The industry standard among renewable energy companies is to value projects using the traditional discounted cash flow method. Unfortunately, discounted cash flows do not incorporate the value of flexibility. This makes the technique unsuitable for valuing investment opportunities related to wind farm development. Such investments are often highly uncertain, making flexibility valuable.
In this thesis we demonstrate the difference between some selected valuation methods, and discuss their suitability for valuing an investment opportunity in a wind power project. The focus is on real options analysis, and our hypothesis is that the use of real options valuation methods will improve the quality of the information available for decision makers, ultimately improving the quality of investment decisions in wind power.
The specific wind power project that we value is owned by TrønderEnergi AS, a renewable energy company based in Sør-Trønderlag county in Norway. The project is called Stokkfjellet Wind Farm, and the site is located in Selbu, a municipality in Sør-Trøndelag. TrønderEnergi has applied for a concession to build, and is expecting a decision by the Norwegian Water Resources and Energy Directorate (NVE) by the end of the summer of 2014.
To investigate this investment opportunity, we first consider the practical conditions affecting the investment decision. We move on to briefly present some theory of discounted cash flow analysis and the contingent net present value (NPV) approach. Then, a thorough introduction to real options theory and our assumptions is given, before we explain in detail the application of the binomial lattice and Monte Carlo valuation methods, which take a real options perspective. We next perform the actual valuations, and analyse the results from these as well as the methods used. The conclusions to be drawn are at last discussed.
For the development of the Stokkfjellet project, we assume there is uncertainty about events and markets. With events, we mean for example the outcome of the licence application processing by NVE. By market uncertainty we refer to the future development of the prices of power and tradable green certificates. Both prices influence the income from a wind farm directly.
For the real options valuations, we have modelled the power price as a mean-reverting process, and the price of tradable green certificates as a geometric Brownian motion. Modelling the electricity price as a mean-reverting process is an advantage as compared to modelling it as a geometric Brownian motion, since electricity prices show a clear tendency to revert back to a long-term level, rather than wander far away from its mean like a random walk. The parameters of the stochastic processes have been estimated from historical electricity price data from January 1 2001 to December 31 2013 for the NO3 Elspot Market available at Nord Pool Spot, and historical prices of tradable green certificates from February 19 2003 to March 5 2014 as listed by NVE. The code developed for the real options valuations has been written in Visual Basic for Applications (VBA) for Excel, and is included in Appendices A-D in this thesis.
Throughout the report, we carefully present all the theory needed in order to perform the calculations for the valuations. For the most intricate mathematical discussions, is assumed that the reader is familiar with calculus, some basic statistics and stochastic processes. We have included a brief introduction to the concept of stochastic calculus in Appendix F.
Other than discuss each method theoretically, we also compare the methods for practical applications. We find that the most severe problem with discounted cash flows is that the method cannot handle the value of flexibility incorporated in wind projects, which could cause us to grossly undervalue a project when we use this method. The contingent NPV method succeeds in handling the uncertainty related to events, but is not by itself a sufficient tool for solving real options because of problems related to calculating the appropriate discount rate in every stage of the project. The binomial lattice approach avoids this problem by using so-called risk-neutral probabilities, which is a more accurate fashion than choosing a discount rate. The binomial lattice permits us to value the project as a compound option, which is beneficial considering that investments in the Stokkfjellet project are done in stages. This method handles both event and market uncertainty. The Monte Carlo valuation is a simple and flexible method that also exploits risk-neutral pricing. However, it is not straightforward to use Monte Carlo simulations to value compound options. This is a drawback with this method and we have to model the investment opportunity as a single European call option when using this method. In addition, this method does not cope well with event uncertainty. There will always be benefits and drawbacks with every method, thus it is in general better to refer to more than one valuation method when making important investment decisions.
The real option we decide to value is the option to invest. We model the real option as a European compound call option and a single European call option for the binomial lattice and Monte Carlo methods, respectively. The investment opportunity is valued at 29 MNOK with the binomial lattice approach, and 47 MNOK with the Monte Carlo method. The difference in the values assigned can be explained by the different option types used to model the project, among other factors. The value of the project has increased from a negative value of -21 MNOK for the contingent NPV. Thus, we demonstrate that valuing the investment opportunity while accounting for flexibility assigns the project a positive value whereas a contingent net present value approach assigns the investment opportunity a negative value. This is critical information that should be considered before making a decision about investing. We therefore recommend that a real options approach to valuing wind projects is adopted by TrønderEnergi.
In addition to our conclusions and the recommendations that we suggest in this thesis, we have created a program that is meant to be used by TrønderEnergi for valuing Stokkfjellet Wind Farm as well as other projects that are subject to similar risks and flexibility. The program is easy to use, flexible and effective, and constitutes an important part of the work done in relation to this thesis. This program is to be handed over to TrønderEnergi at the time when this thesis is submitted.
Institutt for industriell økonomi og teknologiledelse , 2014. , 118 p.