A delay of a mere month in bringing the second round of renewable energy independent power producer (REIPP) projects to financial close represents something of a coup for the South African government. The first round saw the government and state utility Eskom run over schedule by almost half a year as they tried to complete documentation, requiring significant revisions to the programme’s timetable. REIPP’s volume of investment is unprecedented, however, and in hindsight the original targets seem unrealistic.

After closing on 1,416MW of capacity as 28 projects for the first round in November last year, the government signed project documentation with second round developers at the beginning of May. The second bid window featured nine solar photovoltaic (PV) projects with a combined capacity of 417.1MW, seven wind projects of 562.6MW in total, two mini-hydro projects of 14.3MW each and one concentrated solar project of 50MW. Eskom hopes to connect 3,725MW of extra capacity to the national grid following the first three rounds, and has so-far unclaimed allocations for landfill gas, biogas and biomass available in the last of the initial bidding windows.

Financial close for the first two rounds and bid submissions on the third were all originally scheduled to take place in 2012, but local banks, Eskom, the government and advisers all struggled to manage the volume of work. Following close on the first round, the government said that there would be only one bid window each year, for a combined capacity of 1,000MW, and that third round bids would be postponed until August 2013. With this change in pace, programme participants were able to focus properly on the second round. 

Competitive pressure

As with the first set of deals, the second round’s 19 projects feature a range of international and domestic developers. Several sponsors, including Mulilo, SolarReserve, MEMC, Scatec Solar and Solar Capital, have been involved in both rounds, while major international developers such as ACWA Power, Tata Power and GDF Suez have joined the programme for the first time in round two. Some reports suggest Eskom has received as many as 400 requests for grid connections for the next round, and most observers expect between 100 and 120 bids to be submitted in the middle of the year. The utility is only likely to approve around 20 of those bids, which could put further pressure on already squeezed margins.

“The popularity of the renewables programme in South Africa has led to highly competitive bidding to offer the lowest tariff,” explains Nick Buckworth, partner at Shearman & Sterling. “Sponsors will need to be increasingly cost conscious and lenders will be under pressure to reduce margins and fees to maintain sponsor appetite.”

Average tariffs fell for each of the technology categories in round two. Solar PV deals were down from R2.75 ($0.288) per kWh in round one to R1.65 per kWh this time, wind prices declined from R1.14 per kWh to R0.89 per kWh, and CSP prices fell from R2.68 per kWh to R2.51 per kWh. Local banks dominated both rounds, and they say that pricing on debt has fallen between 20bp and 30bp. First round deals tended to have a margin of around 350bp over the Jonhannesburg Interbank Agreed Rate (Jibar). While part of this tightening can be explained by a general improvement in banking liquidity in the South Africa between the rounds, developers have also exerted pressure. Some major sponsors are planning to roll out a fleet of plants over the course of the programme, and banks have been willing to offer cheap margins to capture repeat business.

Many bankers believe that pricing for local banks has bottomed out, but Alastair Campbell, head of power finance at Standard Bank, argues that it is hard to justify the decline in margins, given the risk still present in the REIPP projects. “The programme has not been de-risked. There are no plants up and running yet and no one knows how they are going to perform. There is not a lot of wind data in South Africa, and only time will tell if the pricing is right. We used the S&P, Fitch and Moody’s breeze bonds in Europe as benchmarks to structure our deals off, and there we saw a year-on-year decline over the last decade.” Although solar performance is easier to predict, the country’s data sets only stretch back three years and give little comfort to lenders.

International impact

Standard Bank and Nedbank Capital had the most success in round one. Both have taken a step back to lend to just five projects each this time, and development finance institutions (DFI) such as the IFC, IDC and OPIC have plugged the gap, but local banks still took the lion’s share of the financing opportunities. The main obstacle to international lenders providing debt to the programme is the government’s insistence that all power purchase agreements must be denominated in Rand, making dollar lending difficult. With the South African currency so volatile, few foreign banks have cheap enough access to Rand to lend directly, and would be dependent on South African lender for hedging.

Buckworth adds: “Lenders are typically averse to assuming currency exchange risk and look for hedging solutions. This can be challenging given the timeframes involved in project financing. The IFC and EIB, for example, amongst the DFIs have taken a lead in offering Rand financing. From our discussions with them, several major Asian commercial lending institutions are now considering how they can adapt to lend in Rand.”

More international lenders are expected to participate in round three but they will not just have to get comfortable with currency risk. Financings for REIPP projects have tenors of 15 years plus construction, but most international banks are reluctant to lend beyond 10 years in their own currencies. Given the infancy of the programme, covenants are more conservative than in many other markets. Export credit agencies (ECAs) should have had a larger role in REIPP financings, given the programme’s dependence on foreign equipment. But the guarantee that Hermes extended on Dorper Wind in the first round has turned out to be an exception.

Developers have been wary of the delays to documentation that ECA involvement creates, and ECAs have provided little clarity about whether they would honour guarantees in dollars or rand. So far ECAs have not been much cheaper than local lenders, and it is notable that the easiest second round deal for Nedbank to conclude was the 10MW Kakamas small hydro plant in Neusberg. Kakamas had local equipment suppliers that did not push as hard as their international contemporaries on performance warranties and liquidated damages. Financial and legal advisory mandates, not to mention engineering, procurement and construction contracts, are increasingly likely to go to local firms as domestic expertise improves, and margins will be squeezed.

Bond ambitions

Funding issues are likely to become pressing in future rounds, and concerns over local bank liquidity persist. Mike Peo, head of infrastructure, energy and telecommunications at Nedbank, comments: “We have got away close to R80 billion-worth of projects so far, so the market is nervous about when the local financial institutions will start running out of capacity.” Standard Bank and Nedbank are both committed to participating in rounds four and beyond, which will involve the construction of another 3,200MW of capacity, but are also in the process of trying to sell down their initial debt commitments. Bond financing is likely to play a large part in this process.

Peo explains: “There is a very deep institutional investor market in South Africa, and with a lot of the projects being CPI inflation linked, they are natural fits for these funds. Historically a lot of these institutions did not like construction risk, but every single major asset manager is securing mandates for renewable energy and beefing up their teams to be able to deal with the asset class.”

Because of institutions’ discomfort with construction risk, most of the bond financings for REIPP projects will be used to refinance banks, but local banks are seriously looking at bonds as a primary debt instrument. The Touwsrivier project from the first round (see box above) shows that bonds can be closed in the local market before construction is complete, and lenders are now looking at how they can pool projects to make them more attractive to bondholders. S&P affirmed South Africa’s sovereign rating at BBB in March, meaning that structuring an investment grade bond for the international market will be tricky. All of the investors on the Touwsrivier bond, for instance, were South African.

Although the second round may have been a relatively calm process for REIPP programme participants after the storm of the initial projects, choppy waters lie ahead. On top of funding concerns, the country’s transmission grid urgently requires upgrades and good locations near existing transmission lines are becoming harder to find. South Africa is expected to launch a slate of new cogeneration plants to the market later this year and follow that with a new generation of coal-fired IPPs, putting further pressure on local liquidity. 

Touwsriver’s rapid refinancing

None of the REIPP projects has yet completed construction, but Soitec has already closed a bond refinancing for its first round deal, the 44MW Touwsrivier solar photovoltaic plant. The issue is the first ever bond financing for a renewables project in a country with a relatively undeveloped capital market. The bond did not feature in the developer’s original financing plan, is priced higher than any of the first or second round REIPP project loans, and is likely to be refinanced. But it demonstrates the potential for bond financings to be a part of the programme. Developers may become increasingly interested in alternative funding sources as future bidding rounds approach.

Soitec is the main sponsor for the R2 billion ($210 million) Touwsrivier development, holding 60% of its equity, while local community groups (5%) and broad-based black economic empowerment organisations (35%) own minority shares. Soitec had originally lined up Investec as sole lender for the plant’s financing, with an indicative debt/equity split of 70:30. Negotiations broke down between the bank and the developer last year, however, and Soitec was left with a funding gap as the date for financial close crept closer.

Soitec decided to finance the whole project on balance sheet, meaning it could take part in the closing ceremony with the rest of the first round developers last November. It then began discussions with Standard Bank on a refinancing of part of its shareholder loan in the bond market. A week before the sponsors of the second round deals signed on their financial documents, Soitec closed a R1 billion bond issue with a fixed interest rate of 11%. Moody’s has given the 16-year issue a Baa2 local rating, and will review it on an annual basis. In total, around six investors subscribed to the bond, all of them local.

There is currently strong appetite in the South African market for higher yielding bonds, but institutional investors could still be forgiven for being wary about the construction risk attached to the unfinished Touwsrivier. Standard Bank increased their comfort by structuring a reversion right into the deal. Fundings will only be released, and the bondholders will only assume construction risk, once the project is connected to the grid, is producing power at 50% capacity and has around 70% of construction material on site. This reversion event ties into the government’s requirement for the programme that unless 50% of the name-plate capacity that was bid by sponsors comes online, Eskom can terminate its power purchase agreements.

The Touwsrivier bond is the longest-dated issue ever at its rating level in South Africa, and the total size of its BBB-rated market was only around R11 billion before Touwsrivier closed. The 16-year bonds start amortising after the project hits its reversion event. The first interest payment is due in June 2013, with capital repayments starting two years after that.  Local lenders think reversion rights could be incorporated into other REIPP deals, as long as they have standardised and predictable construction periods. Bonds are more likely to be used to refinance first and second round deals after construction is complete, due to the programme’s infancy. With other first and second round deals coming in at around 300bp above Jibar, Soitec is paying more for its bond debt, but not by much, considering that it has no precedents.