What constrains renewable energy investment in Sub-Saharan Africa? A comparison of Kenya and Ghana

World Development 109 (2018) 85–100

Contents lists available at ScienceDirect

World Development journal homepage: www.elsevier.com/locate/worlddev

What constrains renewable energy investment in Sub-Saharan Africa? A comparison of Kenya and Ghana Ana Pueyo Institute of Development Studies, Library Road, University of Sussex, Brighton BN1 9RE, United Kingdom

a r t i c l e

i n f o

Article history: Accepted 12 April 2018

Keywords: Renewable energy Investment Africa

a b s t r a c t Policymakers in Sub-Saharan Africa face several choices to increase levels of access to electricity under severe budget constraints. First, they need to prioritise technologies that can supply electricity at a low cost. Second, they need to design and implement appropriate policies to attract private investment. On the first choice, renewable energy is becoming increasingly competitive with fossil fuels. Moreover, it contributes to energy security and environmental sustainability, while providing access to new sources of (sustainable) finance. On the second choice, developing countries typically face a multitude of constraints to attract investment to their energy sector. It can be daunting and expensive to address them all at once. This paper presents a new methodology to support policymakers to better target policies for the promotion of commercial-scale renewable energy investment. The methodology, which we call ‘‘Green Investment Diagnostics” draws upon the Growth Diagnostics framework, extensively used in the field of Development Economics to identify the binding constraints to economic growth. It is operationalised with a decision tree analysis that builds cumulative evidence to prioritise some constraints over others, through the review of indicators and validation through expert interviews. We apply this approach to Kenya and Ghana, finding that Ghana’s key constraints to investment in renewable energy are an unreliable off-taker, macroeconomic imbalances, regulatory uncertainty, pressures to keep prices low, as well as insufficient and costly domestic finance. Kenya instead offers generous returns to investment in renewables but faces a low demand, a lack of networking infrastructure and problems of governance and social acceptance, exacerbated by uncertain land property rights and rent-seeking. Ó 2018 Elsevier Ltd. All rights reserved.

1. Introduction Renewables are becoming mainstream sources of electricity in countries at all levels of income in all parts of the world. For the first time, 2015 saw more renewable capacity added globally than fossil fuel sources and in the same year renewable energy (RE) investment in developing countries exceeded that of developed countries (Frankfurt School & BNEF, 2017). Despite this undoubted progress, many developing countries continue to struggle to attract investment for generation capacity in general and for renewables in particular. The gap is particularly acute in Sub-Saharan Africa (SSA), a region hosting more than 55% of the 1.2 billion people without access to electricity globally (World Bank, 2017a). SSA possesses abundant RE resources which, if tapped, could fundamentally change the region’s development prospects. However, the region has been historically neglected by energy investors, and its rich RE resources remain underutilised (Pueyo, Orraca & Godfrey-Woods, 2015). The literature is replete with E-mail address: [email protected] https://doi.org/10.1016/j.worlddev.2018.04.008 0305-750X/Ó 2018 Elsevier Ltd. All rights reserved.

attempts to understand why this situation persists. In a broad sense, shortages of capital, skills, and governance capacity prevent Africa from using its plentiful RE resources (Collier & Venables, 2012). More specific problems include: under-pricing, financial weakness of electric utilities, corruption and patronage, flawed and uncertain regulation, low savings rates, poorly developed financial markets, low and dispersed demand, high transmission losses, constricted power generation planning, and lack of technological knowledge (Gratwick & Eberhard, 2008; Collier & Venables, 2012; Suberu, Mustafa, Bashir, Muhamad & Mokhtar, 2013). As well as highlighting a multitude of constraints, academics and practitioners have proposed a commensurate amount of policies to address them. From the 1990s, development finance institutions promoted the so called ‘‘standard model of power sector reform” in SSA to deal with the failures of centralised power utilities and raise private finance for capacity expansion (Malgas & Eberhard, 2011). The standard model, pioneered in the United Kingdom, the United States, Chile and Norway, prescribed unbundling, liberalisation, competition, and independent regulatory oversight to improve efficiency and transparency. In reality, the

86

A. Pueyo / World Development 109 (2018) 85–100

standard model has not been fully implemented by any African country. What has emerged instead are hybrid power systems, where incumbent state-owned utilities remain dominant but independent power producers (IPPs) are used to fill the financing gaps (Gratwick & Eberhard, 2008). These hybrid systems tend to be very inefficient and have failed to attract private investment at anything like the levels needed. A more recent wave of reform targets RE investment specifically. For example, donors have promoted Feed-in-tariffs (FiT) as cornerstone instruments to guarantee long-term, fixed-price electricity purchase agreements and grid access (Waissbein, Glemarec, Bayraktar, & Schmidt, 2013). FiT have been successfully implemented in many Northern and Southern countries, such as Germany and China, where they have contributed to the widespread deployment and dramatic cost reduction of wind and solar photovoltaic (PV) technologies (Hoppmann, Huenteler, & Girod, 2014). As of 2017, eight SSA countries had enacted FiT (REN21, 2017), but most of them had not succeeded in attracting the intended private investment. In many cases, there was not a single project benefitting from FiT several years after these were approved. The previous examples illustrate how policies that work wonders in some countries can be ineffective in others. This research, therefore, starts from the position that simply attempting to replicate successful policies from other countries is unlikely to bring much needed investment for Africa’s renewable electricity generation. Worse, attempting to do so will see scarce political and financial resources wasted. The same resources could achieve more if they targeted the most important constraints to investment in each country at the particular time when reform is implemented. If this is so, the task is to find a way to prioritise between the many problems that can usually be identified. This rationale underpins the development of the Green Investment Diagnostics methodology, which seeks to systematically identify the binding constraints to investment in RE. Our method is an adaptation of the Growth Diagnostics framework developed by Haussman, Rodrik and Velasco (HRV) (2004) to identify the key constraints holding back economic growth. Their approach was driven by the needs of policymakers facing multiple causes but lacking the ability either to tackle them all at once or to prioritise between them. The solution proposed was to identify and concentrate limited resources on the ‘‘binding constraint”, which would be identified with a tool conceptualised as a decision tree. Growth diagnostics have been widely applied by multilateral organisations to better target support for growth promotion. ‘Green Investment Diagnostics’ adapts this national approach for the sectoral analysis of the electricity sector, with a particular focus on commercial-scale, grid connected generation. The constraints faced by investors in small scale, off-grid generation are different in many respects and would require a separate analysis. This paper pilots the new methodology in Kenya and Ghana, showcasing the choices the SSA region faces as it aims to provide universal access to affordable energy, foster economic growth, and reduce the vulnerability to imported fuels and erratic hydrological resources. Kenya and Ghana offer an interesting comparison, as their power systems face very different challenges. Kenya’s power system has recently achieved financial sustainability, and it has a large share of RE, but demand remains among the lowest in Africa. On the other hand, the share of renewables other than large hydro is negligible in Ghana, and its power system is struggling financially. However, Ghana has reached one of the largest connection rates in SSA, even if the quality of that access has fast deteriorated. The remainder of this paper provides some background about renewable electricity in Kenya and Ghana, in Section 2. Section 3 describes the green investment diagnostics methodology. Section 4 applies the methodology to the target countries. Section 5

discusses and compares the results of the analysis, and Section 6 concludes by inviting policymakers and donors to replicate the approach in other countries and regions. 2. Background of renewable electricity in kenya and ghana The electricity sectors of Kenya and Ghana represent the many challenges that countries in SSA encounter as they try to build green and inclusive economies. Both countries aim at providing universal access to affordable electricity, fostering economic growth, and reducing the vulnerability of their power systems to erratic rainfall patterns and imported fuels. Both have also committed to climate change mitigation as part of the 2016 Paris Agreement. With regards to access to electricity, Ghana performs well with an 84% access rate, compared to an average 42% in SSA and 52% in West Africa (IEA, 2017). The quality of this access is poor, however, with very low consumption levels and prevalent power outages (Gyamfi, Modjinou & Djordjevic, 2015). Insufficient power generation is the major cause of the poor quality of the electricity supply (Fritsch & Poudineh, 2016). This deficit brings significant costs to the economy. For example, national planning documents recognise that the gross electricity supplied is 24% less than required for the Government’s 4% economic growth target (Energy Commission of Ghana, 2017). In like manner, enterprises consider electricity as the second most important constraint to business activities in the country (after access to finance) and estimate 2% losses in production as a consequence of power crises (World Bank, 2013a). Kenya’s rate of access to electricity has been increasing at an impressive pace from 8% in 2000 to 65% in 2016. As a result, Kenya has evolved from being one of the SSA countries with the highest access deficits to one of the few on track to reach universal access by 2030 and the leading country in East Africa (International Energy Agency, 2017). Kenya has shown political commitment and consistency in planning processes, both to increase electrification rates and to expand RE generation (Kapika & Eberhard, 2013). As opposed to many other SSA countries, electricity supply is not considered the major constraint to business growth by Kenyan firms, which are more concerned by issues like informality, corruption, and political instability (World Bank, 2013b). However, the quality of the electricity provided by the national grid is still poor in Kenya, the system is very small, with just 2404 MW for a 46 million population and consumption levels are very low. Transmission and distribution networks are weak and overloaded, and electrical outages are common, which explains why 57.4% of enterprises choose to have their own generator (World Bank, 2013b). The policy mix of Kenya and Ghana is quite similar, with approved FiT, RE targets, fiscal incentives and a more recent interest in RE auctions (Eberhard, Gratwick, Morella & Antmann, 2017). Still, the share of RE in the generation mix, excluding large hydropower, is negligible in Ghana. All private investment in commercial scale power generation capacity committed between 1999 and 2016, for 3.4 Billion USD, targeted natural gas or diesel power plants1, while renewable generation was neglected (World Bank, 2018b). In fact, Ghana was the first recipient in private investment for natural gas based power plants in SSA during 1999–2016. Fossil fuel-based thermal plants have, thus, become the dominant source of electricity, with 57% of installed capacity, after decades of hydropower reliance (Energy Commission of Ghana, 2017). With the expansion of thermal generation capacity, the Government has aimed at improving the reliability of power supply, as recurring 1 Investment figures extracted from the World Bank Private Participation in Infrastructure database cover projects with at least 20% private participation and do not have good coverage of small scale projects due to lack of publicly available information.

87

A. Pueyo / World Development 109 (2018) 85–100 Table 1 Key background indicators of Kenya’s and Ghana’s power sector. Indicator

Kenya

Ghana

SSA

Population (million people) 2016 Income per capita (constant 2010 USD) 2016 Electricity access (%) Installed capacity (GW) 2015

48.4 1143.07 18% (in 2010) 65% (in 2016) 2.4

28.2 1707.7 61% (in 2010) 84% (in 2016) 2.8

Renewable capacity (%) Renewable capacity except large hydro (%) Electricity consumption per capita (kWh pc per year) 2014 Electricity demand growth (%) 2010–2015 System losses (%) 2014 Number of outages per month 2013 Average duration of outages (hours) 2013 Enterprises using own generators (%) 2013 Investment in large scale power generation infrastructure (USD Billion) 1999–2016 Renewable generation infrastructure investment (%) 1999–2016

63% (in 2015) 29% 166.7 6% 17.6% 6.3 5.6 57.4% 2.3 (52 USD pc) 65%

43% (in 2016) 0.1% 354.7 7% 22.6% 8.4 6.6 52.1% 3.4 (122 USD pc) 0%

1033.1 1639.2 32% (in 2010) 42% (in 2016) 92 GW 46 GW without South Africa – – 480.6 – 11.7% 8.6 5.7 52.8% 34.9 (34 USD pc) 20.4 without South Africa 58% 34.3% without South Africa

Sources: World Bank (2018a), IEA (2017), Energy Commission of Ghana (2017); Lahmeyer (2016); World Bank (2013b, 2013c); World Bank (2018b).

blackouts were crippling the economy. Some undesired side effects have been the escalation of power prices and greenhouse gas emissions, as many of these new plants use Heavy Fuel Oil, given gas supply constraints. In spite of the heavy cost to the economy and the environment of using imported oil, renewables do not feature prominently in the Ghanaian Government’s plans to address the energy crisis. Although the Government has committed to increasing their share to 10%, renewables are presented as either too expensive to be competitive or not feasible due to lack of experience with these technologies (Fritsch & Poudineh, 2016; Energy Commission, 2015). ‘‘Clean and affordable” natural gas is presented instead as the best long-term solution to the power crisis by both the Ghanaian government and the World Bank (2015a). However, gas exploitation has been slow and local gas will not be adequate to meet the country’s gas requirements for the medium to long term. Conversely, Kenya enjoys a large share of renewables in its generation mix, with 64% of renewable installed capacity, or 29% if excluding large hydro. Private investment figures reflect the importance of RE in Kenya. Out of the 2.3 Billion USD committed to the power generation sector between 1999 and 2014, 65% targeted wind, geothermal, or biomass power plants (World Bank, 2018b). As opposed to Ghana, renewables are a key element of the Kenyan Government energy strategy (Hoka Osiolo, Pueyo & Gachanja, 2017). A long-term programme for the development of geothermal resources has turned Kenya into the leading geothermal power producer in Africa. Moreover, the Kenyan Government plans to dramatically increase the share of wind power, expecting its share to reach 18% by 2019, from 1% in 2015 (Lahmeyer International, 2016). Notwithstanding the prominent role of renewables, Kenya was the main recipient of private investment in diesel-based generation in SSA during 1999–2014. This shows that, in spite of the abundance of RE resources in Kenya, the power sector is still reliant on fossil fuel imports (Kiplagat, Wang, & Li, 2011). Table 1 provides an overview of key indicators of Kenya’s and Ghana’s power sectors. In summary, both countries have made commendable progress in providing electricity to their population, with higher access rates than in the SSA region as a whole. However, their electricity systems are still small and unreliable, with higher system losses than the regional average and a high prevalence of outages, particularly in the case of Ghana. More than half of the enterprises in both countries cope with the poor quality of power supply by using their own generators, at a high cost. Both countries have pursued a least-cost, local, long-term way out of hydropower dependence to improve energy security. Kenya has

championed geothermal power as a secure, least-cost, and clean energy source and is encouraging a higher penetration of wind power. Ghana has instead sought to develop its domestic gas resources, while renewables have been relegated to a minor role. In the short term, to deal with power supply crises, both countries have locked themselves into contracts with private thermal plants using imported fuel. Kenya has ended many of these contracts, but Ghana is fully embarked on them, and the choice is proving dirty and expensive. With regards to investment inflows, both Kenya and Ghana attract higher private investment inflows in electricity infrastructure per capita than the regional average. However, Ghana’s private investment targets fossil fuel-based generation capacity almost exclusively. Conversely, a large share of private investment in electricity infrastructure in Kenya targets RE. Taking into account the insights of this background section, the focus of the diagnostic of Ghana will be to understand why investment does not flow to renewables. For Kenya, we will look both at the factors behind its success to attract investment in renewable generation and at the prevailing constraints. 3. The green investment diagnostics methodology The original growth diagnostics framework raised the following question: ‘‘For this particular country, at this particular time, what is preventing the country from achieving higher sustained and shared growth?” (Haussmann, Klinger & Wagner, 2008: 4). The methodology is based on the idea that ‘‘there may be many reasons why an economy does not grow, but each reason generates a distinctive set of symptoms” (ibid: 4). The identification of these symptoms through a logical decision tree framework enables a weight of evidence to be built and the identification of the most binding constraint to growth. The first node in the decision tree asks if there is a demand problem – lack of projects delivering a high return – or a supply problem – lack of finance. If the problem is low returns, is this either because of a lack of supporting infrastructure, for example, or because of low appropriability caused by high levels of corruption? Our framework adapts and extends this approach to investment in the RE sector, asking: ‘‘For this country, at this particular time, what is preventing higher levels of investment in a specific technology for which there is an economic rationale?”. To answer this question, we build a decision tree drawing on three sources: the original Growth Diagnostics decision tree (Hausmann et al., 2004); a review of the literature about constraints to investment in RE (Pueyo, Spratt, et al., 2015); and 20 interviews with investors,

88

A. Pueyo / World Development 109 (2018) 85–100

regulators, financiers, academics, and practitioners working in the energy sector. The two initial branches of the decision tree that can explain low levels of investment are either: (i) investment returns are not attractive enough relative to alternatives; or (ii) the sorts of finance needed are not available. Of course, it is possible that both of these problems exist, or that they are interconnected, which would explain the existence of more than one ‘binding constraint’. RE projects may be considered unattractive for two reasons: either returns are too low, or risks are too high. The supply of appropriate finance for renewables (long-term, low-cost debt finance) may be low because either there is not enough capital available or capital is potentially available but is not allocated efficiently by the financial system. Fig. 1 illustrates these initial nodes of the green investment diagnostics decision tree. Decisions about which constraint to investigate further at each node are informed by the country’s performance in several related indicators as compared to other countries or to international benchmarks. As well as direct quantitative comparisons, we also rely on interviews with country stakeholders to interpret these indicators. Qualitative methods are particularly useful in identifying not only the relationships between different constraints, but also, their relative importance from the perspective of investors and other stakeholders, such as regulators or financiers. As potential problem areas are ruled out, we are left with a small number of areas that the evidence suggests may be the most important constraint, in addition to a growing understanding of why. The next two sub-sections look in detail at the possibilities open on each side of the decision tree: project attractiveness and finance availabilities. 3.1. Are renewable energy investments attractive enough? We start by looking at the returns branch of the decision tree. To assess the sufficiency of returns we compare the Equity internal rate of return (IRR)2 of RE generation projects to the opportunity cost of funds. The opportunity cost is the return yield of potential investment alternatives. For an unconstrained international investor, this would be the returns obtainable in the global equity and bond markets. For investors with a regional or sectoral scope, their opportunity costs would be the return of investments in these markets. If we see that returns are not competitive with alternatives, the next question is: Why is this so? Moving down the decision-tree to the next level, we again face two possible explanations: returns may be insufficient either because project costs are too high or because revenues are too low. Fig. 2 shows the different subbranches following signs of low returns in the project. Low returns may result either from high costs of RE generation or from low revenues. High costs of RE generation plants could relate to either the plant itself or the system. To capture project-level costs, we use the levelised cost of energy (LCOE)3 of RE plants and compare it to international benchmarks for RE and to the LCOE of fossil fuelbased plants in the country. If the LCOE appears high as compared to international benchmarks, we will investigate further. System costs are those that a single generation plant imposes on the electricity system as a whole (IEA, 2010). These include the costs of transmission and distribution infrastructure required

2 The methodology to estimate Equity IRR of RE projects is described in Pueyo et al. (2016). 3 The concept of LCOE and the methodology to calculate it is described in Pueyo et al. (2016).

for grid connection and the cost of operating and planning reserve for intermittent renewables.4 To assess this, we look at the following parameters: (i) The generation share of intermittent RE projects in the country (it is often stated that if renewable penetration is higher than 20% by energy it may pose challenges to the system)5; (ii) the geographic and technology diversification of existing RE (more diversification reduces system costs); (iii) the availability of dispatchable back-up capacity and storage in the country (mainly hydro reserves and gas turbines); (iv) the availability of interconnections with neighbouring countries; (v) the reach of the grid infrastructure compared to the location of renewable resources; (vi) and the responsibility for paying the transmission costs of a specific project.6 If the analysis suggests that high project or system costs are not the cause of low returns, the explanation must be that revenues are too low. There are several possible explanations for this, including low electricity tariffs, low demand, poor quality of RE resources, and curtailment of intermittent RE. The latter refers to the possibility that all the electricity produced by intermittent generators cannot be fed into the system. Each of the previous issues would require a very different policy solution. High risks might be the main reason why projects are unattractive. Investment in RE technologies is particularly vulnerable to a risky environment because they tend to be very capital intensive and require the bulk of the costs to be committed at the initial stage of the investment. Six potential risks were selected as particularly relevant for RE, as illustrated in Fig. 3. First, regulatory risks arise for electricity providers when there is uncertainty over their property rights, their right to feed the power they generate to the national grid, their ability to sell electricity at a predictable price that allows for cost recovery, and their right to be paid for their services. Challenges in all these areas are widespread in SSA. Regulatory risks also arise for electricity consumers when the Government is unable to serve demand and when lack of competition or corruption increase the price of supply. Five activities encompass the roles of the regulator to protect the rights of suppliers and consumers. These activities are planning, procurement, contracting, provision of access to land and access to the electricity market. Second, risks related to the creditworthiness of the off-taker arise when the purchaser of electricity – typically national distribution companies – is not financially sound and might be unable to pay generators for the power they feed to the grid. Several financial challenges make African electricity distribution companies unreliable payers. For example, retail electricity prices are often lower than the wholesale price of electricity charged by generators. In addition, customer default rates are often high and billing processes may be ineffective, while large losses can result from poor network system maintenance. Indicators of operational perfor-

4 Reserve capacity is a measure of available capacity over and above the capacity needed to meet normal peak demand levels. The reserve capacity should be available to the system operator within a short interval of time to meet demand in case a generator goes down or there is another disruption to the supply. 5 Countries in Europe such as Denmark, Ireland and Spain have achieved much larger shares than this 20% by energy, stabilising the grid either trading with neighbouring countries or adding flexible generation sources. 6 For example, in Germany the costs of transmission are borne by the transmission system operators and charged to the federal grid agency of Germany ((Klessmann, Nabe & Burges, 2008). In many other countries, grid connectivity to the nearest substation needs to be borne by the project developer and grid access is not always guaranteed, which can significantly increase the costs of the project.

A. Pueyo / World Development 109 (2018) 85–100

89

Fig. 1. First nodes of the GID decision tree. Source: Author’s own, drawing from Haussmann et al. (2004).

Fig. 2. Unattractive investments due to low returns. Source: Author’s own.

Fig. 3. Unattractive investments due to high risks, Source: Author’s own.

mance, liquidity, credit risk and financial structure available in the financial statements of utilities can point to their financial weaknesses (Fritsch, 2011). Third, risks associated with resource supply and technology can arise either from inaccurate assessments of RE resources or from unreliable resource suppliers, in the case of biomass or geothermal, for example. To mitigate these risks, public dissemination of robust resource assessments for each technology may be helpful.

Technology risks are likely to be high when the country has no previous experience in a particular technology, when its industry is not sufficiently developed to provide spare parts and knowledge, and when its workforce has not got the required skills. Macroeconomic risks are particularly important for RE investments in developing countries because the scale of capital requirements is very high compared to what is available in either immature or absent local capital markets. Therefore, finance typi-

90

A. Pueyo / World Development 109 (2018) 85–100

Fig. 4. Insufficient finance due to inadequate supply of savings. Source: Author’s own.

cally needs to be obtained internationally, and debt and equity are denominated in foreign currency. In contrast, project revenues are usually denominated in local currency, creating a significant currency risk. Social and reputational risks can also reduce the attractiveness of RE projects. Community acceptance, in particular, has proven to be a real block to the completion of RE projects, as evidenced by projects experiencing delays or being abandoned due to conflict with local communities.7 Certain features of RE projects make them more exposed in this regard. They are smaller per project than are fossil fuel plants, hence requiring more location decisions for an equivalent capacity. Their visual impact per unit of output (kWh) is also larger, and there is much less flexibility on where they can be located as they need to be where the best resources are (Wustenhagen, Wolsink & Burer, 2007).8 Questions of community acceptance and compensation are complicated when the current users of the land do not hold formal land titles, as is often the case in SSA. The lack of clarity over land rights also creates the possibility of rent seeking from local communities seeking compensation for land they do not use, as highlighted by interviewees in the wind energy sector (Interviews with representatives of Aldwych International, 11th September 2015; and an international EPC contractor in Kenya, 2nd October 2015). Many of the risks identified above are intimately related to governance. As a further piece of evidence pertaining to these areas, quality of governance can also be looked at directly, using the World Bank’s ‘Doing Business’ indicators in the following areas: corruption; security; conflict; property rights and rule-based governance ratings (World Bank, 2017b). Having considered potential constraints relating to the relative attractiveness of particular projects, the next section turns to the supply of finance. 3.2. Is appropriate finance available? Enterprise surveys identify access to finance as a major constraint to investment in Africa (World Bank, 2017b). At the national level a low share of credit to the private sector and high financing costs may provide further evidence of financial constraints. At their heart, financial supply constraints result from either a lack of 7

In Kenya, for example, a 60 MW wind power project recently had to halt construction as local farmers took to the streets and the court, making the site unsafe. Social opposition has delayed or stopped wind farms’ development in other developed and developing countries, such as Mexico (Juárez-Hernandez and León, 2014) and the UK (Cass & Walker, 2009). 8 There are interesting differences between social acceptance in the developed and the developing contexts. For example, opposition to wind farms in developed countries mostly derives from their negative landscape impact. In developing countries the main concern is the impact on livelihoods of changes in access to land (Juárez-Hernandez and León, 2014).

savings (domestic or external) or a failure of the financial system to allocate those savings (Hausmann et al., 2004). We start by looking at the decision tree branch related to an inadequate access to savings. Its ramifications are illustrated in Fig. 4. As was the case when we looked at project attractiveness, a process of elimination results in some areas being ruled out. Where there is evidence to suggest that an area may be an important constraint, we engage in further investigation at more detailed levels of the decision-tree. Finance can be available through domestic and foreign sources. Among foreign sources, we focus on the availability of foreign direct investment (FDI), and foreign aid, due to their importance for the power sector. Foreign aid can both provide funds directly and leverage private finance through blended finance type models and political ‘insurance’ for private investors. Aid also shields the sector from the volatility of private investment and targets countries and technologies either avoided or under-served by the private sector (Pueyo, Orraca, et al., 2015). At the same time, private investment is crucial to cover the financing gap for electrification. FDI could be lower than required in a country as a result of either an inadequate business environment or direct restrictions through capital controls (Hausmann et al., 2004). As well as a shortage of external finance, problems may exist in terms of insufficient domestic finance. High domestic interest rates are a symptom of insufficient domestic finance, as they indicate a scarcity of supply relative to demand. As summarised in Fig. 4 potential explanations range from limited domestic savings to a low level of tax collection, preventing public investment in infrastructure. In this regard, the low level of tax collection in SSA is considered as one of the greatest barriers to the transformation of the power sector (Africa Progress Panel, 2015). Rather than the underlying supply of finance, the inability to access finance on suitable terms may be a problem of financial intermediation. Well-functioning banks are competitive and stable and able to provide long-term finance to promote investment and economic growth. Fig. 5 summarises the issues of financial intermediation in decision-tree form, illustrating the main categories and their more detailed potential causes.

3.3. Narrowing down constraints The diagnostic processes described thus far are designed to highlight areas of potential concern by comparing performance in different areas with international peers and benchmarks. However, even if an area is identified as potentially problematic, this does not mean it is necessarily ‘binding’. A country may perform extremely badly in a particular area, but this may have little impact on investment incentives. It could also be the case that a country

A. Pueyo / World Development 109 (2018) 85–100

91

Fig. 5. Insufficient finance due to poor intermediation. Source: Author’s own.

scores badly in many, or even most, aspects. Which of them should be targeted first? Insights from actual investors, regulators and suppliers in the RE industry of the relevant country are essential to assess how binding the first set of problems identified are. As in Hausmann et al. (2004), discussions with in-country stakeholders will uncover the following ‘‘symptoms” of binding constraints:

owned enterprises, or investors with strong links to the Government or the ruling elite, thrive in the power sector. In a similar way, where cooperatives or community-managed generation projects thrive, social opposition could be the binding constraint.

i. The price of the constraint should be high. ii. Historical movements in the constraint should have produced significant changes in the desired outcome. iii. Agents in the economy should be attempting to either overcome or bypass the constraint. iv. Agents less affected by the constraint are more likely to survive and thrive, and vice versa.

In this section we navigate through our decision tree to understand what constraints investment in renewables in the two target countries. The final discussion compares the results obtained, highlighting similarities and differences in the investment environment of Kenya and Ghana. A collection of indicators comparing the performance of Kenya and Ghana, and that of their regional and income peers, across the different constraints underlies the analysis in this chapter and is presented as an Appendix.

The first symptom uses high prices or shadow prices to signal a binding constraint, by reflecting the scarcity of a resource – e.g. finance, infrastructure, skills or confidence. For example, if the constraint is an insufficient supply of finance, we would expect to see very high real interest rates. If the problem was restricted to longterm finance, then interest rates for this type of finance should be high. If a lack of supporting infrastructure was binding, we would expect the infrastructure that did exist to generate high returns. If the issue was governance, then premiums for political risk insurance would be very high. According to the second symptom, we would expect to see significant impacts on RE investments when a binding constraint is relaxed. For example, if the constraint is low electricity tariffs, we would expect to see an increase in investments if the regulator approves cost-reflective tariffs. If it is the uncertainty of procurement processes, a credible international competitive bid should attract a lot of interest, and if it is the high cost of finance, periods of lower real interest rates should unleash private investment. The third symptom would be the observation of firms trying to overcome a constraint. For example, if poor regulation for grid connected generators was a binding constraint, we would expect to see an explosion of non-regulated off-grid small scale generation, even in locations within reach of the national grid. If an unreliable off-taker was the key constraint, we would observe that power purchase agreements between generators and off-takers would not be enough of a guarantee for investors, who would also require either sovereign guarantees or partial risk guarantees to protect themselves from non-payment. The fourth symptom would point to agents that thrive in the current conditions, expecting that they would be less reliant in the constraint identified. For example, if the key constraint is related to poor governance, we would expect to see either State-

4. What are the binding constraints to renewable energy investment in kenya and ghana?

4.1. Kenya As presented in the background section, Kenya has succeeded in attracting private investment in both renewable and nonrenewable power generation plants. Kenya is positively perceived by donors and investors as a target country for RE investment because of its market-friendly approach to development (Newell & Philips, 2016) and because it has better investment climate than do neighbouring East African countries (Eberhard et al., 2017). For example, an equity investor in the energy sector of the country reflected this optimistic view: ‘‘Kenya is far ahead from the rest of Africa. They have a great track record. They have invested a lot in capacity building, have good advisors, strong political support, and solid commercial relationships. I have a great opinion of Kenya Power. Their regulatory reform has worked and they are rapidly increasing electrification rates. What reassures investors is to see that their competitors are being paid and are getting their investment back. That is why Kenya will go well” (Interview with representative of Aldwych International, 11th September 2015). However, some problems remain, as evidenced by long delays in the implementation of projects and the fact that the National Power Generation Plan warns of ‘‘overcapacity” while levels of access to electricity remain low (Lahmeyer, 2016, p.5). This diagnosis looks both at the factors underlying success in attracting investment in some technologies and at the remaining constraints. The first question posed is: are RE projects in Kenya attractive for investors? Previous research shows that RE projects yield generous returns in Kenya, particularly for geothermal energy (up to 30% nominal return on equity) and wind (between 14 and 20%) (Pueyo, Bawakyillenuo & Osiolo, 2016). The returns for wind

92

A. Pueyo / World Development 109 (2018) 85–100

and geothermal are therefore competitive with other investment alternatives in Kenya, such as treasury bills, which paid nominal rates of 8.6% as of December 2015 (Central Bank of Kenya, 2015). Solar PV and hydropower, however, offer modest returns between 5 and 9% in nominal terms (Pueyo et al., 2016). Healthy returns for wind and geothermal are possible due to their low LCOE, estimated as 7.3 USD cents per kWh for geothermal and 7.5 USD cents per kWh for wind power plants located in sites with high quality wind resource, using a 10% discount rate in both cases (Pueyo et al., 2016). These costs make wind and geothermal competitive with all fossil fuel alternatives in Kenya. Commercial scale solar PV and hydro, however, are not yet competitive with coal and Combined Cycle Gas Turbines in Kenya (Pueyo et al., 2016). In the particular case of wind power, even if project costs are low, system costs, meaning the cost imposed by the project on the whole electricity system, could be high enough to limit the potential of this technology. The fast introduction of a large share of wind power in a single location as will happen in Kenya with the 300 MW Lake Turkana Wind Power Plant (LTWP), can impose high balancing costs on the system. The grid manager could resort to curtailment, with significant costs for wind generators. However, we do not consider balancing costs as a binding constraint, because a recent capacity adequacy assessment of wind in Kenya shows that the large wind projects that Kenya is building are likely to contribute to the generation adequacy of the system, thanks to the complementarity between wind resource and demand and between wind and hydro resources (Edwards, Dent & Wade, 2017). Transmission costs, on the other hand, are a significant constraint to new renewable generation capacity. The latest power generation and transmission master plan recommends considerable expansion, reinforcement and rehabilitation measures to allow the stable transport of energy from power plants to load centres. Furthermore, the long and costly delays experienced by LTWP to start operations, due to late completion of a transmission line also exemplify the importance of transmission and distribution costs in Kenya (Hoka Osiolo et al., 2017). Cost-reflective tariffs and the exceptional quality of Kenya’s RE resources contribute to high returns. Nevertheless, the weak spot for revenue generation in Kenya is the low level of demand. Economic growth in Kenya has been higher than in the rest of SubSaharan African countries, but it is still below the 10% aspiration reflected in the strategy document Vision 2030, which guided the country’s generation expansion plans (Republic of Kenya, 2007). High poverty rates and low incomes per capita indicate a low ability to pay for electricity at the household level, and electricity consumption per capita remains stubbornly low even as access rates accelerate. One of our interviewees, who works in the wind power sector, indicated that ‘‘productive uses are very small. Most people do not have energy intensive appliances in their homes and electricity is used primarily to charge mobile phones, which does not allow the recovery of generation, transmission, and distribution costs”. We conclude therefore that RE projects – wind and geothermal in particular – are attractive in Kenya with regards to the returns they can yield for investors, subject to potentially high transmission costs and a low demand. We now examine if the risks that investors face are commensurate with these returns. Regulatory risks are considered low relative to other SSA countries. Responsibilities for planning, procurement, contracting, and permitting of power generation capacity in Kenya, are generally well defined and efficient (Eberhard et al., 2017). Policymakers are fast to react when there are problems with existing regulations. For example, seeing the proliferation of unsolicited proposals as a result of the FiT regulation, the Government is moving towards an auctioning scheme that sets ex-ante the capacity to be procured. Off-taker risks and macroeconomic risks are also low in comparison to the rest of SSA. The off-taker Kenya Power has been prof-

itable since 2004, after emerging from a period of sustained lossmaking, and has a good track record in honouring Power Purchase Agreements (PPA) (Kapika & Eberhard, 2013; KPLC, 2015). There is some reason for concern with regards to RE resource and technology risks. The Geothermal Development Corporation, which holds steam mining rights in Kenya has not been able to source as much steam as expected. Its perceived technical and financial weakness had to be recently mitigated with partial risk guarantees provided by the African Development Bank to encourage investment in the sector (Eberhard et al., 2017). With regards to technology risks, Kenya shows a lower level of both tertiary education and competitive industrial performance than would be expected for a middle-income country. This lack of local skills increases the cost of projects, which must rely on external staff, and prevents technology transfers to the country. However, we do not consider this constraint as binding, as Kenya has been able in the past to build the capabilities required to develop a complex geothermal power sector and a local off-grid solar sector (Byrne et al., 2014). Social risks are particularly acute in Kenya, as evidenced by several energy generation projects experiencing distress due to disputes with local communities. For example, the geothermal plant Olkaria IV required an expensive resettlement plan for local communities to go ahead (personal communication with Kengen representative, 12th February 2015). Wind power plants at Kinangop could not progress that far and had to be terminated, even after the wind turbines had been shipped to the country. LTWP has successfully engaged with local communities to avoid conflict, but it has, nevertheless, been taken to Court by a local tribe and has experienced significant delays in getting wayleaves for its transmission line (Hoka Osiolo et al., 2017). Behind these conflicts there is lack of clarity regarding land property rights and consultation procedures, as well as unbalanced distribution of costs and benefits between investors and communities. The importance of community issues was raised by several interviewees. For example, the project manager of a large international wind power plant indicated that ‘‘Community issues are a major problem in Kenya. They can kill a project. Protests scare investors and they are a massive reputational issue. This was particularly a problem for the transmission line of more than 400 km”. Similarly, an EPC contractor in the wind power sector said ‘‘The main barriers to investment are associated with the sociopolitical challenges linked to securing land and way leaves”. Governance is the last risk to be considered in evaluating the risk profile of RE projects in Kenya. Kenya performs poorly in the Corruption Perception Index (Transparency International, 2016), and the percentage of firms that declare having to provide gifts to either obtain construction permits or secure government contracts is higher than in the SSA region as a whole (World Bank, 2013b). Corruption and political interference were, in fact, mentioned as important constraints in our interviews. There is a view that ‘‘you need to be on the side of the Government for projects to be successful” (Interview with the director of Quality and Environment of an EPC contractor in Kenya, 2nd October 2015). The evidence also shows that projects requiring heavy infrastructure investments are more likely to thrive in Kenya if they are publicly owned and funded. Projects can bypass important constraints when they have strong political support, which is mainly granted to large-scale, flagship projects that are very visible to the public. Large scale projects are also a source of pride for ruling politicians, and they are aligned with the ambitions set out in Kenya’s national strategy Vision 2030. Some authors have also pointed out that a rent-seeking dimension could explain the preference for large scale projects, in terms of the ability of government officials to maintain control over access to sites, infrastructures, and profits more easily (Newell et al., 2014). We now turn our diagnosis to the supply side of the decision tree, to discern whether there is enough finance available to build

A. Pueyo / World Development 109 (2018) 85–100

these projects. Access to external savings does not appear to be a problem for Kenya. It enjoys higher levels of ODA per capita and per GDP than do its regional and income-level peers (World Bank, 2018a). More specifically, donors have played an important role in the development of a geothermal power sector in Kenya, providing loans at very low interest rates and long maturities (Pueyo et al., 2016). Access to international private finance is not problematic either. Kenya has been able to tap the global capital markets by issuing bonds to finance infrastructure projects and with high participation of foreign investors in the Nairobi stock exchange (World Bank, 2015b). The situation is different with domestic finance, where there is a lack of savings and very few domestic equity providers (interviews with Director of Research and Policy of Kenya Bankers Association, 28th September 2015; and with a country economist at the African Development Bank, 30th September 2015). Nevertheless, we do not find strong signals about the severity of this constraint. Access to finance is only considered a major constraint to growth by 17% of enterprises, as compared to 37% in the SSA region (World Bank, 2013b), and both the supply of private credit by deposit money banks and stock market capitalisation are very high in comparison to other countries in SSA (World Bank, 2015b). Poor intermediation by the banking system is not a severe constraint either. The Kenyan financial sector is the third largest in SSA and has been resilient to domestic and foreign shocks in the last decade (World Bank, 2015b). Bank concentration is lower than in the SSA region and the group of Lower Middle-Income countries, and interest rate spreads are high, but comparable to the SSA region. In summary, finance supply does not pose significant constraints to investment in Kenya and, while RE projects are attractive for investors, there are some constraints that remain with regards to transmission costs, low demand, and high social and governance risks. Fig. 6 represents the outcomes of our decision tree analysis for Kenya. 4.2. Ghana Ghana’s electricity system suffers from underinvestment in renewable generation capacity. In line with the Kenyan case study, the diagnosis of Ghana’s ailments first tries to reveal whether this underinvestment is caused by a lack of attractive projects or, instead, by a shortage of suitable finance. Evidence shows that there is a mixture of both. First, the risk-return profile of RE investments in Ghana is not sufficiently attractive for investors. Expected nominal returns on equity for commercial scale wind (around 14% return) and solar PV (around 9% return) are lower than potential alternatives (Pueyo et al., 2016). For example, Ghana’s Government 91-day Treasury bills provided a 25% return as of the end of September 2015. High bond rates have stabilised in the last two years though, going down to 13% as of September 2017 (Bank of Ghana, 2017), but stakeholders suggest that projects in either the telecommunications or the gas sectors are more profitable than are renewables (interviews with the Executive Director of KITE, 16th October 2015, and with the Project Manager of Aldwych International, 11th September 2015). Two elements explain these insufficient returns. On one side, expensive debt rates keep costs high, while low tariffs keep revenues low. On the other side, by offering high bond rates to sustain a high fiscal deficit, the Ghanaian Government is crowding out private investors. We focus firstly on the low project returns, while the high returns of alternatives are addressed later as part of the description of macroeconomic risks and constraints to the supply of finance. Revenues are lower than required due to low prices. Ghanaian retail tariffs have, historically, been among the lowest in Africa thanks to a reliance on low-cost hydroelectricity (Briceño-

93

Fig. 6. Kenya’s constraints decision tree. Source: Author’s own.

Garmendia & Shkaratan, 2011). For this reason, Ghanaian consumers, regulators, and politicians have grown accustomed to cheap electricity and price rises are met with strong social opposition (Interview with representative from Ghana Capital Partners, 19th September 2015). Despite resistance, the Government has recently implemented cost-recovery tariffs using an Automatic Adjusted Formula. This was done in extremis, to avoid the collapse of a power sector facing gas shortages, low reservoir levels and increasing oil imports. As a result, Ghana’s tariffs in 2016 were among the most expensive in Lower Middle-Income (LMI) countries, after having long been one of the cheapest (Energy Commission, 2017). High tariffs are not likely to last long, though. The most recent energy outlook blames them for Ghana’s economic underperformance and warns of competitiveness losses as a result. Another negative outcome has been that many of the wealthier, and hence, best customers of the national grid, disenchanted by poor service and now by high tariffs, have turned to either diesel or solar PV self-generation (Energy Commission, 2017). A policy intended to improve the financial sustainability of the electricity sector may have had the opposite result as wealthy, intensive users of electricity are essential to sustain a system that subsidises the poorer consumers. Under those circumstances, tariff increases are likely to be reversed soon. In fact, the latest energy outlook prepared by the Energy Commission recommends that the ‘‘average end-user-tariff is reduced from the current rate of 20–21 US cents per kWh to within 10–15 US cents per kWh” to achieve economic growth targets (Energy Commission, 2017, p. ii). If low tariffs were the most binding constraint to investment in renewables in Ghana, FiT would be a reasonable policy to address the problem. Ghana has indeed approved a FiT scheme consisting of a RE purchase obligation, a premium tariff and guaranteed connection to the transmission and distribution systems. FiT were first published in 2013, and then revised in 2014; however, four years later no RE project benefits from them. The design of the scheme is partly to blame, as premium prices are only guaranteed for 10 years and are denominated in Ghanaian currency, without indexation to foreign currency. However, it might be that prices are not the single most binding constraint to investment. In the words of an investor in the wind power sector ‘‘ECG can sign any PPAs that they want, but they do not have any value because they do not have any money”. Off-taker risks are one among many risks that RE investors face in Ghana. The main off-taker in Ghana’s electricity sector, the Electricity Company of Ghana (ECG), has carried an operating loss for at least 4 years. The most recent financial statements show that it is unable to pay either suppliers or debtors in time due to insufficient cash flow. Furthermore, high system losses mean that ECG does not bill for almost a quarter of the electricity it buys, and a significant share of electricity bills are not actually paid for (ECG, 2014). There are several reasons for the poor performance of ECG, including: low tariffs for distribution and retailing; currency devaluation affecting payments in foreign currency to international IPPs and suppliers of imported materials; damage to network assets caused by regular

94

A. Pueyo / World Development 109 (2018) 85–100

load shedding; poor revenue collection and billing, with 65% of debt from the Government and public institutions; and loss of high revenue customers, as they choose to install their own diesel generators or solar panels (Fritsch, 2011; ECG, 2013). ECG’s financial weakness means that the PPAs it signs with potential generators do not provide enough assurance to financiers, regardless of the electricity price set. While this could be addressed through sovereign guarantees, the Government of Ghana is reluctant to offer them to keep its fiscal deficit under control. In addition to the creditworthiness of the off-taker, the analysis of indicators and stakeholder interviews points at regulatory and macroeconomic risks as being particularly worrying in Ghana. Regulatory risks refer to irregularities in the planning, permitting and procurement processes. A symptom of the inefficiency of the planning process is the large number of plants in the permitting pipeline that are not being constructed. As of December 2017, 104 RE projects for more than 6 GW of capacity had obtained provisional wholesale supply and generation licenses.9 Hence, RE plants with provisional permits doubled the capacity of the Ghanaian system and broadly exceeded the caps and targets set in the Renewable Energy Act. This provides evidence of the lack of coordination between power generation planning and procurement in Ghana, creating uncertainty with regards to how much and what type of RE the Government needs. Permitting procedures are also both inefficient and opaque. Project developers need to obtain up to eleven licences, approvals, agreements and clearances from different institutions (Energy Commission of Ghana, 2012). Each of these steps increases transaction costs and provides opportunities for rent seeking, an issue that came up strongly in our interviews with Ghanaian stakeholders. For example, a representative from the Africa Centre for Energy Policy stated that ‘‘to get a PPA in Ghana you need to know a politician or a big person to take you through the process”. In like manner, an investor in the wind energy sector had experienced that ‘‘at every step there are officials trying to get personal advantage of the project, and if you do not meet their demands the project is delayed and delayed and delayed”. As with permitting, procurement processes are fraught with inefficiencies. For instance, it is not clear which institution is in charge of signing PPAs. Distribution utilities, the Ministry of Energy, and the public generation utility Volta River Authority have all separately signed or undertaken commitments to enter into PPAs with private generators, each of them following different procedures (Malgas & Eberhard, 2011). There is nothing like the transparent international competitive bids we referred to in the Kenyan case. Instead, international contracts are allocated behind closed doors and without clear selection criteria. Ghana’s unstable economy creates further risks for investors. In the last few years Ghana has faced major macroeconomic challenges, including a sharp currency devaluation, high inflation and interest rates and large current account and fiscal deficits. The Ghanaian cedi lost 65% of its value against the US dollar between 2011 and 2015 and 18% in 2015 alone. The country also faces persistent high inflation, reaching 18% in 2016 (Bank of Ghana, 2017). Ghana’s increasingly unsustainable fiscal and current account imbalances required a bailout from the International Monetary Fund (IMF) in 2015 and the implementation of a fiscal consolidation programme. Since then, Ghana’s position has improved. The rate of external debt accumulation has slowed down, while the GDP has grown more in 2017 than in previous years. The inflation rate has also moderated, reaching 12% year-on-year in September 2017, while the cedi has continued to stabilise against the dollar. Although Ghana is now in a better place than it was when the

IMF intervened, many challenges remain. The country is still at high risk of suffering debt distress, as debt keeps on expanding. It faces high financing costs in domestic and external markets and high youth unemployment (World Bank, 2017c). Macroeconomic risks, therefore, continue to be a cause of concern for investors in the energy sector. In short, Ghana does not have a wide portfolio of attractive RE projects as a result of pressures to keep prices low and a risky investment environment, mainly caused by regulatory, off-taker and macroeconomic risks. On top of that, it does not enjoy exceptional quality renewable resources that could counteract low prices and high risks through high utilisation rates (Pueyo et al., 2016). Some of the constraints related to project attractiveness in Ghana are intrinsically linked to constraints in the supply of finance, for example, the high cost of debt, the Government crowding out private investors with high yield bonds, and the macroeconomic imbalances. The remainder of our diagnosis focuses on this branch of the decision tree. Ghana’s enterprises single out access to finance as the most important constraint to growth (World Bank, 2013c). The high cost of debt in Ghana is particularly detrimental to renewable generation, which typically requires high upfront investments. For example, the average lending rate in Ghana as of December 2016 was 32% (Bank of Ghana, 2017). In fact, the only private solar PV plant operational in Ghana10 was able to bypass this constraint by accessing favourable financing conditions available to Chinese-owned projects. Ghana’s finance shortage exists despite good access to international savings, with high FDI, portfolio equity investment and aid inflows in comparison to its peers. However, the high foreign debt and current account deficit of Ghana signal that it has used up its access to foreign savings to the limit. Financing problems are, therefore, eminently domestic and arise from a lack of domestic savings and poor intermediation. Very high domestic interest rates show a high willingness to remunerate savings (Bank of Ghana, 2017). On the other hand, there is a shortage of capital with lower gross domestic savings, private credit, and stock market capitalisation than in the average LMI country (World Bank, 2015b). This problem is exacerbated by poor intermediation in the financial sector, as evidenced by a wide spread between bank deposit and lending rates. Ghanaian banks offer a poor service to energy investors, with small amounts of short-term and costly credit (United States Government and Government of Ghana, 2011). In conclusion, Ghana’s constraints to investment in RE arise both from shortages in the supply of finance and the demand of this finance by attractive projects. Fig. 7 illustrates this outcome with the elements of the decision tree that should be prioritised by policymakers.

9 Energy Commission website http://www.energycom.gov.gh, accessed December 2017.

10 20 MW Solar PV plant BXC Solar at Onyandze is owned by the Chinese company BXC Beijing China.

5. Discussion This paper has described the Green Investment Diagnostics framework and applied it to Kenya and Ghana with the aim of revealing their most important constraints to investment in commercial scale RE. Before applying the diagnostics tool, we have described the context for investment in renewables in Kenya and Ghana. There are some similarities, such as the tension of preserving the financial sustainability of electricity utilities while aiming for universal access. Both countries also have a similar RE policy portfolio, with RE targets and FiT. However, in reality, Kenya and Ghana have very different circumstances. Ghana has been immersed in a power sector crisis for several years, with frequent blackouts and national power utilities on the brink of bankruptcy. RE does not feature as a key solution to energy security problems

A. Pueyo / World Development 109 (2018) 85–100

95

Fig. 7. Ghana’s constraints decision tree. Source: Author’s own.

in the country, a role allocated instead to domestic gas resources. However, gas exploitation has been slow, and local gas is not expected to meet the country’s gas requirements for the medium-to-long term. In the short term, to deal with a power supply crisis, Ghana has locked itself into expensive contracts with private thermal plants using imported fuel. Conversely, Kenya has enough generation capacity to meet its current needs at the current tariff and may exceed capacity needs in the medium term. Its power utilities are financially healthy, prices are set at cost-recovery level, and there is both a good track record of international competitive bids to procure generation capacity and a good history of honouring PPAs. Kenya has decisively supported RE, through a long-term geothermal plan and the approval of several wind power projects. The other side of the story has been, until recently, a large share of population without access to electricity, very low electricity demand, long delays in RE projects becoming operational and a level of social discontent that can halt infrastructure projects. Our analysis has then questioned whether potential constraints to investment in Kenya and Ghana come either from a lack of attractive projects (deal flow) or from a lack of appropriate finance. A decision tree, informed by indicators (included as an Appendix) and interviews with stakeholders point to some constraints in the project attractiveness side for Kenya, and in both project attractiveness and finance availability for Ghana. With regards to project attractiveness, the returns for RE in Kenya appear higher than in Ghana, due to very good quality RE resources (mainly wind and geothermal) and access to lower cost finance. Although the cost of renewable electricity is lower in Kenya than in Ghana, it is lower in both countries than is the cost of electricity from some of the procured fossil fuel-based plants. The risks of investing in RE are higher in Ghana than in Kenya. At the core of Ghana’s risks there is an off-taker that cannot pay power producers because its financial health has been damaged by prices below cost-recovery; system losses; poor metering and billing; unpaid bills; inflation: and currency devaluation, in addition to poor management and political patronage. Kenya’s risks are mainly social, infrastructural and political. Both Kenya and Ghana present regulatory constraints to investment in renewables, but these are again more acute in Ghana, where there is not a clear delimitation of responsibilities across institutions and where the rules of the game are less transparent than in Kenya. Ghana’s financial constraints are also more acute than are Kenya’s. There is a lack of domestic savings, and an excess of debt by the State, which has caused currency devaluation, fast inflation and fiscal imbalances. Government bonds offer higher returns than investment in RE and there is an overall shortage of the long-term, lowcost finance needed for electricity generation projects. Even though Kenya has a shortage of domestic savings, it has benefitted from substantial inflows of foreign aid. Foreign aid has, for example, provided debt at near-zero interest rates to geothermal plants, even though these are profitable and offer least-cost electricity.

Ultimately, Kenya presents a more fertile ground for investment in RE, but it must redistribute rents for the provision of electricity to the rural poor, promote income-generating uses of energy, and tackle its deficit on transmission and distribution infrastructure. Ghana is in dire need of long-term affordable finance to achieve a sustainable solution to its generation capacity deficit. The growth of renewables in Ghana will also need financially sustainable electric utilities, a predictable and reliable policy framework, and political will. All in all, policy should target different priorities in Kenya and Ghana, and the policy mix of both countries should reflect these differences. 6. Conclusions and policy implications The vast potential of RE is failing to be realised in many African countries, despite the many pledges made by donors and international financiers. This is not due to a lack of policies supporting investment. Many African countries have RE targets, FiT, auctioning schemes, and import duty exemptions for RE technologies. In some cases, these policies are not fully implemented. In others, they are implemented but, put in the language of this paper, they are not targeting the most binding constraints on investment. Whatever the reason for their lack of success, it is clear that simply introducing formal policies is not enough. In the research described in this paper, we have developed and tested an approach to identify what these constraints might be. The Green Investment Diagnostics methodology is designed to help policymakers identify the most important obstacles to investment in the RE technologies most suited to their countries and deliver effective reforms to remove these obstacles. Constraints could arise from the lack of attractive projects, the lack of appropriate finance, or both. In many cases, a large number of issues may be holding back investment. An important characteristic of the proposed framework is that it helps to narrow down the hundreds of potential constraints to a handful of the most pressing. The method for achieving this relies on robust international comparisons of data drawn from a diverse set of credible sources. We triangulate the evidence with interviews to stakeholders and with diagnostic signals to assess the importance of each constraint. While constraints are likely to be country-specific, there are commonalities. As our analysis of Kenya and Ghana has shown, for example, FiT that provide a guaranteed income to generators for a period of time will only succeed when there is a reliable off-taker that can pay those fees, or a solvent state that can take responsibility through sovereign guarantees. Previous successful FiT also had a market large enough to make it worth the effort of calculating an appropriate tariff. The proliferation of unsolicited proposals in Kenya and Ghana also shows that incentive mechanisms open to everyone, without total capacity caps or a procedure to prioritise proposals, are likely to clog small African power systems. Finally, Kenya’s case brings the question of whether foreign aid should be used to support already profitable RE generation

96

A. Pueyo / World Development 109 (2018) 85–100

technologies instead of either electrification of the last mile or improvement of transmission and distribution networks in cases when there is not a deficit in generation capacity. While there are commonalities, they will manifest themselves in very different ways in different contexts. An approach that targets support to the needs of each particular country is, therefore, likely to be more effective than simply importing policies that have enjoyed success elsewhere. Given this, we invite donors and national policymakers to replicate this exercise to better target their support for green growth in developing countries.

Appendix Comparison of indicators from Ghana and Kenya Tables 2–15

Table 2 Levelised cost of energy in Kenya and Ghana. LCOE (USD cents per kWh)

7. Conflict of interest statement No conflicts of interest to declare. Acknowledgements

Technology

Ghana

Kenya

Wind Solar Hydro Geothermal

14.3 18.7 7.9

10.3 14.8 10.7 7.3

Source: Pueyo et al. (2016). Using social discount rates.

Funding from the Engineering and Physical Sciences Research Council (EPSRC) and the Department for International Development (DFID) of the UK is gratefully acknowledged (grant number EP/L002507/1). In Ghana, I would like to thank Simon Bawakyillenuo (ISSER) for guidance and data collection as well as all participants in the interviews held. In Kenya, I am grateful to Helen Hoka Osiolo and James Gachanja (KIPPRA) for guidance and data collection, as well as to all interviewees. I also thank Stephen Spratt (IDS) and Andrew Barnett (The Policy Practice) for their guidance during the research and the writing process.

Table 3 Returns on equity for investment in renewables in Kenya and Ghana. Returns (%) Technology

Ghana

Kenya

Wind Solar Hydro Geothermal

6.9 Close to 0% 33.5

14.3 5.3 5.3 16.8

Source: Pueyo et al., (2016). Using commercial cost of debt and prices at FiT.

Table 4 Indicators of system costs in Kenya and Ghana. System costs Topic

Indicator

Ghana

Kenya

Share of IRE Diversification of IRE

% IRE capacity High or low with regards to technology and geography % generation from hydro (natural gas, hydropower, HFO and biomass plants as a share of the total

6% (2020 target) High (geographically diversified solar PV below 20 MW and wind) 100%

Connectivity Transmission costs

% of population connected to the grid Responsibility for transmission costs

84% Cost to metering point borne by generator.

Transmission capacity

Enough transmission capacity to transmit projected generation

Yes

18% in 2019, from 1% in 2016 Low (only wind and more than 50% in a single location) Around 44% in 2016, but with a small share of gas turbines and a large share of unreliable hydro and dirty diesel engines 65% Project developer covers shallow costs, but delays in building new transmission lines can add to the cost of the project No

Share of dispatchable generation

Sources: IEA (2017); Ghana: Energy Commission of Ghana (2015), GRIDCo (2014); Kenya: Lahmeyer (2016), KPLC (2015).

Table 5 Indicators of demand in Kenya, Ghana, Sub-Saharan Africa and Lower Middle Income Countries. Demand Topic

Indicator

Ghana

Kenya

SSA

LMI

Growth

GDP growth (%)

Growth volatility

Standard deviation of GDP growth (2006–2014) Poverty headcount ratio at 1.9$ per day (2013) Poverty growth (2006– 2013) kWh per capita (2014) Interconnections exist Exports/total generation (2015)

3.5% (2015) 4.5% (2016p) 7.7% (2017p) 3.15

5.6% (2015) 6% (2016p) 6.1% (2017p) 2.22

3.4% (2015) 3% (2016p) 4% (2017p) 1.24

2.6% (2015) 2.5% (2016p) 3.4% (2017p) 1.22

24.2% (2012 according to national poverty line) 7.7%

50% (2012 based on national poverty line) 1340 +4%

42.7% (2012)

n.a.

n.a.

n.a.

354.7 Yes 4%

166.7 Yes and more being built 0.4%

480.6

769.4

Poverty

Electricity consumption External sales of power

Sources: ‘Growth’ and ‘Growth volatility’: World Bank (2018a); ‘Poverty’: Ghana: GSS (2014); Kenya: KIPPRA (2013); SSA: World Bank (2018a); ‘electricity consumption’: World Bank (2018a); ‘External sales of power’: Ghana: EC (2015); Kenya: KPLC (2015).

97

A. Pueyo / World Development 109 (2018) 85–100 Table 6 Average capacity factors of renewables in Kenya, Ghana, Sub-Saharan Africa and internationally. Average capacity factors Technology

Ghana

Kenya

Africa

International

Wind onshore Solar PV Hydro

25% 17% 50%

45% 20% 55%

32% 22%

30% 20% 50%

Sources: Pueyo et al. (2016).

Table 7 Indicators of regulatory risk in Kenya and Ghana. Regulatory risks Topic

Indicator

Ghana

Kenya

Planning

Number of RE projects in the pipeline (with provisional license) Projects that have been built

82

50 projects for 1139 MW in the pipeline

2 (22.5 MW)

Number of projects and MW procured through competitive bids Number of steps and licences required to generate electricity Length of RE price guarantee Exchange rate and inflation risk

1 solar PV project, 20 MW (bid announced but not awarded) 11

All planned except 360 MW coal and 110 MW wind 7 fossil fuel projects for 512 MW

Access to grid guaranteed by Law Connection cost covered by utility % of generation capacity controlled by the State

Yes No 73%

CPIA property rights and rule-based governance index 2015

4

Procurement

Contracting and market access

10 years Not covered

10 to 12, depending on the technology 20 years Yes (tariff denominated in foreign currency) Yes No 70% of capacity 80% of total electricity produced 3

Sources: World Bank (2018a); Energy Commission of Ghana (www.energycom.gov.gh); Energy Regulatory Commission of Kenya (http://erc.go.ke/).

Table 8 Indicators of off-taker risk in Kenya and Ghana. Off-taker risk Topic

Indicator

Ghana

Kenya

Operational performance Liquidity

EBIT (if positive, EBIT margin) Current liabilities/current assets cash and cash equivalents/ Accounts payable and short term borrowing Short-term debt/equity and LT debt System losses (as % purchases) Revenue collection to sales Trade receivables past due by more than a year

123 Mill cedis 154% 8% 45% 23% 89.9% 13%

12% 60% 91% 19% 17.5%

Receivables from public institutions

55%

Financial structure System losses Credit risk

3% (over a year) 11% (plus impairment)

Sources: KPLC (2015); ECG (2013).

Table 9 Indicators of resource and technology risks in Kenya, Ghana and Lower Middle Income countries. Resource and technology risk Topic

Indicator

Ghana

Kenya

LMI

Resource risk Human resources Industrial development

Resource assessments available Gross enrolment ratio in tertiary education Competitive Industrial Performance Index

Yes (except biomass) 4% 0.01 (108 of 140)

23%

Grid capacity

Enough transmission capacity to transmit projected generation

Yes 14.3% 0.074 119 out of 140 Yes

No

Source: World Bank (2018a); UNIDO Competitive Industrial Performance Reports website (www.unido.org/data1/Statistics/Research/cip.html); UNIDO (2014); GRIDCo (2014); Lahmeyer (2016).

98

A. Pueyo / World Development 109 (2018) 85–100

Table 10 Indicators of macroeconomic risks in Kenya and Ghana. Macroeconomic risk Topic

Indicator

Ghana

Kenya

Currency risk

Change in exchange rate (USD per local currency)

Inflation Fiscal stability

Consumer Price Index year on year Dec 2015 Current account deficit Fiscal deficit

65% (2011–2015) 18% (2015) 17.7% (2015) 9.2% (2014) 10.4% (2014)

13% (2011–2015) 10% (2015) 6.6% 7.1% 7.3% (2015)

Sources: Bank of Ghana website (www.bog.gov.gh/); Central Bank of Kenya website (www.centralbank.go.ke/).

Table 11 Indicators of social and reputational risks in Kenya and Ghana. Social and reputational risk Topic

Indicator

Ghana

Kenya

Social discontent

Precedents of social discontent in renewable energy projects Clear social consultation guidelines CIRI empowerment rights index (2011 is latest value)

Land grabs for growing biofuels No 11 (out of 14)

Yes No 4

Human and workers’ rights

Sources: ActionAid (2012); Danwatch (2016); McGovern (2016); CIRI Human Rights Dataset (2016).

Table 12 Indicators of governance risks in Kenya, Ghana, Sub-Saharan Africa and Lower Middle Income countries. Governance risk Topic

Indicator

Ghana

Kenya

SSA

LMI

Corruption

CPIA transparency, accountability and corruption in the public sector index (1 low–6 high) Transparency International Corruption Perception Index (rank)

3.5 56 out of 167 global 7 out of 47 SSA 35.2% 35.1% 1.7 Low

3 139

2.74

2.97

33.4% 34.6% 7 High

31.7% 29.5% 14.4

5.9

Security

Percentage of firms expected to give gifts to secure a Government contract Percentage of firms expected to give gifts to obtain a construction permit Intentional homicides per 100,000 Threat of terrorism

Source: World Bank (2018a, 2018b, 2013b, 2013c).

Table 13 Indicators of supply of domestic finance in Kenya, Ghana, Sub-Saharan Africa and Lower Middle Income countries. Domestic savings Topic

Indicator

Ghana

Kenya

SSA

LMI

Access to domestic savings

Gross domestic savings (%GDP) 2014 Gross capital formation (%GDP) 2014 Private credit by deposit money banks to GDP (%) 2013 Stock market capitalisation to GDP (%) 2013 GDP per capita (constant 2010 USD) Commercial bank branches per 100,000 people 2013 Bank accounts per 1,000 adults Tax revenue (% GDP)Average last 5 years

17.73 26.24 14.2 7.5 1670.7 6.0 470 13.723

5.2 22 29.1 25.4 1101.2 5.6 652 15.81

17.84 21.31 16.7 22.2 1162.1 3.9 166 15.25

23.23 25.11 32.2 16.6 1969.5 8.2 396 10.98

Low disposable income Access to banking services Tax collection

Source: World Bank (2018a, 2018b, 2015b).

Table 14 Indicators of supply of international finance in Kenya, Ghana, Sub-Saharan Africa and Lower Middle Income countries. Foreign savings Topic

Indicator

Ghana

Kenya

SSA

LMI

Foreign investment

FDI Foreign direct investment, net inflows (% of GDP) 2014 Ease of Doing Business (2016 rank) Portfolio equity, net inflows (% GDP) 2009 Aid inflows, net ODA (%GNI) annual average 2010–2014 Net ODA received per capita (current US$) average 2010–2014

8.7% 114 (11th SSA) 2.1% 4.1% 61

1.5% 108 (9th in SSA)

2.5%

2.1%

1.1% 3.1% 50.2

0.8% 0.9% 16.4

Aid

Source: World Bank (2018a).

5.14% 59.5

99

A. Pueyo / World Development 109 (2018) 85–100 Table 15 Indicators of financial intermediation in Kenya, Ghana, Sub-Saharan Africa and Lower Middle Income countries. Intermediation Topic

Indicator

Ghana

Kenya

SSA

LMI

Poor intermediation Competition

Bank lending-deposit spread 2015 Bank concentration (%) 2013 Boone indicator (2013) Bank overhead costs to total assets (%) Bank return on equity (%, after tax) 2013 Bank non-performing loans to gross loans (%) Stock market returns (%, year on year) Syndicated loan average maturity (years) 2013 Average maturity on new external debt commitments (2014)

16 70.1 0.13 5.7 31.8 12 74.4 1.1 12.1 (2014) 18 (2010–14)

8.7 51.6 0.09 4.7 15.5 5 50 9.5 16.7 (2014) 28 (2010–2014)

8.8 75.4 0.05 5.2 15.5 5.6

6.8 70.1 0.05 3.4 11.2 5.6

4.5

5

Cost Performance

Short-termism

Source: World Bank (2018a, 2015b).

References ActionAid (2012). Land grabbing, biofuel investment and traditional authorities in Ghana (ActionAid Policy Brief). Retrieved from ActionAid website: www. actionaid.org/sites/files/actionaid/policy_brief.pdf. Africa Progress Panel (2015). Power, people, planet: seizing Africa’s energy and climate opportunities (Africa Progress Report 2015). Retrieved from: www. africaprogresspanel.org/publications/policy-papers/2015-africa-progressreport. Bank of Ghana (2017). Monthly Statistical Bulletin. September 2017. Retrieved from: www.bog.gov.gh/statistics/statistical-bulletin. Briceno-Garmendia, C. & Shkaratan, M. (2011). Power tariffs: Caught between cost recovery and affordability (Policy Research working paper; WPS 5904). Retrieved from World Bank website: http://documents.worldbank.org/ curated/en/234441468161963356/Power-tariffs-caught-between-costrecovery-and-affordability. Byrne, R., Ockwell, D., Urama, K., Ozor, N., Kirumba, E., Ely, A., . . . Gollwitzer, L. (2014). Sustainable energy for whom? Governing pro-poor, low carbon pathways to development: Lessons from solar PV in Kenya (STEPS Centre Working Paper 61). Retrieved from: https://cdkn.org/wp-content/uploads/ 2014/07/Energy-Access-online.pdf. Cass, N., & Walker, G. (2009). Emotion and rationality: The characterisation and evaluation of opposition to renewable energy projects. Emotion, Space and Society, 2(1), 62–69. Central Bank of Kenya (2015). Statistical Bulletin. December 2015. Retrieved from: www.centralbank.go.ke/uploads/statistical_bulletin/1947071303_December% 202015.pdf. CIRI Human Rights Dataset (2016). Retrieved from: www.humanrightsdata.com/ p/datadocumentation.html. Collier, P., & Venables, A. J. (2012). Greening Africa? Technologies, endowments and the latecomer effect. Energy Economics, 34(S1), S75–S84. Danwatch (2016). Storm ahead for vestas as Giant Kenyan wind power project is in court. Danwatch. Retrieved from www.danwatch.dk/undersogelseskapitel/ storm-ahead-for-vestas-asgiant-kenyan-wind-power-project-is-in-court/. Eberhard, A., Gratwick, K., Morella, E., & Antmann, P. (2017). Independent power projects in Sub-Saharan Africa: Investment trends and policy lessons. Energy Policy, 108, 390–424. Edwards, G., Dent, C., & Wade, N. (2017). Assessing the potential impact of gridscale variable renewable energy on the reliability of electricity supply in Kenya. IDS Bulletin, 48(5–6), 29–48. Electricity Company of Ghana (2013). 2013 Annual report and financial statements. Retrieved from: http://www.ecgonline.info/index.php/media-center/link-postformat/100-media-center/reports.html. Electricity Company of Ghana (2014). 2014 Annual report and financial statements. Retrieved from: http://www.ecgonline.info/index.php/media-center/link-postformat/100-media-center/reports.html Energy Commission of Ghana (2012). Licence manual for service providers in the renewable energy industry. Retrieved from: www.energycom.gov.gh/files/RE% 20LICENCE%20MANUAL.pdf. Energy Commission of Ghana (2015). 2015 Energy (supply and demand) outlook for Ghana. Retrieved from: http://energycom.gov.gh/data-center/energy-outlookfor-ghana. Energy Commission of Ghana (2017). 2017 Energy supply and demand outlook for Ghana. Retrieved from: http://energycom.gov.gh/data-center/energy-outlookfor-ghana. Frankfurt School-UNEP Centre & BNEF (2017). Global trends in renewable energy investment 2017. Retrieved from: http://www.fs-unep-centre.org. Fritsch, R. (2011). Comparative financial analysis of electricity utilities in West Africa. Energy Policy, 39(10), 6055–6064. Fritsch, J., & Poudineh, R. (2016). Gas-to-power market and investment incentive for enhancing generation capacity: An analysis of Ghana’s electricity sector. Energy Policy, 92, 92–101. GRIDCo (2014). 2014 electricity supply plan (Ghana Grid Company Ltd). Retrieved from: http://www.gridcogh.com/en/press-media/electricity-supply-plan.php.

Ghana Statistical Service (2014). Poverty profile in Ghana 2005–2013. (Ghana Living Standards Survey Round 6). Retrieved from: www.statsghana.gov.gh/docfiles/ glss6/GLSS6_Poverty%20Profile%20in%20Ghana.pdf. Gratwick, K. N., & Eberhard, A. (2008). Demise of the standard model for power sector reform and the emergence of hybrid power markets. Energy Policy, 36 (10), 3948–3960. Gyamfi, S., Modjinou, M., & Djordjevic, S. (2015). Improving electricity supply security in Ghana – The potential of renewable energy. Renewable and Sustainable Energy Reviews, 43, 1035–1045. Hausmann, R., Rodrik, D. & Velasco, A. (2004). Growth diagnostics (John F.Kennedy School of Government, Harvard University). Retrieved from: http://growthlab.cid.harvard.edu/files/growthlab/files/growth-diagnostics.pdf. Hausmann, R., Klinger, B. & Wagner, R. (2008). Doing growth diagnostics in practice: A ‘Mindbook’ (CID Working Paper 177). Retrieved from: https://growthlab.cid. harvard.edu/publications/doing-growth-diagnostics-practice-mindbook. Hoka Osiolo, H., Pueyo, A., & Gachanja, J. (2017). The political economy of investment in renewable electricity in Kenya. IDS Bulletin, 48(5–6), 119–140. Hoppmann, J., Huenteler, J., & Girod, B. (2014). Compulsive policy-making – The evolution of the German feed-in tariff system for solar photovoltaic power. Research Policy, 43(8), 1422–1441. International Energy Agency (2017). Energy access outlook 2017. Retrieved from: www.iea.org/publications/freepublications/publication/weo-2017-specialreport-energy-access-outlook.html. International Energy Agency (2010). Projected costs of generating electricity. Retrieved from: https://www.iea.org/publications/freepublications/ publication/projected_costs.pdf Juárez-Hernandez, S., & León, G. (2014). Energía Eólica en el Istmo de Tehuantepec: Desarrollo, Actores y Oposición Social. Revista Problemas del Desarrollo, 178(45). Julio–Septiembre. Kapika, J., & Eberhard, A. (2013). Power-sector reform and regulation in Africa: Lessons from Ghana, Kenya, Namibia, Tanzania, Uganda and Zambia. Pretoria: HSRC Press. Kiplagat, J. K., Wang, R. Z., & Li, T. X. (2011). Renewable energy in Kenya: Resource potential and status of exploitation. Renewable and Sustainable Energy Reviews, 15(6), 2960–2973. Kenya Institute for Public Policy Research and Analysis (2013). Kenya economic report 2013. Retrieved from: https://kippra.or.ke/kippra-publications-2/ Klessmann, Corinna, Nabe, Christian, & Burges, Karsten (2008). Pros and cons of exposing renewables to electricity market risks – A comparison of the market integration approaches in Germany, Spain, and the UK. Energy Policy, 36(10), 3646–3661. KPLC (2015). Kenya power annual report and financial statements, financial year ended 30 June 2015. Retrieved from: http://kplc.co.ke/img/full/jSsYVq47rObE_ KENYA%20POWER%20ANNUAL%20REPORT%202015%20-%20FOR%20WEB.pdf Lahmeyer International (2016). Development of a power generation and transmission master plan, Kenya: Long term plan – energy efficiency: 2015– 2035. Retrieved from the website of Kenya’s Energy Regulatory Commission: https://www.erc.go.ke/index.php?option=com_content&view=article&id=167& Itemid=680. McGovern, M. (2016). 61MW Kinangop Project Cancelled. Windpower Monthly. 25 February. Retrieved from: www.windpowermonthly.com/article/1385206/ 61mw-kinangopproject-cancelled. Malgas, I., & Eberhard, A. (2011). Hybrid power markets in Africa: Generation planning, procurement and contracting challenges. Energy Policy, 39(6), 3191–3198. Newell, P., & Phillips, J. (2016). Neoliberal energy transitions in the South: Kenyan experiences. Geoforum, 74, 39–48. Newell, P., Phillips, J., Pueyo, A., Kirumba, E., Ozor, N. & Urama, K. (2014). The political economy of low carbon energy in Kenya (IDS Working Paper 445). Retrieved from: http://www.ids.ac.uk/publication/the-political-economy-oflow-carbon-energy-in-kenya. Pueyo, A., Orraca, P. & Godfrey-Woods, R. (2015a). What explains the allocation of aid and private investment for electrification? (IDS Evidence Report 123). Retrieved from: https://opendocs.ids.ac.uk/opendocs/handle/123456789/5944. Pueyo, A., Spratt, S., Schmitz, H., Willenbockel, D., Dent, C., Wade, N. & Crossland, A. (2015b). Green growth diagnostics for Africa: Literature review and scoping study

100

A. Pueyo / World Development 109 (2018) 85–100

(IDS Working Paper 455). Retrieved from: https://www.ids.ac.uk/publication/greengrowth-diagnostics-for-africa-literature-review-and-scoping-study. Pueyo, A., Bawakyillenuo, S. & Osiolo, H. (2016). Cost and returns of renewable energy in Sub-Saharan Africa: A comparison of Kenya and Ghana (IDS Evidence Report 190). Retrieved from: https://opendocs.ids.ac.uk/opendocs/bitstream/ handle/123456789/11297/ER190_CostandReturnsofRenewableEnergyinSub SaharanAfricaAComparisonofKenyaandGhana.pdf. Renewable Energy Policy Network for the 21st Century – REN21 (2017). Renewables 2017 global status report. Retrieved from: www.ren21.net/wpcontent/uploads/2017/06/17-8399_GSR_2017_Full_Report_0621_Opt.pdf. Republic of Kenya (2007). Kenya vision 2030. Retrieved from: http://theredddesk. org/sites/default/files/vision_2030_brochure__july_2007.pdf. Suberu, M. Y., Mustafa, M. W., Bashir, N., Muhamad, N. A., & Mokhtar, A. S. (2013). Power sector renewable energy integration for expanding access to electricity in Sub-Saharan Africa. Renewable and Sustainable Energy Reviews, 25, 630–642. Transparency International (2016). Corruption perception index 2015. Retrieved from: https://www.transparency.org/news/feature/corruption_perceptions_ index_2016. UNIDO (2014). Competitive Industrial Performance Report 2014, Working Paper 12/ 2014. Vienna: United Nations Industrial Development Organization. United States Government and Government of Ghana (2011). Ghana constraints analysis. Partnership for growth, Final Report. Retrieved from: https://20092017.state.gov/documents/organization/202533.pdf. Waissbein, O., Glemarec, Y., Bayraktar, H. & Schmidt, T.S. (2013). Derisking renewable energy investment. A framework to support policymakers in selecting public instruments to promote renewable energy investment in developing countries. Retrieved from UNDP’s website: http://www.undp. org/content/undp/en/home/librarypage/environment-energy/low_ emission_climateresilientdevelopment/derisking-renewable-energyinvestment.html. World Bank (2013a). Energizing economic growth in Ghana: Making the power and the petroleum sectors rise to the challenge. Retrieved from:

http://documents.worldbank.org/curated/en/485911468029951116/ Energizing-economic-growth-in-Ghana-making-the-power-and-petroleumsectors-rise-to-the-challenge. World Bank (2013b). Kenya country profile 2013 (Enterprise Surveys, Washington DC: World Bank). Retrieved from: www.enterprisesurveys.org/~/media/GIAWB/ EnterpriseSurveys/Documents/Profiles/English/Kenya-2013.pdf. World Bank (2013c). Ghana country profile 2013 (Enterprise Surveys, Washington DC: World Bank). Retrieved from: www.enterprisesurveys.org/~/media/GIAWB/ EnterpriseSurveys/Documents/Profiles/English/Ghana-2013.pdf. World Bank (2015a). International development association – project appraisal document on a proposed credit in the amount of SDR 26.3 million (US$37 million equivalent) to the Republic of Kenya for a financial sector support project, 9 April. Retrieved from: http://documents.worldbank.org/curated/en/ 268521468039041661/text/PAD12320PAD0ID060Box391426B00OUO090.txt (accessed 24 April 2017). World Bank (2015b). Little data book on financial development 2015–16. Retrieved from: https://openknowledge.worldbank.org/bitstream/handle/10986/22553/ 9781464805547.pdf. World Bank (2017a). Global tracking framework-progress toward sustainable energy 2017. Washington DC: The World Bank and the International Energy Agency. World Bank (2017b). World bank enterprise surveys. Retrieved from: www. enterprisesurveys.org/data/survey-datasets. World Bank (2017b). Ghana overview. Retrieved from: http://www.worldbank.org/ en/country/ghana/overview. World Bank (2018a) World development indicators. Retrieved from: http://databank.worldbank.org/data/reports.aspx?source=world-developmentindicators. World Bank (2018b) Private participation in infrastructure database. Retrieved from: http://ppi.worldbank.org. Wüstenhagen, R., Wolsink, M., & Bürer, M. J. (2007). Social acceptance of renewable energy innovation: An introduction to the concept. Energy Policy, 35(5), 2683–2691.