Risk Allocation in Wind Projects
The developer of a windpower project must contend with many risks that may not affect a more conventional power plant. These risks range from fickle winds that blow in different ways at varying times and places, to a fickle Federal government that approves for only limited time periods the production tax credit (currently 1.8¢ per kilowatt-hour) that accounts for a significant portion of the value in the project, leading to boom and bust cycles in wind farm development.
As is the case with more conventional power plants, a fairly well-developed risk-allocation regime has evolved for windpower projects that follows the project finance maxim of allocating risks to the persons best able to manage them. This article examines some of the key risks and how they are handled.
Wind is the “fuel” that powers wind farms and like any power project, assuring a reliable fuel supply is essential to developing a viable wind farm. Power purchase agreements for wind farms typically provide for energy-based payments – when the wind blows, electricity is produced and the developer makes money; when the winds are silent, the developer does not. Production tax credits work essentially the same way. In addition, under some power purchase agreements, the offtaker will pay a lower price for energy delivered if the windpower project does not produce enough energy to meet specified targeted amounts. Even a small shortfall in actual wind over a sustained period relative to forecast wind may, from an equity investor’s perspective, make the difference between a good investment and a bad one. Thus, the challenge is to obtain a site for the wind farm where the developer has a high level of comfort that the wind resource is as good, and as well forecast, as possible.
This determination is made by the wind consultant for the project using site-specific and regional data and studies. Wind is a variable resource. It varies by season and by time of day. It also varies from year to year. At a specific wind farm site, it can also vary at different heights and as a result of different terrain and vegetation. Thus, it is essential that the wind consultant collect data that reflects the conditions that the wind turbines can be expected to experience at their various locations on the site. In general, the more data and the higher its quality, the better the accuracy of the forecasts derived from that data. Data collection should be closely monitored to assure that quality is maintained throughout the data collection process. The goal is to obtain recorded consistent trouble-free data from each collection tower and at each tower height where data is collected. Wind resource consultants say the best data is long-term (at least one year and preferably two) and site specific.
The wind consultant collects all this data and uses various analysis techniques and computer models to create a series of probabilistic cases for wind power production (based upon forecast wind and wind farm power curve data) at the relevant site. These cases are likely to reflect 50%, 75%, 90% and 95% confidence levels for one- and 10-year probability forecasts. They may also include a P99 case, which represents a 99% confidence level. Wind data and analysis are also used by wind consultants to advise developers on the optimal siting of each wind turbine at the site and the most suitable turbine to use. Developers may also use wind forecasting to schedule maintenance during periods of anticipated lower wind speeds. Wind forecasts can also be a useful tool in scheduling and dispatching wind farm output.
Lenders to wind projects generally require, as a condition to closing, that a developer provide a wind consultant’s and independent engineer’s report reflecting a range of confidence levels (generally from 50% to 95%) for one- and 10-year probability forecasts. These numbers are plugged into the financial model for the project to determine the projected debt service coverage ratios over the range of confidence levels. Lenders address wind resource risk by typically looking for a minimum debt service coverage ratio of 1.50:1.00 over the life of the loan based upon the P50 case and a minimum ratio of 1.00:1.00 for a P90 or P95 case. In certain instances, lenders have also required developers to demonstrate that the loan can be serviced in any year under a P99 case using a portion of the debt reserve fund.
Wind risk may also be addressed through derivative products. Though not in wide use, wind hedges may be purchased that will protect a developer against low and excessively high wind conditions. Wind hedges are created on a formulaic basis to address wind risk. The formula includes fixed assumptions as to price and the power curve, among other things, with wind speeds as the variable input. The hedge provider is obligated to make payments based upon the formula in the event wind speeds are below or above agreed levels. Of course, wind hedges cost money so that, by purchasing a wind hedge, a developer is essentially reducing its upside while protecting against a downside case.
Transmission risk, and especially curtailment risk (the risk that in certain periods demand for transmission may outstrip supply, resulting in reduced or no transmission service during these periods), is a very significant issue in any windpower project. This is because wind is an intermittent resource that may render the purchase of firm transmission capacity for a windpower project uneconomic. The nature of this risk, and thus how it is addressed, is largely defined by the transmission rules for the transmission system over which the windpower project transmits its energy. (For a discussion of some of these rules, see “Federal Regulatory Issues in Windpower Projects” in the October 2004 NewsWire.)
The optimal means for a developer to address curtailment risk is to persuade the offtaker to accept this risk under the power purchase agreement. A simple means of achieving this result is to provide for an electricity delivery point at or near the wind farm site. However, this issue can be very contentious and developers are not always successful in laying off curtailment risk. This may not render a wind project unfinanceable if curtailment risk is perceived to be low. In this event, the developer may call upon a transmission consultant to analyze curtailment risk (and hopefully reach the desired conclusion) and report its findings to the developer and its lender.
Windpower projects may also have to schedule deliveries of electricity over the transmission grid. If the developer schedules delivery and then the electricity is not produced because of a failure of the wind or otherwise, imbalance charges may result. This makes good wind forecasting important, but scheduling issues may also be addressed in other ways. Again, the nature of the problem and the manner in which it is addressed is largely a function of the rules governing the applicable transmission system. Under certain transmission regimes, windpower projects may be exempted from imbalance charges or such charges may be netted on a monthly basis, making imbalances much more manageable. Under other regimes, imbalance charges are calculated and assessed over much shorter time periods so that it is more important that the offtaker assume this risk. Offtakers may be willing to do so, though sometimes at a price to the developer.
Construction and Technology Risks
The construction risks incumbent in a windpower transaction are similar to those found in any power project and are addressed in familiar ways. Optimally, a developer will contract with an experienced and financially strong contractor under a contract containing clear and strong schedule and performance requirements and incentives, including rigorous performance tests. These provisions should be designed to provide strong contractual assurances that the project will be completed on schedule and meet the production levels contemplated by the project’s power purchase agreement. Construction risk is low relative to a more traditional power project because constructing a wind project is not as technically complex and typically requires only six months substantially to complete.
Technology risk is a more serious concern in wind projects. Windpower technology has made tremendous strides over the last two decades resulting in far larger and more efficient wind turbines. However, this progress has meant that wind turbines have gone through two generations of technology in the last ten years with the inevitable problems that new technologies present. The simplest way to address technology risk is to use commercially-proven technology. However, given the rapid pace of technological evolution, this may not be the most economic option.
Technology and performance risks in windpower transactions are principally allocated to the turbine supplier or contractor through a fairly extensive set of warranties. Key turbine warranties include a warranty against defects in design, manufacture, installation or construction or a failure to comply with applicable specifications or law (a “general warranty”), a power curve warranty, an availability warranty and a serial defect warranty, among others.
In determining the adequacy of the warranty package, industry participants note some key considerations: Is the turbine a commercially-proven model with a good track record or is it a new or relatively new model? If it is a new or relatively new model, does it incorporate significant or only incremental changes compared to existing models? Also, certain turbine manufacturers have many years’ experience bringing out new models but do not have deep balance sheets supporting their warranty obligations, while some manufacturers may have deeper balance sheets than experience introducing new turbines.
The remedies for a breach of a general warranty are essentially the same as are found in a construction contract for project finance power plant transactions. The supplier must repair or replace a defective part with a new or factory reconditioned part and pay all incidental costs associated with the repair or replacement. Warranty terms for general warranties in windpower transactions are somewhat longer than in power projects using more established technology and generally range from two to five years. Repaired or replaced parts are rewarranted for the longer of the remaining warranty period or an agreed period of usually one year. A general warranty will exclude damage due to ordinary wear and tear, deficient maintenance, excess wind turbulence or temperature or force majeure events.
Under an availability warranty, a supplier will warrant the availability of each wind turbine to produce power. Availability warranties typically warrant availability of 95% to 97%, marginally less than the expected availability of 96% to 98%. Availability is measured as a ratio of total hours during a measurement period when a turbine is ready to produce power to the total hours during that period. In calculating total available hours, the supplier is not responsible for hours lost due to events like force majeure, curtailments, interconnection failures or the like, but is responsible for downtime due to turbine defects and excessive scheduled repair times.
Remedies for breaches of an availability warranty are intended to replace revenue losses resulting from the shortfall and, as such, are typically calculated based upon the energy price in the power purchase agreement and loss of production tax credits and other economic benefits (such as renewable energy credits) and may be determined based upon an agreed formula or the actual loss suffered by the developer. The supplier may also earn a bonus for availability above the 95% to 97% range.
A power curve reflects the power output of a wind turbine at specific wind speeds. A supplier will typically warrant that the actual power curve of the turbines comprising the wind farm will equal 95% to 98% of the warranted power curve, as calculated based upon a test of a representative sample of the project’s wind turbines over a specified test period. The supplier generally has the right to modify the turbines and retest them to attempt to achieve a better result. If the supplier fails to meet the warranted power curve, it typically must pay the developer an amount reflecting the economic harm to the developer resulting from such shortfall, usually for 12-month periods. There may be some offset between the availability warranty and the power curve warranty to ensure that the developer is not compensated twice for the same loss. Power curve warranties generally are limited to the same term as the general warranty, although some have extended for longer periods, including up to the useful life of the turbines (typically 20 years), where the turbine is a new, unproven model incorporating significant changes relative to existing proven models.
Suppliers may also offer a “serial defect” warranty especially for a relatively new turbine series. The serial defect warranty is intended to address the situation where a major component of a wind turbine is found to be defective in a significant number of turbines. The serial defect warranty reflects industry experience with a number of turbines that have had to be recalled or reengineered. If a component is found to be defective in a certain percentage of the project’s wind turbines, the supplier may be required to reengineer and replace the defective component for all of the wind farm’s turbine of the same type. The developer may also press the supplier to provide serial defect coverage where a serial defect is found in the same turbine at other wind farms.
Technology risk is also addressed by obtaining independent technical certification of the windpower project from one of the companies providing this service, such as Germanischer Lloyd WindEnergie GmbH or Det Norske Veritas. These companies typically apply wind turbine and wind project certification standards promulgated by the International Electrotechnical Commission (IEC). Wind turbine “type certification” is based upon a design evaluation, type testing (load measurements, black test, power performance, safety and function test), manufacturing evaluation and final evaluation. Project certification includes a site-specific assessment of the turbines and turbine specifications used in the project.
The lender’s approach to technology risk is essentially a due diligence exercise. A lender will assess the turbine selection based upon its operating history and manufacturer. It will also typically require a 20-year site-specific design certification by one of the independent technical certification companies. A lender will also assess whether the warranty package is adequate.
Finally, as is the case in any power project financing, a lender will require that an independent engineer conduct a technical review of the project and prepare a report confirming, among other things, the adequacy of the project’s overall wind farm design and that the wind turbines will operate materially in accordance with their design specifications.
The nature of a wind project makes a strong operations and maintenance regime essential – developers are paid and production tax credits accrue for energy produced. O&M risk in windpower projects is handled in a fairly traditional way. Typically, the developer enters into an O&M contract, either with an O&M affiliate company (where the developer is part of an organization with strong operational experience) or with a third-party company, usually the turbine vendor (where the developer and its affiliates lack operational experience or capacity). A developer may also enter into a long-term service agreement with the turbine vendor pursuant to which the vendor provides routine and non-routine maintenance services. These agreements are usually coterminous with the warranty period under the turbine warranties. Of course, these agreements are at a price to the developer and some developers with strong “in-house” operations and maintenance capabilities may choose to forgo long-term service agreements and instead “self insure.”
A lender will assess whether the operations and maintenance package of contracts and in-house capabilities are sufficient in making its decision whether to lend to a wind project developer. An independent engineer will also monitor the operations and maintenance of the project for the lender to assure that good O&M practices are followed. In some instances, equity investors in wind projects also insist on having an equity owners’ engineer to monitor the O&M provider’s performance. In addition, lenders will require developers to establish maintenance reserves against future scheduled and unscheduled maintenance costs.
Federal production tax credits provide approximately one-third of the capital cost of a wind project. In some states, state production tax credits may provide additional value. Thus it is essential that one of the owners of a windpower project have taxable income against which the production tax credits may be used. Where the developer of a wind project lacks a tax appetite, the developer will typically sell a portion of the project company (a limited liability company or limited partnership) to an investor with such an appetite. The limited liability company or partnership agreement will then allocate substantially all of the tax benefits (ranging from a 90/10 to a 99/1 split) from the project to the investor for the 10-year production tax credit period or until the investor achieves a specified internal rate of return (IRR), at which point there is a “flip” in ownership interests in favor of the developer ranging from (20/80 to 5/95). Some developers structure their transactions so that they receive distributions of cash until their capital is returned, after which all cash and tax allocations flow to the tax-driven investor until a return hurdle is met.
Between a developer and an investor, the risks of the project qualifying for the production tax credits will typically be borne by the developer. This risk allocation is effected through representations made by the developer either enabling a conclusion that production tax credits are available (e.g., the turbines were placed in service by December 31, 2003) or expressing the conclusion itself. The developer’s representations are typically backed by an indemnity and the indemnity is secured by developer security, including potential redirection of developer distributions to the investor.
In a financing, production tax credits are monetized through a capital contribution agreement made by the investor in favor of the project company. A lender views the capital contribution cash flow as an important part of the overall revenue stream supporting the loan and thus, while a lender is willing to take the risk that the wind may not blow and energy may not be generated for sale under the power purchase agreement and the production tax credits may not be created, a lender is not willing to take tax or change-in-law risk. Accordingly, the capital contribution obligation is conditioned only on the wind, and capital contributions must be made on the basis of the value of production tax credits that should have been earned as a result of the wind farm producing power even if the investor does not have the ability to use credits, the credits are repealed by the Congress, or the turbines are not placed in service by the deadline for qualifying for tax credits (currently December 31, 2005). However, an investor may not actually have to make the capital contributions in full to the extent that the capital contributions are not required to achieve specified debt service average ratios.
Although generally perceived as an environmentally-friendly means of producing electricity, wind farms are not free of environmental issues. The most significant concerns are bird fatalities (including migratory birds and raptors) and, more recently, bat fatalities, as well as noise and visual impacts.
Noise and visual impact concerns have not caused significant problems for wind farm developers (although the potential visual impact of a wind farm in development off-shore of Cape Cod has mobilized some prominent local property owners). This is particularly the case for wind farms located on remote sites. In addition, developers obtain noise warranties from the turbine suppliers. Under these warranties, the turbines are warranted not to exceed certain noise levels at specified wind speeds. These warranties are set at noise levels that ensure that any local noise ordinances will not be violated. The construction contract or turbine supply agreement will also contain specifications for the wind turbines, blades and towers that may be intended to produce a less offensive visual impact.
The issue of bird fatalities is typically addressed by the developer commissioning comprehensive avian studies to determine whether the wind farm site is on a migratory bird route or in an area where raptors are found. If bird fatalities are anticipated and those birds are part of a threatened or endangered species, the developer will need to obtain an “incidental take” permit from the US Fish and Wildlife Service (USFWS). Any applicant for an incidental take permit must submit to the USFWS a conservation plan that specifies, among other things, the impacts on the affected species resulting from the construction of the wind farm and the steps the developer will take to minimize and mitigate such impacts.
Technology also plays a role in mitigating against, or potentially increasing, the risk of bird fatalities. The turbine supplier will typically warrant that the wind turbine structures do not have features that are attractive to birds (such as ledges on which birds might perch). Also, the majority of migratory birds fly at heights above 100 meters. This means that larger turbines may pose a greater risk to migratory birds than smaller ones. In addition, under Federal Aviation Administration rules, structures 200 feet or higher within a certain proximity to an airport must have lights. Structures below 200 feet and within 20,000 feet of a runway might also have to have lights, depending on the results of a site survey conducted by the Federal Aviation Administration. Any lights might attract migratory birds, particularly on cloudy nights where natural navigation aids (like the moon and stars) are not visible.