Schmalensee, R., Montero, J. P., &Bailey, E. M. (2000). Markets for clean air. The U.S. acidrain program. USA: Cambridge University Press.

Gaudioso, D., Klink, J., & Lagniss, O. (2007). ESCOs in theItalian and UK energy efficiency schemes. Annex 5—Taskreport work package 5. EuroWhiteCert Project.

Gestore del Mercato Elettrico SpA (2007). Energy efficiencycertificates results. Retrieved July 2007, from http://www.mercatoelettrico.org/GmeWebInglese/Default.aspx.

Greenspan Bell, R. (2005). No magic solutions: what is wrongwith current plans to manage climate change. Resourcesfor the Future. Resources, Fall(159), 11–12.

Hahn, R. W. (1984). Market power and transferable property rights.The Quarterly Journal of Economics, 99(4), 753–765.

Harrington, W., Morgenstern, R. D. & Sterner, T. (Eds.) (2004).Choosing environmental policy: Comparing instrumentsand outcomes in the United States and Europe. USA: RFFPress

House of Commons (2004). Committee of Public Accounts‘Warm Front: helping to combat fuel poverty’ 5th Reportof Session 2003–2004. Retrieved November 2000, fromhttp://www.publications.parliament.uk/pa/cm200304/cmselect/cmpubacc/206/206.pdf.

Johansson, T. B., & Goldemberg, J. (Eds.) (2002). Energy forsustainable development - A policy agenda. New York:UNDP.

Lees, E. (2006). Evaluation of the Energy Efficiency Commit-ment 2002–2005. Report to DEFRA. Eoin Lees Energy, UK.Retrieved May 2006, from http://www.DEFRA.gov.uk/environment/energy/eec/pdf/eec-evaluation.pdf.

Mundaca, L. (2007a). Transaction costs of tradable whitecertificate schemes: the Energy Efficiency Commitment ascase study. Energy Policy, 35(8), 4340–4354.

Mundaca, L. (2007b). Markets for energy efficiency. Potentialimplications of an EU-wide ‘tradable white certificate’scheme. Working paper. International Institute for IndustrialEnvironmental Economics. Sweden: Lund University.

Mundaca, L. (2008). A multi-criteria evaluation framework for‘tradable white certificates’ schemes. Working paper.International Institute for Industrial Environmental Eco-nomics. Sweden: Lund University.

Mundaca, L., & Neij, L. (2006). Tradable white certificateschemes: What can we learn from early experiences inother Countries? A Swedish national report in the contextof the IEA-DSM Task XIV Market Mechanisms for WhiteCertificates Trading. International Institute for IndustrialEnvironmental Economics. Sweden: Lund University.

NERA (2006). Energy efficiency and trading. Part I: Options forincreased trading in the Energy Efficiency Commitment.Report prepared for DEFRA. NERA Economic Consulting.Retrieved May 2006, from http://www.DEFRA.gov.uk/environment/energy/eec/pdf/eec-trading.pdf.

NIA (2006). Submission by the National Insulation Association tothe 2006UKEnergy Review. RetrievedNovember 2006, fromhttp://www.nationalinsulationassociation.org.uk/stakeholder/policy/2006%20Energy%20Review%20response.pdf.

Nordhaus, R. R., & Danish, K. W. (2003). Designing a mandatorygreenhouse gas reduction program for the U.S. RetrievedNovember 2003, from http://www.pewclimate.org/.

OECD (1999). Implementing domestic tradable permits forenvironmental protection. Proceedings OECD workshop.Paris: OECD.

OECD (2002). Implementing domestic tradable permits: Recentdevelopments and future challenges. Proceedings OECDworkshop. Paris: OECD.

OFGEM&EST (2003). A review of the energy efficiency standardsof performance 1994–2002. Retrieved December 2005, fromhttp://www.ofgem.gov.uk/Sustainability/Environmnt/EnergyEff/Documents1/4211-EESoP_report_July03.pdf.

OFGEM (2005). A review of the Energy Efficiency Commitment2002–2005. A report for the Secretary of State for Environ-mental, Food and Rural Affairs. Retrieved August 2005,from http://www.OFGEM.gov.uk/temp/OFGEM/cache/cmsattach/12015_18105.pdf.

OFGEM (2007) A review of the second year of EnergyEfficiency Commitment 2005–2008. Annual report forthe Secretary of State for Environmental, Food and RuralAffairs. Retrieved August 2007, from http://www.ofgem.gov.uk/Sustainability/Environmnt/EnergyEff/Pages/EnergyEff.aspx.

Point Carbon (2007). Press release March 2007. RetrievedNovember 2007, from http://www.pointcarbon.com/getfile.php/fileelement_103925/13_March_2007_EU_ETS_now_significantly_reducing_emissions.pdf.

Philipson, L., &Willis, H. (1999).Understanding electric utilitiesand de-regulation. New York, Basel: Marcel Dekker.

Rohr, C. (2004). Energy efficiency: The government’s plan foraction—Drivers, policies and measures. Presentation at IEA-DSM Task XIV workshop on ‘Market Mechanisms forWhite Certificates Trading’, London, UK.

Rubin, J. (1996). A model for intertemporal emission trading,banking and borrowing. Journal of Environmental Econom-ics and Management, 31, 269–286.

Scottish and Southern Energy (2007). Corporate responsibilitypolicy. Director of Corporate communications. RetrievedJuly 2007, from http://www.scottish-southern.co.uk/SSEInternet/WorkArea/showcontent.aspx?id=890.

Stavins, R. N. (1995). Transaction costs and tradable permits.Journal of Environmental Economics and Management,29(2), 133–148.

Tietenberg, T. H. (2006). Emissions trading: Principles andpractice (2nd ed.). USA: RFF Press.

Urvoas, L. J. (2007). From commodity seller to sustainableenergy supplier—The changing role of utilities. Presenta-tion at the 2007 Summer Study of the European Councilfor an Energy Efficient Economy, Cote d’Azur, France.

Urvoas, L. J., Glachant, d., Lorge, L., Baudry, P. (2007). Theaction of EDF in end-use energy efficiency. Paper presentedat the Summer Study of the European Council for an EnergyEfficient Economy, Cote d’Azur, France.

Voogt, M., Luttmer, M., de Visser, E. (2006). Review and analysisof national and regional certificates schemes. Task reportwork package 2. EuroWhiteCert Project.

Market behaviour and the to-trade-or-not-to-trade dilemma in ‘tradable white certificate’...

Documents

Transcript of Market behaviour and the to-trade-or-not-to-trade dilemma in ‘tradable white certificate’...

The dilemma over minimum wage

Feminists' Dilemma: With or Without the State?

Anomalous Behavior in a Traveler's Dilemma

Hans Thorborgs dilemma-3

CQG TRADABLE SYMBOLS | Qbitia

The Identity Dilemma in The Impressionist

Modeling Moral Dilemma Judgment

Spectrum Brands: The Sales Force Dilemma - baixardoc

Dilemma Demonstration P Duhamel v2

A Chunk of Wood's Dilemma

Zombie Companies in Portugal - The non-tradable sectors of ...

Dilemma ......One-Act Play

Convenience or Nuisance?: The ‘WhatsApp’ Dilemma

Israel's Dilemma: Unity or Peace

A Yablovian Dilemma

FRENCH AND US ARMY'S DILEMMA BETWEEN ATHENA ...

An Asian Dilemma - CiteSeerX

SERBIAN DILEMMA: WATER ECONOMY OR WATER MANAGEMENT?

Plato's Dilemma

History without Periods: Dilemma or Deliverance?