Energy for Sustainable Development 59 (2020) 83–96 Contents lists available at ScienceDirect Energy for Sustainable Development Renewable electricity generation target setting in developing countries: Modeling, policy, and analysis Anthony Afful-Dadzie a, Eric Afful-Dadzie a,⁎, Nii Asafoatse Abbey a, Bright Ansah Owusu a, Iddrisu Awudu b a University of Ghana Business School, LG 78, Legon, University of Ghana, Accra, Ghana b Department of Management, Quinnipiac University, 275 Mt. Carmel Avenue, Hamden, CT 06518, USA ⁎ Corresponding author. E-mail addresses: aafful-dadzie@ug.edu.gh (A. Afful-Da (E. Afful-Dadzie). https://doi.org/10.1016/j.esd.2020.09.003 0973-0826/© 2020 International Energy Initiative. Publish a b s t r a c t a r t i c l e i n f o Article history: Received 15 February 2020 Revised 20 June 2020 Accepted 6 September 2020 Available online 5 October 2020 Keywords: Generation capacity planning Renewable energy Renewable electricity target Unmet demand Budget constraints Developing countries Many countries have set renewable energy targets in their electricity supplymix to encourage investments in re- newable energy technologies. In developing countries, however, many of such targets are either abandoned or fall far short of the target date, primarily due to issues of financing, cost of electricity, and level of unmet demand. This paper presents a Generation Expansion Planningmodel that can be used to assess such issueswhen setting a renewable electricity generation target. In particular, the model can be used by developing countries to set re- newable electricity generation targets that are in line with their financial ability and thus, stand a higher chance of being achieved. Additionally, the analysis offered can inform developing countries of the cost and benefits of a renewable electricity generation target policy. The usefulness of the model is demonstrated using Ghana as a case. The results indicate that Ghana will need a budget of not less than 1% of its GDP for generation capacity in- vestment if it desires to achieve its 10% renewable electricity generation target by 2030while keeping unmet de- mand at reasonable levels. If Ghana however enforces the target at its current capacity investment levels, it risks raising unmet demand levels by an average of 4% per year and cost of electricity provision by about US$224 mil- lion annually between 2019 and 2030 when compared to the absence of a renewable electricity generation target. © 2020 International Energy Initiative. Published by Elsevier Inc. All rights reserved. Introduction The growing call to reduce carbon dioxide emissions alongwith cost of electricity generation from renewable sources becoming more com- petitive (IRENA, 2015a) have energized many countries to set renew- able electricity targets in their electricity supply mix. According to the International Energy Agency (IEA, 2017), renewables are now the fastest growing source of world energy, with consumption growing by 3% per year. Data from the IEA reveal that renewable energy accounted for more than two-thirds of the world's 165 GW of new electricity gen- eration capacity installed in 2016 (IEA, 2017). The IEA further estimates that the increasing growth in renewable energy penetration is expected to continue given the technological advances and the continuous de- cline in capital cost. As a result of these developments, there is a renewed interest by many countries to increase the amount of electric- ity generated from renewable energy sources.Many countries have thus been setting targets to serve as a motivation to focus and direct re- sources to meet such renewable energy goals. For example, as of 2015, dzie), eafful-dadzie@ug.edu.gh ed by Elsevier Inc. All rights reserved there were as many as 164 (four times the number in 2005) countries setting at least one type of renewable energy target. A sizable number (131) of these countries are from developing or emerging economies. Almost all the 28 EU countries have renewable energy targets set for 2020. (Klessmann et al., 2011). The United Kingdom for instance has a target of achieving 30% renewable content of its electricity generation mix by 2020 (IRENA, 2015b). China has set a target of 30% renewable electricity consumption by 2030 (IRENA, 2015b). Developing countries are not left out. Ghana has set a target of 10% renewable content of its electricity generation by 2030 - originally set for 2020 (BFT, 2017; Obeng-Darko, 2019). Table 1 lists some countries, their renewable elec- tricity targets, and their due dates. However, while countries such as China, India, Germany, and U.S.A have made substantial progress (IRENA, 2017), and are stepping up their renewable energy targets, others are either abandoning or revising down their targets. Australia announced to abandon its renewable en- ergy target in favour of coal power generation (Smyth, 2017). Ghana's target of 10% electricity production from renewable electricity (exclud- ing existing large hydroelectric plants) had to be rescheduled to 2030 because, less than 0.3% had been achieved as at the end of 2018 (ECG, 2019c). A number of EU member states such as Ireland, Netherland, and Poland are likely to miss their 2020 renewable energy targets (Hassel et al., 2017; Oyewunmi et al., 2019). . http://crossmark.crossref.org/dialog/?doi=10.1016/j.esd.2020.09.003&domain=pdf https://doi.org/10.1016/j.esd.2020.09.003 mailto:aafful-dadzie@ug.edu.gh mailto:eafful-dadzie@ug.edu.gh https://doi.org/10.1016/j.esd.2020.09.003 http://www.sciencedirect.com/science/journal/ Table 1 Selected countries and their renewable electricity generation target and due Date. Source: Adapted from IRENA (2015b, Page 25). Region Country Target (%) Due Date Europe Iceland 100 2020–2021 Denmark 50 2020–2021 Germany 35 2020–2021 United Kingdom 30 2020–2021 Italy 25 2020–2021 Ukraine 11 2020–2021 North America Mexico ≈30 2023–2025 Costa Rica 100 2020–2021 Cuba 24 2030–2032 Dominican Republic 25 2020–2021 South America Chile 20 2023–2025 Brazil 85 2023–2025 Argentina 20 2025 Africa South Africa 15 2030–2032 Morocco 25 2020–2021 Burkina Faso ≈48 2030–2032 Cape Verde 50 2020–2021 Asia China 30 2020–2021 India ≈15 2020–2021 Japan ≈23 2030–2032 Indonesia 15 2023–2025 Philippines 40 2020–2021 Oceania Australia 20 2020 New Zealand 90 2025 Fiji 80 2020–2021 A. Afful-Dadzie, E. Afful-Dadzie, N.A. Abbey et al. Energy for Sustainable Development 59 (2020) 83–96 A number of reasons could explain a country's need to abandon, or its failure to meet a proposed renewable energy target. First, although the cost of electricity over the lifespan (i.e. Levelized Cost of Electricity (LCOE)) of some renewable energy technologies is now competitive to that of some conventional generators, their capital cost still remains high (EIA, 2018; Green & Staffell, 2016; Sawin et al., 2018; Tang et al., 2012). This fact is particularly important in the context of developing countries, since it potentially discourages investment in large-scale re- newable technologies. This is because, planning for electricity capacity in developing countries sometimes demands that a generator with cheaper capital but relatively higher running cost is favored over a gen- erator with high capital but lower running cost, even if the latter has a lower LCOE than the former. The reason for this stems from the fact that, sinking limited fund for capacity investment in a generator with high investment cost, could lead to capacity shortage in future due to lack of funds (Afful-Dadzie et al., 2017). Shortage in generation capacity in turn leads to high levels of unmet demand, which could be costly to an economy (Africa Progress Panel, 2015; Eberhard et al., 2008). It is es- timated that Ghana for instance, lost about $2.1million per day during a 2014/2015 electricity crisis (Kumi, 2017). Thus, the need to have suffi- cient generation capacity available to avoid high levels of unmet de- mand might mean avoiding renewable energy technologies with higher capital cost. Second, it is a fact that without generous government tax incentives and feed-in tariffs, majority of renewable energy projects, worldwide, would not be financially viable (Green & Staffell, 2016). Germany's re- newable energy success for instance, is largely attributed to a subsidy policy which required utilities to buy electricity from renewable gener- ators at prices far above the cost of electricity generation from conven- tional sources. Such tax incentives and feed-in-tariffs would be unaffordable tomost governments in developing countries and can dis- suade investment in renewable energy technologies (Frondel et al., 2010; Geels et al., 2017; Zhang et al., 2014). Third, whereas thermal plants can provide electricity at virtually any time, the same cannot be said of some renewable energy technologies such as wind and solar, which tend to be intermittent in intensity. To overcome the intermit- tency of renewable energy sources, the practice has been to rely on 84 electricity imports from neighbouring countries and backup generators (Green & Staffell, 2016; Hammons, 2008). However, most developing countries already face capacity shortages (Batinge et al., 2019) and can- not afford to spend much needed funds on back-up generators. Fourth, most renewable generators require that they are sited at specific places with significant concentration of wind or solar. This would require sig- nificant investment in new transmission lines if such places are far away from existing transmission lines (Schaber et al., 2012). Given the reasons above, it is not surprising that Ghana for instance, has barelymade any progress in achieving its renewable electricity gen- eration target. This also means that for a renewable energy target to be realistic, it must be factored into the long-term planning of energy sup- ply. For electricity systems, this means including such target policies in Generation Expansion Planning (GEP)models so that its impact on elec- tricity supply, particularly, on levels of unmet demand, cost of electric- ity, and capacity investment needed in meeting demand is analyzed and taken note of. Such models can serve several purposes. First, they are useful in analyzing impact on levels of unmet demand as well as cost of electricity if a renewable electricity target policy were to be ad- hered. It can also be used to search for a target that meets a desired goal in relation to levels of unmet demand and cost of electricity, given one's financial position. Further, it can as well be used to deter- mine the financial commitment in new generation capacity needed to achieve a proposed renewable electricity target. This research therefore proposes an enhanced GEP model that can be used to study the purposes listed above when a percentage of a country's electricity generation is to be realized from renewable sources by a specific date. Themodel is useful especially in an erawhere the set- ting of renewable energy target is the norm. In addition, the model is also designed with a focus on peculiarities in developing countries, most especially on difficulties in raising financial resources to fund gen- eration capacities, and concerns about access to electricity. Traditional GEP models in general are solved to determine the optimal time, size, and type of power generators to procure, as well as the electricity to be supplied by each generator, in order to minimize the cost of electric- ity supply. The traditional GEP model is typically solved subject to elec- tricity demand, generator capacity, and operational constraints. The proposed model augments the traditional GEP model with two impor- tant constraints – renewable energy target and generation capacity in- vestment budget constraints. In particular, the generation capacity investment budget constraint is added in order to take into account the financial strength of a country in meeting its renewable electricity target and ultimately, the electricity demand. The proposed GEP model is developed in the form of a multi-period stochastic mixed-integer linear program (MILP) in order to capture the time period within which the renewable electricity generation target is to be met, as well as uncertainty in some of the model parameters. Sto- chasticMILP approach is appropriate since solutions from stochastic op- timizationmodels have proven to be superior to other techniqueswhen uncertainties inmodel parameters exist (Awudu& Zhang, 2013; Feng & Ryan, 2013; Jin et al., 2011; Shiina & Birge, 2003). This has led to a num- ber of authors applying stochastic optimization techniques to electricity generation capacity planning. Shiina and Birge (2003) developed a multi-stage stochastic linear integer programming for analyzing elec- tricity generation capacities under uncertainty. Jin et al. (2011) analyzed a long-term GEP problem with uncertainties in future demand and fuel prices based on a multi-period stochastic linear optimization model to determine the type and size of power plants to be constructed over an extended planning horizon. Feng and Ryan (2013), also studied a GEP problem based on a multi-period stochastic linear optimization model under a new scenario generation approach. A GEP model that best describes peculiar conditions faced by electricity generation planners in developing countries is presented in Afful-Dadzie et al. (2017). The authors solved a long-termGEPproblemwith a budget con- straint designed to take into consideration the financial capacity of a country. A. Afful-Dadzie, E. Afful-Dadzie, N.A. Abbey et al. Energy for Sustainable Development 59 (2020) 83–96 Though a number of studies have considered GEP problems with a renewable energy dimension, the focus has not been on electricity pro- duction target. Oree et al. (2017) provided a review of GEP techniques with environmental considerations and highlighted the emerging chal- lenges presented by the intermittent nature of some renewable energy sources on power systems. Muis et al. (2010) presented anMILP model for optimal planning of electricity generation schemes to meet a nation's specified CO2 emission target. Their work is quite different from the renewable electricity production target considered in this work because they focus on selecting generation plants to achieve a cer- tain level of CO2 emission, whereas we select generation plants to achieve a certain level of electricity production from renewable energy sources. Muis et al. (2010) also did not consider budget constraint and uncertainty in electricity demand and fuel price. Nomenclature Indices i index for generator type, i = 1, 2, …, I t index for time period (in years), t = 1, 2, …, T k index for sub-target period (in years), k = 1, 2, …, K n index for renewable generator type, n = 1, 2,…, N s index for scenario, s = 1, 2, …, S Parameters Ci, t TC capital cost of generator type i in period t [$/MW] Ci, t f annualized fixed cost of generator type i in period t [$/MW/ yr] Ct IM cost of imported electricity in period t [$/MWh] Pi, t co2 CO2 emissions cost per MWh of electricity generated by gen- erator i in period t [$/MWh] Ai capacity factor of generator type i [%] Ht total hours in year t [hours] ρt proportion of electricity demand in period t that can be served through imports Bt budget for additional generation capacity investment in pe- riod t [$] Mi, t minimum amount of electricity generation from generator i in period t. Ci, t c annualized capital cost of generator type i in period t [$/MW/ yr] Wi max maximum installed capacity of a new generator type i [MW] Ct U cost of unmet demand in period t [$/MWh] Gi max number of generator type i that can be built over the entire planning period r periodic discount rate [%] γk renewable energy target in period k [%] Ii initial (i.e. t = 1) installed capacity of generator type i [MW] φt rated capacity degradation factor Bt L total of past periods unused budget in period t [$] Uncertain parameters ps probability of the occurrence of scenario s Ci, s, t v marginal cost of generation of generator type i under scenario s in period t [$/MWh] Ds, t electricity demand under scenario s in period t [MWh] Decision variables Gi, t number of new generator type i to build in period t Wi, t total installed capacity of generator type i available in period t [MW] Yi, s, t total electricity generated by generator type i under scenario s in period t [MWh] 85 Zi, t total of new installed capacity of generator type i at time t [MW] Es, t IM total imported electricity under scenario s in period t [MWh] Es, t U total unmet demand under scenario s in period t [MWh] In their work, Arnette and Zobel (2012) provided a multi-objective linear programming (MOLP)model to determine the optimal mix of re- newable energy generation technologies and existing fossil fuel facilities on a regional basis. Their model was formulated to enable a decision maker to minimize both annual electricity generation costs and level of greenhouse gas emissions, but with no focus on achieving a desired level of electricity generation from renewable energy sources. In our related literature search, only Knopf et al. (2015) explicitly consider renewable electricity target in their study. Their model was formulated to find the cost-effective shares of electricity from renew- able energy sources to meet the EU's 27% renewable energy target in final energy consumption by 2030. However, their formulation implic- itly assumes the availability of adequate capital needed to finance the recommended capacity decision. Note that many developing countries struggle to raise funds to finance generation capacity expansions, and therefore failure to take this into account when setting renewable elec- tricity targets has the tendency to render such targets virtually useless. It is no wonder that for instance, Ghana has achieved only 0.3% of its 10% renewable electricity target with less than a year left to the target date (Obeng-Darko, 2019). Ghana's failure could be attributed to the fact that, the target was set with little consideration to the country's fi- nancial ability. Perhaps, the strength of the proposedmodel over that of Knopf et al. (2015) lies in the fact that it can be used to determine a tar- get and the related financial commitment needed to meet a desired level of electricity demand and cost. This paper contributes to the literature on energy transition (partic- ularly with regards to developing countries) by providing a unique GEP model for the setting and evaluation of a renewable electricity genera- tion target policies. The paper also contributes to practice by helping de- veloping countries set realistic renewable electricity target that is commensurate to their financial position. The rest of the paper is orga- nized as follows. The next section presents the proposed GEP model with renewable electricity generation target. This is followed by a case study on Ghana's proposed 10% renewable electricity generation target by 2030, an extensive analysis and discussion of the case study results, and conclusion of the study. Generation expansion planning model with renewable electricity generation target The proposed GEP model with renewable electricity generation tar- get is represented by Eqs. (1)–(8). The definitions of the parameters and variables of the model are also presented in the nomenclature. A brief explanation of the major components of the model is provided next. Minimize∑T t¼1 ∑I i¼1 Ci,t c þ Ci,t f � � Zi,t þ∑S s¼1Ps ∑I i¼1 Cv i,s,t þ Pco2 i,t � � Yi,s,t þ CIM t EIMs,t þ CU t,E U s,t � � 1þ rð Þt−1 2 4 3 5 ð1Þ ∑I i¼1Yi,s,t þ EIMs,t ¼ Ds,t−EUs,t∀s,∀t ð2Þ Yi,s,t≤HtAiWi,t 1−φtð Þ∀i,∀s,∀t ð3aÞ Yi,s,t≥Mi,t∀i,∀s,∀t ð3bÞ EIMs,t ≤ρDs,t∀s,∀t ð4Þ Zi,t ¼ Zi,t−1 þWmax i Gi,t∀i,∀t ð5aÞ A. Afful-Dadzie, E. Afful-Dadzie, N.A. Abbey et al. Energy for Sustainable Development 59 (2020) 83–96 Wi,t ¼ Ii þ Zi,t∀i,∀t ð5bÞ ∑N n¼1Yn,s,k≥γk∑ I i¼1Yi,s,k∀s,∀k ð6Þ ∑I i¼1C TC i,t W max i Gi,t≤Bt þ 1þ rð ÞBL t−1∀t ð7aÞ BL t ¼ Bt þ 1þ rð ÞBL t−1−∑I i¼1C TC i,t W max i Gi,t∀t ð7bÞ Gi,t , Zi,t ,Wi,t ,Yi,s,t , E IM s,t , E U s,t , ≥0,Gi,t integer∀i,∀s,∀t ð8Þ Decision variables The major variables of interest in the proposed multi-period GEP model areGi, t, Yi, s, t, Es, tIM, and Es, t U . In any period, the GEPmodel is solved to determine Gi, t, the number of new generator type i to build in period t. The variable Yi, s, t, is the electricity generated by generator i in a par- ticular uncertain scenario (e.g. demand and fuel price) in period t, whereas Es, tIM , is the amount of electricity imported from neighbouring countries in scenario s in period t. In case the total electricity generated, and the amount imported are not enough to meet the periodic demand Ds, t, the unmet demand is tracked using the variable Es, tU . The remaining variables of the model are Zi, t, and Wi, t which are respectively used to track the accumulated capacity of new and installed generators since the beginning of the planning period. Objective function The objective function of the model which seeks to minimize the cost involved in the provision of electricity over a planned period starting from period t = 1 to period T, is represented by Eq. (1). It rep- resents the total of the present value of the several costs incurred over the planning period. Thefirst termof the numerator captures the capital cost and the fixed operating cost for new generation plants built within the planning period. The second term captures the cost incurred in the production of actual electricity. This cost is determined by a generator's marginal cost of electricity generation Ci, s, t v , emissions cost Pi, tco2, and the amount of electricity generated Yi, s, t. The third and fourth terms respec- tively represent the cost of electricity imported from neighbouring countries, and the indirect cost to a country's economy of any level of demand not met. Model constraints The objective function of Eq. (1) is determined subject to the follow- ing constraints. Electricity demand and supply Constraints on electricity demand and supply are modelled using Eq. (2). The left-hand-side is the electricity available for meeting period t demand, and is made up of actual generation and imported electricity fromneighbouring countries. Somedemandwill not bemet if electricity demanded is greater than the available supply. The right-hand-side is thus the demand met, which is the difference between the actual elec- tricity demanded, Dt, s, and demand not met, Et, sU . Electricity produced with generators owned by the planner is represented by Eqs. (3a) and (3b). Eq. (3a) ensures that for each generator type i, the total electricity produced at any period t is within the maximum possible electricity it can produce. Eq. (3b) helps to enforce generation from firm and non- firm power plants and enforces a minimum level of production on cer- tain power plants. For example, almost all existing hydro units in the Ghanaian power system produce at a minimum level. The amount of 86 electricity imported is modelled with Eq. (4). Although in some cases electricity can be imported from neighbouring countries, this is usually limited. Eq. (4) is thus designed to account for this restriction. Generation capacity The available generation capacity of the generators is modelled with Eqs. (5a) and (5b). The cumulative capacity of new generators installed since the start of the planning period is tracked with Eq. (5a), whilesEq. (5b) does the same as Eq. (5a) but includes capacity that existed before the start of the planning period. Renewable electricity target The renewable electricity generation target to be achieved is repre- sented by Eq. (6), which ensures that total electricity generation from renewable energy sources at the target due date, meets the target. For practical purposes, a target due in period T (i.e. the end of the planning period) is divided into sub-targets due at selected periods k=1, 2,…, K. For example, Argentina aims to achieve a 20% renewable energy pene- tration in 2025. However, this target is divided into sub-targets of 12% by 2019, 16% by 2021, 18% by 2023 and 20% by 2025 (Pink, 2018). The left-hand-side of Eq. (6) ensures that the sum total of all electricity gen- eration coming from renewable technologies meet the target percent- age for a selected period k. One could replace Eq. (6) with Eq. (9) below, if the focus of the target is on generation capacity, and specific megawatts of capacity are to be achieved for individual renewable en- ergy technologies at targeted periods. Here, Wn, k, and Fn, k are respec- tively, the actual attained megawatts, and the targeted megawatts capacity from renewable energy technology n in targeted period k. For example, while Ghana has set an overall renewable electricity genera- tion target of 10% by 2030, its renewable energymaster plan has specific capacity target for each renewable technology. Wn,k≥Fn,k∀n,∀k ð9Þ Capacity investment budget constraints The ability of a country to finance additional generation capacity needed in period t is modelled with Eq. (7a). The equation ensures that at any period t, the money spent on new generators is within the allocated budget for that period. The total budget available in any period consists of the planned budget for the period and the unused funds in the previous period. The unused funds are trackedwith Eq. (7b) and as- sumed to earn interest each period at a rate of r%. The integer and non- negativity constraints expressed in Eq. (8) complete the proposed GEP model. The proposed GEP model is unique in the sense that, in addition to the usual GEP output, it can be used to determine the financial commit- ment needed to bring to fruition a desired renewable target that meets an acceptable level of unmet demand and cost of electricity provision. On the other hand, themodel can be used to determine a target that ac- commodates a country's financial ability, and the tolerance level of unmet demand and electricity generation cost. Most importantly, given a country's financial ability, it can be used to analyze the impact on electricity supply (in terms of cost of electricity and levels of unmet demand) should a target be adhered to.We note here that the proposed GEPmodel only focuses on the generation side of the electric power sys- tem and does not consider transmission constraints and investment. Case study on Ghana's 10% renewable electricity generation target policy We begin the case study with an overview of the electricity genera- tion sector of Ghana in relation to current installed and dependable gen- eration capacity, electricity generation trend, and a brief information of the country's renewable energy master plan. Hydro 6017.36 Thermal 9803.15 Imports 139.7 Fig. 1. Electricity generation in Ghana by Source in 2018. (Source: ECG (2019c)). 1 The 10% electricity generation target excludeshydro generators of 100MWand above, as well as all existing hydro generators. See the footnote on page iv of the Ghana Renew- able Energy Master Plan (ECG, 2019a). A. Afful-Dadzie, E. Afful-Dadzie, N.A. Abbey et al. Energy for Sustainable Development 59 (2020) 83–96 Generation capacity Ghana is a developing country with an electricity generation sector made up of both public and private companies. The public companies consist of two authorities, namely Volta River Authority (in charge of the Akosombo hydroelectric dam, and government owned thermal plants), and Bui Power Authority, in charge of the Bui hydroelectric dam. The private companies consist of a vibrant group of Independent Power Producers (IPP), currently commanding about 32% of the country's installed generation capacity. IPPs usually sell electricity to the country on Take-Or-Pay contract, where the government is obliged to buy a stipulated amount of electricity within an agreed period from the IPP or pay the IPP a penalty. Electricity generated is transmitted to consumers by the Ghana Grid Company Limited (GridCo). As at the endof 2018, the country had an installed anddependable generation ca- pacity totaling 4533 MW and 4048 MW respectively (ECG, 2019c). The installed capacity comprises of 2906MWthermal, 1584MWhydroelec- tric, and 42.5MW solar. As shown in Fig. 1, about 58.14% of the 15,960.3 GWh of electricity generated by Ghana in 2018 came from thermal, followed by hydroelectric with 39.6%, and 2.26% by imports (ECG, 2019c). There is also a total of 530MWof dependable committed capac- ity of combined circle gas turbines for 2019 (ECG, 2019c). Plans are far advanced for the construction of a 700 MW coal-fired plant with a po- tential to be increased to 2000 MW. Electricity demand trend The annual electricity generated in Ghana from 2001 to 2017 is as shown in Fig. 2. Altogether, the average annual growth rate of electricity generatedwithin the stated period stood at roughly 4.4%. However, this data does not includedemandnotmet, especially during themajor elec- tricity crisis in 2003, 2007, and 2015 when the annual growth rates were negative. The average annual growth rate is 10.52% if one excludes data from 2003, 2007, and 2015. Also, this data does not include the fre- quent load shedding and blackouts experienced whenever demand ex- ceeds supply. Adding demand not met, and the fact that 17.5% of the population did not have access to electricity in 2016, the average annual growth rate could be far higher. In fact, the 2018 projected demand growth rate by the national grid operator was 15.01% (GridCo, 2017). The Ghana renewable energy master plan Ghana has set a target to generate at least 10% of its electricity from renewable technologies by 2030 (originally by 2020). This target fo- cusesmainly on solar,wind, and biomass technologies aswell as smaller hydroelectric plants. To achieve the 10% target, the country has drawn up an ambitious renewable energy master plan (as presented in Table 2) that aims to add about 200 MegaWatt-peak (MWp) from 87 smaller hydro and wave plants, 741.3 MWp from solar, 327 MWp from wind, and 122 MWp from biomass technologies. In the master plan, the country expects to have in place a total of 194.74 MWp of re- newable energy capacity by 2020, 936.64 MWp by 2025, and 1353.63 MWp by 2030 towards electricity generation. More than 80% of the planned capacity is targeted at utility scaled renewable technologies, whereas at least 8% is targeted at rooftop solar installations. However, as at the end of 2018, less than 1% (42.5MWp) of the country's installed capacity came from renewable generators (excluding existing large hy- droelectric plants). In fact, out of the 15,906.36 GWh of electricity gen- erated in 2018, there was virtually no contribution from renewable energy technologies that qualifies under the renewable energy target1 (ECG, 2019c). The meager renewable energy capacity achieved so far, prompted the government to extend the 10% renewable electricity gen- eration target to the year 2030. Model parameters This section presents the parameters used in running the model for the case study, startingwith data on electricity demand, followed by in- formation on candidate generators, costs, budget for additional capacity investment, and assumed parameters. Electricity demand Electricity demand exhibits an increasing trend over time mainly due to population and economic growth. The demand trend ismodelled using the trend formula in Eq. (10), where demand in period t, Ds, t, is aewt percentage increase over Ds, t−1, the demand in period t− 1. The pa- rameter ewt is the percentage increase in demand from period t − 1 to period t. In practice, ewt is unknown and captures theuncertainty in elec- tricity demand. Although unknown, it is usually modelled with a prob- ability distribution based on past data. Based on Ghana's electricity generation from 2000 to 2017, and adjusting to account for the electric power crisis in 2003, 2007, and 2015, the best fitted distribution for ewt was a triangular distribution of Triang (1%, 7%, 15%). This is interpreted to mean that the annual percentage increase in electricity demand is most likely to be 7%, but could be as low as 1%, and as high as 15%. To capture the uncertainty in the electricity demanddata over theplanning period, we follow the Monte-Carlo approach in Breeden and Ingram (2010) to generate demand forecast. To do this, and for each period, a random draw is taken from Triang (1%, 7%, 15%) for ewt which is then used in Eq. (10) to obtain a demand forecast. The set of demand forecast for each period together forms one scenario. The process is repeated several times depending on the number of scenarios needed to charac- terize the uncertainty. This case study used 20 scenarios with equal probability of occurrence to characterize the demand uncertainty. The Mean Absolute Percentage Error (MAPE) of the scenarios was less than 8% signifying a highly accurate forecast in relation to the national forecast. Ds,t ¼ 1þ ewtð ÞDs,t−1 ð10Þ Table 3 presents five electricity demand scenarios from 2019 to 2027. These were determined using the recorded actual electricity sup- ply of 15,960.36 GWh in 2018 as found in ECG (2019c) as the starting value for Ds, t−1. Also shown in the last two columns of Table 3 are the national annual average forecasts for 2019–2027 (ECG, 2019c, page 62–63). The 2019 national annual average forecast of 17,237.8 GWh is a base case demand forecast representing a percentage growth of 8% over that of 2018. The corresponding high case forecast is 18,013.96 GWh, representing a growth rate of 12.9%. As shown in Table 3, the gen- erated scenarios show a reasonable deviation from the national average Table 2 Electricity generation capacity expansion plan for renewable energy technologies in Ghana. Source: ECG (2019a). Technology Project Units 2015 2020 2025 2030 Cumulative (2030) Solar Utility scale MWp 22.5 130 195 100 447.5 Distributed Solar PV MWp 2 18 80 100 200 Standalone Solar PV MWp 2 8 5 5 20 Solar Street/Community lighting MWp 3 4 4 14 25 Solar irrigation MWp 2.8 6 20 20 48.8 Wind Utility scale MWp 0 0 275 50 325 Standalone Wind Systems MWp 0.01 0.1 0.9 1 2 Biomass Utility-scale/co-generation MWp 0 0 72 0 72 Waste-to-Energy Utility Scale MWp 0.1 0 30 20 50.1 Small/Medium Hydro Plants MWp 0 0.03 80 70 150.03 Hydro/Wave Power Wave Power MWp 0 5 0 45 50 Hybrid Mini- Grid Mini/Micro-grid MWp – – – 12 12 Total MWp 32.41 171.13 761.9 437 1402.43 5,000 7,000 9,000 11,000 13,000 15,000 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 E le ct ri ci ty G en er at io n (G W h) Year Electricity Generation Trend in Ghana Fig. 2. Trend chart of annual electricity generation in Ghana. Source: ECG (2013, page 7), and ECG (2019b, page 8). A. Afful-Dadzie, E. Afful-Dadzie, N.A. Abbey et al. Energy for Sustainable Development 59 (2020) 83–96 annual forecast, and better capture uncertainty in expected demand compared to the two national forecasts. In other words, any of the 20 scenarios is likely to occur. The strength of the proposed stochastic model approach is that all 20 scenarios are used to derive a solution that is better suited to the demand uncertainty present. Generator technical information The types of electric generators considered in the case study along with their technical and cost data are presented in Table 4. Seven differ- ent kinds of generators were considered. These were hydroelectric (HYDRO), Combined Circle Gas Turbine (CCGT1 and CCGT2), Open Cir- cle Gas Turbine (OCGT), Coal (COAL1 and COAL2), Onshore Wind Table 3 Example forecast of possible electricity demand for Ghana. Data under Scen-3, Scen-7, Scen-11, Scen-16, and Scen-19 are possible electricity demand path for Ghana from 2019 to 2027. Data under ‘National (Base)’ and ‘National (High)’ are the officially reported base and high demand case scenarios from the 2019 Ghana electricity supply plan issued by the Ghana Energy Commission (ECG, 2019c). Year Scen-3 Scen-7 Scen-11 Scen-16 Scen-19 National (base) National (high) 2019 17,019 17,518.8 18,141.1 17,258.4 17,047.3 17,237.80 18,014.00 2020 17,768 19,365.7 19,924.2 18,736.2 18,261.2 18,249.50 20,070.50 2021 19,327.4 20,372.1 21,547.7 19,564.7 19,604.9 18,918.00 21,398.70 2022 20,214.1 21,856.4 23,347.9 20,203.2 20,931.5 21,108.30 24,189.90 2023 21,240.6 24,294.5 24,689.3 21,940.5 22,521.3 22,186.90 25,899.70 2024 22,645.8 26,873.5 26,560.2 24,298.6 24,605.2 24,931.40 28,999.10 2025 23,513.9 28,720.1 29,267.3 26,402.4 27,337.4 25,690.50 30,391.00 2026 25,497.7 30,799.7 32,487.8 28,243.9 29,461.3 26,476.10 31,857.00 2027 27,850.3 32,775 33,721.4 29,646 30,269.1 27,214.60 33,332.60 88 (WIND1 and WIND2), Solar (SOLAR1 and SOLAR2), and biomass (BIO- MASS), for a total of eleven generator types. Out of these, only HYDRO, WIND1, WIND2, SOALR1, SOLAR2, and BIOMASS are eligible to be se- lected tomeet the 10% renewable electricity target. Themodel assumes not more than four hydro generators of size 60 MW each, can be con- structed in the planning period. This is in accordance with the remain- ing significant hydroelectric opportunities in Ghana. To account for this assumption, Eq. (11) was added to the proposed model for the case study. Although Ghana currently has thermal capacities of various sizes from as low as 22 MW to as high as 470 MW, we considered only the capacity sizes of 300MW and 450MW for the Combined Circle Gas Turbines and150MWfor theOpen Circle Gas Turbines to reflect the recent capacity investment trend. For this reason, existing CCGT plants of capacities less than 300 MW were grouped together under CCGT1 and those above under CCGT2. Note that the Dependable capacity in Table 4 include committed capacity of 190 MW and 340 MW respec- tively for CCGT1 and CCGT2 for 2019 (ECG, 2019c). The capacity factor for the hydro and thermal plants were estimated based on supply data from ECG (2019c), while that of wind and solar were taken from IRENA (2018), and biomass from EIA (2019b). ∑ T t¼1 GHYDRO,t≤4∀i,∀t ð11Þ Capital, fixed operating and maintenance (O&M), and emissions costs The Capital, FixedO&M, and Emissions costs for the generators at the beginning of the planning period (i.e. 2019) are presented in Column 6 to 8 of Table 4. With the exception of capital cost for WIND1, WIND2, Table 4 Parameters for the beginning period (2019) of the GEP planning horizon for eleven candidate generation plants. These include dependable committed capacity of 190MW for CCGT1 and 340 MW for CCGT2. Source: (GridCo, 2011; EIA, 2019; EIA, 2019b; EIA, 2018; EIA, 2016; EIA, 2013; IRENA, 2018; IRENA, 2012; ECG, 2019c). Generator type Rated Capacity (MW) Dependable Capacity (MW) Capacity Factor Generator Life span (yrs) Capital cost (1000 $/MW) Fixed O&M cost ($/MW/year) Emissions Cost ($/MWh) Variable O&M Cost, excluding fuel ($/MWh) Generator Degradation Factor HYDRO 60 1365 47.5% 60 5000 15,000 0 2 0.003 CCGT1 300 888 40.5% 25 1057 10,100 6.7 2 0.003 CCGT2 450 1300 40.5% 25 1104 10,100 6.7 2 0.003 OCGT 150 1039.6 40.5% 25 678 6870 8.2 10.7 0.003 COAL1 250 0 50% 35 3818 42,100 16.2 4.6 0.003 COAL2 350 0 50% 35 3636 42,100 16.2 4.6 0.005 WIND1 50 0 16.5% 20 1877 39,700 0 9.5 0.005 WIND2 100 0 16.5% 20 1657 47,470 0 9.5 0.016 SOLAR1 20 42.5 19.5% 20 2671 23,400 0 0 0.016 SOLAR2 50 0 19.5% 20 2434 22,020 0 0 0.005 BIOMASS 50 0 50% 30 3837 112,150 8.1 5.6 0.005 A. Afful-Dadzie, E. Afful-Dadzie, N.A. Abbey et al. Energy for Sustainable Development 59 (2020) 83–96 SOLAR1, and SOLAR2, the values for the rest of the planning period (i.e. 2020–2030) for these costs were determined by increasing that of an immediate past year by 0.5%. On the other hand, the capital cost for the periods after 2019 for WIND1, WIND2, SOLAR1, and SOLAR2 were determined by decreasing that of the immediate past year by 0.5%. The decrease in this case is to reflect the continuous decline in investment cost of wind and solar plants. The emission cost is based on the assumption of a $15 per ton of CO2 emission (roughly half of what is charged in most high income countries). This is set to reflect the economic valuation in developing countries relative to developed countries. Also, the emission cost of biomass is set at half that of coal to accommodate either side of the argument for and against the notion of biomass as a carbon-neutral fuel. Marginal cost of generation Themarginal cost of generation is dependent on the amount of elec- tricity generated, which is made up of two parts namely, marginal cost of fuel, and variable O&Mcost. The Variable O&Mcost for the generators in the beginning of the planning period is given in column 9 of Table 4. The Variable O&M cost for the subsequent periods after the beginning period were determined in the same way as that of the other costs by increasing the immediate past year's cost by 0.5%. The variable O&M cost for onshore wind in 2019 is given as $9.5/MWh according to data in the generation master plan of Ghana (GridCo, 2011), and similar es- timates as found in (IRENA, 2012). The marginal cost of fuel component is assumed to be uncertain for the generators. However, that of generator types WIND1, WIND2, SOLAR1, and SOLAR2 are assumed to have a marginal fuel cost of $0/ MWh,while that of HYDRO is set at $2/MWhbased ondata from the op- erator of the largest hydroelectric dam in Ghana. Table A.1 in Appendix A gives a breakdown of how the marginal fuel cost for CCGT1, CCGT2, OCGT, COAL1, COAL2, and BIOMASS were generated. Capacity investment budget constraint It is assumed that the financial resource available for investment in newgeneration capacity is limited. This assumption is important, other- wise achieving a renewable target is trivial. Public utility companies in developing countries are hardly able to finance newgeneration capacity Table 5 Additional model parameters. Parameter Value Cost of unmet energy ($/MWh) 225 Cost of imported electricity ($/MWh) 150 Annual interest rate for discounting (%) 12 Number of hours in a year 8760 89 through revenues from electricity generated because of low electricity prices to make electricity affordable to the populace. As a result, these public utilities are generally supported by the government each year in the financing of new generation capacity. We assume that a portion of the country's annual GDP is allocated for the financing of new gener- ation capacity. GDP values are assumed to increase each year over the planning period by 4% based on Ghana's 2017 GDP value of $58.99 Bil- lion. Four different types of budget constraints are considered in order to analyze the impact of budget limitation on the level of unmet de- mand and cost of electricity provision. These are budgets equivalent to 0.5%, 0.75%, 1.0%, and 1.25% of Ghana's annual GDP. It is worth noting that currently, Ghana spends around 0.75% of its GDP on new genera- tion capacity investment.2 Other assumptions The model assumes that the 10% renewable target is spread within the planning period, and that at least 1%, 4%, 7%, and 10% of total elec- tricity production in 2020, 2023, 2026, and 2030 respectively should come from renewable energy sources, as is done in practice. The re- maining assumed parameters used to run the GEP model are given in Table 5. The cost of unmet demand at the beginning of the planning pe- riod is set at a moderate value of $225/MWh. However note that this cost could be higher given that Ghana lost an estimated $2.1 million a day during the 2014/2015 electricity crisis (Kumi, 2017). In fact, the Ghana Energy Commission estimates the cost of unmet demand at $1000/MWh (GridCo, 2011). Also, Ghana and Ivory Coast trade electric- ity with each other when the need arises. We assume the amount of electricity imported cannot exceed 4% of Ghana's annual demand. This percentage is in agreement with the annual importation figures in the electricity supply plan of Ghana. Electricity is imported at an assumed cost of $150/MWh. We also assume that existing hydro generators operates exclusively as base plants (as is primarily the case in Ghana) and sets theminimum production at 80% of the total energy equivalent to the right hand side of Eq. (3a). Also, OCGT plants are assumed to be peaking plants. In addi- tion, it is assumed that no generator (including existing ones) would be retired within the planning period. Lastly, electricity demand is the sum of all demand throughout the entire year. Case study results and analysis The stochastic MILP-GEP model with renewable electricity produc- tion target was programmed using the General Algebraic Modelling 2 The 0.75%GDP spendingwas estimated based on data on new generation capacity ad- ditions provided in the annual Electricity Supply Plan of Ghana provided by the Ghana En- ergy Commission. See (page 27, Table 22 of ECG, 2019c). A. Afful-Dadzie, E. Afful-Dadzie, N.A. Abbey et al. Energy for Sustainable Development 59 (2020) 83–96 System (GAMS) optimization software and solved using the ILOG CPLEX 12.6.0.0 solver. The result of the case study is presented next, starting with new generation capacity additions and percentages of electricity generation from renewable energy generators, followed by levels of unmet demand, and lastly total cost of electricity provision. The results are provided for both the case of when a renewable electricity produc- tion target policy is in place (Target), and when there is no target policy in place (No Target).We consider four budgetary policies for new gener- ation capacity investment, with the aim of understanding the impact of budget constraint on installed generation capacity, levels of unmet de- mand and cost of electricity provision. We also consider the imaginary case where there is sufficient financial resource to put into perspective the advantage richer countries have over poorer countries in the energy transition process. New generation capacity additions When available generation capacity is not enough to meet demand, the resulting demand not met leads to economic loss to a country. As such, the impact of a renewable electricity generation target policy on generation capacity must be analyzed, especially when a country is constrained financially inwhat it can allocate for new generation capac- ity investment. Table 6 presents the result on cumulative new installed generation capacity within the planning period for four different capac- ity financing budget allocations and the case when there is sufficient budget under both Target and No Target policies. In general, unless dur- ing or prior to a due date of a sub-target, there ismore generation capac- ity available under the No Target policy than under the Target policy. Also, and as expected, generation capacity in a given period increases with increasing budget. Fig. 3a–c also gives a breakdown of the cumulative new installed ca- pacity for the selected generator types at the sub-targets dates under both Target and No Target policies for different capacity financing budget allocations. As shown in Fig. 3a, under the No Target policy, no renewable energy generator is considered for new generation capacity at budget allocations of 0.5% and 0.75% of GDP. It can also be seen in Fig. 3b that even at higher budget allocations of 1% and 1.25% of GDP, the size of renewable energy generation capacity recommended as part of the required new generation capacity under the No Target policy is quite insignificant when compared to non-renewable generation ca- pacity. This is a clear sign that under current conditions, non- renewable energy generators are perceived to be more cost-effective than renewable energy generators when a country faces challenges raising the needed financial resource for new generation capacity in- vestment. This conclusion is supported by the recommended genera- tion mix under the Target case in Fig. 3a and b, where the renewable Table 6 Cumulative new installed generation capacity (in MW) within the planning period when a 10 Target) at annual capacity investment funds of 0.50%, 0.75%, 1.0%, and 1.25% of Ghana's GDP, and Year 0.5% of GDP 0.75% of GDP 1.0% of G Target No Target Target No Target Target 2019 0 0 0 0 0 2020 50 300 400 300 350 2021 350 600 700 900 1100 2022 800 900 1300 1350 1700 2023 900 1200 1450 1950 1870 2024 1200 1650 2050 2400 2620 2025 1500 1950 2500 3000 3220 2026 1670 2250 2700 3600 3440 2027 1970 2850 3300 4200 4190 2028 2420 3300 3900 4800 4990 2029 2520 3750 4060 5420 5440 2030 2670 4220 4310 6170 5690 90 energy generators recommended aremainly for the purpose of meeting the renewable electricity target and nothing more. Looking closely at Fig. 3a and b also reveals that the size of renewable energy generators (albeit insignificant) increases with increasing budget allocations under both the Target and No Target cases. This leads to an interesting observation when there is sufficient budget for capacity financing as shown in Fig. 3c. In this case, virtually all new gen- eration capacities come from the renewable energy generation plant, SOLAR2, whether under the Target or No Target policy. Note also that the generation capacity needed in this case is almost twice that needed under the 1.25% budget allocation which is the closest budget needed to ensure all demand ismet when a renewable electricity tar- get is enforced. This is not unusual given the low capacity factor of solar plants. However, the total capital cost needed to finance such size of generation capacity will be enormous, and quite frankly impractical. The preference for renewable energy generators under a sufficient budget in Fig. 3c also gives an indication that renewable energy genera- tors are not favored under the other four budget allocations purely be- cause of financial constraint. This goes to support the argument that at current cost estimates, some renewable energy technologies are cost ef- fective if one is not constrained financially. On the other hand, if one is constrained financially, then given the current high capital cost of re- newable technologies, very little capacity will be available to satisfy de- mand if renewable generators are preferred over conventional generators. Result in Fig. 3a–c can be used as a guide in selecting the cost effec- tive renewable energy generators to fulfil a planned renewable electric- ity target when facing a budget constraint. In general, the optimal decision under the Target policy when facing generation capacity in- vestment budget of 0.5% and 0.75% of GDP is to go for biomass. Just like the Ghana Renewable Energy Master (REMP) plan, a mix of bio- mass, wind, solar and small hydro is recommended at relatively higher budget allocations of 1.0% and 1.25% of GDP for new capacity invest- ment. When there is sufficient financial resources available however (see Fig. 3c), solar appears to be the best option among all renewable energy generators. Percentage of electricity generated The pattern in the generation capacity additions is also reflected in the percentage of electricity generation coming from renewable energy generators as seen in Table 7. It can be observed under the Target policy that, under all four budget allocations, renewable energy generators ap- pear to be procuredmainly for the sake ofmeeting the target. This state- ment is supported by the fact that only the sub-target year of 2020 appears to have had a renewable electricity content significantly more % renewable electricity generation target is enforced (Target), and when not enforced (No when there is sufficient budget. The sub-target due dates are 2020, 2023, 2026, and 2030. DP 1.25% of GDP Sufficient Budget No Target Target No Target Target No Target 0 0 0 2190 2190 600 700 750 3240 3240 1200 1450 1500 4190 4190 1800 2350 2250 5390 5390 2550 2530 3000 6540 6540 3300 3280 3240 8090 8090 4050 4090 4140 9790 9790 4560 4590 5040 10,940 10,940 5070 5490 5940 12,190 12,190 5820 6540 6890 14,140 14,140 6720 6990 7640 15,190 15,190 7620 7540 8640 16,740 16,740 A. Afful-Dadzie, E. Afful-Dadzie, N.A. Abbey et al. Energy for Sustainable Development 59 (2020) 83–96 thanwhat is desired under the Target policy. Much of the renewable en- ergy content attained in 2020 is due to the 42.5 MW of solar capacity that existed before the start of the planning period. Also, the percentage of renewable electricity achieved under any Target policy in Table 7 ap- pears to be roughly the same even when the allocated budget increases (from 0.5% to 1.25% of GDP), which is an indication that renewable elec- tricity is only considered just for the sake of meeting the target. The percentage of renewable electricity achieved under a No Target policy in Table 7 appears quite insignificant. However, though small, it can be seen that the content appears to increase with increasing budget allocations. For example, the renewable electricity percentage under a No Target policy is 1.1% under a budget allocation of 1.0% of GDP compared to 3.7% under a 1.25% GDP budget in the year 2026. Sim- ilarly, it is 1.5% under a 1.0% GDP budget but 4.7% under a 1.25% GDP budget in the year 2030 under a No Target policy. In linewith the fact that virtually only renewable generators are con- sidered for new capacity additions when there is sufficient budget for generation capacity investment (see Fig. 3c), the percentage of renew- able electricity increases drastically within the planning period under the Sufficient Budget case, whether under a Target or theNo Target policy (from 33.1% in 2020 to as high as 72.8% in 2030). Level of unmet demand The percentage of unmet demand (i.e. electricity demanded but not met due to insufficient generation capacity) under Target and No Target policies for the different budget allocations are given in Table 8. As a 0 200 400 600 800 1000 1200 1400 G en er at or C ap ac ity (M W ) Generator Type Cumulative New Capacity Additions under Renewable Target at 0.5% GDP 2027-2030 2024-2026 2021-2023 2019-2020 0 500 1000 1500 2000 2500 3000 3500 G en er at or C ap ac ity (M W ) Generator Type New Capacity Additions under Renewable Target at 0.75% GDP 2027-2030 2024-2026 2021-2023 2019-2020 a Fig. 3. a. Cumulative capacities of recommend new generator types during the sub-target yea capacity investment equivalent of 0.5% and 0.75% of Ghana's annual GDP. b. Cumulative capac and 2030 under both Target and No-Target policies for generation capacity investment equiv new generator types during the sub-target years of 2020, 2023, 2026, and 2030 under suffici and No Target policies. 91 direct effect of the results on capacity additions from Sub-section 4.1, level of unmet demand is generally higher under the Target case than under the No Target case. For example, at budget allocation of 0.5% of GDP under the Target policy, the level of unmet demand is 6.2% in 2020, reaching 32.2% in 2030 (which is the target due date). Simi- larly, the level of unmet demand is 3.4% in 2020 and rises to 14.0% by 2030 under the Target policy for a budget allocation of 0.75% of GDP. In contrast, the level of unmet demand reduces by more than half under the No Target policy for the same budget allocations of 0.5% and 0.75% of GDP from 2021 to 2030. From capacity investment levels of 1.0% of GDP onwards, the level of unmet demand reduces drastically as more generation capacity becomes available. Even then, the levels under the Target policy are still more than under the No Target policy. Graphs comparing levels of unmet demand between the Target and the No Target policies at the sub-target due dates are shown in Fig. 4. In conclusion, Table 8 and Fig. 4 make it clear that in order to keep unmet demand at a reasonable level, Ghana might need to have in place an amount of not less than 1% of GDP towards financing new generation capacity if it were to follow through with its 10% renewable electricity target policy. Note that without the renewable target in place, an equiv- alent budget allocation of 0.75% of GDP guarantees reasonably low levels of unmet demand comparable to a developing country such as Ghana. The 0.75% GDP is of particular interest since it closely matches the required investment for the trend of additional generation capacity needs reported in the annual electricity supply plan of Ghana (based on investment cost of a natural gas generation plant). 0 1000 2000 3000 4000 G en er at or C ap ac ity (M W ) Generator Type New Capacity Additions under No Renewable Target at 0.5% GDP 2027-2030 2024-2026 2021-2023 2019-2020 0 1000 2000 3000 4000 5000 6000 G en er at or C ap ac ity (M W ) Generator Type New Capacity Additions under No Renewable Target at 0.75% GDP 2027-2030 2024-2026 2021-2023 2019-2020 rs of 2020, 2023, 2026, and 2030 under both Target and No Target policies for generation ities of recommend new generator types during the sub-target years of 2020, 2023, 2026, alent of 1.0%, and 1.25% of Ghana's annual GDP. c. Cumulative capacities of recommend ent budget for generation capacity investment. The graph is the same under both Target 0 1000 2000 3000 4000 G en er at or C ap ac ity (M W ) Generator Type New Capacity Additions under Renewable Target at 1.0% GDP 2027-2030 2024-2026 2021-2023 2019-2020 0 1000 2000 3000 4000 5000 G en er at or C ap ac ity (M W ) Generator Type New Capacity Additions under No Renewable Target at 1.0% GDP 2027-2030 2024-2026 2021-2023 2019-2020 0 500 1000 1500 2000 2500 3000 3500 G en er at or C ap ac ity (M W Generator Type New Capacity Additions under Renewable Target at 1.25% GDP 2027-2030 2024-2026 2021-2023 2019-2020 0 1000 2000 3000 4000 5000 6000 7000 8000 G en er at or C ap ac ity (M W ) Generator Type New Capacity Additions under No Renewable Target at 1.25% GDP 2027-2030 2024-2026 2021-2023 2019-2020 b Fig. 3 (continued). A. Afful-Dadzie, E. Afful-Dadzie, N.A. Abbey et al. Energy for Sustainable Development 59 (2020) 83–96 Cost of electricity provision The total cost of providing electricitywithin the planningperiodunder both the Target and the No Target policies for the four budget allocations, and when there is sufficient budget are presented in Table 9. The term Total Cost gives the total cost of electricity provision (including cost of unmet demand) for the Target and No Target policies. The annual ex- penses for capacity investment was not considered in the Total Cost anal- ysis since it is meant (in theory) to be recovered by the end of the plant's life through the capacity charge component in the objective function. A 0 4000 8000 12000 16000 20000 G en er at or C ap ac ity (M W ) Generator T Cumulative New Capacity Addit Sufficien c Fig. 3 (cont 92 plot of the total costs for the selected budget allocations, and under suffi- cient budget is shown in Fig. 5. It can be seen that total cost of electricity provision is higher under the Target policy thanunder theNoTargetpolicy under all budget allocations, signifying an increase in cost for the 10%elec- tricity generation target policy. The term Extra Cost for Enforcing the Target gives the difference between the total cost of electricity provision under the Target and No Target policies. These are extra cost incurred for enforcing the renewable target policy. The extra cost in an annual basis using the discounted rate of 12% per year is also given in column 5 of Table 9. It can be observed that the extra costs are significant enough to ype ions under Renewable Target at t Budget 2027-2030 2024-2026 2021-2023 2019-2020 inued). Table 7 Percentage of electricity generation from renewable generators at sub-target due dateswhen a 10% renewable electricity generation target is enforced (Target), andwhennot enforced (No Target) at annual capacity addition funds of 0.50%, 0.75%, 1.0%, and 1.25% of Ghana's GDP. Year 0.5% of GDP 0.75% of GDP 1.0% of GDP 1.25% of GDP Sufficient budget Target achieved (%) Target achieved (%) Target achieved (%) Target achieved (%) Target achieved (%) Target No target Target No target Target No target Target No target Target No target 2020 1.7 0.4 1.3 0.4 1.7 0.4 1.2 0.4 33.1 33.1 2023 4.0 0.3 4.1 0.3 4.5 0.3 4.2 0.3 50.3 50.3 2026 7.0 0.3 7.0 0.3 7.2 1.1 7.2 3.7 66.1 66.1 2030 10.0 0.3 10.1 0.3 10.3 1.5 10.6 4.7 72.8 72.8 Table 8 Percentage levels of unmet demandwhen a 10% renewable electricity generation target is enforced (Target), and when not enforced (No Target) at annual generation capacity addi- tion funds of 0.50%, 0.75%, 1.0%, and 1.25% of Ghana's GDP. Year 0.5% of GDP 0.75% of GDP 1% of GDP 1.25% of GDP Target No target Target No target Target No target Target No target 2019 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 2020 6.2 3.7 3.4 3.7 3.4 2.6 2.5 2.1 2021 5.6 2.8 2.4 0.8 0.5 0.4 0.2 0.1 2022 5.8 4.4 1.3 0.8 0.0 0.0 0.0 0.0 2023 19.1 6.5 5.3 1.4 1.5 0.4 1.0 0.0 2024 13.1 7.9 4.5 2.2 1.0 0.1 0.1 0.1 2025 13.7 8.6 4.0 1.5 0.7 0.0 0.0 0.0 2026 27.0 10.8 11.3 1.8 3.9 0.5 1.8 0.2 2027 18.2 9.4 5.4 1.3 1.1 0.0 0.0 0.0 2028 20.0 11.6 6.4 1.8 1.2 0.0 0.0 0.0 2029 23.5 12.4 9.3 2.3 1.9 0.0 0.0 0.0 2030 32.2 14.6 14.0 2.3 4.0 0.0 0.6 0.0 0% 10% 20% 30% 40% 2020 2023 2026 2030 Pe rc en t o f D em an d U nm et (G W h) Year Comparison of Percent of Demand Unmet at Annual Capacity Financing of 0.5% GDP 10% RE Target No RE Target 0% 2% 4% 6% 2020 2023 2026 2030 Pe rc en t o f D em an d U nm et (G W h) Year Comparison of Percent of Demand Unmet at Annual Capacity Financing of 1.0% GDP 10% RE Target No RE Target Pe rc en t o f D em an d U nm et (G W h) Pe rc en t o f D em an d U nm et (G W h) Fig. 4. Comparison of percentage levels of unmet demand when a 10% renewable electricity ge addition financing budget of 0.75%, 1.0%, 1.25%, and 1.5% of Ghana's GDP during the sub-target A. Afful-Dadzie, E. Afful-Dadzie, N.A. Abbey et al. Energy for Sustainable Development 59 (2020) 83–96 93 discourage a developing country from attempting to increase the renew- able content of its electricity consumption. However, it can also be seen that the total cost (and also extra cost) decreases with increasing budget allocations. The last row of Table 9 gives the cost of electricity provision when there is sufficient financial resource to finance new generation capacity additions. In this case, the cost of electricity provision is the same under both the Target and No Target policies. However, this cost is less than that of 1.25% GDP budget allocation. The difference is as high as US $1396.9 Billion and US$1167.6 Billion when under the Target and No Target policies respectively over the planning period. This translates to a savings of US$225.5 Million and US$188.5 Million annually over the 1.25% GDP budget allocation. These savings are realized partly from the zero operational cost of the massive renewable energy generators recommended under sufficient budget. However, note that this can only be realized if one can source the enormous financial resources needed to purchase the needed capacity to meet the expected demand. In 2020 for instance, the capacity financing required will be almost US $2.54 Billion, roughly 3.8% of Ghana's projected GDP used in this study. 0.0% 0.5% 1.0% 1.5% 2.0% 2.5% 3.0% 2020 2023 2026 2030 Year Comparison of Percent of Demand Unmet at Annual Capacity Financing of 1.25% GDP 10% RE Target No RE Target 0% 2% 4% 6% 8% 10% 12% 14% 16% 2020 2023 2026 2030 Year Comparison of Percent of Demand Unmet at Annual Capacity Financing of 0.75% GDP 10% RE Target No RE Target neration target is enforced (Target), and when not enforced (No Target) at annual capacity years of 2020, 2023, 2026, and 2030. Table 9 Cost of electricity provisionwhena 10% renewable electricity generation target is enforced (Target), and when not enforced (No Target) at annual capacity addition funds of 0.50%, 0.75%, 1.0%, and 1.25% of Ghana's GDP, and at sufficient budget, during the planning period from 2019 to 2030. Annual capacity financing Total cost (millions) Extra cost for enforcing target (millions) Target No target From 2019 to 2030 Annual 0.5% of GDP $16,262.24 $14,639.91 $1622.33 $261.90 0.75% of GDP $14,035.56 $12,648.40 $1387.16 $223.94 1.0% of GDP $12,389.61 $11,663.64 $725.97 $117.20 1.25% of GDP $11,446.86 $11,217.63 $229.24 $37.01 Sufficient budget $10,050.00 $10,050.00 $0.00 $0.00 A. Afful-Dadzie, E. Afful-Dadzie, N.A. Abbey et al. Energy for Sustainable Development 59 (2020) 83–96 In summary, and according to current cost figures, the cost incurred for transitioning from conventional energy to renewable energy isman- ageable if a country has sufficient financial resources for generation ca- pacity investment. This is not a surprise, since the main hindrance to renewable energy adoption is capital cost. This also supports the evi- dence on the ground where (perhaps with the exception of Brazil and India), it appears that the countries making significant headways in the energy transition process are those having relatively stronger econ- omies (e.g. Germany, United States, China, and countries from the European Union). Conclusion and policy implications This paper developed an electricity Generation Expansion Plan- ning (GEP) model for the setting of renewable electricity generation target in developing countries. To do this, two important constraints in the form of a target on the percentage of electricity generation from renewable energy sources, and a limit on funds for new gener- ation capacity investments were incorporated into a standard GEP model. The model was developed in the form of a multi-period sto- chastic mixed integer linear program (MILP) to allow for long-term planning as well as to account for uncertainty in model parameters (electricity demand and fuel price). Additionally, the objective func- tion was designed to help determine cost of electricity provision, in- cluding cost of unmet demand when a renewable electricity target is adhered to. Applying the model to the case of Ghana's 10% renewable electricity target policy, revealed insightful lessons. First and foremost, the cost of electricity from some renewable generators are on the whole cost $8,000 $10,000 $12,000 $14,000 $16,000 $18,000 $20,000 0.5% of GDP 0.75% of GDP 1% of GDP 1.25% of GDP Sufficient Budget C os t o f E le ct ri ci ty P ro vi si on ($ M ill io ns ) Annual Budget for New Generation Capacity Cost of Electricity Provision under Target and No-Target condition Target No Target Fig. 5. Plot of cost of electricity provision under a 10% renewable electricity generation target and under a no target policy at annual generation capacity addition financing budget equivalent to 0.5%, 0.75%, 1.0%, and 1.25% of Ghana's GDP, and at sufficient budget, during the planning period from 2019 to 2030. 94 competitive with conventional energy generators. This is in agree- ment with the current Levelized Cost of Electricity (LCOE) estimates from the International Renewable Energy Agency (IRENA). However, this is only true when there is sufficient financial resource available to fund the needed generation capacity. Otherwise, if unable to raise such funds, one risk facing high levels of unmet demand (due to capacity shortages), as well as increase in cost of electricity provi- sion when factoring in economic losses due to unmet demand. In the case of Ghana, it might need to allocate at least 1.0% of its GDP (which is approximately 25% increase of current annual capacity addition expenses) to capacity financing in order to meet its 10% renewable electricity generation target by 2030. This is the only way possible if it wishes to keep levels of unmet demand and cost of electricity provision at reasonable levels. Sticking to current expenditure levels while attempting to achieve the target can lead to levels of unmet demand that could be as high as 14% in 2030, and cost of electricity provision rising by as much as $224 million per year. It is therefore not a surprise that Ghana has not been able to achieve even 0.3% of its original 10% target set for 2020 (now scheduled to 2030). In sum, any developing country contemplating on increasing the share of renewable electricity in its electricity supply mix must be adequately prepared to increase the amount it spends on financing new generation capacity additions. The second insight follows from the first, which is that, without enough financial resources the only way for developing countries to improve their already low electricity access rate, as well as meet the increasing growth in electricity demand is to, rather unfortu- nately, continue to use conventional energy sources. This places de- veloping countries at a disadvantage in the world's quest to transition from fossil fuel based energy to renewable energy. This paper contributes to literature, particularly in the research area of energy transition by providing a unique GEP model for the setting and evaluation of a renewable electricity generation target policies. The paper also provides a cost-benefit analysis procedure for estimating the additional cost or benefit related to the cost of electricity provision for a proposed renewable electricity production target. The proposed GEP model also offers a number of practical contribu- tions. The model can be used to search for the reasonable target that is commensurate to a country's financial position, and the level of unmet electricity demand and extra cost it is willing to tolerate. Such a target will have a high chance of being achieved. Similarly, for a proposed tar- get in mind, the model will allow for the assessment of the financial commitment needed, and the expected level of unmet electricity de- mand and cost of electricity provision. The special features of the pro- posed model not only fits electricity planning practices in developing countries, but also brings to bare matters of greater concern in these countries as far as energy transition is concerned - the inclusion of a budget constraint captures the struggle to raise the needed capital for new generation capacity, whereas the provision for the estimation of level of unmet energy helps to foresee the impact on the already dire levels of unmet demand. Finally, the model's output in terms of the rec- ommended generator types, their size, and timing will serve as a guide for cost-effective electricity planning towards the achievement of a planned renewable electricity target. A limitation of this research is that due to the unavailability of local cost data, most plant costs were taken from international reports such as the Energy Information Administration of the U.S. Plant costs data such as capital cost, fixed cost, and operational cost tend to be higher in developing countries than in the U.S., and could lead to higher elec- tricity cost overall. Yet, we also note that renewable energy auctions in some countries such as South Africa have led to the procurement of renewable energy capacities at competitive and at times, much lower costs than previously assumed (Eberhard, 2013). Thus, obtaining credi- ble local plant cost data is important for policy analysis using the pro- posed model. A. Afful-Dadzie, E. Afful-Dadzie, N.A. Abbey et al. Energy for Sustainable Development 59 (2020) 83–96 A potential related future work could look into directly modeling the intermittency of the energy sources for the renewable generators to estimate the level of back-up capacity needed to prevent system failure. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influ- ence the work reported in this paper. Acknowledgement We thank the Queen Elizabeth Scholarship program for Climate Change and Societal Transformation, the host university, Carleton Uni- versity, Ottawa, Canada, and the Centre for Climate Change and Sustain- ability Studies (C3SS) at the University of Ghana, for the research mobility scholarship that facilitated the preparation of this manuscript. We also thank the University of Ghana Business School for the award of the research grant #05570919 that provided the needed funding for this research. Appendix A. A description of the estimation of the marginal cost of fuel for electricity generation, and plots of cumulative capacities of recommend new generator types for electricity generation Table A.1 Summary of electricity price estimation for coal, natural gas, and biomass. Fuel Price estimation Explanation and conversion to $/MWh Coal prices ($/tonne) 27 + 101 ∗ BETA (0.655, 0.848) There is a proposed 700 MW coal plant project between government of Ghana and Shenzhen Energy group of China. A Chinese company will build the proposed coal plant: therefore, coal price (in $/tonne) data from 2000 to 2016 were obtained from the BP Statistical Review of World Energy (2017) using the China Qinhuagdao coal prices (which includes delivery price). Using the Arena Input Analyzer, the best distribution fit for coal price was a Beta distribution of the form 27 + 101 ∗ BETA (0.655, 0.848). This value is then converted to $/MWh by multi- plying by 0.538. Gas prices ($/MMBtu) Market Price: 2 + GAMMA (1.89, 1.28) Delivery: 4.8 The natural gas price is made up of two components; market price and delivery cost. The market price component was derived based on the Henry Hub monthly natural gas prices from 1997 to 2017 (EIA, 2017). Using the Arena Input Analyzer, and the annual average of the monthly prices, the best distribution fit for the market price was a Gamma distribution of the form 2 + GAMMA (1.89, 1.28). The delivery cost is based on data from the Ghana Energy Commission and was taken to be $4.8/MMBtu. The total estimated price was then converted to $/MWh by multiplying by an amount of 6.6. Biomass prices ($/MWh) NORM (50, 5) Biomass prices were assume to follow a normal distribution with a mean of $50/MWh and a standard deviation of $5/MWh based on projected imported biomass feedstock prices for the UK heat sector (E4tech, 2010). References Afful-Dadzie, A., Afful-Dadzie, E., Awudu, I., & Banuro, J. K. (2017). Power generation ca- pacity planning under budget constraint in developing countries. Applied Energy, 188, 71–82. Africa Progress Panel (2015). Power people planet: Seizing Africa’s energy and climate opportunities. Africa Progress Report, 2015. 95 Arnette, A., & Zobel, C. W. (2012). An optimization model for regional renewable energy development. Renewable and Sustainable Energy Reviews, 16(7), 4606–4615. Awudu, I., & Zhang, J. (2013). Stochastic production planning for a biofuel supply chain under demand and price uncertainties. Applied Energy, 103, 189–196. Batinge, B., Musango, J. K., & Brent, A. C. (2019). Sustainable energy transition framework for unmet electricity markets. Energy Policy, 129, 1090–1099 (23). BFT (2017, June). Government shifts renewable energy target; 10 percent target shifted to 2030. The Business & Financial Times Online. [accessed 03.18.19]. Breeden, J. L., & Ingram, D. (2010). Monte Carlo scenario generation for retail loan portfo- lios. Journal of the Operational Research Society, 61(3), 399–410. E4tech (2010, January). Biomass prices in the heat and electricity sectors in the UK. Doc- ument prepared for the Department of Energy and Climate Change. http://www. e4tech.com/wp-content/uploads/2016/01/100201Biomass_prices.pdf> [accessed 01.15.19]. Eberhard, A. (2013). Feed-in tariffs or auctions? Procuring renewable energy supply in South Africa. Viewpoint. Public policy for the private sector. Note, (338). http:// documents.albankaldawli.org/curated/ar/990081468305682402/pdf/ 779710BRI0Box300Tariffs0or0Auctions.pdf [accessed 06.16.2020]. Eberhard, A., Foster, V., Briceño-Garmendia, C., Ouedraogo, F., Camos, D., & Shkaratan, M. (2008). Underpowered: The state of the power sector in Sub-Saharan Africa. ECG (2013). National energy statistics 2000–2012. Ghana: The Energy Commissionhttp:// www.energycom.gov.gh/files/Ghana_Energy_Statistics_2012_AUG.pdf [accessed 04. 20.19]. ECG (2019a). Ghana Renewable Energy Master Plan.Ghana: The Energy Commissionhttp:// www.energycom.gov.gh/files/Renewable-Energy-Masterplan-February-2019.pdf [accessed 04.27.19]. ECG (2019b). National energy statistics 2008–2017. Ghana: The Energy Commissionhttp:// www.energycom.gov.gh/files/ENERGY_STATISTICS_2018_FINAL.pdf [accessed 04.11. 19]. ECG (2019c). The 2019 electricity supply plan for the Ghana power system. Ghana: The En- ergy Commissionhttp://www.energycom.gov.gh/files/2019%20Electricity%20Supply %20Plan.pdf [accessed 04.27.19]. EIA (2013, April). Updated Capital Cost Estimates for Utility Scale Electricity Generating Plants. Energy Information Administrationhttps://www.eia.gov/outlooks/capitalcost/ pdf/updated_capcost.pdf(Accessed 14 May 2019). . EIA (2016, November). Capital cost of utility scale electricity generating plants. Energy Infor- mation Administrationhttps://www.eia.gov/analysis/studies/powerplants/capita lcost/pdf/capcost_assumption.pdf [accessed 05.12.19]. EIA (2017, October). Henry Hub Natural Gas Spot Prices. https://www.eia.gov/dnav/ng/ hist/rngwhhdm.htm [accessed 20.10.17]. EIA (2018, August). Construction cost data for electric generators installed in 2016. En- ergy Information Administration. [accessed 04.13.19]. EIA (2019, January). Cost and performance characteristics of new generating technolo- gies. energy information administration, Annual Energy Outlook 2019. [accessed 03.18.19]. EIA (2019b). Electric power monthly. https://www.eia.gov/electricity/monthly/current_ month/epm.pdf; 2019. [accessed 4 April 2019]. Feng, Y., & Ryan, S. M. (2013). Scenario construction and reduction applied to stochastic power generation expansion planning. Computers & Operations Research, 40(1), 9–23. Frondel, M., Ritter, N., Schmidt, C. M., & Vance, C. (2010). Economic impacts from the pro- motion of renewable energy technologies: the German experience. Energy Policy, 38 (8), 4048–4056. Geels, F. W., Sovacool, B. K., Schwanen, T., & Sorrell, S. (2017). The socio-technical dynam- ics of low-carbon transitions. Joule, 1(3), 463–479. Green, R., & Staffell, I. (2016). Electricity in Europe: exiting fossil fuels? Oxford Review of Economic Policy, 32(2), 282–303. GridCo (2011). Generation master plan study for Ghana. Ghana Grid Company Limited Re- trieved from http://www.gridcogh.com/media/photos/forms/annual/2011% 20GRIDCo%20Annual%20Report.pdf. GridCo (2017, January). Electricity supply plan for the Ghana power system. Ghana Grid Company Limitedhttp://www.gridcogh.com/media/photos/forms/supplyplan/2018_ Electricity_Supply_Plan.pdf [accessed 23.03.19]. Hammons, T. J. (2008). Integrating renewable energy sources into European grids. International Journal of Electrical Power & Energy Systems, 30(8), 462–475. Hassel, A., Egenhofer, C., Nicolescu, R., Nica, A., & Elisei, S. (2017). Fulfilment of National Ob- jectives under the Renewable Energy Directive: State of play and projections. CEPS policy insights no. 2017/04, February 2017. IEA (2017, October). Renewables 2017: Analysis and forecast to 2022. International Energy Agencyhttps://www.iea.org/Textbase/npsum/renew2017MRSsum.pdf [accessed 04. 16.19]. IRENA (2012). Renewable energy technologies: Cost analysis series (wind power). Inter- national Renewable Energy Agency (IRENA) Working Paper Series 5(5). [accessed 04.12.19]. IRENA (2015a). Renewable power generation costs in 2014. International Renewable En- ergy Agency (IRENA)https://www.irena.org/documentdownloads/publications/ irena_re_power_costs_2014_report.pdf [accessed 04.18.19]. IRENA (2015b). Renewable energy target setting. International Renewable Energy Agency (IRENA)http://www.irena.org/-/media/Files/IRENA/Agency/Publication/2015/ IRENA_RE_Target_Setting_2015.pdf [accessed 04.23.19]. IRENA (2017). Renewable capacity statistics 2017. International Renewable Energy Agency (IRENA), Abu Dhabi. Retrieved April 20 2019 from https://www.irena.org/ http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0005 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0005 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0005 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0010 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0010 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0015 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0015 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0020 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0020 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0025 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0025 https://www.ghanaweb.com/GhanaHomePage/business/Government-shifts-renewable-energy-target-10-percent-target-shifted-to-2030-544475 https://www.ghanaweb.com/GhanaHomePage/business/Government-shifts-renewable-energy-target-10-percent-target-shifted-to-2030-544475 https://www.ghanaweb.com/GhanaHomePage/business/Government-shifts-renewable-energy-target-10-percent-target-shifted-to-2030-544475 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0030 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0030 http://www.e4tech.com/wp-content/uploads/2016/01/100201Biomass_prices.pdf%3e http://www.e4tech.com/wp-content/uploads/2016/01/100201Biomass_prices.pdf%3e http://documents.albankaldawli.org/curated/ar/990081468305682402/pdf/779710BRI0Box300Tariffs0or0Auctions.pdf http://documents.albankaldawli.org/curated/ar/990081468305682402/pdf/779710BRI0Box300Tariffs0or0Auctions.pdf http://documents.albankaldawli.org/curated/ar/990081468305682402/pdf/779710BRI0Box300Tariffs0or0Auctions.pdf http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0040 http://www.energycom.gov.gh/files/Ghana_Energy_Statistics_2012_AUG.pdf http://www.energycom.gov.gh/files/Ghana_Energy_Statistics_2012_AUG.pdf http://www.energycom.gov.gh/files/Renewable-Energy-Masterplan-February-2019.pdf http://www.energycom.gov.gh/files/Renewable-Energy-Masterplan-February-2019.pdf http://www.energycom.gov.gh/files/ENERGY_STATISTICS_2018_FINAL.pdf http://www.energycom.gov.gh/files/ENERGY_STATISTICS_2018_FINAL.pdf http://www.energycom.gov.gh/files/2019%20Electricity%20Supply%20Plan.pdf http://www.energycom.gov.gh/files/2019%20Electricity%20Supply%20Plan.pdf https://www.eia.gov/outlooks/capitalcost/pdf/updated_capcost.pdf https://www.eia.gov/outlooks/capitalcost/pdf/updated_capcost.pdf https://www.eia.gov/analysis/studies/powerplants/capitalcost/pdf/capcost_assumption.pdf https://www.eia.gov/analysis/studies/powerplants/capitalcost/pdf/capcost_assumption.pdf https://www.eia.gov/dnav/ng/hist/rngwhhdm.htm https://www.eia.gov/dnav/ng/hist/rngwhhdm.htm https://www.eia.gov/electricity/generatorcosts/ https://www.eia.gov/outlooks/aeo/assumptions/pdf/table_8.2.pdf/ https://www.eia.gov/outlooks/aeo/assumptions/pdf/table_8.2.pdf/ https://www.eia.gov/electricity/monthly/current_month/epm.pdf; https://www.eia.gov/electricity/monthly/current_month/epm.pdf; http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0075 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0075 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0080 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0080 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0080 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0085 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0085 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0090 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0090 http://www.gridcogh.com/media/photos/forms/annual/2011%20GRIDCo%20Annual%20Report.pdf http://www.gridcogh.com/media/photos/forms/annual/2011%20GRIDCo%20Annual%20Report.pdf http://www.gridcogh.com/media/photos/forms/supplyplan/2018_Electricity_Supply_Plan.pdf http://www.gridcogh.com/media/photos/forms/supplyplan/2018_Electricity_Supply_Plan.pdf http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0105 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0105 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0110 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0110 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0110 https://www.iea.org/Textbase/npsum/renew2017MRSsum.pdf https://www.irena.org/documentdownloads/publications/re_technologies_cost_analysis-wind_power.pdf https://www.irena.org/documentdownloads/publications/re_technologies_cost_analysis-wind_power.pdf https://www.irena.org/documentdownloads/publications/re_technologies_cost_analysis-wind_power.pdf https://www.irena.org/documentdownloads/publications/irena_re_power_costs_2014_report.pdf https://www.irena.org/documentdownloads/publications/irena_re_power_costs_2014_report.pdf http://www.irena.org/-/media/Files/IRENA/Agency/Publication/2015/IRENA_RE_Target_Setting_2015.pdf http://www.irena.org/-/media/Files/IRENA/Agency/Publication/2015/IRENA_RE_Target_Setting_2015.pdf https://www.irena.org/DocumentDownloads/Publications/IRENA_RE_Capacity_Statistics_2016.pdf A. Afful-Dadzie, E. Afful-Dadzie, N.A. Abbey et al. Energy for Sustainable Development 59 (2020) 83–96 DocumentDownloads/Publications/IRENA_RE_Capacity_Statistics_2016.pdf [accessed 04.10.19]. IRENA (2018). Planning and prospects for renewable power: West Africa. Abu Dhabi: Inter- national Renewable Energy Agencyhttps://www.irena.org/-/media/Files/IRENA/ Agency/Publication/2018/Nov/IRENA_Planning_West_Africa_2018.pdf 2018. [accessed 8 May 2019]. Jin, S., Ryan, S. M., Watson, J. P., & Woodruff, D. L. (2011). Modeling and solving a large- scale generation expansion planning problem under uncertainty. Energy Systems, 2 (3–4), 209–242. Klessmann, C., Held, A., Rathmann, M., & Ragwitz, M. (2011). Status and perspectives of renewable energy policy and deployment in the European Union—what is needed to reach the 2020 targets? Energy Policy, 39(12), 7637–7657. Knopf, B., Nahmmacher, P., & Schmid, E. (2015). The European renewable energy target for 2030–an impact assessment of the electricity sector. Energy Policy, 85, 50–60. Kumi, E. N. (2017). The electricity situation in Ghana: Challenges and opportunities. Wash- ington, DC: Center for Global Development. Muis, Z. A., Hashim, H., Manan, Z. A., Taha, F. M., & Douglas, P. L. (2010). Optimal planning of renewable energy-integrated electricity generation schemes with CO2 reduction target. Renewable Energy, 35(11), 2562–2570. Obeng-Darko, N. A. (2019). Why Ghana will not achieve its renewable energy target for electricity. Policy, legal and regulatory implications. Energy Policy, 128, 75–83. Oree, V., Hassen, S. Z. S., & Fleming, P. J. (2017). Generation expansion planning optimisa- tion with renewable energy integration: a review. Renewable and Sustainable Energy Reviews, 69, 790–803. 96 Oyewunmi, T., Heffron, R. J., & Crossley, P. (2019). OGEL special issue on “energy law and regulation in low-carbon and transitional energy markets”. Oil, Gas & Energy Law Journal (OGEL), 17(1). Pink, R. M. (2018). The climate change crisis: Solutions and adaption for a planet in peril. Springer. Sawin, J. L., Rutovitz, J., & Sverrisson, F. (2018). Renewables 2018 global status report (2018th ed.). Paris: Canadian Electronic Library. Schaber, K., Steinke, F., & Hamacher, T. (2012). Transmission grid extensions for the inte- gration of variable renewable energies in Europe: who benefits where? Energy Policy, 43, 123–135. Shiina, T., & Birge, J. R. (2003). Multistage stochastic programming model for electric power capacity expansion problem. Japan Journal of Industrial and Applied Mathematics, 20(3), 379–397. Smyth, J. (2017, October 17). Australia to Abandon Clean Energy Target. Financial Timeshttps://www.ft.com/content/37854ff0-b2fd-11e7-a398-73d59db9e399 [accessed 01.18.19]. Tang, A., Chiara, N., & Taylor, J. E. (2012). Financing renewable energy infrastructure: Formulation, pricing and impact of a carbon revenue bond. Energy Policy, 45, 691–703. Zhang, H. L., Van Gerven, T., Baeyens, J., & Degrève, J. (2014). Photovoltaics: reviewing the European feed-in-tariffs and changing PV efficiencies and costs. The Scientific World Journal, 2014. https://www.irena.org/DocumentDownloads/Publications/IRENA_RE_Capacity_Statistics_2016.pdf https://www.irena.org/-/media/Files/IRENA/Agency/Publication/2018/Nov/IRENA_Planning_West_Africa_2018.pdf https://www.irena.org/-/media/Files/IRENA/Agency/Publication/2018/Nov/IRENA_Planning_West_Africa_2018.pdf http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0135 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0135 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0135 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0140 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0140 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0140 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0145 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0145 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0150 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0150 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0155 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0155 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0155 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0155 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0160 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0160 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0165 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0165 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0165 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0170 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0170 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0170 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0175 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0175 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0180 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0180 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0185 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0185 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0185 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0190 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0190 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0190 https://www.ft.com/content/37854ff0-b2fd-11e7-a398-73d59db9e399 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0200 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0200 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0200 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0210 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0210 http://refhub.elsevier.com/S0973-0826(20)30295-7/rf0210 Renewable electricity generation target setting in developing countries: Modeling, policy, and analysis Introduction Nomenclature Indices Parameters Uncertain parameters Decision variables Generation expansion planning model with renewable electricity generation target Decision variables Objective function Model constraints Electricity demand and supply Generation capacity Renewable electricity target Capacity investment budget constraints Case study on Ghana's 10% renewable electricity generation target policy Generation capacity Electricity demand trend The Ghana renewable energy master plan Model parameters Electricity demand Generator technical information Capital, fixed operating and maintenance (O&M), and emissions costs Marginal cost of generation Capacity investment budget constraint Other assumptions Case study results and analysis New generation capacity additions Percentage of electricity generated Level of unmet demand Cost of electricity provision Conclusion and policy implications Declaration of competing interest section33 Acknowledgement Appendix A. A description of the estimation of the marginal cost of fuel for electricity generation, and plots of cumulativ... References