Private Investment in Wind Power in Colombia*

David Robinson, Oxford Institute for Energy Studies

Alvaro Riascos, Quantil SAS David Harbord, Market Analysis Ltd

SP 27

July 2012 * A report commissioned by the UK Foreign and Commonwealth Office's Latin America Prosperity Fund

     

ii    

The contents of this paper are the authors’ sole responsibility. They do not necessarily represent the views of the Oxford Institute for Energy Studies or any of its members.

Copyright © 2012 Oxford Institute for Energy Studies

(Registered Charity, No. 286084)

This publication may be reproduced in part for educational or non-profit purposes without special permission from the copyright holder, provided acknowledgment of the source is

made. No use of this publication may be made for resale or for any other commercial purpose whatsoever without prior permission in writing from the Oxford Institute for Energy Studies.

ISBN

978-1-907555-56-5

     

iii    

Preface This study was carried out under a grant from the UK Foreign and Commonwealth Office Prosperity Fund for Latin America to the Oxford Institute for Energy Studies (OIES) for the implementation of the project entitled: Colombia: Developing a Framework to Promote Renewable Power. David Robinson (OIES) directed the project team. Alvaro Riascos (Quantil SAS) led the financial modelling work and David Harbord (Market Analysis Ltd.) advised on the Colombian firm energy market. We thank Lucía Martínez for her research on Colombian land and indigenous peoples issues, and Ivan Cadena and Mauricio Romero of Quantil for their research assistance with the financial model. We also thank the following people for reading and commenting on parts of the report: Malcolm Keay, Adam Mantzos, Michael Tennican and Charles Donovan. Finally, we thank staff from the Colombian Regulatory Commission for Electricity and Gas (CREG), the Colombian Ministry of Environment, Housing and Territorial Development, the Colombian Ministry of Mines and Industry, Empresas Públicas de Medellín, Isagen, EMGESA, the Global Wind Energy Council and other anonymous reviewers for meeting with us and/or for helpful discussions. The contents of this paper are the authors’ sole responsibility. They do not necessarily represent the views of the Oxford Institute for Energy Studies or any of its Members.

     

iv    

Contents 1. Introduction and executive summary ............................................................................. 1

2. The Colombian power sector and the case for wind power .......................................... 6

2.1 Supply side of the Colombian power sector .............................................................. 6

2.2 Why consider additional non-conventional renewable power? ................................. 7

3. Two policy instruments: ENFICC and CER ................................................................ 12

3.1 Colombian firm energy market ...................................................................................... 12

3.2 CER payments ............................................................................................................... 23

4. Financial analysis of a wind project under the current regulatory regime ............... 25

4.1 What the model calculates ............................................................................................. 25

4.2 The logic of the model ................................................................................................... 26

4.3 Main elements of the financial modelling ..................................................................... 27

4.4 Benchmarking the model ............................................................................................... 28

4.5 Results of the main policy scenarios .............................................................................. 29

4.6 Sensitivities .................................................................................................................... 31

4.7 Risk Analysis ................................................................................................................. 35

4.8 Other financial considerations ....................................................................................... 36

4.9 Conclusions from modelling .......................................................................................... 38

5. Other risks and opportunities facing investors ............................................................ 40

5.1 Commercial opportunity and risk .................................................................................. 40

5.2 Political considerations .................................................................................................. 42

5.3 Regulatory risk ............................................................................................................... 43

5.4 What most investors look for from policy and regulation ............................................. 43

6. Conclusions ...................................................................................................................... 47

Annex 1: Information on Colombian Electricity System 2010-2011 .................................. 49

Annex 2: Colombian commitments on climate change mitigation ..................................... 51

Annex 3: Requirements to execute a wind project in Colombia, in Indigenous Territory53

Annex 4: Quantil Model ......................................................................................................... 58

     

v      

Tables Table 1: Installed capacity Colombian integrated electricity system, 31/12 2011. .................... 6 Table 2: Generation output 2010-2011 for the Colombian electricity system ............................ 7 Table 3: Levelized costs of energy from different technologies ................................................. 9 Table 4: ENFICC % for different technologies ........................................................................ 13 Table 5: Outcome of May and June 2008 auctions for firm energy in Colombia .................... 14 Table 6: Outcome of December 2011 auctions for firm energy in Colombia .......................... 15 Table 7: Calculation of Wind Capacity Credit Factors in USA ................................................ 21 Table 8: CREG ENFICC methodology for wind applied to El Niño periods .......................... 22 Table 9: PJM methodology to determine ENFICC for wind in Colombia ............................... 23 Table 10: World Bank parameterization with Quantil model ................................................... 28 Table 11: Comparing results of WB and Quantil model – Project IRR ................................... 29 Table 12: Results of reference scenario and of changes in ENFICC ........................................ 30 Table 13: Results of reference scenario and of changes in ENFICC and CER ........................ 30 Table 14: Results of sensitivity case for changes in Benchmark Investment cost (BIC) ......... 32 Table 15: Results of sensitivity case for changes in wind speed .............................................. 32 Table 16: Results of sensitivity case for changes in energy price ............................................ 33 Table 17: Results of sensitivity case for changes in interest rate ............................................. 34 Table 18: Results of sensitivity case for changes in exchange rate .......................................... 35 Table 19: Results of Monte Carlo Analysis – Mean Reversion ................................................ 35 Table 20: Results of Monte Carlo Analysis – Model 2 ............................................................ 36 Figures Figure 1: LCOE for different generation technologies: Q2-2009 to Q4-2011 .......................... 9 Figure 2: Estimating the Effective Load Carrying Capability (ELCC) .................................... 19 Figure 3: Average surface wind speed en January in Colombia (m/s) ..................................... 41

   

1. Introduction and executive summary Colombia currently has a very low penetration of non-conventional (non large-scale hydro) renewable energy sources. This reflects the dominance of large-scale hydro projects within the generation sector. However, increasing concerns over the impact of periods of serious drought in El Niño periods, most recently in 2009/10, have meant that Colombia is investing heavily in new thermal power capacity, increasing the country’s carbon footprint. A study by the World Bank suggests that onshore wind power may be a cleaner and economically viable alternative, as a compliment to the substantial hydro resources in Colombia1. We understand that the Colombian Government is also considering whether and how to develop a range of alternative energy sources, including solar, biomass and geothermal power. Many other large developing countries, including Brazil, Peru and Mexico, are already encouraging private investment in a range of renewable energy sources, especially wind power. However, under current regulatory arrangements in the Colombian power sector, wind power appears not to be financially viable. Colombia is currently considering changes in the regulatory framework that would extend the payment for ‘firm’ energy to wind power and other non-conventional renewable sources of generation. The question addressed in this paper is whether the change in regulation will make private investment in wind power attractive in Colombia. This report is the main output of a project supported by the UK Foreign and Commonwealth Office (FCO) Prosperity Fund for Latin America. It examines the feasibility of private investments in wind power in Colombia within the current regulatory framework. By reference to a specific wind power project, it estimates the financial gap between what investors might require and what they can expect, and considers how this gap could be reduced under the current regulations. Although our terms of reference do not include detailed research on alternative regulations that might be introduced in Colombia, we have drawn on international experience to identify some ideas that deserve further exploration if the government wishes to encourage private investment in non-conventional sources of renewable energy. In their letters related to the Prosperity Fund, officials from the Colombian energy regulator (CREG) and the Ministry of Environment both indicated a particular interest in research on the firm energy payment (cargo de confiabilidad) for non-conventional renewable power generation, including wind.2 Since then, the CREG has made a proposal that estimates the

                                                                                                               1 Vergara, W., Alejandro Deeb, Natsuko Toba, Peter Cramton and Irene Leino Wind Energy in Colombia: A Framework for Market Entry (World Bank, July 2010). In our paper, we refer to this as the “World Bank report”, or simply “WB”. 2 In his letter, Javier Augusto Diaz of the CREG indicated an interest in the “desarrollo de una metodología para medir la energía firme que pueden aportar proyectos de generación eléctrica con base en recursos renovables.” In her letter of 2 June 2011, Andrea García Guerrero of The Ministry of the Environment, Housing and Territorial Development wrote, “Nos gustaría se analizara el mecanismo para que proyectos de las energías

     

2      

firm energy factor (ENFICC) from wind power at between 6% and 7.3% of the wind plant’s capacity. This compares to an ENFICC of over 90% for coal and gas-fired plants, and between 30% and 50% for hydro plants. The central policy issues analysed in the report are: (a) the methodology for determining the ENFICC for wind power; and (b) the financial implications for private investors of the choice of methodology for setting the ENFICC. The analysis is relevant for the regulation and remuneration of wind power and of other non-conventional sources of renewable power, for instance including geothermal, biomass and solar. The report has five sections, in addition to the introduction and executive summary. Section 2 provides basic background to the Colombian electricity sector and the reasons to consider the development of non-conventional renewable power, such as wind. The heavy reliance on hydroelectricity explains Colombia’s relatively small carbon footprint in the sector by comparison to most countries in Latin America. However, during extended periods of dry weather (specifically those referred to as El Niño periods), hydro generation falls and the system requires alternative sources of energy. In this context, there is potentially a role for non-conventional renewable sources of power, including wind power, provided these technologies are economically viable. Since reliability is a particularly important issue in the Colombian electricity system due to the problems associated with El Niño, a central question is what contribution these alternatives sources of power would make to the system’s reliability, and what the payment for that reliability should be. Section 3 considers two policy variables that are key determinants of profitability. The most important of these is the methodology for determining the contribution of wind plant to system reliability; in other words, for determining the firm energy factor (ENFICC) for a wind power station. We draw on international experience to analyse different ways of estimating ENFICC for wind power. In our view, the CREG’s methodology probably undervalues the contribution of wind to system reliability and possibly by a significant margin. Whereas the CREG estimates of ENFICC for the Jepírachi plant are between 6% and 7.3% of the plant’s capacity, our estimates suggest that the ENFICC is at least 15% and possibly in the range of 27% to 33%. The difference between our conclusions and the CREG’s proposals are firstly related to the fact that the CREG’s methodology does not focus on El Niño periods, when the system is under stress. Furthermore, its use of percentiles (i.e. to estimate the probability that wind power will be available) probably understates the contribution of wind power to system reliability and is conservative by international standards. In addition to the analysis of the ENFICC, Section 3 introduces an international policy variable, the payment for Certified Emission Reductions (CERs) under the UNFCCC Clean Development Mechanism (CDM). This variable influences the profitability of wind projects in Colombia and illustrates the importance of valuing externalities – in this case, through compensation for avoiding the negative global externality related to CO2 emissions.

                                                                                                                                                                                                                                                                                                                                                       anteriormente mencionadas [renovables aparte de hidroelectricidad] pueden ser consideradas dentro del esquema actual de cargo por confiabilidad de la Comisión de Regulación de Energía y Gas (CREG)”.  

     

3      

Although the CER is an international policy mechanism, it has domestic policy implications. This is largely because eligibility under the CDM (and hence CER income) is conditional upon demonstrating that the project’s economic viability requires these additional payments; but economic viability depends in large part on the way that wind power is compensated under Colombian regulations. Section 4 uses a simplified financial model to analyse the viability of private investment in a specific wind power in Colombia under the current regime, as well as the impact of changing the ENFICC and the CER3. For different estimates of ENFICC, we estimate the real internal rate of return (IRR) on equity, the payback period and the debt service cover ratio (DSCR) for a 302 MW wind power project in the Guajíra region of Colombia, financed by a 70/30 ratio of non-recourse debt to equity. We carry out sensitivities on a number of non-policy parameters (e.g. investment cost, interest rates, exchange rates, wind speed, and energy prices) and use Monte Carlo simulations to examine specific risks related to energy prices and wind speed. Under the current CREG proposal for ENFICC, the project fails by a significant margin to meet the main financial tests for viability. The financial gap between the required real return on equity (14%) and the expected return (3.5%) is substantial. When we increase the ENFICC to reflect the higher end of our estimates of firm energy (30%), the equity IRR increases to almost 8%, but the financial gap compared to an IRR of 14% is still large. In Section 4, we also analyse the impact of different CER prices to compensate wind power for the avoidance of CO2 emissions through the displacement of more carbon intensive generation at the margin. Under all reasonable scenarios (of ENFICC and CERs), additional income from CERs is insufficient to reach the 14% IRR threshold for the project. We have not attempted to model the impact of internalizing other environmental and social costs related to the development of large hydro and coal projects, and we are not sure of the degree to which these are already fully taken into account, but these costs could be very important.4 Section 5 identifies some risks and opportunities that investors, especially foreign investors, are likely to take into account when they are considering whether to invest in wind power in Colombia. These include the potential upside from entering the market early, as well as political and regulatory risks and related costs. Some of these are specific to Colombia, for instance the size of the potential wind resource and the costs of dealing with social and environmental concerns in the Guajíra Indigenous Territory, where the project is located. Some risks are common to most countries, in particular the volatility of revenue streams for intermittent renewables like wind power, the availability of information about wind resources

                                                                                                               3 The model is not intended as a basis for making financial decisions for a specific investment. It is rather a simplified model that aims to explore the implications of changing two key policy assumptions as well as key variables, assuming a given wind turbine technology and wind conditions that apply in the Guajíra region of Colombia. It does not attempt to analyze all the relevant policy issues, technologies or locations for wind power in Colombia. It draws on the data and the logic of a more detailed model developed in the World Bank Report to study wind power in Colombia; that report analyzes a wider range of policy issues. 4 The recent controversy over the development of the Quimbo hydroelectric project illustrates the fact that hydroelectricity projects have social costs.  

     

4      

and the cost and speed of transmission connections. As background for possible future research and consideration by the government, we identify some of the policies and regulations that have been adopted in other countries to attract investment in wind power, without advocating any one in particular. Section 6 summarizes our conclusions, which we reproduce here. First, the result of the financial modelling is that wind power does not appear to be an immediately attractive investment in Colombia. Under the reference scenario that sets ENFICC at 6%, the financial gap is substantial (target IRR of 14% v. expected IRR of 3.5%). Increasing the ENFICC to 30% from 6% narrows the financial gap substantially (14% v. 8%), and adding CER income of $10/tCO2 reduces the gap even further (14% v. 9.4%). However, only with optimistic assumptions, or subsidised finance, do we see a prospect of eliminating the gap entirely within the current regulatory system and with our estimated electricity prices. It is important to stress that our conclusion is based on the conditions that apply to a particular power station, similar in turbine technology to the Jepírachi station, although much larger (302 MW v. 19.9 MW). To the extent that new turbine technologies are more efficient, for instance at lower wind speeds, the financial appraisal will be more positive, provided the improved efficiency more than compensates for any additional costs. Furthermore, wind conditions may be better at other sites. We have carried out a number of sensitivities, including ones that are optimistic about the wind speed, investment costs, CER payments, interest rates and energy prices. Even then, the financial gap is large when we use the CREG’s proposed ENFICC. Our conclusion is that the regulatory treatment of the ENFICC is a key barrier in Colombia, even for wind power projects that might face better conditions than the project we have analysed. Second, as the fixed costs of wind power decline and the need for additional capacity grows, it becomes increasingly important to calculate the ENFICC properly for wind and other non-conventional renewables sources of power. This is important to ensure that the system is building the optimal mix of plant, which won’t happen if the ENFICC is measured inaccurately. For investors, an increase in the ENFICC for alternative technologies such as wind power increases the mean return on investment in wind power. In particular, it reduces the uncertainty of cash flows because the firm energy payment will be guaranteed for a period of 20 years for new plants that are successful in the firm energy auctions. Third, the calculation of capacity credit factors (i.e. like ENFICC) for wind and other non-conventional power is a recent development around the world and there is no universally accepted method. However, our recommendation is that the CREG reconsider the methodology for defining the ENFICC for wind power, and also consider using a similar (revised) methodology for other sources of non-conventional power that have potential in Colombia, including geothermal, biomass and solar energy. One well-recognized methodology for measuring a plant’s contribution to system reliability is known as the effective load carrying capability (ELCC). Alternatively, we would encourage the CREG to consider the approach used by PJM, adapted to reflect relevant peak hours in El Niño periods (as summarized in our Table 9). According to our provisional calculations, using this

     

5      

adapted PJM methodology would increase the ENFICC (for conditions at Jepírachi) from the CREG’s value of 6% to our estimate of about 30%. This would make an important contribution to project revenues and help to finance the projects at lower cost. Fourth, the CERs are important for a number of reasons. One is that CER income contributes to closing the financial gap faced by investors. Although the contribution is relatively small at today’s CER prices, there is a reasonably good prospect that the CER price will rise, mainly because the EU is determined to raise the price of CO2 emission allowances in the EU Emission Trading Scheme (ETS), which effectively drives the price of CERs. Second, project developers typically reach an agreement to sell CERs for a number of years; this makes CER income predictable and therefore helps investors to raise finance for these projects. Third, CERs are a way of recognizing the economic cost of externalities, in this case by compensating the avoidance of CO2 emissions. Although the CER is an international policy mechanism, it has domestic policy implications since eligibility under the CDM is conditional upon demonstrating that the project’s economic viability requires these additional payments. When projects are economically viable without CER payments, the Government will need to decide whether to compensate new projects (i.e. those which are not eligible for CER payments) for the avoidance of CO2 emissions. We would encourage the government to support domestic and international mechanisms that penalize the negative externalities associated with some power stations (e.g. coal power), and/or compensate power stations (e.g. wind power) that are able to avoid these externalities. Fifth, the risk assessment suggests that investors could decide that uncertainty about revenues is so great that they require a higher expected return to compensate for that risk. The risk and hence the required return will be lower if a larger share of the revenue stream can be guaranteed. This conclusion underlines the importance of both policy variables (ENFICC and CER) since they both offer revenue guarantees. Finally, we have drawn on international experience to identify some of the issues that investors will be considering, in addition to those studied in our financial analysis. Investors stress the importance of having reliable sources of information about wind speeds at 60 meters and about the potential and geographical distribution of other renewable resources in Colombia, in particular geothermal, biomass and solar. They also stress the importance of a clear policy with respect to the development of renewable energy sources and of legislation to carry it out. In addition, they emphasize the timing and cost of gaining access to transmission networks. This international experience provides background for possible future research and consideration by the Colombian Government should it decide that it wants more actively to promote investment in these new technologies. If so, auctions, accompanied with long-term contracts, are particularly useful mechanisms for promoting competition among investors.

     

6      

2. The Colombian power sector and the case for wind power When one speaks about the prospects for wind power in Colombia, the first question is: why bother with wind power and other non-conventional renewables, when the country has so much hydroelectric capacity? This is a very good question. Hydro normally accounts for about 80% of Colombia’s electricity; this explains why Colombia is in the bottom four Latin American countries in terms of carbon intensity in electricity5. In order to reply to the question, we first introduce some essential features on the supply side of the Colombia power sector.

2.1 Supply side of the Colombian power sector The Colombian system relies very heavily on hydroelectric power, but thermal generation also plays an important role6. Table 1 indicates that about 64% of generation capacity is large hydro, with 32% thermal and especially gas-fired. Today, wind power, all from the Jepírachi plant, accounts for only 0.1% of total generation capacity. Table 1: Installed capacity Colombian integrated electricity system, 31/12 2011.7

In a normal year, the reliance on hydro generation is especially evident: it accounts for about 80% of electricity output. This poses problems for reliability during periods of el Niño. Drought substantially reduces hydroelectric generation, with potentially serious economic and political consequences. More generally, demand for water for all uses is growing and this raises the value of water and reduces its availability for hydroelectric generation. This underlines the importance of having backup generation to replace hydro during periods of el Niño.                                                                                                                5 Colombia has a carbon intensity of 136 grams CO2/kWh from electricity and heat; only Paraguay, Brazil and Bolivia have lower carbon intensity. See IEA CO2 Emissions from Fuel Combustion 2011 edition, page II-69. 6 Annex 1 summarizes the regulatory framework for renewable power and reproduces tables that provide more detail on the Colombia electricity system, as reported by XM (the system operator) on its website. 7 XM (System Operator) website: http://www.xm.com.co/Pages/DescripciondelSistemaElectricoColombiano.aspx  

     

7      

Table 2 illustrates both how el Niño affects the supply of hydroelectricity and the important role of thermal power. In 2010, which included the effects of El Niño at the beginning of the year, hydroelectric power accounted for 67% of output, and thermal plants for 27%. In 2011, not an El Niño year, hydro generation increased by 20%, thermal generation fell by 40%, and their respective shares of output were 78% and 16%.

Table 2: Generation output 2010-2011 for the Colombian electricity system8

2010 GWH

2010 2011 2011 CHANGE GROWTH % GWH % GWH %

Hydro 38,088.60 67 45,583.10 78 7,494.50 19.7 Thermal 15,590.70 27 9,383.70 16 -6,207.00 -39.8 Minor plants 2985.6 5 3,336.70 6 351 11.8 Cogeneration 222.7 1 316.9 1 94.1 42.3 TOTAL 56,887.60 100 58,620.40 100 1,732.80 3

The question now is what additional generation capacity should be built in order to supplement hydro in normal years and to provide backup in very dry El Niño years. Should the system rely more on thermal power or diversify to include non-conventional renewables, or other sources of power and demand response? As explained more fully later in Section 3, Colombia holds auctions to select the new plants that will be built to provide system reliability, and to determine the price for firm energy. In the last auction (December, 2011), thermal plants accounted for over 70% of the additional new capacity and 89% of firm energy. We understand that most thermal plants will provide backup to hydro. However, some new thermal plants could also operate base-load or mid merit, thereby increasing the carbon footprint (i.e. absolute level of emissions) of Colombia’s electricity sector. Wind power would partly displace thermal plants at the margin, as confirmed by an analysis of the potential for Jepírachi plant to avoid CO2 emissions, and thereby limit growth of total emissions9.

2.2 Why consider additional non-conventional renewable power?

There are a number of reasons why Colombia may wish to consider non-conventional renewable sources of power, like wind and solar, as alternatives to coal and gas-fired plants. Here are some of the arguments frequently offered in support of this view.                                                                                                                8 XM (System Operator) website: http://www.xm.com.co/Pages/DescripciondelSistemaElectricoColombiano.aspx 9 AENOR, “CDM Validation Report Renewal Of Crediting Period, International Bank For Reconstruction And Development As The Trustee Of The Prototype Carbon Fund, Validation Of The Project Activity: Jepírachi Wind Power Project”, 3 March, 2011, page 7.

     

8      

First, as we have just illustrated, Colombia risks an increasing carbon footprint, which is difficult to limit in the sectors where it is most important, especially in transport10. Colombia has a relatively low carbon footprint compared to other countries in Latin America11. However, the absolute level of emissions has been growing and is expected to grow further, along with economic growth.12 Colombia has made a unilateral commitment to the UNFCCC to address the challenges of climate change, including that “at least 77 per cent of the total energy capacity installed by 2020 will be generated from renewable sources”.13 International experience shows that it is easier to reduce or at least manage growth in CO2 emissions through the electricity sector than in other carbon-emitting sectors, such as transport. So, even if the Colombian electricity sector has a relatively small carbon footprint today, the government may want to reduce it further or at least ensure that it does not grow too much.

Second, there appears to be a natural complementarity (or hedge) between hydro on the one hand, and wind generation on the other: during periods of el Niño, less rain appears to coincide with stronger wind. Any assessment of the backup sources to hydro generation should reflect this complementarity. In particular, the payment for firm energy from wind and solar should reflect the contribution of these technologies to system reliability at times of shortage, especially during el Niño periods. Of course, the complementarity has to be demonstrated for the specific plant in question, and the ENFICC should be plant specific. Third, including non-traditional renewable energies in its energy portfolio may make the Colombian electricity sector and the economy less exposed to volatile hydrocarbon prices. While this portfolio argument has stronger appeal in a country without its own hydrocarbon resources, it may also apply in Colombia since domestic market prices for natural gas and international prices for coal are volatile. Fourth, the costs of renewable power are declining and the cost of fossil-based generation is likely to rise. Bloomberg New Energy Finance (BNEF) has tracked the changes in the levelized cost of energy (LCOE) for wind power, other renewables and fossil fuel based plant over the past three years.14 The costs of onshore wind power and solar PV have been on a

                                                                                                               10 Transport accounted for a third of Colombian CO2 emissions in 2009 and electricity for about one sixth. See IEA CO2 Emissions from Fuel Combustion 2011 edition, page II-173.  11 1.33 tonnes CO2/capita in 2009, compared to the Latin American average of 2.16. See IEA CO2 Emissions from Fuel Combustion 2011 edition, page II-57. 12 CO2 emissions grew about 35% between 1990 and 2009. See IEA CO2 Emissions from Fuel Combustion 2011 edition, page II-173. 13 Annex 2 reproduces the relevant pages from the “Compilation of information on nationally appropriate mitigation actions to be implemented by Parties not included in Annex I to the Convention, UNFCCC, Note from the Secretariat, Ad Hoc Working Group on Long-term Cooperative Action under the Convention”, FCCC/AWGLCA/2011/INF.1, 18 March 2011. 14 LCOE is a way of comparing the price of electricity that would enable an investor to achieve a required rate of return to justify investment. It is helpful as a way of measuring changes in the relative cost of certain technologies. On the other hand, it would be misleading to suggest that wind power and thermal power have the same value when they have the same LCOE. This is because, in any given power system, the value will depend on when the generation occurs. Wind power may have the same LCOE as gas-fired generation, but could be worth more or less than the CCGT depending on when the plant was actually generating. If the average price is $60/MWh, and wind generation runs when prices are less than that, then wind is worth less than the average. If wind runs mainly at times of system peak, it would earn more than the average system price and could be more valuable than the CCGT. For a more detailed review of the difficulty of comparing the costs of

     

9      

downward trend that has accelerated recently. For wind power, the decline in costs reflects increasing scale of wind turbines, more efficient turbines, less expensive manufacturing methods and intense competition. Meanwhile, according to BNEF, the cost of generation from coal and gas-fired plants has been rising. Figure 1 illustrates the trend (assuming a target of 10% IRR on equity), with the dashed-yellow line representing wind power.

Figure 1: LCOE for different generation technologies: Q2-2009 to Q4-2011 15

Table 3 summarizes the LCOE from different technologies in Q4-2011, using BNEF figures and assuming a 10% IRR on equity. It illustrates the wide range from high to low estimates of LCOE for onshore and offshore wind, whereas the LCOE for coal and gas-fired plants do not vary much between high and low estimates. It also suggests that onshore wind can sometimes be competitive with gas and coal-fired plants.

Table 3: Levelized costs of energy from different technologies16

Technology Low ($/MWh) Base ($/MWh) High (S/MWh) Coal 70.68 76.51 81.3 Gas CCGT 57.34 62.24 66.81 Wind onshore 59.02 80.04 117.22 Wind offshore 143.14 232.20 331

                                                                                                                                                                                                                                                                                                                                                       conventional plant and intermittent plants, see Paul Joskow, “Comparing The Costs Of Intermittent and Dispatchable Electricity Generating Technologies”, September 27, 2010 (Revised February 9, 2011) DISCUSSION DRAFT. 15 Bloomberg, New Energy Finance model, assuming a 10% IRR on equity. 16 Ibid, assuming 10% IRR on equity.

     

10    

In Colombia, the comparison between wind power and other plants will reflect different technical and cost conditions, but the relative decline in the costs of wind technologies will nevertheless improve the competitive position of wind power17. So, even if these non-conventional technologies do not seem competitive with coal and gas under today’s market conditions, they are likely to become increasingly competitive. Fifth, the negative externalities and the long lead times required for large hydro and coal plants contrast with relatively limited externalities and flexibility offered by non conventional renewable sources of power. Whenever a new hydroelectric project is under consideration, it is likely to involve complicated and controversial impact assessments, especially when the project involves deviating rivers and flooding valleys where people live. This can delay the building of these large projects, which in any case can take many years. When new coal based generation is built, the impact on the environment and on the people who live in the area is also potentially significant, and the large scale of these plants means that they also require a relatively long time to build. By comparison, wind power and other non-conventional renewables have a relatively benign impact on local communities and environment. Furthermore, they are quick to build and therefore have the advantage of flexibility 18. Sixth, non-conventional renewable power provides energy in areas that are isolated and badly (if at all) connected with the national grid. Non-conventional renewables (like solar panels) may be the preferred alternative (economically) compared to the use of diesel generators or kerosene, wood or candles. Furthermore, these new technologies can be combined with mobile communications, with the latter offering a means of paying for the energy, while the electricity provides a way to recharge mobile phones. Seventh, the small scale of non-conventional renewable energy facilitates entry and competition. Large-scale hydro and large coal plant require very significant capital and this limits the number of potential competitors. Renewable energy projects like wind and solar energy are, by contrast, relatively small and require limited amounts of capital. The evidence throughout the world, including in Latin America, is that competition among renewable energy suppliers is very effective in driving down prices required by investors. There are other reasons why governments consider non-conventional renewable power, including: employment creation; industrial policy aimed at promoting innovation in technologies with promising international markets; and regional policy to support economic development in rural areas. Faced with these and other considerations, a key question for policy makers is: should the government actively promote non-conventional renewable power, if these technologies

                                                                                                               17 In countries with a high penetration of wind power, the cost of wind energy could rise when the best locations have been exploited.  18 However, wind power projects do have an environmental and social impact. Annex 3 of this report summarizes some of the requirements and costs associated with developing a wind park in the Guajíra region.

     

11    

would not be chosen in a technology-neutral regulatory system? This requires a full cost-benefit analysis that we have not undertaken. However, our study contributes to the decision making process by answering a more limited set of questions related to the current regulatory regime: • Within the current regulatory framework for electricity, are all technologies treated in a

comparable way, especially with respect to the payment for firm energy (and the measurement of ENFICC)? If not, what adjustments are warranted?

• Are the returns sufficient to attract private investors to wind power projects under the current regime? If not, what is the financial gap between the expected return and the required return? Can policy reduce this gap without actively promoting specific technologies?

• What other risks or opportunities in Colombia will influence investment decisions,

especially by foreign investors, and what policies have been adopted in other countries to attract investment in wind power and in other non conventional sources of renewable energy?

The remainder of the report deals with these questions, before drawing conclusions.

     

12    

3. Two policy instruments: ENFICC and CER

This section analyses the two policy instruments that we have modelled in our report. One of these policies is domestic: the payment for firm energy. The other is an international policy instrument: the CER payment for avoided CO2 emissions.

3.1 Colombian firm energy market

In 2006, the CREG introduced a new scheme to ensure the long-term reliability of the electricity supply in Colombia, and in particular to guarantee that there is always sufficient capacity available to meet peak demand during El Niño periods, when hydro resources are significantly reduced.19 The scheme allocates “firm energy obligations” (“OEFs”) to new and existing generation plant at prices determined in competitive auctions. OEFs are “option contracts” that commit generating companies to supply given amounts of energy at a predetermined Scarcity Price whenever the spot price in the electricity market rises above the Scarcity Price.20 They receive the spot price for any additional generation above their firm energy obligation, and pay a penalty if they cannot meet their firm energy obligation, equal to the difference between the spot price and the scarcity price on the OEF quantity not met in any hour. In return for agreeing to supply at the Scarcity Price, generators allocated OEFs in the auctions receive a fixed annual option fee (the firm energy price, or Cargo por Confiabilidad) for each capacity unit contracted. This option fee makes an important contribution to the recovery of fixed costs for generating plants that sell very little in normal times, such as the CCGT plants in central Colombia that generate infrequently outside of El Niño periods. The maximum amount of firm energy that a generator may offer in a firm energy auction is known as its ENFICC (Energía Firme para el Cargo por Confiabilidad). ENFICC refers to the amount of energy a generator of a given type can reliably and continually produce during periods when hydro generating capacity is at a minimum.21 Table 4 shows the typical ENFICCs for different generation technologies in Colombia as a percentage of a plant’s CEN, or effective net capacity.

                                                                                                               19 See Harbord, D. And M. Pagnozzi, “Review of Colombian Auctions for Firm Energy”, 25 November 2008. The paper was commissioned by the CREG. 20 The Scarcity Price is established by the CREG and updated monthly based on the variation of the Fuel Price Index. In March 2012 it was approximately US$248/MWh. The Scarcity Price has a double purpose. On the one hand, it indicates the time when the different generation units or plants will be required to fulfill their firm energy obligations, which happens when the Spot Price exceeds the Scarcity Price; on the other hand, it is the price at which this energy will be paid.

21 See CREG RESOLUCIÓN 071 DE 2006: “Energía Firme para el Cargo por Confiabilidad (ENFICC): Es la máxima energía eléctrica que es capaz de entregar una planta de generación continuamente, en condiciones de baja hidrología, en un período de un año.”  

     

13    

Table 4: ENFICC % for different technologies

Technology Maximum ENFICC (%)

Hydro with storage 55

Hydro without storage 30

Coal 97

Natural Gas 93

Fuel Oil 88

Wind 6

If a coal plant, for example, has an ENFICC of 97%, the maximum annual OEF for a 100 MW plant would be 100 MW*0.97*8760 hours = 849,720 MWh. Hence the maximum firm energy payment the plant could receive in a year is 849,720 MWh multiplied by the auction-determined firm energy price (currently US$13.998/MWh), or $11,894,381 (US). In 2015 this will increase to $13,340,604 (US) when the option fee set by the 2011 firm energy auction (US$15.7/MWh) will be applied. The first firm energy auctions were held in May and June 2008 and allocated OEF’s for up to twenty years beginning in December 2012. About 9,300 GWh per year of OEF’s were allocated to new resources, including 1,117 GWh from new coal plant and 1,678 GWH from new gas-fired generation plant at an auction-determined option fee of $13.998/MWh. Existing plant will receive this option fee until December 2015, and the new plants are guaranteed this fee for up to 20 years. Table 5 below shows the results of these auctions.

     

14    

Table 5: Outcome of May and June 2008 auctions for firm energy in Colombia22

Project Company Market

Entry Date

Generation

Type

Capacity

(MW)

OEF

Assigned

(Gwh/Año)

6 May Auction

Gecelca III Gecelca Dec 2012 Coal 150 1117

Amoyá ISAGEN Dec 2010 Hydro 78 214

Termocol Poliobras Dec 2012 Gas 200 1678

13 June Auction

Pescadero Ituango

Pescadero Ituango

Dec 2018 Hydro 1200 1085

Isagen Sogamoso Dec 2014 Hydro 800 2350

Emgesa Quimbo Dec 2014 Hydro 396 1650

EPM Porce IV Dec 2015 Hydro 400 961

Promotora Miel II Dec 2014 Hydro 135 184

EPSA Cucuana Dec 2014 Hydro 60 50

The second firm energy auctions were held in December 2011. Table 6 summarizes the allocation of 3,700 GWh of OEFs to five new generation projects, with a total capacity of 575 MW. The option fee (firm energy price) resulting from this auction was $US15.7/MWh. As noted earlier, over 70% of the selected capacity was thermal plant, and over 89% of the OEFs assigned. The option fees will be paid to new plant for 20 years, starting on December 1, 2015.23

                                                                                                               22 http://www.acolgen.org.co/jornadas3/expansion2012.pdf 23 http://www.creg.gov.co/html/i_portals/index.php?&p_origin=plugin&p_name=news&p_id=N-341&p_options  

     

15    

Table 6: Outcome of December 2011 auctions for firm energy in Colombia24

Project Company Location Type of Generation

Capacity MW

OEF

Assigned

(Gwh/Año)

Proyecto Hidroeléctrico del Río Ambeima

Energía de los Andes S.A.S. E.S.P.

Tolima Hydro 45 75

Central Hidroeléctrica Carlos Lleras Restrepo

Hidralpor SAS ESP Antioquia Hydro 78 200

San Miguel La Cascada SAS ESP Antioquia Hydro 42 123

Gecelca 32 Generadora y Comercializadora de Energía del Caribe S.A.

Córdoba Thermal 250 1,971

Tasajero II Termotasajero S.A. E.S.P.

Norte de Santander

Thermal 160 1,331

To pay the option fees for the OEFs allocated in the auctions, the regulatory system collects revenue through the CERE. This is a tax per unit of electricity generated. Each generator contributes in proportion to the energy generated. The generators (above 20 MW) treat this as a cost, which raises the price of electricity.

The aim and the effect of the auction and the resulting option fees are to help finance plants, especially those that run very little but are considered important for system reliability. In particular, the ENFICC for CCGT plant corresponds to 93% of its rated capacity. In this way, the system provides a revenue guarantee that justifies building plants that primarily provide backup to hydro generation during El Niño periods. At the new firm energy price of $15.7/MWh, a 100 MW CCGT would be guaranteed an annual revenue stream of $12.8 (US) million per year, even if the plant never generates electricity. In the words of the CREG,

‘Así mismo, aseguró que para los generadores los beneficios se concretan en ingresos fijos asociados a la obligación de energía firme hasta por 20 años, lo cual se traduce en una estabilización de su flujo de caja y reducción de sus riesgos de inversión.’25

                                                                                                               24CREG website: http://www.creg.gov.co/html/i_portals/index.php?&p_origin=plugin&p_name=news&p_id=N-341&p_options. See also, Obligaciones De Energía Firme Asignadas En La Subasta De OEf 2015-2016, XM (htp://www.xm.com.co/Resultados %20Subasta/OEF_Asignada_Subasta.pdf)    

     

16    

3.1.1. Calculating ENFICCs

CREG Resolution 071 of 2006 (Annex 3) describes in detail how the CREG calculates ENFICCs (i.e. the maximum amount of “firm energy” a generator can sell in the auction) for the hydro and thermal plants that receive firm energy payments. For thermal plant this is essentially CEN*(1-IHF) where CEN = “effective net capacity” (the generation capacity of the plant) and IHF = the historical probability of forced (i.e. unplanned) outages.26 The ENFICC of hydro plants is calculated using a computational model that maximizes the minimum energy that a hydro generation plant can produce monthly during dry periods.27 The model incorporates historical data on average monthly water inflows; discharges and restrictions in the water conduction systems; characteristics of the generation plants including the average efficiency and their minimum and maximum generation; water reservoir data and other uses of water like aqueduct or irrigation and environmental restrictions; historical unavailability due to forced outages; and flow constraints. The minimum production numbers are then ordered from least to greatest, and the lowest is defined as the plant´s ENFICC BASE, or the amount of energy the plant can be relied upon to produce with 100% probability. In other words, ENFICC BASE corresponds to the minimum monthly energy supply obtained from the maximization model. The CREG also defines the ENFICC 95% PSS - the amount of energy the plant can be relied upon to produce with 95% probability. 28 As noted in Table 4 above, while thermal plant ENFICCs (expressed as percentages of effective net capacity) tend to exceed 90%, hydro plant ENFICCs typically range between 30% and 55%. Until recently, wind power was not eligible for a firm energy payment in Colombia. In July 2011 however, the CREG released a proposal for measuring ENFICCs for wind plants based upon the historical experience of EPM´s Jepírachi plant.29 Following a broadly similar methodology to that applied to hydro plant, the CREG used historical generation data from 2004 to 2011 to estimate monthly capacity factors for the Jepírachi wind farm, and derived an ENFICC BASE of 6% and an ENFICC 95% PSS of 7.3%.30

                                                                                                                                                                                                                                                                                                                                                       25 Ibid. 26 The CREG also takes account of fuel availability and constraints on the supply and transport of natural gas, for generators that use natural gas for their energy generation. 27 See also CREG, Firm Energy for the Reliability Charge: Hydraulic Plants (http://www.creg.gov.co/cxc/english/enficc/plantas_hidraulicas.htm) 28 Hydro plants which wish to bid their ENFICC 95% PSS into firm energy auctions in order to receive firm energy payments on this larger amount must demonstrate that they have signed contracts with other generators to supply the difference between these two amounts. 29 CREG Document 075, 7 July 2011, Energía Firme para Cargo por Confiabilidad de Plantas Eolícas.  30 Rather than measuring Jepírachi energy production during dry periods or “en condiciones de baja hidrología”, the CREG appears to use the entire seven-year production history of the Jepírachi plant to arrive at the 6% figure. We show below that this ENFICC changes if only El Niño periods are used for this calculation.

     

17    

In its July 2011 document and in its subsequent draft Resolution 148 of October 2011, the CREG suggest two alternative methods for calculating ENFICCs for wind plants: one for plants that have less than 10 years of information on wind resources; and another for plants that have at least 10 years of information. In the first case, they use the operating experience from Jepírachi as the basis for determining the ENFICCs for a new wind power plant, i.e. 6% ENFICC BASE and 7.3% ENFICC 95% PSS. For plants for which there is more than 10 years of wind data, they use the following formula.

E = min (24*1000*k*v3, 24*1000*CEN*(I-IHF))

Where:

E : energy (kWh/day)

k: conversion factor for wind plants, reflecting the number of turbines [MW/(m3/s3)]

v: average monthly wind speed (m/s)

IHF: historic forced outage rate

CEN: effective net capacity (MW)

With this formula, the CREG constructs a probability distribution curve, from the lowest to the highest level of firm energy, using monthly values. The lowest firm energy factor corresponds to a 100% probability of it being exceeded and the highest value has a 0% probability of being exceeded.

The World Bank study, on the other hand, suggested measuring ENFICCs for wind plants using the following exponential smoothing formula under which the “firm energy rating” (the ENFICC) is updated annually:

Firm energy rating in t+1 = ½ (firm energy rating in t)+ ½ (energy produced in year t),

The firm energy rating for the initial year t could be based on recent data; for instance, plants located on the northern coast could use the period of generation recorded by Jepírachi. According to the World Bank, the firm energy rating will adjust quickly to the long run average level of firm energy capability, even if the initial estimate is wrong.31 Applying their formula to a 24-year series of monthly wind and production data related to the Jepírachi plant, the WB estimated an average annual firm energy rating of 38%, with a range between 25% and 47%. They also estimated a firm energy rating for dry seasons of 40%, with a range from 30% to 47%. Their summary financial results refer to a maximum 36%. The difference of these two approaches (WB and CREG) is significant when measured in terms of the financial consequences. Following the CREG approach to estimating ENFICCs for wind energy, a 100 MW wind plant would earn approximately $735,734 per annum at the

                                                                                                                                                                                                                                                                                                                                                         31 WB, pages 33-34.  

     

18    

current firm energy price ($13.998/MWh). Using the World Bank approach (at 36% firm energy rating) the wind plant would earn approximately $4.4 million in annual firm energy payments. As shown in Section 4 below, this makes a large difference to the potential financial viability of wind farms in Colombia.

3.1.2. International practice in calculating firm energy/capacity credits for wind energy

There is no universally accepted method for calculating the contribution of intermittent generating technologies (such a wind) to system reliability. However, there are some basic principles that guide the methodology to be used, as well as experience in the application of this methodology. The basic principle, as described by Cramton and Stoft (“A Capacity Market that Makes Sense”, Electricity Journal, 18, 43-54, August/September 2005) is to only reward capacity that contributes to system reliability as demonstrated by its ability to supply energy or reserves during shortage hours. According to Cramton and Stoft, a major flaw in many US electricity capacity markets has been that they pay for capacity based on average availability, which may or may not contribute to energy reliability when it is actually required.32 Similarly, Cramton and Ockenfels (“Economics and design of capacity markets for the power sector,” 30 May 2011) point out that capacity that qualifies for the capacity market needs to be accurately defined, verified and rated. In particular, they suggest that the contribution of wind and solar energies to system reliability is usually smaller than the contribution of thermal units, and should attract reduced capacity credits. The correct principle is that the capacity market should only elicit the construction of, and pay for capacity that contributes to system reliability or firm energy when it is actually needed, and on the appropriate time scales. The idea behind the CREG's concept of ENFICC is consistent with that principle: it should measure the contribution of a plant to the system’s reliability when it really matters. One internationally recognized methodology for measuring a plant’s contribution to system reliability is referred as the effective load carrying capability (ELCC), which is conceptually the same as the term “capacity credit” as used in the UK.33 In the simplest terms, the ELCC reflects the effect on system reliability of adding the new plant to the system, where reliability is measured in terms of the loss of load probability (LOLP), or the expectation of lost load (LOLE). In other words, when a new plant is added, the ELCC refers to the additional load that can be sustained by the system without a change in reliability. This is the                                                                                                                32 Cramton and Stoft give an example of a “dog” plant which with a long start up time which would never be called upon to relieve short-term energy shortages but which has a high average availability when running, as an example of a plant that should receive very low, or even zero, capacity payments. 33 For a definition of the ELCC, See Milligan M, and K. Porter, “Determining the Capacity Value of Wind: An Updated Survey of Methods and Implementation”, National Renewable Energy Laboratory Conference Paper NREL/CP-500-43433 June 2008.  

     

19    

same as the following UK definition of capacity credits: “a measure of the amount of load that can be served on an electricity system by intermittent plant with no increase in the loss-of-load probability (LOLP), which is often expressed in terms of conventional thermal capacity that an intermittent generator can replace."34 In Figure 2, the new plant adds 400 MW of load carrying capacity, while maintaining the same LOLE (10%). If the information is available to calculate the ELCC of wind and other types of generating plant it would make sense for the CREG use it to measure the ENFICC for all plants on the system. Figure 2: Estimating the Effective Load Carrying Capability (ELCC)35

While the formal methodology has been widely used for system planning, it has seldom been used to define the capacity factors for wind power stations. This seems mainly to reflect the absence of sufficient data, but there may be other reasons, including the complexity of the ELCC calculation and a preference for simpler, established methodologies. The alternatives are approximations to the ELCC and, consequently, have various drawbacks. Nevertheless, they are widely used and serve as the basis for analysing the alternatives that the CREG should consider. There are two types of methodology: risk-based or time period-based. The risk-based methods approximate the system’s LOLP, whereas the time period-based methods capture risk indirectly by assuming a high correlation between hourly demand and LOLP. Although there are drawbacks to time period-based approaches, they are simpler to calculate and frequently used in the USA and elsewhere and we focus on them below.36

                                                                                                               34 “The Costs and Impacts of Intermittency: An assessment of the evidence on the costs and impacts of intermittent generation on the British electricity network,” A report of the Technology and Policy Assessment Function of the UK Energy Research Centre, with financial support from the Carbon Trust, March 2006, page 6. 35 Milligan M, and K. Porter, “Determining the Capacity Value of Wind: An Updated Survey of Methods and Implementation”, National Renewable Energy Laboratory Conference Paper NREL/CP-500-43433 June 2008, page 8.  36 M. R. Milligan “Modeling Utility-Scale Wind Power Plants Part 2: Capacity Credit,” March 2002, p. 18, finds that although “capacity factor (i.e. risk-based or time period-based) methods are not as accurate as ELCC methods for calculating

     

20    

In time period-based estimates of the ELCC, the idea is to measure a plant’s contribution to the system’s reliability when it matters. The World Bank ENFICC calculations for wind in Colombia concentrate on annual average output. Arguably, these do not provide sufficient guarantees that the wind energy will be there when it is needed, in particular when the system is under stress during El Niño periods.37 The most straightforward approach to time period-based methods for approximating the ELCC or capacity credit is to calculate the wind capacity factor (i.e. the ratio of the mean to the maximum energy output) during times of high system demand. Many US regulators and utilities use this method. Each system has different hours of shortage, and each wind power station within a system will have output that coincides more or less with those shortage hours. To the extent that wind generation is higher at times of shortage, the plant will have a higher capacity credit factor. If generation occurs mainly during off-peak hours and little during shortage hours, the capacity credit factor will be much lower. There are different ways to use the time period-related data to approximate the wind capacity credit factor (some of which have been adopted in the USA, as summarized in Table 7). A common but arbitrary method is to order the wind generation output (or estimates based on wind speed) that has been observed, and then to adopt a threshold percentile limit. For instance, the regulator in the Southwest Power Pool in the US sets the capacity factor by reference to the level of wind generation that has occurred at least 85% of the time. This arguably ignores the statistical independence of outages and wrongly suggests that all generation below the threshold makes a zero contribution to system reliability. Consequently, it potentially underestimates the plant’s contribution to system reliability. 38 The CREG’s approach (ENFICC Base) is especially conservative because it ignores the contribution of all generation that cannot be guaranteed 100% of the time. While we would not recommend the use of percentiles as a methodology, we would at least encourage the use of a lower percentile figure than the CREG has adopted.

                                                                                                                                                                                                                                                                                                                                                       capacity credit,” time-period based methods provide “a reasonable trade-off between accuracy and effort, either early in project assessment or if it is not possible to calculate the ELCC”. 37 Although as noted above, the World Bank did calculate firm energy factors for wind power during dry periods, which more closely corresponds to the time-period based approach. 38 “The use of a percentile arbitrarily discounts reliability contributions that are achieved at levels below the percentile value. These approaches are based on fallacious use of probability theory, and they ignore the statistical independence of outages and the fact that system reliability can be achieved at a very high level (such as 1 day in 10 years LOLE) even though every unit in the system is somewhat unreliable.” [Milligan, M. and K. Porter, Determining The Capacity Value Of Wind: A Survey of Methods and Implementation, NREL, May 2005, page 20, http://www.nrel.gov/docs/fy05osti/38062.pdf.]  

     

21    

Table 7: Calculation of Wind Capacity Credit Factors in USA

PJM 13,00% Based on average three year output in peak summer hours.

NYSO 10% - 30% Based on previous years output in peak hours

SWPP 10,00% Based on 85% percentile of outputs in highest 10% of load hours monthly

Minnesota 26,70% Based on sequential Monte Carlo study Pacifcorp 20,00% Based on sequential Monte Carlo study

ERCOT Texas 2,90% Based on wind generation during peak summer hours

Nebraska 17,00% Unknown

Idaho 5,00% Based on 79% percentile of outputs in peak summer hours

Pacific Northwest

15,00% Unknown

California 15% - 60% Three year average of monthly hourly peak production

Colorado 12.5% Based on hourly wind eerngy production from 1996-2005

The alternative methodology (to the use of percentiles) is to average the wind-related generation over the relevant shortage periods. The PJM, for instance, uses the following methodology. 39

A. Sum all of the “hourly outputs” for each of the summer calculation hours (3PM - 7PM) in the year that is three years prior to the current year.

B. Then, for each of those same summer calculation hours, sum the Net Maximum Capacity values (the manufacturer’s output rating less the Station Load).

C. The quotient of the summed summer calculation hour outputs (a) divided by the summed summer calculation hour Net Maximum Capacities (b) will yield a single year capacity [credit] factor for that year.

                                                                                                               39 PJM is a regional transmission organization (RTO) that coordinates the movement of wholesale electricity in all or parts of 13 states and the District of Columbia. For information on capacity factors, see PJM Manual 21 Rules and Procedures for Determination of Generating Capability Revision: 09 Effective Date: May 1, 2010 http://pjm.com/~/media/documents/manuals/m21.ashx )  

     

22    

D. If there is no or incomplete operating data for one or more of the summers (immature Intermittent capacity resource) then the single year capacity [credit] factor for each of those years is assigned the value of the Class Average Capacity [Credit] Factor.

E. Repeating steps (A) through (D) above for each of the two intervening years (current year minus 2, and current year minus 1) will generate two more single year capacity [credit] factors, one for each of those years.

F. The Capacity [Credit] Factor to be used in the current year is the mean (arithmetic average) of the three single year capacity [credit] factors calculated in steps (C) and (D) above.

G. Capacity [credit] factors shall be calculated annually following the summer peak period and be applicable for the delivery year beginning the following June.

If a formal ELCC cannot be measured for wind power stations in Colombia, we would suggest consideration of a methodology similar to the PJM approach, adapted to the relevant shortage hours that will need to be defined to reflect Colombian conditions. The relevant shortage periods in Colombia are mainly El Niño periods, especially during peak hours40. In the absence of information to calculate the ELCC, we have estimated the ENFICCs for wind power using two different methodologies, from hourly generation data for the Jepírachi plant from April 2004 to April 2011.41 The first estimates (Table 8) use the CREG (percentile) methodology but applied to El Niño periods, without reference to peak hours. This yields an ENFICC BASE of 14.8%, compared to 5.8% in all periods (approximately the CREG estimate of 6%).

Table 8: CREG ENFICC methodology for wind applied to El Niño periods

Source: Quantil calculations

Table 8 also reports ENFICCs calculated at lower percentile levels. On this basis, for instance, the ENFICC for wind power with 75% probability would be about 29.2% in El Niño periods and 25.4% using all periods.42

                                                                                                               40 There may also be periods outside el Niño periods when system reliability is under stress. A detailed study of ENFICC for specific plants should reflect availability during the times when reliability is threatened. 41 One reviewer suggested that our estimates of ENFICC may be biased downwards because they include a period when Jepírachi plants were still in a learning phase. 42 In Table 8 we find the “All Periods” ENFICC 95% PSS to be 12.8%, compared with the CREG's 7.3%. We have been unable to determine how the CREG arrived at this lower figure.

ENFICCs El  Niño  Periods All  PeriodsENFICC  BASE 14,8% 5,8%ENFICC  95%  PSS 18,4% 12,8%ENFICC  75%  PSS 29,2% 25,4%

     

23    

The second estimate (Table 9) uses the PJM methodology, applied to peak hours during the last three El Niño periods43. This yields estimates of ENFICC Base between 27% and 33%.

Table 9: PJM methodology to determine ENFICC for wind in Colombia

Source: Quantil calculations We note that the estimates of ENFICC between 15 – 30% are consistent with the international experience that we have reviewed, including US and UK studies44. Based on the analysis above, in Section 4 we model the impact on IRR for a prototype wind project under six different measures of ENFICC: 0%, 6%, 15%, 20%, 30% and 36%.

3.2 CER payments The second policy variable is the price of a CER, which is paid for a certified emission reduction related to projects like Jepírachi. The procedure for determining whether a project is eligible for CERs and how many CERs it is entitled to, is defined by the Clean Development Mechanism (CDM), under the Kyoto Protocol.45 A project owner may secure investment or a reliable source of revenue through the forward sale of CERs in agreements with (mainly) Annex 1 parties under the Kyoto Protocol. Although the CER is an international policy mechanism, it has domestic policy implications. This is largely because eligibility under the CDM (and hence CER income) is conditional upon demonstrating that the project’s economic viability requires these additional payments; but economic viability depends in large part on the way that wind power is compensated under Colombia regulations. This is not an issue as long as wind power is uneconomic even with the CER payment, but it would become an important domestic policy issue when wind power becomes economically viable without CER payments.

                                                                                                                                                                                                                                                                                                                                                         43 Niño 1 from 31/07/04 to 31/03/05; Niño 2 from 30/09/06 to 28/02/07; and Niño 3 from 30/06/09 to 31/05/10. 44 A UK report says that: "There is a range of estimates for capacity credits in the literature and ... the range of findings relevant to British conditions is approximately 20 – 30% of installed capacity when up to 20% of electricity is sourced from intermittent supplies". “The Costs and Impacts of Intermittency: An assessment of the evidence on the costs and impacts of intermittent generation on the British electricity network,” A report of the Technology and Policy Assessment Function of the UK Energy Research Centre, with financial support from the Carbon Trust, March 2006, page v. 45 For details of the procedures, see the documents that apply to the Jepírachi application and approval for CDM credits. See Jepírachi CDM Project Design Document http://cdm.unfccc.int/Projects/DB/SGS-UKL1135244574.04. Also, for a more general introduction, see this website: http://www.tfsgreen.com/global-markets/clean-development-mechanism/cer-pricing.php,  

ENFICCs  BASE 5-­‐7AM;  6-­‐8PM All  dayNiño  1 24,13% 29,05%Niño  2 27,32% 37,02%Niño  3 30,34% 33,78%Average 27,26% 33,29%

5-­‐7AM;  6-­‐8PM All  dayAll  3  Niños 27,52% 32,58%

     

24    

With current CER prices below $10/tCO246, investors will give little value to this revenue component in their investment appraisal. However, international policy decisions will influence the level of the CER prices. For instance, the EU could tighten the EU ETS market through retiring emission allowances or by raising the EU commitment to reduce emissions from 20% to 25% compared to 1990 levels. The creation of California and Australia cap and trade regimes could likewise increase the demand and the price for “offset” products. The compensation from CERs depends not only on market prices for CER’s, but also on the “emissions factor” that has been attributed to the project47. The emissions factor refers to the amount of CO2 avoided for each MWh generated by the plant that is earning CERs. That factor depends on what generation is being substituted at the margin. In the most recent decision with respect to the Jepírachi plant, the emission factor for wind power was set at 0.44tCO2/MWh, which suggests that wind power stations are displacing gas fired stations. We have modelled the effect on the IRR of three different levels of CER per ton of CO2 avoided: $0, $10 and $30. Current CER market conditions suggest a price between $0 and $10, and the $30 price is optimistic. We have used an emissions factor of 0.35tCO2/MWh, so our IRR results will be slightly biased downward.

                                                                                                               46 Forward prices for CER range from €3.52/tCO2 (2012) to €5.96/tCO2 (2020). See http://www.eex.com/en/Market%20Data/Trading%20Data/Emission%20Rights/Certified%20Emission%20Reductions%20Futures%20|%20Derivatives 47 See Jepírachi CDM Project Design Document http://cdm.unfccc.int/Projects/DB/SGS-UKL1135244574.04.

     

25    

4. Financial analysis of a wind project under the current regulatory regime

We have developed a financial model to compare the financial viability of investment in wind power under the different regulatory scenarios defined in the last section. The analysis assumes that the plant under investigation has similar technical characteristics and wind conditions to the Jepírachi plant that belongs to Empresas Públicas de Medellín (EPM). We have also made simplifying assumptions about financing (e.g. leverage). Although we carry out a number of sensitivities, the results are specific to that plant and would naturally differ if the technology, location, costs, financing and wind conditions were different. Furthermore, we have chosen to focus on two specific policy issues (ENFICC and CER), but recognize that other policies are also important (e.g. taxation)48. This section summarizes the model and results, including: 1. What the model calculates 2. Logic of the model 3. Main elements of the financial modelling 4. Benchmarking the model 5. Results of the main policy scenarios 6. Sensitivities to non policy parameters 7. Risk analysis 8. Other financial considerations 9. Conclusions from the modelling

4.1 What the model calculates49

The model calculates the internal rate of return (IRR) associated with investment in a wind power plant in the Guajíra region of Colombia, under a range of different regulatory assumptions. The IRR is the discount rate that yields a net present value of cash flows equal to zero. The higher the IRR, the more attractive the project is likely to be. Generally, a company should consider investing in a project when the IRR exceeds the company’s risk-adjusted cost of capital. In our modelling, we have calculated the equity IRR, on the assumption that the project will be financed with 70% debt and 30% equity. We use three different financial metrics to assess the attractiveness of the project for private investors. First, following the lead of the WB in its report, we have adopted a real 14% IRR on equity (pre-corporation tax) as the minimum threshold for investing in the assumed wind project at this level of gearing. We refer to the difference between the equity IRR and the 14% threshold as the financial gap – the bigger the gap, the less likely the private investment.

                                                                                                               48 See World Bank Report for a review of a wider range of policy issues.  49 The model is not intended as basis for making financial decisions for a specific investment. It is rather a simplified model that aims to explore the implications of changing key policy assumptions and key variables. More details about the modeling are in a separate document prepared by Quantil, in Annex 4.

     

26    

Second, we have measured the payback period. Investors in wind power project normally expect payback periods of less than 10 years. Third, we estimate a debt service cover ratio (DSCR) to ascertain whether annual cash flows are sufficient to cover annual payments for interest and principal on debt. Typically lenders of non-recourse debt expect the borrower to maintain a DSCR of at least 1.2. There are two parts to the modelling. The first estimates the IRR using the forecast mean of prices and wind speed, with general uncertainty incorporated into the 14% IRR threshold (i.e. higher than the risk-free rate of return). The other uses Monte Carlo simulations to explore two specific uncertainties, related to wind speed and energy prices. This second approach offers insights into the probability distribution of returns and is an additional piece of information that could raise or lower the investor’s perception of risk and hence the required rate of return.

4.2 The logic of the model

The model enables us to estimate the IRR of a wind project, with the following standard

formulation: if !∗ = !"" then  0 = !"!!!!!∗ !

!

!!!, where !"!! is the net cash flow at time t

and !"!! is the initial investment. To understand the logic and the key policy issues affecting profitability, we can look at the main determinants of revenue and of cost, especially those determinants that are outside the investor’s control.

4.2.1 Revenue

On the revenue side of the equation, there are three main sources of revenue in the Colombia electricity system for a wind project. The first is sales in the Colombian spot market (or short term contracts that reflect the spot market) for the life of the project. Estimated revenue will therefore depend primarily on the forecast output of the plant and on the expected price of electricity. Second, the model assumes that the wind generator also receives a fixed payment per kWh (the firm energy payment, or option fee) for estimated firm energy based on the approved ENFICC. The generator receives the fixed payment regardless of actual output, but there are special rules that apply when the spot price is above the scarcity price (which would normally only be reached in periods of el Niño). For any given generator, the fixed payment depends on the price for firm energy, as determined in a system-wide auction, and the ENFICC. The higher the ENFICC, the higher the generator’s certain revenue stream. A third source of revenue is related to externalities. One externality is reflected in our model: CER income reflects avoided CO2 emissions in the case of a wind project. Our model

     

27    

underestimates slightly the contribution from the CER because it uses an emissions factor of 0.35tCO/MWh, rather than 0.44tCO/MWh. A second externality has been characterized as being related to a project’s contribution to “sustainability”, with a positive contribution earning additional revenue and a negative externality involving a penalty payment. As far as we know, the current regulations do not include this source of revenue for wind power, and we have not included this concept in our model.

4.2.2 Cost

On the cost side, apart from the investment and operating costs that are specific to the plant itself, there are two taxes that apply to output. FAZNI is a tax (per kWh) that helps to finance investment projects in areas not connected to the national grid. CERE is a charge (per kWh) that funds the payment of the system’s reliability charges. The total amount of money collected by CERE payments must be equal to the total amount of money paid out in firm energy payments. For plants below 20 MW, these taxes are not levied, even though this would appear to contravene the revenue-neutrality principle. Since we will be analyzing projects larger than 20 MW, the model assumes that the plant does pay these taxes on output. The model sets to zero the fixed O&M costs and variable costs. Although they are relatively small and do not affect our basic conclusions, they do mean that our estimated IRR are biased upward by a small amount. On the other hand, as noted, we have used a slightly lower emissions factor in our model and this lowers our IRR estimates.

4.3 Main elements of the financial modelling We have split the variables between parameters, which are variables that we will not consider as risk factors, and risk factors. The parameters, divided into policy instruments and non-policy instruments, are assumed to be fixed50. Risk factors, on the other hand, are modeled using Monte Carlo simulations in order to quantify the uncertainty associated with the internal rate of return (IRR). We make the following classification of variables. Parameters:

1. Policy: a. Firm energy factor (ENFICC) as % of capacity b. CER (Certified emission reduction) in $/tCO2 avoided

2. Non-policy:

a. The power curve that determines output b. Investment per KW c. Scarcity price

                                                                                                               50 Although parameters can vary over time, we assume they are deterministic for any given model run.

     

28    

d. Firm energy price (cargo por confiabilidad). e. Variable costs f. Operational costs g. Energy in excess of firm energy sold in the spot market when market

prices exceed the scarcity price h. Exchange rate (firm energy and scarcity price are denominated in dollars) i. Interest rate.

Risk factors: 1. Spot price 2. Wind speed.

4.4 Benchmarking the model

Our model uses a similar methodology to the one used by the World Bank (WB). In order to calibrate our model and confirm that our results seem robust, we have prepared a benchmark scenario that also has broadly comparable inputs with the WB’s. The resulting IRRs should be comparable. Table 10 summarizes the main inputs of our model, attempting to approximate World Bank parameters for investment in a 302MW (230 Nordex N60 mills) wind farm in the Guajíra region of Colombia. Note that the left hand column identifies the parameter in the “natural currency”; the other two columns express the parameter by reference to the two reports: WB value is in US$ and our equivalent value is in Colombia pesos, unless stated otherwise. In other words, we model investments in pesos, and the World Bank is modeling in $US. Table 10: World Bank parameterization with Quantil model

Parameter (natural currency) WB Value (US $) Our Value (Colombian Pesos)

Exchange rate (pesos/USD) 1,900 Spot Price (pesos/KWh) 0.0506 96.14 (based on WB) Installation costs (USD/KW) 1,800 3,420,000 (based on WB) FAZNI (pesos/MWh) 1000 pesos 1,000 CERE (pesos/MWh) 20,000 pesos 20,000 Firm energy Price (USD/MWh) 14 28,000 (26,600 WB) Wind speed (m/s) Not certain 8.75 Time horizon (years) Not certain 25

Table 11 summarizes the project IRR corresponding to our model and to the WB model under different firm policy scenarios.

     

29    

Table 11: Comparing results of WB and Quantil model – Project IRR

ENFICC (%) Project IRR-Quantil (%)

Project IRR-WB (%)

0 5.38 5.8 20 7.40 7.8 30 8.42 8.7 36 9.01 9.35

The two models estimate approximately the same project IRR for all the scenarios they have in common. Quantil estimates are consistently about 0.3-0.4% lower than the World Bank estimates. We are not certain about what explains the difference since we do not know all of the World Bank assumptions. However, we think it may be due to a combination of factors, including a different wind speed or power curve. In any case, we consider that our model is robust in the sense that it is giving comparable results to the World Bank’s.

4.5 Results of the main policy scenarios

The reference policy scenario is our understanding of the current regulatory regime that would apply to a new wind power plant (like Jepírachi, but) larger than 20 MW, assuming 70 debt-financing for fifteen years at a 9% real rate of interest. It assumes an ENFICC of 6% for wind power and a CER of $0/tCO2. It does not reflect low-cost finance, nor does it take into account any avoided externalities. We then study the following policy scenarios

• First, we estimate the impact on IRR of a higher level of ENFICC than the CREG has attributed to wind power; we explore the implications for the IRR of ENFICC for wind power of 15%, 20%, 30% and 36% (compared to the CREG’s 6%).

• Second, we estimate the impact of positive CER prices: $10 and $30/tCO2 for avoided CO2 (compared to $0 in the reference scenario), assuming an emission factor of 0.35tCO2/MWh.

For the reference scenario and the alternative policy scenarios, we estimate the real equity IRR, pre-corporation tax and in real terms, as well as two financial ratios, the payback period and the debt service cover ratio (DSCR). For the investment to be attractive, we are assuming that it should earn 14% on equity pre corporation tax, that the payback period should be less than 10 years and that the debt service ratio should be at least 1.2.

4.5.1 Variations of the ENFICC

Table 12 summarizes the results for the reference scenario and for different firm energy factors (ENFICC). In the reference scenario (no CER income, ENFICC 6%) the IRR on equity is only 3.5%. Compared to the threshold 14% IRR, there is a large financial gap. The

     

30    

payback period and the DSCR are correspondingly unacceptable to investors. The equity IRR rises to a maximum of 8.9% (at 36% ENFICC), with a payback period of 16 years and a debt service cover ratio of 1.16. Our own estimates for the ENFICC range from 15% to 30%, which would yield an IRR on equity of between 5.0% and 7.8%. Although the increase in returns compared to the reference case illustrates the importance of the ENFICC payment, the financial gap is still substantial. Table 12: Results of reference scenario and of changes in ENFICC

ENFICC Equity

IRR Pay Back

(years) Debt Service Cover Ratio

0% 2.4% 21.7 0.86

6% 3.5% 20.5 0.91 15% 5.0% 18.9 0.99 20% 5.9% 18.2 1.03 30% 7.8% 16.8 1.11 36% 8.9% 16.0 1.16

4.5.2 Different measures of CER

Table 13 summarizes the results of introducing a CER payment of different amounts, assuming that each MWh of wind power avoids 0.35tCO2. The IRR on equity reaches 14.2% at a CER of $30/tCO2 and an ENFICC of 36%; both of those assumptions are unrealistic. With a CER of near $10 and an ENFICC of 30%, the IRR is close to 10%. Table 13: Results of reference scenario and of changes in ENFICC and CER

ENFICC Equity IRR CER $0

Equity IRR CER $10

Equity IRR CER $30

0% 2.4% 4.0% 7.2% 6% 3.5% 5.0% 8.3% 15% 5.0% 6.6% 10.0% 20% 5.9% 7.6% 10.9% 30% 7.8% 9.4% 13.0% 36% 8.9% 10.6% 14.2%

     

31    

4.6 Sensitivities Tables 14-18 summarize the results of sensitivities of the IRR to changes in key non-policy parameters, in particular wind speed, energy price, investment cost, exchange rates and interest rates. They assume a zero CER and different values of ENFICC51. The sensitivities help to understand which variables are critical and also to identify variables that require additional analysis in our risk assessment.

4.6.1 Investment cost

Investment costs are to a large extent known at the time of investment and are clearly an important determinant of returns. Our reference scenario assumes a benchmark investment cost (BIC) of $1800/kW. This was the low scenario used by the World Bank in its study (2010). Fixed costs may have fallen since then; at least that seems to be the evidence reflected in industry estimates. For instance, BNEF currently has three capital cost scenarios for wind power plants. Their estimates range from $1,450/kW to $1,950/kW, including development costs, construction cost and capital investment. Their figures, and the recent outcome of auctions in other countries (e.g. Brazil, Peru) suggest that the fixed costs of a new wind power station have declined recently. On the other hand, companies with whom we have spoken in Colombia have suggested that the capital cost could be closer to or even higher than $2000/kW. These investment costs do not include network related costs. One company in Colombia estimated that the investment cost could reach $2400/kW if the full effect of transmission and connection costs were borne by a new 300 MW plant. We have not included these costs in our BIC estimates for this plant, partly because we do not know what they will be, and also because we anticipate that they will be shared with the system users, Jepírachi and with other new projects in the area52. In Table 14, we report the estimated impact of investment costs that are 75% and 125% of our benchmark costs. Our conclusion is that an unlikely decline of 25% in the assumed investment cost (from $1800/kWh to $1350/kWh) would yield almost 14% IRR on equity with an ENFICC of 20%. If costs were 25% higher ($2250/kW) than the benchmark cost, returns on equity would drop substantially, to no more than 3.8% when ENFICC is 36%.

                                                                                                               51 The IRR results in the sensitivities are biased downward because we assume zero CER revenue. But the conclusions are unchanged, unless we assume CER prices of $30/tCO2, which we consider unrealistic. 52 A recent article in Scientific American suggests that new projects are being considered in the Guajíra region, “including a 200-megawatt wind farm in the peninsula's Ipapure region and a 20-megawatt site at Joutkai, which would be close enough to Jepírachi to share the same substation feeding Colombia's national power grid.” See http://www.scientificamerican.com/article.cfm?id=wind-power-colombia-guajira.    

 

     

32    

Table 14: Results of sensitivity case for changes in Benchmark Investment cost (BIC)

ENFICC Equity IRR BIC

Equity IRR 75% BIC

Equity IRR 125% BIC

0% 2.4% 8.4% -1.2% 6% 3.5% 10.0% -0.3% 15% 5.0% 12.4% 0.9% 20% 5.9% 13.7% 1.6% 30% 7.8% 16.7% 3.0% 36% 8.9% 18.6% 3.8%

Although we do not anticipate that investment costs will be as low as $1350, we note that investment costs for turbines have been declining. If costs of this order were achieved, the payback periods would fall below 10 years and the debt service cover ratio would rise to about 1.4 when the ENFICC is above 20%.

4.6.2 Wind output

Average wind output is uncertain. Our reference scenario assumes an average of 8.75 m/s, based on historical data from Jepírachi generation and confirmed in discussions with industry. Table 15 shows that, if wind speed is higher, the IRR on equity rises sharply; with 10 m/s, equity IRR rises above 14% for an ENFICC of 20%. At that wind speed and a 20% ENFICC, payback periods fall below 10 years and debt service ratios rise above 1.3. On the other hand, a relatively small reduction in average wind speed (from 8.75 to 8.0 m/s) reduces the IRR to a range of -2.0% to 4.4%. These results explain why investors pay special attention to prospective wind speeds when considering alternative locations. We use Monte Carlo analysis below to attempt to understand better the risks associated with wind speed and price, obviously without entering into the detail that would be required prior to making an investment. Table 15: Results of sensitivity case for changes in wind speed

ENFICC Equity IRR 8.75 m/s

Equity IRR 10 m/s

Equity IRR 8 m/s

0% 2.4% 10.1% -2.0% 6% 3.5% 11.3% -0.9% 15% 5.0% 13.1% 0.7% 20% 5.9% 14.2% 1.6% 30% 7.8% 16.4% 3.4% 36% 8.9% 17.8% 4.4%

     

33    

4.6.3 Price of energy

The spot price of energy is unknown at the time of investment and is probably the most significant source of uncertainty. For instance, the price of energy in the spot market has recently increased from a range of 70-90 pesos/kWh in the fall of 2011 and early 2012, to over 120 pesos/kWh in March 2012. The scarcity price is well in excess of 400 pesos/kWh. Although a significant amount of energy is traded through contracts, our understanding is that they are relatively short. Unless investors are able to sign a purchase power agreement (PPA) for at least 15 years, they will face the uncertainty related to spot prices and short-term contracts. Our reference case assumes a constant real price of 96.14 pesos/kWh53. Table 16 shows that if the average energy price turns out to be 120 pesos/kWh, the equity IRR would rise substantially, reaching 14% on equity with an ENFICC of 30%. On the other hand, if average energy prices were 85 pesos/kWh, the maximum IRR is about 6%. This serves as an illustration of the extent of uncertainty facing investors who are not able to lock in a long-term revenue stream through the firm energy auction. Table 16: Results of sensitivity case for changes in energy price

ENFICC Equity IRR 96.14

pesos/kWh

Equity IRR 120

pesos/kWh

Equity IRR 85

pesos/kWh

0% 2.4% 8.2% -0.3% 6% 3.5% 9.3% 0.8% 15% 5.0% 11.0% 2.4% 20% 5.9% 12.0% 3.2% 30% 7.8% 14.1% 5.0% 36% 8.9% 15.4% 6.1%

We are conscious that the price earned by wind power may be different than the average system price54. This will depend on daily wind patterns: if wind is strongest at times when demand and system prices are lower than average, then wind power will earn less than the average system price, and vice versa. With the data available, we have studied this effect in el Niño periods and in other periods for the Jepírachi plant. Our conclusion is that there is no statistically significant difference between the weighted average price for that plant (assuming it sells in the spot market) and the average system price.

                                                                                                               53 This is based on the WB report. One approach to long term price forecasting would be to estimate the long run marginal cost of electricity in Colombia. 54 See Paul Joskow, “Comparing The Costs Of Intermittent and Dispatchable Electricity Generating Technologies”, September 27, 2010 (Revised February 9, 2011) DISCUSSION DRAFT.

 

     

34    

4.6.4 Interest Rates

We have assumed that the project is financed through non-recourse debt at a 9% real rate of interest, with a ratio of 70% debt and 30% equity55. Table 17 illustrates that the results are not very sensitive to relatively small changes in the interest rate. However, if the project receives a loan at a subsidized rate of interest, this could change the conclusion. We estimate that a 4% interest rate would increase the equity IRR to approximately 13%, with a payback of less than 10 years and a debt service cover ratio of over 1.5. Table 17: Results of sensitivity case for changes in interest rate

ENFICC Equity IRR Interest

9%

Equity IRR Interest

8%

Equity IRR Interest

10%

Equity IRR Interest

4%

0% 2.4% 3.1% 1.6% 6.6% 6% 3.5% 4.3% 2.6% 7.8% 15% 5.0% 5.9% 4.2% 9.7% 20% 5.9% 6.8% 5.1% 10.8% 30% 7.8% 8.7% 6.8% 13.0% 36% 8.9% 9.9% 7.9% 14.4%

4.6.5 Exchange Rates

Our model estimates the IRR based on revenue and cost streams in Colombian pesos, but some costs will be influenced by the exchange rate, either before the investment is made, or after. The main effect of a different exchange rate before the investment is made will be through the capital costs. After the investment, exchange rates may affect profitability, for instance through changes in the firm energy price. The reference scenario assumes an exchange rate of 1900 pesos/US$. Table 18 illustrates that an increase or decrease of about 10% in the exchange rate has very little effect on the IRR.

                                                                                                               55 The level of leverage will also influence the return on equity. However, examining the implication of different leverage is not straightforward, among other reasons because higher leverage raises the risk facing lenders and can raise the cost of debt.

     

35    

Table 18: Results of sensitivity case for changes in exchange rate

ENFICC Equity IRR 1900

pesos/$

Equity IRR 2100

pesos/$

Equity IRR 1700

pesos/$

0% 2.4% 2.4% 2.4% 6% 3.5% 3.6% 3.3% 15% 5.0% 5.3% 4.7% 20% 5.9% 6.3% 5.5% 30% 7.8% 8.3% 7.2% 36% 8.9% 9.6% 8.2%

4.7 Risk Analysis

We have modeled the implications of introducing uncertainty in the market price of energy and in wind speed. Using two different Monte Carlo simulation models, we derive a distribution of IRR’s for the wind project, under our reference scenario and under different levels of ENFICC for wind. Since the returns in our reference scenario (and under variations in ENFICC) are systematically too low to attract investment, a key question is what probability there is of earning a 14% return on equity.

4.7.1 Basic model

The basic model uses a standard Ornstein Uhlenbeck mean reverting process estimated in logarithms of the spot price. The model captures basic uncertainty in the future spot price. All results are reported using 1000 simulations for 25 years. The mean IRR on equity is approximately the same as in our basic financial model, in this case between 2.7% and 9.9%. The results reported in Table 19 suggest that there is a 5% probability of earning much less than the mean, but since the mean is too low anyway, what matters is whether there is a reasonable probability of earning 14%. Not so, according to this Monte Carlo analysis. Even with a 36% ENFICC, the probably of earning 14% on equity is no greater than 6.9%. Table 19: Results of Monte Carlo Analysis – Mean Reversion

ENFICC Equity IRR Mean

Equity IRR 5%

Quantil

Probability IRR>14%

0% 2.7% -0.3% 0.1% 6% 3.9% 0.8% 0.1% 15% 5.7% 2.5% 0.3% 20% 6.7% 3.5% 0.9% 30% 8.7% 5.3% 2.7% 36% 9.9% 6.4% 6.9%

     

36    

4.7.2 Model 2

The second model is based on several historical variables (anomaly, energy demand, lagged prices, and others). It captures the seasonal effects of the spot price, but not el Niño. It introduces much more uncertainty about future prices than the mean reverting model. The model implies much more uncertainty about future spot prices than the basic model, and consequently wide variations in IRR. As reflected in Table 20, it predicts a much higher future average price (122.08 pesos/kWh) than our forecast based on historic levels (96.14 pesos/kWh). However, because it forecasts relatively low average prices for the first five years (77.61 pesos/kWh) and higher average prices for the remainder of the period (133.20 pesos/kWh), the mean IRR is significantly lower than in the basic model, and ranges more widely with changes in the ENFICC, from -7.2% to 5.0%. The uncertainty is reflected in the very high potential losses at a 5% probability, and the reasonably high probability of achieving 14% IRR on equity. We would expect investors to reflect the significant risk in their investment appraisal, either by lowering the anticipated cash flows, or by raising the required return to reflect risk. Table 20: Results of Monte Carlo Analysis – Model 2

ENFICC Equity IRR Mean

Equity IRR 5%

Quantil

Probability IRR>14%

0% -7.2% -64% 14% 6% -3.5% -26% 16% 15% 0.1% -18% 20% 20% 1.7% -16% 21% 30% 3.9% -12% 23% 36% 5.0% -10% 25%

4.8 Other financial considerations

Investors will consider a number of other financial issues, apart from expected return on equity. They will typically be seeking non-recourse loans, with minimum debt service cover ratios. Lenders will also seek “bankable” revenue streams. And equity investors will be seeking acceptable payback periods, which are unlikely to be much more than 10 years.

4.8.1 DSCR

Virtually all private investors in wind power assess projects on the basis that the project will be heavily leveraged, typically with 70-85% of non-recourse debt. Many project developers will borrow the funds from banks. But even companies that finance projects off their balance sheets assess projects as if they were borrowing the funds. It is therefore fundamental to know how private lenders will react to the risk associated with the project.

     

37    

First, lenders naturally want to make sure that the project will be able to return the debt along with annual interest payments. They usually attach annual debt service cover ratio (DSCR) conditions to their loans, allowing them to demand repayment of the loan if the conditions are breached. For example, a DSCR of 1.2 requires $120 cash to cover every $100 in obligations and annual payments. The higher the probability that cash flow may not cover the ratio, the harder it will be to raise debt finance and the higher the cost of capital. We have analysed the DSCR of our wind project. In the reference scenario and most realistic sensitivities, the ratio is below 1.2 and often well below 1. This confirms the unattractiveness of the project under current conditions.

4.8.2 Bankable revenue streams

Second, private lenders to renewable power projects almost always seek a bankable income stream as a condition for what are typically non-recourse loans. The bankable revenue stream may be in the form of long-term power purchase agreements (PPA) with a credible buyer, or a regulated feed in tariff. Lenders will also lend for renewable energy projects when some element of the revenue stream is variable (as they do, for instance in the UK and Spanish electricity sectors). However, the cost of debt is generally higher when the revenue stream is more uncertain and, at some point, there will be no willing commercial lenders. Indeed, in the UK, the difficulty of raising debt to finance renewable projects within the current regulatory regime has led the government to introduce new regulations involving long-term contracts (misleading referred to as feed-in-tariffs) for low carbon sources of generation, including nuclear power, renewable power and carbon capture and storage projects involving coal. The British Government argues that long-term revenue security will make more capital available for investment and lower the cost of capital56. To the extent that a wind project in Colombia cannot rely on a bankable revenue stream, the prospects of raising debt finance are significantly worse. There are three ways of providing some guaranteed revenue in Colombia. The first is to sign a long-term contract with a large customer, which seems unlikely at current costs of wind power, unless the project is being subsidized in some way. The second is to be selected in the firm energy auctions, which would guarantee the payment of an option fee for 20 years. This would not generate much revenue with the current ENFICC for wind. However, if the ENFICC were significantly higher, the additional revenue could improve the prospects for debt finance. For instance, if the ENFICC were 30% and the option fee was $15.7/MWh, a 100 MW wind power station would receive an annual income in excess of $4.1 million. The third source of stable revenue is the CER, provided the project is eligible. These three revenue streams may or may not be sufficient to meet the minimum threshold of profitability for the equity owner, but they would help to lower the cost of borrowing funds.

                                                                                                               56 It is arguable that the real benefit is not for the country, but rather for the investors who are able to deduct interest payment from their taxable income, which is effectively a cost to the general taxpayer. In any case, providing a steady stream of income does facilitate project financing.

     

38    

4.8.3 The payback period

Investors in wind parks with whom we have spoken say that they expect a payback period of less than 10 years. Our modelling suggests that this is not possible under the reference scenario, even at high levels of ENFICC.

4.9 Conclusions from modelling

The main conclusion is that investment in this specific wind power project in Colombia is unlikely to be attractive to private investors under the current regulatory regime, which in our reference case assumes a 6% ENFICC and no CER income. The financial gap is substantial between the assumed minimum threshold equity return for private investors (14% real pre corporation tax) and the expected IRR in the reference case (3.5%). We have examined the impact of an increase in the ENFICC, in line with our own analysis of the methodology in the last section, which suggests that the ENFICC for wind power should be in the range of 15-30%. At the higher end of that range, the expected IRR rises to almost 8%, but this still leaves a substantial financial gap. We also consider the implications of a positive CER. With a CER of about $30t/CO2 and a 30% ENFICC, the expected IRR could reach almost 13%. However, we do not consider this to be a realistic assumption for the CER. Our sensitivity analysis confirms that the equity IRR is very unlikely to reach levels that make this wind park an attractive investment. Furthermore, Monte Carlo risk analysis suggests that there is a substantial potential downside to the returns. Although one of the simulations suggests a reasonable probability of achieving a 14% IRR, in that case the risks on the downside are so great that we are confident that investors would raise their threshold required return above 14%. We also examined some of the other financial criteria that equity investors and project lenders would seek, including evidence of sufficiently high debt interest cover ratios, reasonable payback periods and bankable revenue streams. On all of these criteria, the project would appear to be unattractive to private investors. Our overall conclusion is that private investment in this wind power station in Colombia is unlikely under current regulatory and market conditions, even when the ENFICC rises to the level that we think is reasonable. Recognizing the avoided CO2 emissions through CER payments improves the returns, but not sufficiently to change the conclusion. We also note that the uncertainty of spot prices could well raise the threshold cost of capital, and that the risk is reduced through payment mechanisms (e.g. firm energy, CERs) that provide greater certainty about cash flows.

     

39    

A fuller analysis of the environmental and social externalities associated with large hydro projects and coal based generation would probably favour wind and other non-conventional renewable sources of energy, but we have not carried out that analysis here. It is important to stress that this financial analysis applies to a particular power station. To the extent that new turbine technologies are more efficient, or other locations have better wind conditions, the financial appraisal will be more positive, provided the improved efficiency more than compensates for any additional costs. Nevertheless, having carried out a number of sensitivities, the project does not appear to be attractive when we use the CREG’s proposed ENFICC. Our conclusion is that the regulatory treatment of the ENFICC is a key barrier in Colombia, even for wind power projects that might face better conditions than the project we have analysed.

     

40    

5. Other risks and opportunities facing investors

In addition to the financial appraisal, investors will be interested in considering a number of other opportunities and risks. In this section, we review some of the opportunities and risks that would influence investor perceptions, in particular foreign investors. To offer ideas that may be worth considering further, we also summarize some lessons from other countries that have sought to promote investment in wind power.

5.1 Commercial opportunity and risk

On the positive side, investors may be interested in investing early to benefit from the potential for significant growth of wind energy in Colombia. There are differing views on the potential for wind power in Colombia, with some studies focusing more on technical potential and others more on the commercial potential. At least one study suggests “not so large” potential57, but others are more positive. A number of companies with whom we spoke stressed the importance of having reliable and detailed information about wind and other renewable energy resources, in order to promote efficient investment in the sector. In particular, they noted the need for accurate information on wind speed at 60 metres. A study by UPME and CorpoEma reports that Colombia has the potential for 99 GW of wind power, of which 25 GW in the Guajíra region could generate 81,216 GWh per year (more than total Colombian electricity demand)58. However, much of the capacity is in areas of the country that are remote from markets. The Atlas de Viento y Energía Eólica de Colombia59 includes data and maps (see Figure 3) on wind speeds in different areas of the country. It concludes that the areas with the greatest potential for wind power are in the Guajíra peninsula, the Island of San Andrés, parts of Boyacá and along the Caribbean coast in the Department of Bolívar. The potential for growth in wind power generation will depend also on the availability and costs of alternative generation technologies, especially when existing capacity needs to be replaced or supplemented to cope with high demand growth. The report by UPME and CorpoEma suggests that significant incremental capacity will be needed especially after 2021, and that this will be the timeframe in which to plan for an expansion of non-conventional energy sources.60                                                                                                                57 “Renewable Energy Potential Of Latin America”, Global Energy Network Institute (GENI), Dec. 2009, page 38.

58 Formulación De Un Plan De Desarrollo Para Las Fuentes No Convencionales De Energía En Colombia (Pdfnce), Volumen 1: Plan De Desarrollo Para Las Fuentes No Convencionales De Energía En Colombia (Pdfnce) Ttp://www.corpoema.com/pdf/vol 1 Plan Desarrollo.pdf page 3-15 and 3-16.

59 http://www.upme.gov.co/Atlas_Viento.htm. 60 Formulación De Un Plan De Desarrollo Para Las Fuentes No Convencionales De Energía En Colombia (Pdfnce), Volumen 1: Plan De Desarrollo Para Las Fuentes No Convencionales De Energía En Colombia (Pdfnce) Ttp://www.corpoema.com/pdf/vol 1 Plan Desarrollo.pdf page 0-5.

 

     

41    

In a competitive market, the primary commercial challenge facing potential investors in wind power is competition from alternative energy sources. Our report illustrates the apparently uneconomic nature of wind power today, as measured by reference to our project. Even if the cost of wind power fell to levels that would be competitive against other technologies, investors in wind power also face the risk that the cost of fossil fuel technologies might fall, for instance with low cost unconventional shale gas. Figure 3: Average surface wind speed en January in Colombia (m/s)61

                                                                                                               61 UPME, Atlas de Viento y Energía Eólico, http://www.upme.gov.co/Atlas_Viento.htm, page 25.

     

42    

5.2 Political considerations Foreign investors will naturally consider a number of political factors, including those at the following levels: international, regional, national and local. Foreign investors will be influenced by the general perception of the country as a place to do business. A recent presentation in Madrid by the Colombian Ambassador to Spain62 suggests that Colombia is now ranked quite highly on this criterion. Generally, wind power sector investors with whom we have spoken recently consider Colombia to be a country that has a stable political system with promising economic prospects and where investment is generally safe. Foreign investors will be specifically interested in the nature of the relationship between the governments of Colombia and of the countries where the investors are based. One way of measuring this is through the existence of bilateral trading and investment agreements. For instance, Colombia and the EU have almost concluded a free trade agreement. If it is finally signed, this will open the door to a variety of potential benefits. On the other hand, the EU had not yet approved the free trade agreement when we wrote this report; we understand that there is opposition to signing the agreement, mainly on the grounds of inadequate protection of human rights in Colombia.63 The EU decision is expected this summer and could influence investment. Foreign investors will also be influenced by the perception of regional political stability. The relationship between Colombia and Venezuela is obviously important because the Guajíra region borders Venezuela. The evidence from a general review of the press suggests that the relationship between the two countries is significantly better than a few years ago, when there was talk of war. Nevertheless, foreign investors in the Guajíra region will inevitably be watching this relationship closely, including the upcoming elections in Venezuela. At the national level, there are a number of political considerations, including concerns over security, corruption and the challenges of dealing with the drugs trade and the FARC. None of this is new, but developments on these fronts are bound to affect the perception of investors. At the local level, investors need to understand what are the requirements for developing a wind power station in Colombia, especially in Indigenous Territories, like Guajíra. This development involves obligations to study and mitigate environmental impact as well as to deal with the social impact of the investment. It is also important to understand the expectations and the perceptions of the people in the area with respect to this development and how they will benefit from it. We have commissioned a short paper that summarizes                                                                                                                62 Ambassador Orlando Sardi de Lima in a lunch seminar organized by InforPress in Madrid, March 13, 2012. See http://www.livemint.com/4BECAF7A-FE00-43CC-A88D-3900964C51BBArtVPF.pdf. Also see http://viewswire.eiu.com/index.asp?layout=VWCountryVW3&country_id=1510000151&rf=0 63 The opposition comes from different quarters, including some EU Parliamentarians. See http://www.dailymotion.com/video/xl1qfu_growing-eu-opposition-to-free-trade-with-colombia_news  

     

43    

these requirements, their associated costs and some concerns expressed by local groups in the Guajíra region64. We note that in Mexico, the central government has argued that investment in wind parks is making an important contribution to employment and to economic development in the regions where they are based.

5.3 Regulatory risk A more narrowly defined risk has to do with the stability and predictability of the regulatory regime and whether it is generally welcomes foreign investment in wind power. Generally, the Colombian regulatory system for electricity is stable. In Annex 1, we quote a report that recently reviewed the case for extending the payments of CERs to the Jepírachi plant. That report points out that the regulatory framework has not changed fundamentally at least since the original CERs were approved in 2006. It also stresses that the system is based on competitive markets, at least with respect to the payment for wholesale electricity. The development of a firm energy market is a notable change since 2006, but it has confirmed that Colombia’s electricity system prices are fundamentally market-based. Regulatory “predictability” is always a desirable goal, but investors can never be absolutely certain that regulation won’t change in a way that damages the value of their assets. Nevertheless, the greater is the perception of regulatory stability, the more confident investors will be in assessing the risks that are built into the regulation. We would note that frequent changes to the regulations are a source of concern in all countries, and that it is particularly important to have documentation that gives investors a clear and up to date picture of the entire regulatory framework. The less clear this picture is, and the more it is subject to change, the greater the perceived risk on the part of investors, especially foreign investors.

5.4 What most investors look for from policy and regulation

Investors are obviously interested in the prospects of earning a return that compensates adequately for risk. They will argue for policies that make this more likely, including guaranteed markets, revenues and subsidies if they can get them. Obviously, governments have a range of objectives and will need to decide whether they wish to “promote” particular sources of energy, or simply provide a level playing field to allow for effective competition among the alternatives. Below, we offer some information on what investors are seeking and the sorts of policies that governments have been adopting to encourage investment in wind and other nonconventional renewables. We do not advocate specific policies, but report them because some may be worth further exploration by the Colombian authorities.

                                                                                                               64 See Annex 3.

     

44    

First, investors stress the importance of a clear government plan and legislation to back it up. Of course, governments don’t always achieve their targets, but investors seek evidence about the role of new renewables in the energy system, the scale of renewable investment, the timetable and political commitment. • Where does renewable energy fit within the wider context of government policy?

Renewable energy is normally placed within the wider context of sustainable economic development, and meeting international commitments as well as national, regional and social objectives.

• Is renewable power for the integrated system or isolated areas, or both? Investors are keen to know whether the government sees wind, solar, biomass, geothermal and other non-conventional renewable energies as feeding in to the integrated system, or as a means of meeting energy needs in areas that are poorly connected to that network. Of course, the government may want renewable power to play a role in both of these ways, in which case the plan should be clear for each.

• Scale. Is the scale of investment sufficient to warrant a commitment from international

investors? Private investors in wind power appear to be willing to offer very competitive prices for wind parks in countries where there is significant potential for growth, for instance in Brazil, Peru, Mexico, India and China.

Second, investors are especially attracted by legislation that guarantees markets for specified amounts of non-conventional renewable energy. Most countries that have developed renewable power have done this through legislation that places renewable obligations on generators (Chile and China), distribution/retail companies (e.g. US, UK) or the national system operator or nationalized utility. Governments may also establish a requirement to buy certain amounts of specific types of generation from time to time, without formally defining a long-term quota. Generally investors are attracted to the clarity of government’s long-term intentions, as well as to the fact that specific actors in the system have obligations. Third, countries that have successfully developed wind power have offered revenue streams that were relatively predictable and stable. In most cases, this has involved long term contracts signed by the wind power producer with utilities (as in Brazil, Peru), or guaranteed feed-in tariffs set by the government (as in China). Wind projects are also being developed by the private sector in Mexico based on long-term contracts between industrial customers and wind developers. The central point is that the cost and the ability to finance projects are influenced by the degree of certainty of revenue streams. This is true especially in the case of wind, because most costs are fixed and recovery of those costs through spot sales is very uncertain. Fourth, private investment in wind power does not necessarily require special treatment, in other words subsidies. Two recent examples are illustrative:

     

45    

• In Mexico most private investment in wind power proceeds without special treatment or subsidies. This is possible due to relatively high electricity prices for industrial customers, who see commercial benefit to buying electricity from wind producers. The latter are only willing to invest if they have a long-term contract that offers reasonable guarantees of a risk-adjusted return on investment.

• In a recent auction, the Brazilian regulator invited bids from all technologies, and over 1000 MW of wind were selected. In order to be selected, the wind power stations offered prices (around $60-65/MWh) that were in some cases below the bids from CCGT plant.

Fifth, auctions have become the preferred method for selecting renewable projects and for determining long-term contract prices65. Auctions for long-term (option or forward) contracts are now the standard means of ensuring supply security in the electricity sector throughout most of Latin America. They have also revealed a willingness to offer wind power and other generation technologies at prices that are significantly less than expected.

• Brazil recently held a number of different auctions to select power projects and to determine their prices; some involved competition among all technologies and others were specific to wind.

• Peru also holds auctions, with prices for wind power between $65-85/MWh (against a reserve price of $110/MWh).

Sixth, investors with whom we have spoken express a growing preference for private long-term PPA contracts, rather than regulated feed in tariffs. Both feed in tariffs and PPAs provide some long-term revenue guarantees that are bankable. Recently governments have changed the terms of feed-in-tariffs in ways that are highly controversial and have led to litigation. This is the case, for instance, in the UK and in Spain. This explains the preference for contracts that are not subject to political manipulation and that are subject to the protection normally provided by commercial contracts. Seventh, investors stressed the importance of having accurate information (e.g. maps) about the geographic availability of wind (at 60 meters) and other natural resources, and about the cost and timing of gaining access to the network. We understand that the Colombian Government is undertaking detailed mapping studies. We would recommend that the government also consider how it regulates electricity transmission. In the UK, for instance, the government has introduced a new regulatory regime (RIIO) whose purpose is to provide incentives for the transmission company to invest in networks to encourage sustainable development, for instance the connection of renewable energy sources.

                                                                                                               65 See the following report for a review of auctions and other mechanisms for acquiring renewable energy in Latin America: http://www.iit.upcomillas.es/docs/IIT-11-012A.pdf

     

46    

Last, development banks and the international climate policy regime offer a number of mechanisms to share the costs of developing renewable energy. In our modelling, we have largely ignored concessionary financial terms because we wanted to stay within a more strictly technology neutral framework. But one of our sensitivities illustrated that an interest rate below 4% could make the wind project financially viable in Colombia, if the ENFICC was high enough. Concessionary finance is available for renewable energy projects and has certainly played an important role in the promotion of renewable power in other developing countries, including Brazil. The Colombian Government will decide whether and to what extent it wishes to support the development of non-conventional renewable power sources. The experience of other countries that have done this successfully in Latin America suggests the need for a convincing plan and legislation to back it up. It also suggests that auctions, accompanied with long-term contracts, are particularly useful mechanisms for promoting competition among investors.

     

47    

6. Conclusions

First, the result of the financial modelling of the project we studied is that wind power does not appear to be an immediately attractive investment in Colombia. Under the reference scenario that sets ENFICC at 6%, the financial gap is substantial (target IRR of 14% v. expected IRR of 3.5%). Increasing the ENFICC to 30% from 6% narrows the financial gap substantially (14% v. 8%), and adding CER income of $10/tCO2 reduces the gap even further (14% v. 9.4%). However, only with optimistic assumptions, or subsidised finance, do we see a prospect of eliminating the gap entirely within the current regulatory system and with our estimated electricity prices. It is important to stress that our conclusion is based on the conditions that apply to a particular power station, similar in technology to the Jepírachi station, although much larger (302 MW v. 19.9 MW). To the extent that new turbine technologies are more efficient, for instance at lower wind speeds, the financial appraisal will be more positive, provided the improved efficiency more than compensates for any additional costs. Furthermore, wind conditions may be better at other sites. We have carried out a number of sensitivities, including ones that are optimistic about the wind speed, investment costs, CER payments, interest rates and energy prices. Even then, the financial gap is large when we use the CREG’s proposed ENFICC. Our conclusion is that the regulatory treatment of the ENFICC is a key barrier in Colombia, even for wind power projects that might face better conditions than the project we have analysed. Second, as the fixed costs of wind power decline and the need for additional capacity grows, it becomes increasingly important to calculate the ENFICC properly for wind and other non-conventional renewables sources of power. This is important to ensure that the system is building the optimal mix of plant, which won’t happen if the ENFICC is measured inaccurately. For investors, an increase in the ENFICC for alternative technologies such as wind power increases the mean return on investment in wind power. In particular, it reduces the uncertainty of cash flows because the firm energy payment will be guaranteed for a period of 20 years for new plants that are successful in the firm energy auctions. Third, the calculation of capacity credit factors (i.e. like ENFICC) for wind and other non-conventional power is a recent development around the world and there is no universally accepted method. However, our recommendation is that the CREG reconsider the methodology for defining the ENFICC for wind power, and also consider using a similar (revised) methodology for other sources of non-conventional power that have potential in Colombia, including geothermal, biomass and solar energy. One well-recognized methodology for measuring a plant’s contribution to system reliability is known as the effective load carrying capability (ELCC). Alternatively, we would encourage the CREG to consider the approach used by PJM, adapted to reflect relevant peak hours in El Niño periods (as summarized in our Table 9). According to our provisional calculations, using this

     

48    

adapted PJM methodology would increase the ENFICC (for conditions at Jepírachi) from the CREG’s value of 6% to our estimate of about 30%. This would make an important contribution to project revenues and help to finance the projects at lower cost. Fourth, the CERs are important for a number of reasons. One is that CER income contributes to closing the financial gap faced by investors. Although the contribution is relatively small at today’s CER prices, there is a reasonably good prospect that the CER price will rise, mainly because the EU is determined to raise the price of CO2 emission allowances in the EU Emission Trading Scheme (ETS), which effectively drives the price of CERs. Second, project developers typically reach an agreement to sell CERs for a number of years; this makes CER income predictable and therefore helps investors to raise finance for these projects. Third, CERs are a way of recognizing the economic cost of externalities, in this case by compensating the avoidance of CO2 emissions. Although the CER is an international policy mechanism, it has domestic policy implications since eligibility under the CDM is conditional upon demonstrating that the project’s economic viability requires these additional payments. When projects are economically viable without CER payments, the Government will need to decide whether to compensate new projects (i.e. those which are not eligible for CER payments) for the avoidance of CO2 emissions. We would encourage the government to support domestic and international mechanisms that penalize the negative externalities associated with some power stations (e.g. coal power), and/or compensate power stations (e.g. wind power) that are able to avoid these externalities. Fifth, the risk assessment suggests that investors could decide that uncertainty about revenues is so great that they require a higher expected return to compensate for that risk. The risk and hence the required return will be lower if a larger share of the revenue stream can be guaranteed. This conclusion underlines the importance of both policy variables (ENFICC and CER) since they both offer revenue guarantees. Finally, we have drawn on international experience to identify some of the issues that investors will be considering, in addition to those studied in our financial analysis. Investors stress the importance of having reliable sources of information about wind speeds at 60 meters and about the potential and geographical distribution of other renewable resources in Colombia, in particular geothermal, biomass and solar. They also stress the importance of a clear policy with respect to the development of renewable energy sources and of legislation to carry it out. In addition, they emphasize the timing and cost of gaining access to transmission networks. This international experience provides background for possible future research and consideration by the Colombian Government should it decide that it wants more actively to promote investment in these new technologies. If so, auctions, accompanied with long-term contracts, are particularly useful mechanisms for promoting competition among investors.

     

49    

Annex 1: Information on Colombian Electricity System 2010-2011 Regulatory regime applying to small wind power in Colombia “The national electricity market in Colombia is ruled according to the Law 143 of 1994. The law assigns the coordinating and regulatory role to the Commission on Energy and Gas (CREG), and unbundles the provision of transmission and power generation and commercialization services. The only natural monopoly remains in the provision of transmission service, the other services being competitive. The wholesale market is administered at the National Dispatch Center, which belongs to the company in charge of transmission through the national grid.

Small power generators (less than 20 MW) have preferential access to the market, and are always first dispatched. Price is the result of demand and supply for private transactions, and of the auction for the wholesale market (safe some charges for securing the minimum generation level). Small generators supplying to the wholesale market have to use the prevailing price at that market. The central dispatch is based on the lower cost to attend daily demand, creating a merit order system. Small generators and low cost, such as Jepirachi Wind Power Project, are always first dispatching.”66

Data on capacity, operations and market variables67

                                                                                                               66 AENOR, “CDM Validation Report Renewal Of Crediting Period, International Bank For Reconstruction And Development As The Trustee Of The Prototype Carbon Fund, Validation Of The Project Activity: Jepírachi Wind Power Project”, 3 March, 2011, page 9.

67 XM (System Operator) website: http://www.xm.com.co/Pages/DescripciondelSistemaElectricoColombiano.aspx  

     

50    

     

51    

Annex 2: Colombian commitments on climate change mitigation68 Colombia

45. Colombia communicated that it is undertaking studies on its mitigation potential and on abatement cost curves for the transport, agriculture, energy, waste management and industrial sectors as part of its national strategy of low-carbon emissions development.

46. The Party communicated the following preliminary mitigation actions in three categories:

(a) Unilateral actions: Colombia will guarantee that at least 77 per cent of the total energy capacity installed by 2020 will be generated from renewable sources. These are actions that Colombia commits to undertake using its own resources, both private and public; the Party would not require any international or market-based funding;

(b) Actions with financial support:

(i) Colombia will reduce deforestation in the Colombian Amazon rainforest to zero by 2020;

(ii) Colombia will stimulate the growth of biofuel production, such as ethanol and biodiesel, without endangering the natural forests or the food security of the Colombian people, and by promoting the use of these fuels in the national market with the aim of achieving a 20 per cent share of total national fuel consumption by 2020. These are actions that Colombia is interested in undertaking and willing to do, but it lacks the necessary resources or capacity, and will therefore require financial support for their implementation;

(c) Actions related to carbon markets:

(i) Colombia supports the use of market-based mechanisms in order to contribute to GHG mitigation actions in developing countries. Colombia has made use of the existing flexibility mechanisms under the Kyoto Protocol, especially the CDM, for which Colombia has a project portfolio with an estimated annual emissions reduction potential of 17.4 Mt CO2;

(ii) Colombia has great potential to reduce emissions from deforestation (REDD) through the protection of endangered forests and the inclusion of new protected areas in the national parks programme;

(iii) Colombia has estimated a total emissions reduction of up to 54.8 Mt CO2 by 2020 through the implementation of the CDM in the energy, forest, industrial, transport and waste management sectors. Up to now, eight projects in Colombia have been accredited with

                                                                                                               68 Compilation of information on nationally appropriate mitigation actions to be implemented by Parties not included in Annex I to the Convention, UNFCCC, Note from the Secretariat, Ad Hoc Working Group on Long-term Cooperative Action under the Convention, FCCC/AWGLCA/2011/INF.1, 18 March 2011, pages 11-12.

 

 

     

52    

763,371 certified emission reduction (CER) units from activities related to wind power generation, mass transport systems and hydroelectric power generation;

(iv) Colombia will encourage commercial reforestation through the use of Forest Incentive Certificates;

(v) Colombia has great mitigation potential that could be realized through the implementation of the existing flexibility mechanisms under the Kyoto Protocol and the future mechanisms that may arise from negotiations, which would help the Party to achieve greater emission reductions by deviating from the ‘business as usual’ scenario.

47. The implementation of these mitigation actions could be supported by carbon market-based mechanisms. These actions include: (a) the capacity to measure, report and verify emission reductions for their subsequent sale; (b) the availability of measuring and monitoring tools (similar to CDM mechanisms); (c) the possibility of achieving the financial closure of the projects or activity programmes with the incentive of selling the emission reductions.

     

53    

Annex 3: Requirements to execute a wind project in Colombia, in Indigenous Territory

Requisitos para ejecutar un proyecto de energía eólica en Colombia, en un territorio indígena. Prepared by Lucía Martínez El Parque Eólico Jepírachi esta ubicado en la Península de la Guajira sobre territorio Wayuu y probablemente cualquier proyecto de energía eólica en Colombia se realizará en dicha región. Se debe tener en cuenta que para la ejecución de proyectos en territorios indígenas en Colombia hay que cumplir con ciertas reglas que protegen a dichas comunidades. Además hay que cumplir con otras regulaciones ambientales. A continuación describimos brevemente los pasos a seguir para la obtención de las licencias requeridas y el proceso para la ejecución del proyecto. La información que sigue fue extraida del documento de Empresas Publicas de Medellin EPM, Parque Eólico Jepirachi, informe final.69 1. Proceso de Gestión Ambiental. En Colombia rigen el decreto 1753 de 1994 del Ministerio del Medio Ambiente sobre Licencias Ambientales. EPM previamente a la ejecución del proyecto realizó los siguientes trámites.

• Obtención de permiso de estudio de recursos naturales: Este permiso le impuso a

EPM la obligatoriedad de suministrar información de los resultados de los estudios tanto de vientos como de la gestión social (Resolución # 00562 del 6 de Abril de 2000 y resolución # 002001 del 28 de septiembre de 2000).

• Obtención de permisos ambientales para la localización de tres estaciones climatológicas en el área de influencia del Proyecto por parte de CORPOGUAJIRA (Resoluciones #00562 del 6 de Abril de 2000, Resolución # 2271 del 8 de noviembre de 2000).

• Concepto del Ministerio del Medio Ambiente sobre requerimiento de Licencia Ambiental para el parque eólico (Diciembre 27 de 2001). Aclaración a Ministerio del Medio Ambiente sobre autoridad ambiental competente para el otorgamiento de Licencia Ambiental (Febrero 1 de 2002).

• Consulta a Ministerio de Minas y Energía sobre la aplicabilidad del artículo 45 de la Ley 99 sobre transferencias del sector eléctrico a Municipios y Corporaciones Autónomas Regionales (CAR) por generación de energía de proyectos eólicos (Febrero de 2002). Respuesta del Ministerio de Minas y Energía sobre la no aplicación del régimen de transferencias para los proyectos eólicos (Marzo de 2002).

• Presentación de propuesta de Términos de Referencia para Estudio de Impacto

                                                                                                               69 Empresas Publicas de Medellín, Parque Eólico Piloto Jepirachi Estudio de Impacto Ambiental Informe Final Medellín, Junio de 2002.

     

54    

Ambiental (EIA) del Proyecto de Aprovechamiento Eólico a CORPOGUAJIRA (febrero 18 de 2002).

• Solicitud ante CORPOGUAJIRA de inicio de trámite para obtención de Licencia Ambiental para el Proyecto de aprovechamiento eólico (febrero 18 de 2002). Solicitud a Ministerio del Interior de certificación de existencia de comunidades indígenas y negras en la zona del proyecto (mayo 31 de 2002).

• Solicitud a CORPOGUAJIRA de realización de reunión de protocolización de Consulta previa (junio 2002).

Según EPM los gastos de la gestión ambiental fueron 1.606.580.000 de pesos (de 2002 - aproximadamente $845,000 US) en total, que corresponden al 3,32 % de los costos totales del proyecto. Los costos fueron:70

ITEMS COSTO DE 2002 (COL $) Plan de manejo ambiental 1.233.545.000 Plan de monitoreo y seguimiento 223.035.000 Plan de contingencias 150.000.000 TOTAL 1.606.580.000

2.Proceso de Gestión Social. Según el decreto 1320 de 1998 para minorías étnicas, de la Corte Constitucional de Colombia 71 , EPM debía realizar una socialización con las comunidades ya que la zona del proyecto está ubicada en territorio indígena, se realizó un estudio de estrategia social, dicha gestión social esta articulada con los estudios técnicos y ambientales. 72 Todo el proceso Social fue dirigido a las comunidades, autoridades municipales, ambientales e institucionales.

EPM en el proceso de gestión mediante reuniones informo, consulto y concertó con los diferentes entes territoriales, acerca de los estudios, alcances, tiempo de ejecución del proyecto, actividades de campo, avances, políticas ambientales de EPM y decisiones sobre el proyecto. Este proceso se realizó entre 1999 y 2002. Para un mayor entendimiento con las comunidades indígenas Wayuu se contrato los servicios de un traductor ya que no es la lengua oficial de dicha comunidad, de estas reuniones se elaboraron actas firmadas por los asistentes. Entre noviembre de 2001 y abril de 2002, se presento a todos los entes interesados las características del proyecto y se consulto sobre los impactos a las comunidades.

                                                                                                               70 Todos los costos están en pesos Colombianos, a menos que dice explícitamente que están en US$.

71 Convenio No 169 de 1929 sobre Pueblos Indígenas y Tribales en Países Independientes de Organización Internacional del Trabajo, Colombia lo incluyo en su Constitución en el año 1998.

72 http://www.presidencia.gov.co/prensa_new/decretoslinea/1998/julio/13/dec1320131998.pdf

 

     

55    

Se requirió de los servicios de varios profesionales del área social los costos son los siguientes:

Presupuesto de recursos técnicos, físicos, humanos y económicos: ITEMS UNIDAD COSTOS (COL $) Comunicador social 2.500.000 11.250.000 Antropólogo 2.500.000 11.250.000 Tecnólogo 1.500.000 6.750.000 Traductor Global 1.500.000 Recurso logístico Material Impreso Global 5.000.000 Talleres-visitas- guiadas- Global 15.000.000 Reuniones Glbl 1.000 TOTAL 60.750.000 Mediante los talleres de socialización se evidencio las necesidades y expectativas de dichas comunidades que esperaban la solución a varios problemas, como: • Suministro de agua (plantas desalinizadoras, jagüeyes, carrotanque) • Mejoramiento de los servicios de educación y la salud a través de la dotación de

instalaciones • Empleo y capacitación de personal • Alimento para los animales • Energía eléctrica

Durante la etapa de construcción del proyecto, se dictaron talleres de Educación Ambiental, de Participación y Fortalecimiento Comunitario, Programa de Generación de Empleo y un Programa de Implementación y Seguimiento a Medidas Compensatorias. 2.1 Medidas Compensatorias. En el programa de implementación y seguimiento a medidas compensatorias se acordó con las comunidades lo siguiente, con un costo total de 350.000.000: • Proyecto Construcción de una Planta Desalinizadora: Costo 150.000.000, por parte de

Las Empresas Públicas de Medellín E.S.P., para beneficiar a las comunidades de Kasiwolin, Arutkajui en el área de influencia del proyecto.

• Escuela, operación y mantenimiento lo asumido por el Municipio de Uribía y las comunidades, ampliación y dotación Escuela kamusuchiwo'w y dotación puesto de salud de Media Luna: Costo $120.000.000.

• Construcción de dos jagüeyes en los territorios Arutkajui, Media Luna. Limpieza y adecuación de dos jagüeyes en kasiwolin (los jagüeyes son depósitos artificiales de aguas, construidos mediante excavación) Costo $80.000.000.

• Cercar el cementerio de Arutkajui con la participación de la comunidad. y cerramiento de cementerio.

     

56    

ITEMS COSTO (COL $) Planta desalinizadora 150.000.000 Escuela y Puesto de Salud 120.000.000 Jagüeyes 80.000.000 TOTAL 350.000.000 Entonces el costo total de ese proceso era 2.017.330.000 pesos colombianos, o aproximadamente $1.061.753 (US) y ha tomado aproximadamente dos/tres años. 2.2 El proceso de socialización durante la construcción del proyecto Parque Eólico Jepírachi, tuvo varia etapas que se resumieron en programas dirigidos a las comunidades wayuu que habitan la zona del proyecto, las comunidades son Arutkajui, kasiwolin y Media Luna. Resumen de costos por Programa Programa Costo (COL $) Programa de Información y Comunicación 60.750.000 Programa de Generación de Empleo 20.125.000 Programa de Educación Ambiental 40.125.000 Programa de Participación y Fortalecimiento Comunitario 50.375.000 Programa de Información y Entrenamiento para empleados 17.000.000 Programa de Compensación 350.000.000 Programa de Divulgación Tecnológica 53.875.000 Programa de Establecimiento de Servidumbres 50.000.000 Programa de Monitoreo y Rescate Arqueológico 84.435.000 TOTAL 726.685.000

3.Resultado/Problemas “Informe de la Expedición Energética a la Guajira”73 publicado en 2005

Las comunidades al día de hoy [2005] se muestran inconformes, debido al poco conocimiento y entendimiento real de el significado de un proyecto de está envergadura en su territorio; sienten que no supieron negociar y no entienden como teniendo un parque eólico que produce energía ellos aun continúan su vida sin este servicio. Además la Alcaldía de Uribía prometió hacerse cargo del mantenimiento de la planta desalinizadora, pero dicha planta se encuentra fuera de servicio.

Todo esto se puso en evidencia en la expedición realizada por Censat Agua Viva y que estuvo acompañada por integrantes de diferentes comunidades indígenas y de ONGs, los cuales realizaron entrevistas a los miembros de las comunidades cercanas al Parque Eólico Jepírachi; esta expedición dio como resultado un informe.

                                                                                                               73 http://www.censat.org/ http://www.odg.cat/documents/enprofunditat/Transnacionals_espanyoles/Informe_Guajira.pdf http: //www.earthpeoples.org/blog/?page_id=296

 

     

57    

A principios de 2011 EPM y la Alcaldía de Uribía pusieron en funcionamiento la Planta Desalinizadora que EPM había construido como medida compensatoria para las comunidades del área de influencia del Parque Eólico Jepirachi; dicha planta había fallado debido a el daño de algunas piezas y a las inclemencias del clima. Pasaron 8 años para el cumplimiento de ese compromiso con las comunidades.74

4.Políticas de Banco Mundial y Banco Interamericano de Desarrollo sobre desplazamiento

“Anexo D: El Reasentamiento Involuntario de Pueblos Indígenas (ver el párrafo 3b del informe del Banco Mundial75). En los proyectos que implican la restricción involuntaria del acceso a parques y áreas protegidas y designados legalmente, la naturaleza de las restricciones, así como el tipo de medidas necesarias para mitigar los impactos adversos, serán determinadas durante el diseño y la ejecución del proyecto con la participación de las personas desplazadas. En tales casos, el Prestatario deberá preparar un cronograma del proceso, aceptable para el Banco, en el que se describa el proceso de participación mediante el cual:

(a) se preparen y ejecuten los componentes específicos del proyecto; (b) se definan los criterios de elegibilidad de las personas desplazadas; (c) se identifiquen las medidas de asistencia para las personas desplazadas en su esfuerzo por mejorar sus vidas, o al menos restaurarlas, en términos reales, al tiempo que se mantiene la sostenibilidad del parque o área protegida; y (d) se resuelvan los conflictos potenciales que involucren a las personas desplazadas.

En el marco del proceso también se deberá incluir una descripción de los arreglos acordados para ejecutar y supervisar el proceso”.

                                                                                                               74http://www.elmundo.com/portal/noticias/servicios_publicos/epm_les_llevo_agua_potable_a_los_vecinos_de_jepirachi.php http://www.elcolombiano.com/BancoConocimiento/L/los_wayuu_calman_su_sed/los_wayuu_calman_su_sed.asp

75 http://siteresources.worldbank.org/INTINDPEOPLE/922146-1112796444814/20454193/SumExtConsult-100802-SP.pdf

 

   

Annex 4: Quantil Model

FINAL REPORT Wind Energy Financial Model PRESENTED TO Oxford Institute for Energy Studies July 30, 2012

 

     

2      

Index  1.   Introduction  ............................................................................................................................................  3  2.   Financial Model  ....................................................................................................................................  3  3.   Preliminary Results: Benchmark  .....................................................................................................  5  4.   Risk Assessment  .................................................................................................................................  34  5.   Wind, El Niño and Complementarities  .......................................................................................  36  6.   References  .............................................................................................................................................  49  Appendix A  ....................................................................................................................................................  50  Appendix B  .....................................................................................................................................................  53  Appendix C  .....................................................................................................................................................  54  

     

3      

1. Introduction This report puts together the main elements of the financial model for a wind farm. These are:

A. Parameters.

B. Risk factors.

C. Data, model and simulation of risk factors.

D. Benchmark financial model.

E. Equity IRR and NPV.

F. Empirical equity IRR and WACC.

The main objective is to have a tool to address the implications, relative to the benchmark financial scenario, of different policy proposals to make financially feasible a wind energy project in Colombia.

2. Financial Model We would like to find the internal rate of return (IRR) of wind project. This is the interest rate that makes the net present value of future net cash flows equal to zero. In other words if

!∗ = !"" then  0 = !"!!!!!∗ !

!

!!!, where !"!! is the net cash flow at time t and !"!! is the

initial investment. In general, the net cash flow will have two or three components depending on what scenario we are considering. The first one is the operational cash flow which depends on how much energy is sold in the spot market. The other two, when they apply, are the cash flows associated with financial obligations or CDMs. In the first case these are interest rate payments plus capital amortizations and, in the second case, these are positive inflows that compensate wind generation because of carbon reductions. In any period, if the spot price is under the scarcity price, the operational cash flow is equal to the reliability payment (i.e. firm energy payment) plus the earnings for selling energy in the spot market. The reliability payment is equal to !!"#$!!"#$. Where !!"#$ is the amount of firm energy and !!"#$ it’s the “cargo por confiabilidad”, or the amount of money paid for firm energy. If the market price is under the scarcity price then earnings area equal to:

!!"#$!!"#$ + !! !! − !"#" − !"#$% + !"# − !" − !", where !! is the amount of energy produced, !! is the spot price, CERE is the payment that must be done to fund the reliability payments. The CERE is paid by every kWh of generated energy, unless the generators capacity is under 20 MW, in which case it does not have to pay this charge. The total amount of money collected by CERE payments must be equal to the total amount of money compensated for reliability payments. FAZNI is a tax to finance investment projects in none connected areas. CDM stands for Clean Development

     

4      

Mechanism and is a compensation given to the production of energy avoiding the emission of CO2; it is measured in dollars per avoided ton of CO2. VC is the variable or unit cost of every kWh generated by the wind farm and OC is the operational or fixed cost. If the market price is over the scarcity price then it is mandatory for generators to sell their firm energy at the scarcity price. If a generator produces more energy than its firm energy, it can sell the surplus at the spot price; since the firm energy auctions are designed to match the amount of firm energy to the expected demand, it’s reasonable to think that only a proportion X of this surplus is actually sold. However, if less energy is produced than its firm energy, the generator has to buy this energy at the spot price from someone else. The period net cash flows when the spot price exceeds the scarcity price is equal to: min  (!!"#$ ,!!) !!"#$"%&' − !"#" − !"#$% + !"# − !" − !!max(0,!!"#$ − !!)

+ !max(!! − !!"#$ , 0) !! − !"#" − !"#$% + !"# − !" − !"

From these equations it follows that the net cash flow depends on:

1. The amount of energy produced, which depends on the wind speed and the power curve.

2. The spot price. 3. The scarcity price, which depends on fuel prices and the exchange rate. 4. Investment per KW. 5. The price of firm energy (cargo por confiabilidad). 6. Variable costs. 7. Operational costs. 8. Taxes (FAZNI, CERE) and compensations (CDM). 9. Firm energy. 10. Exchange rate. 11. The amount of extra energy that is sold when the spot price is over the scarcity price

(X).

We split these variables between parameters, those that we will not consider as risk factors, and risk factors. Risk factors are going to be modeled in order to quantify the uncertainty associated with the internal rate of return (IRR), while parameters are going to be fixed (although parameters can vary along time, we assume they are deterministic). We make the following classification of variables. Parameters:

1. Policy Instruments:

a. Amount of firm energy (ENFICC).

b. Taxes (CERE, FAZNI).

c. Compensations (CDM-EMR).

     

5      

2. Non-Policy Instruments:

a. The power curve.

b. Investment per KW

c. Scarcity price.

d. Firm energy price (cargo por confiabilidad).

e. Variable costs.

f. Operational costs.

g. Energy in excess of firma energy sold in the spot market when.

h. Exchange rate (firm energy, scarcity price and CDM are denominated in dollars).

Risk factors:

1. Spot price.

2. Wind speed.

El Niño is an underlying risk factor since it affects the spot price and the wind speed.

3. Preliminary Results: Benchmark

A. World Bank Parametrization We consider a benchmark calibration similar to the World Bank (WB) study on wind energy for Colombia. Consider a 302MW (230 Nordex N60 mills) wind farm. For these exercises we fix the exchange rate at 1,900 (pesos/USD) and an average price of 96.14 pesos per kWh (similar to the 50 USD/MWh of the WB). An investment cost of $ 3,420,000 pesos per kW (similar to the 1,800 USD per Kwh of the WB study). A FAZNI charge of 1,000 pesos/MWh and a CERE of 20,000 pesos/MWh are used. The firm energy payment is set to 28,000 pesos/MWh. The time horizon is set to 25 years. The wind speed is set to 8.75 m/s (which is similar to the average wind speed in our simulations). For this exercises we fixed the scarcity price in 400 pesos per kWh.76 We also assume that the variable and operational costs are equal to zero, and that the percentage of extra energy produced, besides from firm energy that can be sold at spot price, when spot price is above scarcity price is 10%. With these parameters and assuming no firm energy and that the investment is financed without a loan, the IRR is 5.3%. Figure 1 and Table 2 show the sensitivity of the IRR to the amount of firm energy assumed.

                                                                                                               76 This calibration needs to be improved. There is no reason to assume that in the long run the scarcity price is unrelated to the spot price.

     

6      

Parameter Value

Exchange rate (pesos/USD) 1,900 Spot Price (pesos/KWh) 96.14 Installation costs (pesos/KW) 3,420,000 FAZNI (pesos/MWh) 1,000 CERE (pesos/MWh) 20,000 Firm energy Payment (pesos/MWh) 28,000 Wind speed (m/s) 8.75 Time horizon (years) 25 Table 1: World Bank study parametrization.

Figure 1: IRR sensibility to the amount of firm energy.

Firm Energy IRR (WB parametrization) 0% 5.38% 5% 5.90% 6% 6.00% 10% 6.41% 15% 6.92% 20% 7.41% 25% 7.89% 30% 8.43% 35% 8.92%

Table 2: IRR sensibility to the amount of firm energy.

     

7      

B. Equity IRR for different debt levels

If we assume that a proportion of the initial investment is financed with a loan that matures in fifteen years with a 9% annual interest rate and monthly payments then Figure 2 shows the Equity IRR at different firm energy levels. As can be seen, the bigger the proportion of initial investment financed with a loan, the steeper the curve is.

Figure 2: Equity - IRR sensibility to the amount of firm energy and the proportion of initial investment financed with a loan (at 9% annual interest rate). A typical cash flow is shown in Figure 3. These figures show the cumulative cash flow for a project where 50% of the initial investment is financed with a loan that matures in 15 years, has a 9% annual interest rate and has monthly payments. As can be seen, there is an initial negative flow (the initial investment) and then there curve slowly raises, until the debt is completely paid and then the rate at which the curve rises increases.

     

8      

Figure 3: Cumulative cash flow for a proportion of the initial investment financed with a loan of 50% and a firm energy level of 6%. The red line indicates the month when the loan is completely paid off.

C. Project Value for different debt levels

Figure 4 shows the net present value (discount rate is return to equity) for a 302 MW plant, and assuming the same values for all the parameters as before. The amount of firm energy is set at 6% of the capacity. Figure 5 shows how the net present value changes for different levels of debt. As can be seen de net present value of the project is the same whenever the discount rate is equal to the interest rate for the debt. Figure 6 and Figure 7 show the net present value for different firm energy levels. As expected, the greater the amount of firm energy, the higher the net present value.

     

9      

Figure 4: Net present value assuming the initial investment is financed without debt and 6% of firm energy.

     

10    

Figure 5: NPV assuming 6% of firm energy, for different debt levels.

     

11    

Figure 6: Net present value assuming the initial investment is financed without debt for firm energy levels.

     

12    

Figure 7: Net present value for firm energy levels. 50% of the initial investment is financed with a loan that matures in 15 years, has an annual interest rate of 9% and monthly payments.

D. Effect of Clean Development Mechanism on NPV and IRR

In this section we assess the NPV and the IRR of the project when a clean carbon compensation is given to the energy producers. In this case we assume, according to UPME’s (Unidad de Planeación Minero Energética) regulations, that wind produced energy saves 0.35 tons of CO2 per MWh and that for each non-emitted ton of CO2 the producers receive a US$10.00 subsidy. According to these parameters we obtain:

     

13    

Figure 8: Net present value when the Wind Mill Project receives US$10 per non-emitted CO2 ton.

.

Figure 9: Internal Rate of Return when the Wind Mill Project receives US$10 per non-emitted CO2 ton. As can be seen in the two previous graphics, both measures (IRR and NPV) improve, as expected, with the compensation for avoided CO2.

     

14    

E. Effect of CERE and FAZNI taxes on Equity IRR

In order to find out how the financial gap can be filled, the effect of taxes, established by the Colombian Government, must be understood. In particular, we study the effects of FAZNI and CERE. We focus on calculations of Equity IRR for a 302 MW wind farm (see Appendix for a 19.6 MW wind farm). Table 3 and Figure 10, show how a 302 MW wind farm IRR is affected by changes in the CERE, leaving the FAZNI at $ 1,000 COP. Table 4 and Figure 11 show how a 302 MW plant IRR (these scenarios assume that the initial investment is not financed with debt, therefore IRR is the same as Equity IRR) is affected by changes in the FAZNI, leaving the CERE at $ 20,000 COP. If we assume that the initial investment is financed 50% with a loan that matures in 15 years with an annual interest rate of 9% and monthly payments, then the IRR improves and the effect of removing is bigger, as shown in Figure 12 and Figure 13. Appendix A shows the results for a 19.6 MW wind farm.

CERE\Firm Energy 0% 6% 12% 20% 30% 36% $ 0 COP 8.13% 8.72% 9.30% 10.06% 11.01% 11.49% $ 5,000.00 COP 7.43% 8.09% 8.65% 9.42% 10.37% 10.94% $ 10,000.00 COP 6.76% 7.36% 7.94% 8.77% 9.73% 10.30% $ 15,000.00 COP 6.08% 6.69% 7.28% 8.14% 9.08% 9.66% $ 20,000.00 COP 5.38% 6.01% 6.62% 7.41% 8.43% 9.01% $ 25,000.00 COP 4.66% 5.30% 5.93% 6.74% 7.73% 8.36% $ 30,000.00 COP 3.89% 4.58% 5.23% 6.06% 7.07% 7.65%

Table 3: IRR according to different CERE and firm energy levels (FAZNI fixed at $1,000 COP). Initial investment financed without a loan.

     

15    

Figure 10: IRR according for different CERE and firm energy levels (FAZNI fixed at $1,000 COP). Initial investment financed without a loan.

FAZNI\Firm Energy 0% 6% 12% 20% 30% 36% $ 0 COP 5.52% 6.14% 6.75% 7.54% 8.56% 9.14% $ 500.00 COP 5.45% 6.08% 6.68% 7.47% 8.49% 9.08% $ 1,000.00 COP 5.38% 6.01% 6.62% 7.41% 8.43% 9.01% $ 1,500.00 COP 5.31% 5.94% 6.55% 7.34% 8.36% 8.95% $ 2,000.00 COP 5.24% 5.87% 6.48% 7.28% 8.29% 8.88%

Table 4: IRR according for different FAZNI and firm energy levels (CERE fixed at $20,000 COP). Initial investment financed without a loan.

     

16    

Figure 11: Equity IRR according for different FAZNI and firm energy levels (CERE fixed at $20,000 COP). Initial investment financed without a loan.

     

17    

Figure 12: Equity IRR according for different FAZNI and firm energy levels (CERE fixed at $20,000 COP). 50% of the initial investment is financed with a loan that matures in 15 years, has an annual interest rate of 9% and monthly payments.

     

18    

Figure 13: Equity IRR according to different CERE and firm energy levels (FAZNI fixed at $1,000 COP). 50% of the initial investment is financed with a loan that matures in 15 years, has an annual interest rate of 9% and monthly payments.

F. The effect of CERE and FAZNI taxes on the net present value.

Figure 14, Figure 15 and Figure 16 show how changes in the FAZNI and CERE tax affect the net present value. As in the case of the IRR, changes in the FAZNI have a relative small effect since it is a relatively small tax, on the other hand changes in the CERE have a bigger impact on the net present value.

     

19    

Figure 14: NPV for different FAZNI levels. Initial investment financed without a loan. Assuming 6% of firm energy.

     

20    

Figure 15: NPV for different FAZNI levels. 50% of the initial investment is financed with a loan that matures in 15 years, has an annual interest rate of 9% and monthly payments. Assuming 6% of firm energy.

     

21    

Figure 16: NPV for different CERE levels. Initial investment financed without a loan. Assuming 6% of firm energy.

G. IRR sensitivity to different parameters.

Figure 17, Figure 18 and Figure 19 show how the IRR is affected by moving the wind speed, the spot price and the “cargo por confiabilidad”. As expected the effect on the IRR of changes in the “cargo por confiabilidad” increases as the amount of firm energy increases. Figure 20, Figure 21 and Figure 22 show the effect on the IRR if the initial investment is financed with a loan. The results show that the IRR seems to respond more to changes in the parameters when the initial investment is financed by loans. This means that favorable scenarios would lead to higher IRR in this scenario, compared to the case where no loan is made, but also than adverse scenarios will lead to a lower IRR.

     

22    

Figure 17: IRR according to different wind speed and firm energy levels. Initial investment financed without a loan.

     

23    

Figure 18: IRR according to different spot prices and firm energy levels. Initial investment financed without a loan.

     

24    

Figure 19: IRR according to different levels of “Cargo de Confiabilidad” and firm energy levels. Initial investment financed without a loan.

     

25    

Figure 20: Equity IRR according to different wind speed and firm energy levels. 50% of the initial investment is financed with a loan that matures in 15 years, has an annual interest rate of 9% and monthly payments.

     

26    

Figure 21: Equity IRR according to different spot prices and firm energy levels. 50% of the initial investment is financed with a loan that matures in 15 years, has an annual interest rate of 9% and monthly payments.

     

27    

Figure 22: Equity IRR according to different “Cargo de Confiabilidad” and firm energy levels. 50% of the initial investment is financed with a loan that matures in 15 years, has an annual interest rate of 9% and monthly payments.

     

28    

Figure 23: Equity IRR according to different percentages of the benchmark investment cost (BIC) level. 50% of the initial investment is financed with a loan that matures in 15 years, has an annual interest rate of 9% and monthly payments.

H. NPV sensitivity to different parameters.

The following figures show how the net present value changes when the wind speed, the spot price or the Cargo por Confiabilidad changes. As can be seen the Cargo por Confiabilidad has a small effect which is due to the fact that the amount of firm energy is set at 6%. Also, in order to have a viable project the wind speed must be over 8% as longs as the amount of firm energy is set at 6%.

     

29    

Figure 24: NPV according to different wind speed and discount rate levels. Initial investment financed without debt.

     

30    

Figure 25 NPV according to different spot price levels. Initial investment financed without debt.

     

31    

Figure 26 NPV according to different Cargo de Confiabilidad levels. Initial investment financed without debt.

     

32    

Figure 27 NPV according to different wind speed and firm energy levels. 50% of the initial investment is financed

with a loan that matures in 15 years, has an annual interest rate of 9% and monthly payments.

     

33    

Figure 28 NPV according to different spot prices. 50% of the initial investment is financed with a loan that matures

in 15 years, has an annual interest rate of 9% and monthly payments.

     

34    

Figure 29 NPV according to Cargo de Confiabilidad levels. 50% of the initial investment is financed with a loan that

matures in 15 years, has an annual interest rate of 9% and monthly payments.

4. Risk Assessment

We now address the potential riskiness of investing in wind power by simulating the two main risk factors of our model (wind speed and spot price) and its implication on equity IRR. We also do sensitivity analysis to firm energy and other parameters of our model. For doing this we consider to independent models. The basic model (benchmark simulation) relies on:

1. A time series model of wind speed describe in the following section. 2. A basic Orsntein - Uhlenbeck mean reverting model of the spot price (see next

section). Note that, for mean for the benchmark simulation exercise (subsection A), the mean of spot price and wind speed simulations are equal to the benchmark calibration of previous sections. Therefore the differences that we will observe in terms of mean IRR, to some extent, are explained by nonlinearities in the cash flow and present value formulas. Other than that, it is the exercise of the firm energy option when the spot price is over the scarcity price that plays a relevant role through to alternative channels: when the option is exercised and there is enough wind speed to generate firm energy, then we assume only a percentage of excess energy (over firm energy) is sold in the spot price. When the option is exercised and there is not enough wind speed to generate firm’s energy, then the generator has to pay a penalty.

     

35    

To check how sensitive is financial models results to simulations of risk factors, we consider a second model of the spot price. This model is a more sophisticated model (alternative model) that takes into account energy demand, lag prices, anomaly, and other variables to model the spot price of energy (see section below). Hence, our second set of simulations is based on:

1. A time series model of wind speed describe in the following section (the same model as in the benchmark case).

2. An alternative more sophisticated model of the spot price (see next section). Subsection B shows results for the alternative simulating model. In this case the main driving force that explains the differences with the benchmark case is the following. The alternative model forecasts an average price of $122 pesos/KWh (much higher than the benchmark model), but it forecasts a low average price the first five years ($77.6 pesos/KWh) and higher after that ($133 pesos KWh).

A. Benchmark Simulation Analysis

Firm Energy CDM Mean IRR 5% Quantile

Probability of IRR 14% or more

0 0 2.7% -0.3% 0.1%

0.06 0 3.9% 0.8% 0.1%

0.15 0 5.7% 2.5% 0.3%

0.2 0 6.7% 3.5% 0.9%

0.3 0 8.7% 5.3% 2.7%

0.36 0 9.9% 6.4% 6.9%

0 10 4.6% 1.4% 0.2%

0.06 10 5.8% 2.5% 0.3%

0.15 10 7.6% 4.2% 1.5%

0.2 10 8.6% 5.1% 2.7%

0.3 10 10.7% 7.0% 11.0%

0.36 10 11.9% 8.0% 20.2%

0 30 8.4% 4.8% 2.6%

0.06 30 9.6% 5.9% 6.1%

0.15 30 11.6% 7.6% 17.1%

0.2 30 12.7% 8.5% 28.0%

0.3 30 14.9% 10.4% 54.5%

0.36 30 16.3% 11.5% 72.5%

0 50 12.5% 8.2% 27.0%

0.06 50 13.9% 9.4% 42.0%

0.15 50 15.9% 11.0% 68.0%

0.2 50 17.1% 12.0% 80.6%

0.3 50 19.6% 14.0% 94.9%

0.36 50 21.1% 15.3% 98.3% Table 5: IRR histogram mean and quantiles for different firm energy levels and CDM price. All other variables fixed at their benchmark values. Investment financed with 70% debt. Basic model.

     

36    

B. Alternative model

Firm Energy CDM Mean IRR 5% Quantile

Probability of IRR 14% or more

0 0 -7.3% -64% 14% 0.06 0 -3.5% -26% 16% 0.15 0 0.1% -18% 20% 0.2 0 1.7% -16% 21% 0.3 0 3.9% -12% 23%

0.36 0 5.0% -10% 25% 0 10 -5.9% -54% 17%

0.06 10 -1.6% -25% 19% 0.15 10 1.8% -18% 22% 0.2 10 3.3% -15% 23% 0.3 10 5.5% -12% 26%

0.36 10 6.6% -10% 28% 0 30 0.8% -27% 21%

0.06 30 3.3% -20% 24% 0.15 30 6.0% -14% 28% 0.2 30 7.3% -12% 31% 0.3 30 9.4% -8% 36%

0.36 30 10.7% -6% 37% 0 50 6.4% -13% 29%

0.06 50 8.1% -11% 32% 0.15 50 10.5% -7% 38% 0.2 50 11.7% -6% 40% 0.3 50 13.8% -3% 43%

0.36 50 15.1% -2% 46%

Table 6: IRR histogram mean and quantiles for different firm energy levels and CDM price. All other variables fixed at their benchmark values. Investment financed with 70% debt. Alternative model.

5. Wind, El Niño and Complementarities The amount of electricity produced by a wind turbine depends primarily on the wind speed,77 since it determines the total kinetic energy that can be converted into mechanical energy. In order to assess the amount of firm energy that a wind farm could bring to the system one should study how the speed behaves historically and how, if so, it is related to dry periods, as those associated to El Niño.

                                                                                                               77 Another important factor is air density. The power curve below is valid for standard air density. Air density in la Guajira is presumably lower than the standard. This would require a correction.

     

37    

A. Firm Energy Figure 30 shows the amount of energy generated at Jepirachi as a percentage of the available capacity. The red line represent 6%, which is the amount of firm energy the CREG set for wind farms. As can be seen, Jepirachi reached, or almost reached, this level three times. However, this is taking the entire history of Jepirachi. If we look at El Niño and La Niña events (Figure 31 and Figure 32) , there seems to be correlation between low energy generation and La Niña, and high energy generation and El Niño. If we do not take into account the La Niña years there is only one case where the generation is near 6% (Figure 33 ). If we do not take that month into account the month with least amount of generation as a proportion of the capacity is 13.11%, indicating that the amount of firm energy should be around this number.

Figure 30: Amount of energy generated as a percentage of the capacity at Jepirachi. The red line represents 6%.

     

38    

Figure 31: Generation vs. Actual Capacity. The shaded areas represent La Niña events.

Figure 32: Generation vs. Actual Capacity. The shaded areas represent El Niño events.

     

39    

Figure 33: Generation vs. capacity at Jepirachi without taking into account La Niña years.

B. Wind Model All public meteorological stations in Colombia measure wind speed at a height of 10 meters. But wind mills performance depends on wind speed at 60 meters. Using data from a meteorological station at Puerto Bolivar, close by where the EPM’s wind farm, Jepirachi, is located, we will infer the relationship there is between wind speed at 10 meters and at 60 meters using the historical energy output of Jepirachi and the power curve of the turbines used there, the Nordex N60. The power curve tells us the energy output for a particular wind mill according different wind speeds; Figure 35 shows the Nordex N60 power curve. Figure 33 shows the historical wind speed at Puerto Bolivar (10 meters).

     

40    

Figure 33: Wind speed at Puerto Bolivar (10 meters).

     

41    

Using information on the historical generation of energy at Jepirachi (Figure 36) and the power curve, we can estimate what should have been wind speed at Jepirachi at 60 meters. However, the record for energy generation is relatively short (January 2004 - July 2011) and it would be useful assemble a longer record. In order to do this, the relationship between the wind speed at a height of 10 meters in Puerto Bolivar and at 60 meters in Jepirachi is established by means of OLS, assuming a linear relationship: !!" = ! + !!!", where !!" it’s the wind speed at a height of 60 meters in Jepirachi (estimated using the power curve and historical energy generation at Jepirachi as describe above), and !!" is wind speed at a height of 10 meters in Puerto Bolivar. Using OLS we find that  ! = 3.20 and ! = 0.84 (both coefficients are highly significant with an R – squared of 0.58). Finally, the estimated wind speed at 60 meters is found.

     

42    

Figure 35: Nordex N60 power curve (Source: Nordex).

Figure 36: Jepirachi historical generation of energy (Source: XM)

With the longer assembled time series for wind speed at 60 meters, better statistical models to simulate future wind speeds can be estimated. Since wind is a key risk factor in the financial model, it is essential to have simulations that reflect historical wind behavior. To specify the statistical model we first separate the wind speed time series in seasonal and non-seasonal components. First, the seasonal part is assumed to be a linear composition of

     

43    

sinusoids varying at fundamental frequencies.78 In order to get the fundamental frequencies, a fast Fourier transform is done which enable us to construct the periodograms shown in Figure 37. The peak shows us where the fundamental frequencies are. Then a linear regression is fitted using sinusoids with the selected frequencies. The fitted values of this model are the seasonal component of wind speed. The actual wind speed minus the seasonal component is the non seasonal part. We model the non seasonal component with an ARMA(3,2). The final forecasts are estimated as the sum of the seasonal forecast and the non seasonal one. In order to simulate future wind speed, random shocks are applied to the MA component of the non seasonal component. Figure 38 shows the result of simulating the wind speed fifteen years ahead.

                                                                                                               78 Fourier decomposition.

     

44    

Figure 37: Wind speed periodogram . The figure on the right takes out the first frequency.

Figure 38: Wind speed simulation 15 years ahead.

         

     

45    

C. Complementarities To assess the complementary between hydro and wind energy we would like to explore the relationship between El Niño and wind speed. According to NOAA, El Niño (La Niña) is a weather phenomena that occurs when for at least five consecutive months, the three month running average mean of the sea surface temperature anomaly in the region between 5°N-5°S and 150°W-90°W is above (below) 0.5 (-0.5). The correlation between the sea surface temperature anomaly and the wind speed is 0.274 (p-value of: 1.2!10!!), indicating that there is indeed a significant correlation between dry periods and higher wind speeds. Figure 39 shows the relationship between anomaly and wind speed.

Figure 39: Wind speed and Anomaly

D. Spot and scarcity prices The historical data, in Figure 40, shows the market price and scarcity price. These two factors are crucial for the financial feasibility of a wind energy project. The market price tells us how much are the generators going to earn for the energy produced, and when the market price is above the scarcity price, the firm energy options are triggered and generators are required to sell their firm energy at the scarcity price.

     

46    

Figure 40: Market Price and Scarcity Price. Soruce: XM.

To forecast the market price we use two basic models. Forecasting Model 1: Ornstein Uhlenbeck This is a standard Ornstein Uhlenbeck type of model estimated on the log spot price. This model is our basic model and is suggested as the simplest type of model that is able to capture core uncertainty in the spot market. Figure 41 shows 1000 simulated paths for this model (prices are in July 2011 constant values).

     

47    

Figure 41: Simulated spot prices. Basic model.

Forecasting Model 2: Linear Model The second model is a linear model. The independent variables are: the sea surface anomaly, the level of reservoirs in the system and the demand for energy; twelve lags of each of these variables are used as well and twelve lags of the energy prices. Two specifications are used and their forecast is combined to obtain the final estimation. Backward selection is used for eliminating all variables with p-values over 0.01 and 0.05 to reduce the number of independent variables. The two estimated specifications are:

!!! = !!!!! + !!! !!! = !!!!! + !!!

The superscripts 1 and 5 refer to the p-value used in the backward selection. !!! and !!! are the array of explanatory variables remaining after backward selection is applied. The final forecasts, is the average of the forecasts obtained with the two models:

!! =12 !!! + !!! +  !!

In order to simulate future market prices bootstrapping is done. For each simulation a random sample, with replacement, of the centered errors is taken. Simulations are built incorporating the errors to the forecasts in a recursive manner (since the forecast is recursive).

!!!∗ = !!!!! + !!!∗ !!!∗ = !!!!! + !!!∗ !!∗ =

!!!!!∗ + !!!∗ ,

     

48    

where !!∗ are the bootstrap errors. Figure 42 shows 1000 simulations of the market 25 years ahead, for the latter model. Each line represents a simulation. Prices are in July 2011 constant values.

Figure 42: Simulated spot prices. Alternative model.

Since the price used is the monthly average in a given month it may be the case that the spot price is over or under the scarcity price for one or a few days. In order to introduce some flexibility over this, the expected number of days where the price is over the scarcity price is calculated. This is done by assuming the distribution of the daily price has a normal distribution with median equal to the average monthly price and standard deviation equal to the historical average deviation from the price to the monthly average. One can calculate the probability of the price of a given day being over the scarcity price. The expected number of days where the price is over the scarcity price for each month is the thirty times this probability.

     

49    

6. References Pinilla, A., Rodriguez, L., R. Trujillo. 2009. Performance evaluation of Jepirachi Wind Park. Renewable Energy. 34, 48 – 52. Dyner, I., Olaya, Y. , C. Franco. An enabling framework for wind power in Colombia: what are the lessons from Latin America. Comisión Reguladora de Energía y Gas. 2011. Resolución No. 092. Julio 7 de 2011. Comisión Reguladora de Energía y Gas. 2011. Energía Firme para el Cargo por Confiabilidad de Plantas Eólicas. Documento CREG – 075. Julio 7 de 2011. Vergara, W., Deeb, A., Toba, N., Crampton, P. and I. Leino. 2010. Wind Energy in Colombia: A Framework for Market Entry, World Bank.

     

50    

Appendix A The following figures represent changes in the IRR due to changes in the amount of firm energy or the level of the FAZNI and the CERE taxes. Since generator with a capacity under 20 MW do not pay the CERE, changes in its level do not affect the IRR. If we assume that a proportion of the initial investment is financed with a loan that matures in fifteen years with a 9% annual interest rate and monthly payments then Figure 42 shows the IRR at different firm energy levels. As can be seen, the bigger the proportion of initial investment financed with a loan, the steeper the curve is.

Figure 42: IRR sensibility to the amount of firm energy and the proportion of initial investment financed with a loan

(at 9% annual interest rate).

     

51    

A. 19.6 MW WIND FARM. INITIAL INVESTMENT FINANCED WITHOUT LOANS.

Figure 43: IRR according for different FAZNI and firm energy levels (CERE fixed at $20,000 COP). Initial

investment financed without a loan.

Figure 44: IRR according for different CERE and firm energy levels (FAZNI fixed at $1,000 COP). Initial investment

financed without a loan.

     

52    

B. 19.6 MW WIND FARM. 50% OF INITIAL INVESTMENT FINANCED WITH A LOAN AT A 9% ANNUAL INTEREST RATE AND A 15 YEARS MATURITY.

Figure 45: IRR according for different CERE and firm energy levels (FAZNI fixed at $1,000 COP). 50% of the initial investment is financed with a loan that matures in 15 years, has an annual interest rate of 9% and monthly payments.

Figure 46: IRR according for different CERE and firm energy levels (FAZNI fixed at $1,000 COP). 50% of the initial investment is financed with a loan that matures in 15 years, has an annual interest rate of 9% and monthly payments.

     

53    

Appendix B We assume that if a loan is made to finance the initial investment, then this loan has monthly payments of equal size. The amount you have to pay each month is given by the formula:

! =! ∙  !! 1+ !! !

1+ !! ! − 1

Where N is loan amount, T is the loan term in months and !!is the monthly interest rate. This formula can be deduced by means of recursion taking into account that each month the amount of money you owe is equal to: !! = !!!! 1+ !! − ! And that !! = 0. Finally, since the interest rate is given on an annual basis, then the annual interest rate can be calculated by mean of the following formula:

!! = 1+ !! ! !" − 1

     

54    

Appendix C The following figure shows the full set of simulated prices based on the more complex model (no truncation of very high prices as in previous figure).

Figure 47: Full set of simulated prices.