1. SLED scenario assessment for Montenegro László Szabó, PhD – András Mezősi PhD Regional Centre for Energy Policy Research Podgorica, Montenegro October 27, 2015

2. Outline of the presentation 1. Modelling methodology 2. Scenario definitions ‣Main assumptions on demand supply and taxation ‣Input data 3. Model results ‣Prices ‣Generation mix ‣Carbon emissions ‣RES support costs ‣Investment costs 2

3. Methodology The SLED analysis is based on assessing three scenarios: Reference scenario (REF); Currently Planned Policies (CPP); Ambitious Climate Scenario (AMB). Scenario assumptions were related to six dimensions: ‣carbon value; ‣energy/excise tax; ‣environmental standards; ‣deployment of renewable energy technologies; ‣deployment of conventional generation technologies; and ‣electricity demand (integrating assumptions on end-use energy efficiency improvement). Main tools: Electricity Market Model and Network model 3

4. 1. European Electricity Market Model and EKC network model 4

5. Introduction Market impacts in the three analysed scenarios of SLED (REF, CPP, AMB) are modelled with REKK European Electricity Market Model (EEMM) Network impacts with EKC network model Highlights: ‣Electricity trade is modelled within the whole EU ‣Hydro generation is modelled under average rainfall conditions, but in the sensitivity assessment the impacts of dry years are also simulated ‣Benchmark costs on investment, RES supports are calculated 5

6. 66 ► The map shows the main results of the model: ► Competitive market equilibrium prices by countries ► Electricity flows and congestions on cross- border capacities ► 36 countries are handled in the model. ► Morocco, Tunisia, Turkey, Moldova, Russia and Belarus are considered as exogenous markets ► In these markets the net export position are equal with the fact in 2013 (assumed a baseload flow) ► The model is calculating the marginal cost of around 5000 power plant blocks and sets up the merit order country by country. ► Taking into consideration the merit order and exports/import, the model calculates equilibrium prices. ► Power flow is ensured by 85 interconnectors between countries. Comments: Model functionality

7. 7 Basic economics in the model Competitive behavior by power generators ‣„if someone is willing to pay more for my energy than what it costs me to produce it, then I will produce” Prices equalize supply and demand Efficient cross-border capacity auctions ‣„we export electricity to wherever it is more expensive and import from wherever it is cheaper” Capacity limits ‣in production and cross-border trade Large country prices around the region are exogenous to the model, the rest are determined by the model

8. 8 Economic description and main assumptions ► The applied model is a partial equilibrium microeconomic model in which a homogeneous product is traded in several neighboring markets. ► Production and trade are perfectly competitive, there is no capacity withholding by market players. ► Production takes place in capacity-constrained plants with marginal costs and no fixed cost. ► Electricity flows are modeled as bilateral commercial arrangements between markets with a special spatial structure. ► Power flows on an interconnector are limited by NTC values in each direction. ► Fuel prices reflect power plant gate prices, transportation/ transmission costs are taken into consideration. ► Only ETS countries buy CO 2 allowances Main model assumptions Main inputs and outputs of the model ► The model calculates regional power supply – demand balance at certain capacity and import/export constraints ► Demand evolution, power plant capacities, availability and cross border power flow defines market price ► Fuel prices are estimated based on available information

9. 9 Model characteristics In a year 90 reference hours are modelled, representing well the daily, weekly and seasonal variations Power plant data comes from international database (PLATTS), but modelled country capacity data are coming from national sources of information Future capacity expansion are from national strategic documents Fossil fuel prices are based on international forecasts of EIA and IEA. ‣ Natural gas price projections depend on the country: TTF Spot price (Western Europe) OIL index price Mix of oil index and spot price

10. 10 Components of marginal cost

11. 11 Efficiency parameters, utilization rates Taken from literature; dependent on the commission year and the type of the PP Availability/utilization rates: ‣Hydro availabilities: dependent on country and season (based on historical utilization rates) ‣Wind an PV: taken from JRC

12. 12 Determining short-term marginal cost Short term marginal cost = Fuel cost + CO 2 cost + Variable part of the OPEX + Energy tax

13. Merit order curves - examples 13

14. 14 Modelled baseload prices in 2015 (€/MWh), and the yearly trade flows

15. 15 Modelled baseload prices in 2025 (€/MWh), and the yearly trade flows

16. 16 Model output Equilibrium price in a demand period Baseload and peakload prices Electricity trade between countries ‣Price of cross border capacities Production by plants ‣Gas consumption ‣CO 2 emission

17. Network modelling EKC network model was used for the assessment Representatives hours of years 2020 and 2025 were modelled, to assess the network impacts on the whole region The following assessments were carried out: ‣Steady-state and contingency analyses ‣Evaluation of net transfer capacity ‣Transmission grid losses 17

18. 18 2. Scenario definitions

19. Outline Main information sources The consultation process Scenario definitions Main input data to the models 19

20. Main information sources Energy Efficiency Action Plan of Montenegro for 2013- 2015, Ministry of Economy, November 2013 NREAP of Montenegro. Ministry of Economy, 2014 Montenegrin Energy Strategy up to 2030 (Startegija Razvoja Energetike Crne Gore do 2030. GODINE (Bijela Knjiga)) 2014 Update / Upgrade Of the “Energy Development Strategy of Montenegro By 2030” (Green Book and draft White Book). Ministry of Economy, 2012 Most important information source were the two stakeholder consultation with Ministry representatives held in November 2014 and July 2015 20

21. The consultation process A two-phase feedback-loop was built in the SLED project: 2014: consultation on main scenario assumptions 2015: preliminary results were delivered - further alignment of assumptions and data – to reflect the INDC process of the country Timeline: 21

22. SLED Scenario definition- Reference 22 Scenario assumptionsReference GHG scenario (REF) Taxation Introduction of EU ETS ETS to be introduced in 2025 Introduction year of minimum excise duty Year of introduction: 2020 Electricity supply Enforcement of environmental standards (LCP Directive) Due to requirement of LCPD directive Pljevlja I closes in 2023. RES-E deployment NREAPs : 826 MW Hydro, 151 MW wind, 10 MW PV and 29 MW Biomass by 2020. By 2030: 826MW Hydro, 190 MW wind, 32 PV and 39 MW Biomass Conventional capacity developments Pljevlja II comes online in 2023 (254MW) Pljevlja I closes in 2023. Maoce TPP will not be built. FOR LCPD: Pljeva I will operate till 2023 (20000 hours between 2018 and 2023) Electricity demand Electricity demand KAP According to 2014 May Strategy (KAP operates with two lines at 100% capacity from 2019) Means 100% total presently installed capacity (A and B line).

23. SLED Scenario definition- CPP, AMB 23 Scenario assumptions Currently planned policies GHG scenario (CPP) Ambitious GHG policy scenario (AMB) Taxation Introduction of EU ETS CO 2 cost in 2020 is 40 % of the ETS price, from 2025 ETS is introduced ETS to be introduced in 2020 Introduction year of minimum excise duty Year of introduction: 2020Year of introduction: 2018 Electricity supply Enforcement of environmental standards (LCP Directive) Due to requirement of LCPD directive Pljevlja I closes in 2023. RES-E deployment NREAPs : 826 MW Hydro, 151 MW wind, 10 MW PV and 29 MW Biomass by 2020. By 2030: 826MW Hydro, 190 MW wind, 32 PV and 39 MW Biomass NREAPs : 826 MW Hydro, 151 MW wind, 19 MW PV and 29 MW Biomass by 2020. By 2030: 1267 MW Hydro, 229 MW wind, 32 PV and 64 MW Biomass Conventional capacity developments Pljevlja II comes online in 2023 (254MW) Pljevlja I closes in 2023.. For LCPD: Pljeva I will operate till 2023 (20000 hours between 2018 and 2023) Pljevlja II comes online in 2023 (254MW) Pljevlja I closes in 2023. For LCPD: Pljeva I will operate till 2023. 10 % biomass utilisation rate is assumed for Plejva II. Electricity demand KAP: 50% of the total installed capacity, according to the agreement on July 2015 stakeholder meeting. Only one line operating at 100%.

24. Electricity consumption Reference: consumption forecast of Energy Strategy (2014) is used In CPP and AMB scenarios KAP operates only at 50% of its total capacity (one production line) from 2018, which drives down electricity demand (asssumption agreed on the July 2015 meeting) 24

25. Renewable electricity assumptions Till 2020 we stick to the draft NREAP (2014) values in the various RES-E technologies Between 2020-2030: REF scenario: Hydro kept constant, rest of the technologies according to the Green book on Energy Strategy CPP scenario: Hydro kept constant, rest of the technologies according to the Green book on Energy Strategy AMB scenario: Hydro is allowed to further grow (Green Book assumptions), together with biomass In this way capacity development is determined, while production is forecasted by the model up till 2030 assuming country specific utilisation hour (solar and wind) and average rainfall for hydro 25

26. RES-E capacities 1 26 REFERENCE and Currently Planned Policy scenario (CPP) capacity values (MW) REF Scenario 20152016201720182019202020252030 Hydro* 661744753821826 Pumped storage 00000000 Geotherma l 00000000 Solar 36789102232 Wind 0118126 151 172190 Biomass 79141819293339

27. RES-E capacities 2 AMBITIOUS scenario capacity values (MW) 27 AMB Scenario 20152016201720182019202020252030 Hydro* 661744753821826 1 0471 267 Pumped storage 00000000 Geotherm al 00000000 Solar 36789102232 Wind 0118126 151 172190 Biomass 79141819295764

28. Present cross-border capacity 28 HU RS BA MKME AL BG GR HR IT 429 507 758 689 250162 250 488403 440 483 400 0 0 253 491 215 96 329151 223 583 540

29. Planned cross-border capacities 29 HU RS BA MKME AL BG GR HR IT 600 1000 400 RO 800 1000 500 600 500 Under construction and approved categories are used in the model runs till 2030. IT-AL is not realised in the modelling period.

30. Assumed capacities I. 30 Present installed capacity New, planned non RES-E capacities In Montenegro the Plevlja II plant is built in all scenarios with 254 MW capacities. In the AMB scenario 10 % biomass co-firing is assumed. Start year of operation: 2023.

31. 3. Scenario Assessment Results 31

32. Outline Wholesale price impacts Generation mix, CO 2 impacts Impacts on system costs: ‣Investment costs, ‣RES support costs Sensitivity assessment: Impacts of reduced rainfall Network impacts ‣Contingencies ‣NTC valuations ‣Network loss impacts 32

33. Modelling result – baseload electricity price, €/MWh in real term 33

34. Modelling result – peakload price, €/MWh in real term 34

35. Wholesale price evolution Both baseload and peakload electricity wholesale prices have a significant drop between 2015-2020, followed by a slight increase in the later period. The main factors influencing the wholesale price developments in Montenegro are the followings: ‣Generation expansion in the fossil based generation in the region is high. Over 7000 MW capacity (mainly lignite and coal) is built in the countries: AL; BA; BG; GR; HR; HU; ME; MK; RS; RO according to the national plans ‣New RES capacities above 12000 MW are also contributing to the price drop till 2020. ‣Higher interconnectedness in the region also allows trade of electricity (higher NTC) These new capacity expansion is illustrated in the following slide for the region 35

36. New PPs in the wider region* 36 Region includes the following countries: AL; BA; BG; GR; HR;HU; ME; MK; RS; RO ; New coal-based power generation, MW New RES-E generation capacity, MW

37. Electricity mix 37

38. Generation mix and CO 2 emissions Montenegro is characterised by expanding hydro capacities and significant net import share till 2030 in REF and CPP scenarios to satisfy increasing demand for electricity Other than hydro RES-E capacities appear in the all scenario from 2020, however biomass makes their contribution significant in the AMB scenario. Changes in the AMB scenario makes Montenegro net exporter. This is mainly due to demand reduction and higher hydro contribution to the electricity mix. Significant drop in CO 2 emissions is observable only in the AMB scenario, when 10% biomass co-firing is assumed at the Pljevlja II plant. Still, Montenegro is characterised by higher carbon intensity than the ENTSO-E average in all years. 38

39. CO 2 emissions 39

40. Total investment cost of new PPs, m€, 2015-2030 40 Source of investment cost: Serbian Energy Strategy and Fraunhofer (2013) There is a significant investment cost need in the various scenarios: The Reference scenario has a 1.2 Billion € investment need over the following 15 years period, increasing over 2.3 Billion in the AMB scenario due to the higher RES expansion and to the Pljevlja II plant. If this latter one is avoided. In this case investment cost would be below 2 Billion €. The main contributing part is still hydro investments, but these are still the most economical RES options in the country.

41. Calculation of the RES-E support budget Support budget = (LCOE t -P)*Generated electricity ‣LCOE t : Levelized cost of electricity generation of technology t ~ average cost of electricity production ‣P: Modelled baseload electricity price (except PV, where peak load electricity prices are taken into account) LCOE figures are based on literature data (Ecofys, 2014) ‣55 €/MWh for hydro ‣90 €/MWh for wind ‣110 €/MWh for biomass ‣105 €/MWh for PV ‣80 €/MWh for geothermal Baseload and peakload prices are the results of the modelling RES fee = RES support budget/ electricity consumption 41

42. Yearly RES-E support need, m€/year 42

43. Unit RES-E support, €/MWh 43

44. RES-E support For comparison: Germany has a support level of over 60 €/MWh, Czech Republic, Portugal: over 12 €/MWh in 2012. LCOE values show that this level of support will be sufficient to cover Hydro based generation, but other types of RES-E would require higher rates. The higher rates for the AMB scenarios shown in previous figure is due to the new RES capacities in biomass, PV and wind, so careful timing of these capacities should be planned. In PV and Wind high cost saving could still appear due to the technology learning effect. 44

45. RES-E support vs CO 2 revenues 45

46. Sensitivity runs: dry years In order to check the impacts of a dry year sensitivity runs were carried out on all scenarios: A severe drought is modelled (lowest precipitation of last 8 years) Droughts assumed to take place in the whole region of South-East Europe Capacity values are the same as in the original scenarios, but hydro availability reduced according to the reduced rainfall 46

47. Impacts of reducing rainfall 1 47

48. Impacts of reducing rainfall 2 48

49. Impacts of reducing rainfall 3 49

50. Network modelling results - contingencies 50 The increasing consumption level and new generation pattern does not cause problem in the transmission network of Montenegro.

51. NTC change with neighbours 51 Higher RES penetration has varying impact on NTC The highest positive difference could be observed with RS, so trade opportunities increase in this directions, while in the other directions the impacts are mixed 2025 winter 2025 summer

52. Transmission losses An increase in capacities and consumption levels generally increases losses over the modelled period, although the results also show that the AMB scenario, with an increased level of distributed generation, will reduce the overall loss level compared to the CPP scenario. 52 201520202025 WinterSummerWinterSummerWinterSummer Equivalent duration time of maximum losses [h] 313321713133217131332171 Transmission losses [MW] REF19.316.538.438.739.134.4 CPP -- 35.432.634.930.8 AMB -- 35.738.534.129 Yearly transmission losses [GWh] REF 96.3 204.3197.2 CPP -- 181.7176.2 AMB -- 195.4169.8

53. Conclusions 1 The stringency of climate policy commitments has limited impact on wholesale price development. The wholesale price is dependent on regional generational capacity expansion rather than on the ambition level of climate policy. Montenegro is currently a net importer of electricity and continues to be an importing country in the REF and CPP scenarios. However, in the AMB scenario the country could become an exporter of electricity by 2030. This is due to the lower demand (KAP) and the significant increase in hydro-based generation CO 2 emission intensity remains higher than the electricity system in Europe, in spite the high share of hydro-based generation. The modelling results also show that if the country introduces a price tag for carbon emission, the government revenues could roughly finance the required RES-E support budget after 2020. Montenegro could still develop significant capacities in hydro generation, which could be a very valuable asset for the future operation of the electricity system. At present, further deployment is constrained by security of supply considerations: the country would like to reduce its dependence on hydro, which is very sensitive to meteorological conditions (precipitation levels and patterns). 53

54. Conclusions 2 Our sensitivity assessment confirms that Montenegro is sensitive to meteorological conditions: in the short term, severe droughts could drive up prices by EUR 8/MWh, and in the long term by EUR 3-4/MWh. In such a year the country would still rely heavily on imports, but in the AMB scenario imports could be significantly reduced. If Pljevlja II is built, the main options to reduce carbon emission are the expansion of RES based generation, and use biomass co-firing at Pljevla II. The hydro sensitivity assessment also points to an important future policy direction for the country. If further cooperation is enhanced within the region and with EU member states, the country could further utilise its hydro potential. In this case Montenegro could be very supportive towards a stricter EU renewable policy, as it would create more demand for its hydro-based generation. The assessment of network impacts shows that the Montenegrin electricity transmission system would not require further network reinforcements in the future to cope with the planned RES capacity increase in the scenarios. If the planned network additions are not built, contingencies would not appear in the system. 54