Introduction

Electricity generation has been closely linked to the economic development of Ghana since its independence from colonial rule. At the heart of this is the 1,020 MW Akosombo Dam, operated by the state-owned Volta River Authority (VRA) since 1965. Originally, VRA's primary customer was Volta Aluminium Company Ltd (Valco), initially owned by Kaiser and Reynolds Metals, USA, which took up to 60% of its output. VRA also served a number of industrial customers as well as other domestic customers through its sales to Electricity Company of Ghana (ECG), which was set up to be the primary electricity distributor in Ghana. A further substantial part of VRA's output was exported to neighboring countries. Domestic prices were kept below marginal costs, subsidised mainly by export sales. Over time domestic consumption in Ghana increased, leading to a reduction in exports (even though these produced higher revenues than domestic sales). In 2004 the Government of Ghana (GoG) purchased a majority interest in Valco and the smelter closed down. As consumers, now using a much higher proportion of the generated power, were not paying its real cost, this led to wasteful consumption, and in effect the more power it sold, the more ECG faced financial difficulties. Normalising of tariffs began in the mid-1990s but little progress was made because it was too political, an issue (at a time when Ghana was returning to democracy). To meet increasing demand, in 2000 VRA completed Ghana's first major (330 MW) thermal power station at Takoradi. As of 2015, with several other projects completed since Takoradi, the notional installed capacity in Ghana was 2,450 MW, but actual availability was around 2,000 MW. Factors behind this included the low level of water at Akosombo, meaning that it was operating at only 40% of its capacity. Load shedding had been taking place for several years. Project Development In the early 2000s it was already evident that signifcant further generation capacity was needed, and that the scale of investment required meant that private sector needed to get involved. In 2003 Reltub, a Ghanaian investment company, made an unsolicited proposal to GoG to build a 340 MW power plant, and set up Cenpower Generation Co. Ltd (Cenpower) to develop this independent power project (IPP).

At the same time, plans for the West Africa Gas Pipeline (WAGPCO), intended to distribute Nigerian gas to other parts of West Africa, were reaching fruition following the signing of an intergovernmental agreement in 2000. As it was intended that the new power station would use some of this gas, a project site was chosen in the Tema industrial zone in the municipality of Kpone, some 24 km east of Accra, close to where the pipeline (completed in 2006) was to come onshore. In 2006, eleQtra, a UK-based developer acting on behalf of Infraco Africa (a newly formed project-development fund owned primarily by European DFIs), became a development partner in the project. By 2007 the project site had been secured, technical studies and design work undertaken and a generating license issued. A memorandum of understanding for a PPA was signed with ECG, covering 60% of the plant's capacity, the intention being that the balance would be supplied directly to mining companies. At that time, ECG had no external advisers, but these were later provided by USAID. The initial version of the PPAfor which there then was no standard form in Ghanatook two years of detailed negotiation thereafter and was signed in 2009. Having reached this stage in 2009, with a considerable sum already spent in development costs, the original developers were obviously under some financial pressure. The recently-created Africa Finance Corporation (AFC), a DFI based in Nigeria, therefore took up the role of lead developer in 2010, paying its share of development costs to date (as well as in the future). AFC made two major changes to the project:

Firstly, it had become apparent that no gas would be available in the foreseeable future from WAGPCO. Although originally only 25% of its capacity was earmarked for VRA, supply was restricted because of a deficit of gas from Nigeria. (Gas was being exported elsewhere by sea.) Thus, all the gas available for Ghana had already been committed. On the other hand, in 2007 Tullow Oil made a significant discovery of offshore gas in Ghanaian waters, so it could be expected that in due course there would be gas for the project. AFC therefore took the project forward as a dual-fuel plant, using light crude oil (LCO) until gas becomes available.

Secondly the mining companies lost interest in committing to a long-term PPA, since they considered it would be too expensive using LCO, and they believed that gas would be cheaper (which it would have been had there been a supply of gas). AFC's involvement was valuable not only because it relieved the development-cost burden on Reltub and Infraco, but also because the arrival of a regional player made GoG more comfortable with the project. In particular, this resulted in an agreement that all the power from the project would be sold to ECG, and that GoG would provide a guarantee for ECG's financial obligations, which it had previously refused to do. However, it took another three years to renegotiate the PPA to ensure bankability

The Power-Purchase Agreement Key terms of the PPA, signed in 2012, included:

Cenpower is responsible for designing, building, financing and operating the project for the 20-year term of the PPA.

The power-sale tariff consists of three main elements:

A fxed payment, calculated to cover debt service and return on equity

A variable payment, indexed against a combination of engineering and wider economic indices, calculated to cover operating costs

A fuel payment, calculated to cover the cost of fuel consumed, whether LCO, natural gas, or distillate. (As discussed below, ECG takes no risk on fuel supply, only the price.)

The tariff is denominated in US dollars but payments are made in Ghana cedi. The initial tariff is approximately 22/kWh, based on LCO at $100/barrel. The tariff will reduce if there is a reduction in the price of LCO (and vice versa), or if the project is able to switch to using gas.Cenpower is incentivized to switch to gas within two years from the start of commercial operations as its return on equity then increases. Conversely, if this has not taken place after three years, Cenpower's return suffers until the switch is made. ECG took no risks on whether the project could be completed on-time, on budget, or to the required specifications. There are penalties due to ECG for late completion, supported by a performance bond. Similarly, if O&M costs are higher than expected this is Cenpower's problem. If the PPA is terminated because of a default by ECG (e.g. non-payment), ECG is required to repay the debt and compensate the equity investors. ECG has an option to purchase the plant for $1, plus any outstanding sums due to Cen power, after 20 years. GoG guaranteed ECG's offtake and termination payment obligations. The EPC Contract and Debt Finance as a result of the change to a dual-fuel plant, the EPC contract was retendered. Luckily, this followed the 2008 financial crisis, when many projects were cancelled and bids were very competitive, so the EPC contract ended up some 30% cheaper than the previous estimates. It was a key requirement of the bids that bidders should be in a position to offer debt finance for the project. The bid was won by Group Five, a major South African contractor. Export-credit finance was arranged by Rand Merchant Bank (RMB), with insurance from Export Credit Insurance Corporation of South Africa (ECIC). This was the first time that ECIC had provided cover for an IPP project. The RMB-arranged bank syndicate were insured 100% for political risks and 85% for commercial risks. Additional export-credit finance could have been obtained by Group Five's major subcontractors, the turbine supplier (GE, USA) and the boiler supplier (Siemens, Germany), but FMO (the Dutch DFI) had already arranged the balance of the financing required ($110m) from a syndicate of DFIs on attractive terms, and it was considered to be simpler to proceed with this rather than at least two other separate financings.

Interest-Rate Risk the interest rate on the debt raised for the project was based on the short-term money market rate (the London Interbank Offered Rate [LIBOR] for US dollars) because the lenders are using short-term funding. This means that the interest rate is reset to the then-current market rate every six months. This is a standard way of providing long-term finance in US dollars but it leaves the project company open to the risk that there could be an increase in interest rates that could jeopardize its financial stability. Lenders therefore generally require the project company to enter into interest-rate swaps to fix their interest rate, for which there is also an active market. The swap providerusually a commercial bank that is also one of the lenderspays the project company the difference between the fixed rate and the LIBOR rate if the latter is above the fixed rate, and vice versa. But this means that the swap provider has a credit risk on the project company if the swap fixed rate is above the LIBOR rate. This proved to be a difficult issue in Cen power's case, as no-one was willing to take such a risk over the 20-year life of the project, given concerns on the credit of ECG, and the insurance from ECIC did not cover this swap risk. RMB finally stepped in to solve this problem by making use mainly of interest-rate caps instead of swaps. In an interest-rate cap the project company is paid the difference if the LIBOR rate goes above the capped rate, but does not pay anything if it is below the capped rate. The problem with caps is that, unlike swaps, the project company has to pay an initial lump-sum fee, which of course adds to the project's initial costs, and hence it's financing requirements, but this was the only realistic solution in this case. Fuel Supply Another key building block in the project structure was the fuel-supply agreement (FSA). The developers originally asked ECG to take responsibility for the supply of LCO, i.e. to sign a tolling contract, under which ECG would supply fuel and pay Cen power for processing it, but ECG felt unable to do this as it had no experience in this respect. (Had ECG done so it would have removed the problems over the fuel-supply arrangements) So Cen power concluded an FSA with Vitol Group (Netherlands), one of the world's largest oil traders. (Another key reason for appointing Vitol was that it is extensively involved with Ghana's prospective offshore gas Felds.) The FSA is a take-and-pay agreement, i.e. Cen power is not obliged to purchase LCO from Vitol, but if it does so it pays the market price (and ECG takes the risk on market-price movements). But this apparently straightforward arrangement required considerable negotiations with the lenders, who were concerned about the certainty of the fuel supply, especially as Vitol was subject to limited penalties for failure to supply. (Given its limited profit margin on this type of contract, it would make no commercial sense for Vitol to accept a substantial liability for non-delivery, despite the adverse effect on Cenpower.).

The main requirement was for the construction of a fuel-storage facility holding 50 days' supply of LCO so that short-term interruptions in deliveries would not stop the project running. Such a major facility would not have been required for a gas fired project supplied by a pipeline, and the project site did not have enough room to construct it. However, there is a tank farm immediately adjacent to the project site and it was agreed that the necessary storage would be constructed on this site. This raised some complex security issues, as the site did not belong to the project. Furthermore, the construction of the tanks was not part of the EPC contract, but was managed by Vitol on behalf of Cenpower. This added significantly to the project's capital cost: not only did the cost of the fuel tanks have to be paid for, but the first fill of the tanks, and the LCO in transit, also had to be financed. (Thereafter the fuel used would be paid for through the tariff.) Moreover, the lenders required the cost of this initial LCO supply to be hedged, as otherwise, if the price had gone up, there would not have been enough finance available to cover the cost. This was done when the price of oil was around $100. The total additional financing required was $145m: FMO arranged a further DFI-funded $93m mezzanine loan (i.e. subordinated to the other debt), known as the Fuel Finance Facility, with the balance being covered by additional equity.

ECG Credit Risk

The other key concern was the credit risk on ECG as power purchaser. ECG's latest published financial statements, for 2014, show that its net loss before tax increased from GH22m in 2009 to GH148m in 2014. Like many other state power-distribution companies in sub-Saharan Africa, ECG was suffering from a combination of low tariffs, power theft and non-payment (including non-payment by the government, its largest customer). Recent government policy has begun to remedy this situationas a result ECG made a small operating profit in 2014 for the first time for many years. It is understood that a commercially-functioning electricity sector is essential for economic development and there is a political consensus in this respect (instead of the opposition party always objecting to tariff rises). ECG's tariffs have therefore been substantially increased and are now not far below actual cost. GoG has also announced plans to inject private capital and management into ECG. It is evident why lenders required a GoG guarantee of ECG's obligationsbut such a guarantee should only be the second way out. The lenders were therefore concerned to build in cash-flow buffers to cover delays in payment by ECG. These took the form of a bank letter of credit, plus a cash reserve account (another addition to project costs) initially covering 125% of the next six months' debt service (principal repayments and interest payments). So long as the project runs on LCO its cost to ECG is relatively high. It will reduce if the project switches to gas. However, when considering the cost of the project, the following must be borne in mind:

Cenpower is building the sub-station, which would usually be built by the grid operator, which adds to the project's costs, as do the fuel storage arrangements.

The estimated tariff of 22/kWh assumes LCO at $100 per barrel: the current price is well below that level, and if this continues the tariff will be lower.

The alternative source of power would be temporary diesel-based generation, with a cost higher than Cenpower's. It is reasonable to expect that the conversion to gas-firing will take place in the next few years. Financial Close It took until 2014 for all these arrangements to be completed (including parliamentary approval of the GoG guarantee for ECG). At that point Infraco dropped out of the project and new shareholders came into the pictureSumitomo Corporation of Japan (who will operate and maintain the project in a joint venture with AFC and Cenpower's holding company), a South African infrastructure investment fund and FMO. The project is close to the schedule for completion in 2017. When operational, Cenpower will provide approximately 10% of Ghana's installed capacity and 20% of its thermal capacity. PPP Bill 2016. In 2016 there was a slight cloud on the horizon for Cenpower, in the form of a PPP Bill that received its second reading in the Ghana parliament in July. The proposed PPP Act included 'transitional provisions' for PPPs signed after 2011, when the National PPP Policy (the Policy) was publishedas is the case with this project. In such cases PPP's have to be in accordance with the Policy, or if not they need to be regularized by a new Ghana PPP Agency (the Agency).

The Policy is vague on the subject of unsolicited proposals, requiring that they should be considered only on the basis of guidelines 'to be issued', and should not relate to projects already on the relevant authority's project list. They are also subject to further criteria to be set out in a PPP Manual (not yet published). The Ministry of Finance and Economic Planning issued an interim toolkit for unsolicited PPP proposals in 2012. It reflects the Policy's requirements, but also includes a further requirement that unsolicited proposals should be subject to a competitive procurement process including the proposer (cf. Lekki Expressway, where this was done). It therefore seems possible that if the PPP Bill is passed as currently drafted, the Cenpower project, which did not go through a competitive procurement, could be considered not to be in accordance with the Policy, and so could be subject to regularization by the Agency. It is unclear what this means, but it could be interpreted as giving the Agency power to make changes in the PPA without Cenpower's agreement. Moreover, a project that is not regularized in the period specified by the Agency could be subject to a new procurement.

Policy Points General

Legal/institutional framework: The development of the project took longer than it might have done because the framework in which it was to operate was not settled. A lot of time was spent on complex negotiations with the mining companies, for example, that came to nothing. Similarly, specific parliamentary approval was required for the GoG guarantee for ECG's liabilities, which significantly delayed financial close. Capacity building. Negotiations were also prolonged because ECG had no experience in project finance, or of negotiating a PPA (although the USAID support helped to strengthen ECG's capacity in this respect). However, the experience gained has been used to good effect in developing a standard-form PPA that is being used in subsequent procurements. Project Structuring Interface risk. It is usually difficult to develop a project that relies on the successful development of another. The initial decision to structure the project based on gas supply via WAGPCO was understandable, but flawed in this respect. In a similar way, the construction of the fuel storage tanks was not part of the EPC contract, and thus also raised interface risks (e.g. what would happen if they were not completed on time?), but fortunately, these risks were considered to be acceptably limited.

Procurement

Development risk: It is important for public authorities to understand the high level of financial risk undertaken by project developers. This project took more than 10 years to develop and ran through development costs totaling $39m from three developers. This cost is not especially high considering the overall size of the project$900mand the length of time it took to develop. The developers also provided a substantial development bond, which would have been forfeited had the project not reached financial close. If the project had not been successfully developed the loss to the developers would clearly have been substantial, and the developers could not be sure that they would succeed until all the elements of the project were in place.

Africanisation

One of AFC's key aims was to Africanise the project, so the majority of the equity and debt are provided by African sources; the EPC contractor is African; and similarly, the O&M Agreement provides for Sumitomo Corporation to train the local staff of the O&M company to enable the latter to gradually take a larger role in operation and maintenance.

Unsolicited bid

There was no competitive pressure on the developers, which meant that ECG just had to swallow the development costs of the project as they mounted up, and increase the tariff accordingly. Had a competitive procurement been undertaken ECG might have had more control on costs. However, the EPC contractthe largest element of the project costwas procured on a competitive basis. The other problem with unsolicited bids is lack of transparency, which makes them liable to attract attacks, especially after a change in government. It seems that Cenpower may be vulnerable in this respect, given the draft provisions of the 2016 PPP Bill.

Finance

Lender requirements. As discussed in the Overview, the lenders are the parties with the most money at risk, and unlike the investors, they have no 'upside' if the project goes well, since they earn a fixed rate of interest only, but they do face a 'downside'a loss on their loanif it goes badly. Not surprisingly, therefore, the lenders are the most conservative party at the negotiating table. In this case, this resulted in significant extra project costs to satisfy the lenders' requirements relating to the fuel-supply arrangements and the credit risk on ECG. Ideally, the PPP contract and key subcontracts for construction and O&M should not be signed until they have been cleared by the lenders, since forced renegotiation of these to meet lenders' requirements is highly likely. This is exactly what took place in this project.

A guarantee is not the first way out. The weak financial condition of ECG certainly makes a GoG guarantee of its payment obligations necessary, but it is not in the interest of either the project company or ECG or GoG for the guarantee to be the first way out (i.e. as soon as there is a shortage of cash to pay the monthly bill, the project company has to go straight to GoG). GoG may well have difficulty covering ECG's overdue payments at short notice. Hence the importance of the liquidity arrangements the lenders put in place, covering short-term payment problems.

Interest-rate risk.

Some DFIs may be able to provide projects with fixed interest rate debt, but commercial banks, whose deposits and money-market funding are on a short-term basis, cannot normally provide long-term debt at a fixed interest rate. Interest-rate hedging (usually through an interest-rate swap) is usually needed to cover the risk of increased interest rates (but cf. Platinum Highway). Currency risk. Cf. Bujagali Hydropower for a discussion of the long-term currency risk inherent in pricing the tariff in US dollars whereas ECG's revenue is in cedi. As can be seen from the historical exchange rates between the US dollar and the cedi in the Fact Sheet, there has been a very large depreciation of the cedi in recent years. If this continues it could add an extra burden on ECG's already weak finances (although it has to be borne in mind that the weakness of the cedi will cause an increase in inflation, making it easier for ECG to pass this extra cost on to its customers so long as the cedi depreciates reasonably gradually). There is also a short-term currency risk between the time Cenpower bills ECG in GH based on the US dollar exchange rate on the monthly billing date, and the time ECG pays the bill, as the exchange rate will probably have changed by then. This is dealt with by an adjustment to the following month's bill.

Sale of shareholdings. Although acceptable in this case, given that the particular role of Infraco is as a DFI that takes the risk on developing projects, in general it is undesirable for sponsors/developers to be allowed to sell their equity investment, at least until construction is complete and it is operating normally. Developers are usually able to charge their development costs to the project company at financial close, as well as a development fee to compensate for their risk during the development phase. They should then be expected to continue as significant equity investors to ensure that has been no temptation to cut corners in the project development just to earn a development fee (cf. Bujagali Hydropower, DTI Campus, Platinum Highway). Construction Phase

Late completion.

Should a PPP impose penalties and bonding for late completion, as in this case? In general, the main penalty for late completion of a PPP is that the project company loses revenue, and adding penalties and bonding to this just adds to the costs of the project. In this case, however, ECG could argue that if Cenpower does not come on stream when expected it may have to use more expensive power from other sources, and this marginal cost should be covered.

Question from the case study:

Question 1

Re-arrange the case-study under the following order of topics: 1.1 Corporate Objectives, 1.2 Background; Mission, Vision, and Core values, 1.3 Statement of Requirement, 1.4 Solution, 1.5 Do Nothing, 1.6 Project Build-Method Plan, 1.7 Operational Configuration, and 1.8 Risk Analysis)

Question2

Discuss how your project team can perform effectively and efficiently to ensure that the business case you are presenting in question 1, will be accepted by the top management.