Impact of financial risks on the economic feasibility of investment in offshore wind
Rijksuniversiteit Groningen Faculty of Economics and Business
Masters’ thesis
Name: Marieke Beijlevelt Student number: S2888270 Supervisor: prof. M. Mulder
January, 2021
Abstract
Offshore wind farms in The Netherlands play a key role to achieve the renewable energy targets of the Dutch government. The level of investments in offshore wind farms depends, among others, on the cost of capital. However, the future cost of capital is uncertain, through the uncertain value of the input variables. This could cause that investors do not want to invest because of volatile and uncertain revenues in the future. A Monte Carlo simulation incorporates this uncertainty, by conducting an analysis to determine the probability of economic feasibility of an offshore wind farm. We find that the probability that an investment in offshore wind is economic feasible is 84.8%. Furthermore, the market risk premium has the largest impact on the probability of economic feasibility. This variable is heavily dependent on factors underlying the market risk premium, such as volatility and economic stability. The government can help to achieve offshore wind farms’ targets by creating a stable policy framework with policies, measures and subsidies influencing the Weighted Average Cost of Capital (WACC) input variables directly and reducing the cost of capital.
Keywords: Offshore wind, stochastic variables, uncertainty, Monte Carlo simulation, probability, economic feasibility
1. Introduction
One of the climate-policy targets of the Dutch government is that 14% of the energy used in The Netherlands should come from renewable energy sources in 2020, and in 2030 this should be a minimum of 27%. Additionally, in 2050, the Dutch government targets zero CO2
emissions from energy supply. Offshore wind farms play an important role in the energy transition to achieve the targets of the Dutch government. This industry has grown tremendously in the recent years and is expected to further grow in the upcoming years.
According to the plans of the Dutch government, offshore wind farms in The Netherlands should have a capacity of approximately 11.5 gigawatt (GW) in 2030. In 2019, the capacity was approximately 1 GW (see appendix 1). The offshore wind farms will equal the energy supply of 3.3% in 2023 and of 8.5% in 2030 in The Netherlands (Government of The Netherlands, 2020).
The costs of an offshore wind farm can be expressed through the Levelized Cost of Energy (LCOE). This defines the lifetime average total cost for the electricity produced. The LCOE shows the average revenue per unit required to earn back the required rate of return of the investment, over the life of the investment. The rate of return is equal to the discount rate, which is the WACC. According to BVG associates (2019), in the past decade, the LCOE of offshore wind has decreased significantly. This is affected by three elements: first, the reduction of costs for building and operating a windfarm caused by innovation, increase of economies of scale, and decrease in the price of commodities (e.g. steel). Second, increase of energy produced, capacity factors, and productivity. For example, larger blades increase the electricity output per wind turbine. Finally, changes in financing through the reduction of project risk and lower cost of debt financing. The reduction in LCOE resulted in a significantly drop of tender bids for offshore wind farms in The Netherlands: Borssele I&II was awarded with a strike price of EUR 72.7 per MWh and Borssele III&IV a strike price of EUR 54.5 per MWh (Netherlands Enterprise Agency, 2017).
From 2015, the Stimulation of Sustainable Energy Production (SDE+) tender system was the subsidy programme for offshore wind and other renewables in The Netherlands (PWC, 2020).
The tender was granted to the developer who expected to generate electricity at the lowest fixed cost amount per MWh (strike price). The SDE+ programme used a floating Feed-in- Premium scheme, where the electricity generated by an offshore wind farm is subjected to a price cap, which is used as base for the subsidy received. The owners of the farm were compensated for the difference in costs (LCOE) between the price cap (base) and the electricity price. The large, mainly external, cost decreases for offshore wind farms, were the
reason for the Dutch government to drop the subsidies. This led to recent tenders based on zero-subsidy bids for offshore wind farms in The Netherlands. The winner(s) of the tender are allowed to build on the designated location and are selected based on a set of criteria to determine which bid will be awarded with the permit to build the windfarm. Selection criteria are experience, knowledge, capacity of the wind farm, social costs, quality of design, measure to ensure cost efficiency, risk analysis and risk management (operational and financial risk) (PWC, 2020). In a system with zero-subsidy bids, the electricity price is of course the most important and largest source of revenue for an offshore wind developer. The capture price is the revenue from the wholesale electricity market that can be realized from offshore wind. It is the proportion of the electricity price a producer of power captures and is expressed as the capture rate (Stead, Tarasewicz, and Husband 2020).
The capture price and the LCOE are factors which can determine the investment decision for offshore wind. Where the capture price is equal to the revenue for an offshore wind farm and the LCOE includes the costs and production volume of an offshore wind farm and the cost of capital. The investment decision for an offshore wind farm can be categorized into (I) the hurdle rate, and (II) the expected return on investment. The hurdle rate is the return at which an investor is willing to invest in the project. It consists of the cost of capital and the investors view on project risk, thus the opportunity costs of the investment. The cost of capital shows the rate of return than an investor should earn on the invested capital, to compensate for the risks that are related to the cashflows during installation, operation, maintenance, and the delay of timing during the project. The expected return on investment is the return based on cost expectations and revenues. If the expected return is higher than the hurdle rate, the investment is considered to be profitable and a return on capital can be obtained (Stead, Tarasewicz, and Husband 2020). Therefore, if the LCOE is smaller than the expected capture price, an offshore wind farm is economic feasible and a return on capital can be obtained exceeding the hurdle rate.
A lot of research has been done about the uncertainty of electricity prices and production volumes, and therefore focusing on the expected return of the investment. Additionally, studies about capital and operational expenditures and costs of the technology suppose a constant discount rate. This implies a constant cost of capital required by investors. Yet, in this Master Thesis the focus is on the hurdle rate of the investment, the cost of capital, with a constant expected return. Since, the values of the variables determining the cost of capital are very volatile and uncertain in the future, which are related to financial risks in the future, and influence the return on capital the investors can obtain. Therefore, the investment
decision and the economic feasibility of offshore windfarms are heavily dependent on variables determining the cost of capital. These variables are stochastic input variables to calculate the WACC, which is the cost of capital used to calculate the LCOE. Therefore, it is interesting to analyse the variables determining the cost of capital influenced by future volatility, uncertainty and risk. The question of this Master’s thesis is:
What is the impact of financial risks on the economic feasibility of investment in offshore wind?
This question is answered with the aim of the following sub questions:
- What is the probability that offshore wind is economically feasible?
- Which stochastic input variable of the WACC has the largest impact on the economic feasibility of offshore wind?
- What role can the government play to increase the economic feasibility of offshore wind?
This Masters’ thesis analyses the impact of five stochastic variables determining the WACC:
gearing, debt premium, tax rate, market risk premium and beta. We analyse the probability of economic feasibility and the impact of the stochastic variables by conducting a Monte Carlo analysis. The results of running a Monte Carlo simulation shows a probability that the LCOE of an offshore wind farm is below the capture price. If the LCOE is below the capture price, offshore wind is economically feasible. We find that the probability that an investment in offshore wind is economic feasible is 84.8%. This means that in 15.2% of the cases the investments will not be financed by investors, as it is not profitable to invest. Furthermore, the market risk premium has the largest influence on the probability of economic feasibility.
This variable is heavily dependent on factors underlying the market risk premium, such as volatility and economic stability. The government can help to achieve its targets by creating a stable policy framework with policies, measures and subsidies influencing the WACC input variables directly, and decreasing the cost of capital.
In section 2, an overview of literature based on this subject by other researchers is included.
Section 3 provides the hypothesis and an explanation about the model and the methodology used to give an answer to the research question. Section 4 clarifies the data and the descriptive statistics of the data. Furthermore, in section 5 the results are presented and explained. Finally, section 6 provides the conclusion and discussion.
2. Literature review
2.1 Renewable energy technologies
2.1.1 Investment in renewable energy technologies
The market for renewable energies has experienced a strong growth globally over the past 15 years. The commitments of governments to long-term renewable energy targets together with policy frameworks, such as subsidies, have created an attractive investment climate for renewable energy technologies compared to conventional power plants. However, the return on equity expected by investors is higher for technologies with a lower maturity (e.g.
renewable energy technologies) compared to established technologies. It is expected that the financial conditions over time will be equalized among technologies and the risk premium will reduce, if the installed capacity and experience of these renewable technologies increase (Kost, et al., 2018). Arnold and Yildiz (2015) showed that the implementation of renewable energy technologies resulted in economic growth and contribute towards protection of the environment. However, there are also technical, economic and financial barriers that delay the implementation of these technologies.
The investment decision can be represented in a model as a function of return, risk and policy to understand what determines the current levels of investment for renewable energy. In finance theories, the fundamental drivers for an investment are return and risk. Rational investors compare the level of return and risk of possible investment opportunities and choose the opportunity that gives the best or optimal return for a certain level of risk.
Renewable energy investment tends to have a disadvantage compared to conventional sources, regarding the combination between risk and return. Environmental externalities can cause this disadvantage, which can be corrected through energy policies, such as subsidies.
This reduce the level of risk and increase the return of a renewable energy investment (Wüsthagen and Menichetti, 2012). Gross, Blyth, and Heptonstall (2010) argued that risk factors and price are dependent on the market structure and the type of investment being considered. This may affect the way how an investment is financed, and therefore the cost of capital of an investment.
2.1.2 Cost of capital
The cost of capital is a function of gearing, aggregated project risk, provided securities and credit rating of the borrower. In general, an increase in aggregated project risk leads to an increase in interest rates requested for the loan. Institutional and equity investors link the risk of a project to their expectations about return on investment (Arnold and Yildiz 2015). Kost,
et al. (2018) argued that project-specific risk and risk of default influence the return on equity.
A higher risk of default leads to a higher return on equity asked by an investor and the amount of debt will reduce. Yet, a high share of debt with a low interest rate is desirable in order to keep the cost of capital low. Gross, Blyth, and Heptonstall (2010) showed that the fundamental component in the cost of capital and the expected level of return of an investment is gearing. The level of gearing is dependent on the risk profile and the type of the project. Debt providers are paid out before equity providers, therefore the return on debt is significantly lower than the return on equity. Thus, developers want as much debt financing as they can.
The difference in sensitivity to the WACC depends on the level of investors understanding the real risk profile of different kind of assets. In Africa, fossil fuels are less sensitive to changes in the WACC than renewable energy sources (Sweerts, Longa, and Zwaan, 2019). Vertianen, et al. (2019) showed that, the LCOE for solar photovoltaics doubles if the WACC increase from 2% to 10%. Bachner, Mayer, and Steininger (2019) showed that technologies which are highly capital intensive with large upfront investment costs, benefit stronger from lower interest rates than technologies which are less capital intensive. The WACC of renewable energy technologies seems to be riskier, but offers potentials for de-risking.
2.2 Wholesale electricity market 2.2.1 Price and utilisation uncertainty
Wholesale electricity prices make up the largest part of revenues for offshore wind farms. The owner of the farm earns back his investment through electricity prices which are uncertain and unpredictable. With lower electricity prices, the payback time of investments will be longer and more difficult (Bartjes, 2020). Whether to invest and develop an offshore wind farm, depends on the feasibility of the business case and thus the expected future revenue streams. Next to price uncertainty, there is also uncertainty around the actual supply. This comes through the uncertainty around wind variability, since every year the average wind speeds at a specific site differ through variation in meteorological conditions and through wind intermittency. Therefore, each year the annual energy production change. The realised annual production is not constant, and varies around the expected production during the operational life of the offshore wind farm (Deloitte, 2014).
2.2.2 Supply and demand of the electricity market
The effect on future electricity prices through the rise in energy supply, for example through the expansion of offshore wind farms, depends on different situations and is uncertain. If the
demand of electricity stays equal or decrease, the electricity prices will fall or even become sometimes negative. For example, the COVID-19 pandemic leads to a decrease of electricity consumption in The Netherlands. There are also arguments for an increase in consumption through the increase of electric cars, electrification of homes, and a greener industry. This leads to an increase in energy demand and higher prices. However, the increase of wind farms offshore and onshore and solar panels, will lead to increasing periods of electricity surpluses.
The option for storage of electricity is a potential solution for these periods of electricity surpluses (Grol, 2020).
2.2.3 Capture rate
The revenue stream for offshore wind farms depends on two factors as described by Stead, Tarasewicz, and Husband (2020). First, the baseload wholesale electricity prices (the price of supplying electricity covering all hours a day), which equal the average wholesale electricity price from all periods within a year. Second, the capture rate expressed as the part of the baseload wholesale electricity prices accomplished by offshore wind farms. The capture rate accounts for the proportion of value that exists at periods when offshore wind farms generate electricity. The average of these periods is the capture price. Therefore, the capture rate can be expressed as the annual average capture price as part of the baseload electricity price.
Furthermore, Stead, Tarasewicz, and Husband (2020) expected that the capture prices for offshore wind in The Netherlands will remain at a similar level as historic prices. This is because capture rates decreased rapidly the past decade through the increased levels of offshore wind farms and wind generation in The Netherlands. These installation and generation will occur in the same time frame, which has a downward effect on prices, known as the cannibalization effect of wind. However, a rising concern with the increase of wind energy is when wind outputs are the highest, wholesale electricity prices tend to be driven lower.
2.2.4 Risk mitigation
Through the Feed-in-Premium subsidy systems, such as the SDE+ programme in The Netherlands, a large part of the market risk caused by electricity prices were mitigated. This provided a continuous revenue scheme which ensured a fixed income to the developer of the offshore wind farm (Netherlands Enterprise Agency, 2017). The owners of zero-subsidy offshore wind farms face a larger electricity price market risk and have a higher risk profile than offshore wind farms with SDE+ subsidy. The zero-subsidy offshore wind farms do not have the option to mitigate the market price risk, and are therefore exposed to a larger risk.
The electricity generated by the farm is sold in the wholesale market and the revenue stream
received is the selling price of electricity to the wholesale market, the capture price. The development of the capture prices is uncertain, developers of these farms have to secure their revenues streams through other options, for example hedging the risk through contracts such as Power Purchase Agreements (PPA) (Steen, Paberzs, and Prinsen, 2019). A PPA is a contract between the consumer (off-taker) and a producer of power (owner of an offshore wind farm), where the consumer agrees to buy a certain volume of electricity at a certain price level for a specific period of time. From the perspective of an investors, PPAs could increase the predictability and stability of the revenue streams of electricity production from offshore wind farms. This could reduce the cost of capital through a decrease in risk for equity and debt providers and an increase of the available capital. Therefore, PPA’s are an important instrument for hedging the exposure to merchant price risk, in a zero-subsidy environment (PWC, 2020).
2.3 Offshore wind industry
2.3.1 Investors view on offshore wind
To convince investors to invest in offshore wind farms, the market should be more attractive than other available technologies. Over the past decade, the technological risk decreased through a growing and more experienced industry, which increased the attractiveness of the industry. Government support schemes in Europe gave price support to offshore wind farm developers which led to revenue certainty, a lower risk profile and the growth of the offshore wind sector in Europe. Moreover, it led to a reduction in risk exposure and more comfortability of investors with the technology. This led to trends in Europe showing a continuing shift towards debt financing in the past decade (Mone, et al., 2017). PWC (2020) showed that in recent years the level of leverage has increased from 60% to 75% lowering the cost of capital, considering that debt is in general cheaper than equity. The spread, asked for debt financing, between the base rate (LIBOR) and interest rate for offshore wind projects, gives an indication for the perceived risk of debt. Higher spreads correspond to a higher level of perceived risk. Since 2011 this spread has significantly decreased in Europe, from approximately 325 bps above LIBOR in 2011 to 150 bps above LIBOR in 2019. Therefore, the interest rate on debt for offshore wind farms in Europe have thus decreased over the past decade. The increase in leverage and reduction in the spread show that the offshore wind industry has become more attractive for investing and financing.
2.3.2 The WACC and the input parameters
The NPV of an offshore wind farm is highly sensitive to the WACC. If the WACC decreases by 20%, the value of the NPV is more than doubled. Furthermore, the NPV is sensitive to variable
parameters influencing the WACC. The NPV increases from highest to smallest by decreasing the return on equity, interest rate on debt, and debt to equity ratio, and corporate tax rate.
The NPV decreases by decreasing inflation rate as well. The observation from the debt to equity ratio highlights the importance of financing costs on the investment’s feasibility. As the debt to equity ratio increases, the WACC is expected to reduce as the cost of equity is expected to be higher than the cost of debt (Ioannou, Agnus, and Brennan, 2018). Ebenboch, et al. (2015) argued that the offshore wind industry is a very capital-intensive industry.
Therefore, the LCOE is significantly impacted by the WACC, which is again affected by several other factors, such as the overall availability of capital, relative attractiveness, and risks of construction and operation of offshore wind farms compared to other assets. The overall cost of capital is lowered by a moderate level of debt. This is because debt is a cheaper form of capital than equity. However, the benefit of high debt levels is balanced against the probability of financial distress and the costs associated with this. Developers of wind farms will increase the level of debt as much as possible until the cost of financial distress become significant (Levitt, et al., 2011).
2.3.3 Policies and cost of capital
Over the past decade in Europe, monetary policies have afforded developers more favourable terms of financing. Through appropriate campaigns explaining the benefits and the low risk of investments in wind energy, policy makers encourage banks to increase debt financing and lower interest rates. Additionally, in the form of preferential capital access and subsidized interest rates, policy makers make funds available for development of wind energy projects.
All together to make a larger access to capital finance for investment of wind energy.
However, by creating a stable policy framework to improve the prediction of revenue streams of windfarms, is the best policy measure so far since revenue certainty for a long period is crucial for investment in wind energy (Blanco, 2009).
Tighter monetary policies lead to a higher cost of capital, because tighter liquidity could increase the debt margin and return on equity since subsidy-free offshore wind farms are exposed to a greater merchant-pricing risk. Therefore, investors require a higher return on equity and the debt lender demand more restrictive terms of lending, which increases the WACC. For an offshore wind farm to break even, a 1% increase in the projects’ WACC would require wholesale electricity prices to rise by 7%. If the reliance on these prices increases, the WACC will further increase (CRA, 2018). The WACC input variables are related directly to the risk perception and risks. Changing the return on equity and the return on debt have the largest impact on the LCOE. Both variables can directly be influenced by measures, policies
and support schemes. Therefore, predictable policies considered by lenders can decrease the cost of capital of renewable energy technologies significantly. A reliable and stable policy and market context result in a lower return on equity and in longer debt terms, reducing the cost of debt. The impact on the cost of capital is significant (Jager and Rathmann, 2008). Kitzing and Weber (2014) argued that debt shares are a crucial assumption, as they have a significant impact on the results of offshore wind farm. They expected that a wind park with a support scheme can achieve higher share of debt than a wind park without support scheme, because of the more stable cash flows and less risk.
2.4 Deterministic analysis and Monte Carlo simulation
Deterministic models can support decisions related to the operation and development of offshore wind. However, the shortfall of these models is the ability to systematically take in consideration the uncertainty related to input parameters, when predicting the economic feasibility of offshore wind. Therefore, a stochastic approach or probabilistic model can significantly improve the value of the output and the analysis. A method suitable for developing stochastic model is the Monte Carlo simulation. This method simulates the effects of fluctuating stochastic input parameters and their volatility and uncertainty. When modelling different possible paths, a better and more informed range of values of stochastic variables can be provided. This avoids under or over stating of the mean and the variance of the variables according to Ioannou, Angus, and Brennan (2020). They run a Monte Carlo simulation to assess the impact of uncertainties on the performance of the offshore wind energy investments. They showed that if the standard deviation of key variables increases, this results in higher investment risk, and a reduction in the profitability of an investment.
Caralis, et al. (2016) used the Monte Carlo simulation to treat uncertain and a large combination of factors to produce a distribution of probability for the profitability. They showed that the return on equity and return on debt have the largest impact on changing the NPV of an offshore wind farm investment followed by share of equity and finally corporate tax rate. The profitability of offshore wind farms is based on a random outcome from the variability of every single uncertain factor resulting in a joint effect. The uncertain factors considered include: wind potential, distance to the coast, water depth, farm capacity, cost breakdown, financial aspects, support schemes, economic environment and wind farm design. All the independent and involved factors together, varying simultaneously, affect the profitability of an offshore wind farm.
3. Methodology and hypothesis
3.1 Introduction
The traditional approach of estimating the WACC uses point estimates for the input variables of the WACC. In this analysis, the point estimates are the average of a range of values. The input variables are gearing, debt premium, tax rate, market risk premium and beta. In reality the future values of these input variables are associated with uncertainty and volatility. The Monte Carlo simulation can be used to address the uncertainty related to these variables and their influence on financial risk and the cost of capital (Berry, Betterton, and Karagiannidis, 2014). For this Masters’ thesis first, a deterministic model is resolved to see what the sensitivity is of the input variables to the LCOE using point estimates. Next, a Monte Carlo simulation is run to determine the probability of economic feasibility of an investment in offshore wind. Different variants are created by increasing the average of one of the stochastic input variables by 10%, keeping the others equal, and running the Monte Carlo simulation again to get the new probability. By comparing the probabilities of these variants, it can be determined which input variable has the largest impact on the economic feasibility.
Finally, a T-test is done to check if the mean of the variants is significantly different from the capture price and of the mean of the variants compared are significantly different from each other.
3.2 LCOE
The Levelized Cost of Electricity (LCOE) (1) considers all the costs throughout the entire lifecycle of the wind farm divided by the total amount of energy generated throughout this lifecycle. The LCOE of one offshore wind farm remains constant for the operational lifetime of the farm and is therefore equal to the value at the year of installation. Bosch, Staffell, and Hawkes (2019) argued that the economic feasibility of an offshore wind farm can be determined with the use of LCOE. This indicates what the minimum capture price should be in order to obtain a return on capital. The LCOE for windfarm i is calculated as:
𝐿𝐶𝑂𝐸% =∑
()*+,-,/
(123)-
5-61 8 ∑5-619*+,-,/(123)-
∑ :-,/
(123)- 5-61
(1)
Where 𝐶𝐴𝑃𝐸𝑋>,% are capital expenditures: the costs for installation and construction, 𝑂𝑃𝐸𝑋>,%
are the operational expenditures: the costs for maintenance and operation. Furthermore, 𝐸>,%
is the energy generation, determined through specific and local conditions of the plant, such as the life time of the plant, wind conditions, full load hours (FLH), number of wind turbines
and rotor diameter. Finally, r is the discount rate, equal to the WACC (2), and shows the costs of financing, t indicates the year.
3.3 WACC
The WACC (2) shows the weighted average cost of capital from debt and equity financing. The rate to discount an investment is determined through the source of capital and the estimation of the financial risks from the investments (Ioannou, Agnus and Brennan, 2018). The WACC after taxes is calculated as:
𝑊𝐴𝐶𝐶 =@A@ ∗ 𝑅A+@E@ ∗ 𝑅E ∗ (1 − 𝑇𝐶) (2)
Where @A@ is the proportion of equity financing, 𝑅A is the return on equity, @E@ is the proportion of debt financing, 𝑅E is the return on debt and 𝑇𝐶 is the tax rate. Table 1 shows the relation between the input parameters of the WACC.
Table 1 – Relation between input parameters of the WACC
The table shows the input variables of the WACC and their relation with each other. It shows what the effect is on the input variables of the WACC at the first-row if one of the variables in the upper left column increases. So, if the variable in the first-row decrease (¯), increase () or stay the same (x) when the variable in the upper left column increase in value. The WACC shows the weighted average cost of capital from debt and equity financing. The WACC after taxes is calculated as 𝑊𝐴𝐶𝐶 =@A@ ∗ 𝑅A+@E
@ ∗ 𝑅E∗ (1 − 𝑇𝐶). Where @A@ is the proportion of equity financing, 𝑅A is the return on equity,
@E
@ is the proportion of debt financing, 𝑅E is the return on debt and 𝑇𝐶 is the tax rate. Gearing describes the financial ratio between debt and equity, the return on debt consists of the debt premium and LIBOR-yield. The return on equity is calculated with the Capital Asset Pricing Model (CAPM): 𝑅I = 𝑅J+ 𝛽 ∗ 𝑀𝑅𝑃. The CAPM shows that the return on equity (𝑅I) is equal to beta (𝛽) times the market risk premium (𝑀𝑅𝑃) plus the risk-free rate (𝑅J).
Input variables Gearing Debt premium
Tax rate Market Risk Premium
Beta
Gearing - x x
Debt premium ¯ - x x
Tax rate x - x x
Market Risk Premium x x - x
Beta ¯ x x -
Risk-free rate x x x
3.3.1 Gearing
The gearing describes the financial ratio between debt and equity. When the gearing of a company increases, the risk of default rises and thus the uncertainty of the future earnings of a company. Therefore, the debt premium rise as well. Debt providers are paid before equity
providers. Raising the gearing will also increase the return on equity because the risk increases and thus the cost of financing. The increase in risk reflected from the gearing is not associated with market or industry risk, but company specific risk. Therefore, the beta of a company will increase if the gearing is higher. Gearing will not be considered in evaluating and ranking the input variables in order of sensitivity and probabilities. After all, debt shares are a crucial assumption for the WACC, because of their significant impact (Kitzing and Weber, 2014). Ioannou, Agnus, and Brennan (2018) argued that the observation from debt to equity ratio highlights the importance of financing costs on the investment’s feasibility since cost of equity is expected to be higher than cost of debt. Additionally, Gross Blyth, and Heptonstall (2010) argued that the fundamental component in the cost of capital is gearing. Therefore, it is of more interest to say something about the other four input variables. However, the exact future level of gearing is uncertain and volatile and is considered as stochastic input variable in the Monte Carlo simulation.
3.3.2 Cost of debt
The return on debt is quoted in two parameters: (I) the risk-free rate based on expectations of the market about behaviour of a variable-rate index, e.g. LIBOR, (II) a debt premium, the price paid for debt expressed in basis points (bp). The debt premium incorporates the perceived risks of an investment; riskier investments requires therefore higher returns through higher debt premiums. In general, interest payments are tax deductible and therefore implemented in the WACC on an after-tax basis (Angelopoulos, et al., 2016). If the cost of debt increases, debt financing will become more expensive and this could lead to a decrease in gearing and also the other way around. Because if the cost of debt financing increase, the difference between the return on equity and return on debt will become smaller, and gearing leads to a higher risk of default, therefore gearing will decrease.
Furthermore, raising the cost debt premium will make the costs of debt higher and thus the beta of a company will also increase, because of the higher level of risk.
3.3.3 Tax rate
The tax policy has a significant effect on decisions about the capital structure of a firm. The interest rate on debt can be deducted from the profit when calculating taxable profits through the corporate tax rate. Equity related payments such as dividends are not allowed to be deducted. Therefore, the tax advantage derived from having debt, would lead that firms would be entirely financed with debt in the optimal world (Muthee, Adudah, and Ondigo, 2016). If the tax rate increases, the level of gearing will increase as well, as with a higher gearing and a higher tax rate a larger tax shield can be achieved.
3.3.4 Return on equity
The return on equity reflects the minimum required rate of return that investors expect from their investments. Additionally, it quantifies the level of risk of specific investments: higher values of the return on equity illustrates higher risk levels. A method to calculate the return on equity (3) is the Capital Asset Pricing Model (CAPM) (Angelopoulos, et al., 2016). The CAPM shows that the return on equity (𝑅I) is equal to beta (𝛽) times the market risk premium (𝑀𝑅𝑃) plus the risk-free rate (𝑅J).
𝑅I = 𝑅J+ 𝛽 ∗ 𝑀𝑅𝑃 (3)
The market risk premium is equal to the expected return of the market minus the risk-free rate. The price of risk underlying the equity premium can be influenced through the maturity of financial markets, income volatility and other macroeconomic indicators. The size of risk in the economy can be indicated through the volatility. In periods of high economic volatility, the return on equity tend to be higher, and the other way around. This is a statically significant positive relation (Ewijk, Groot, and Santing, 2012). If the market risk premium increases, the gearing of a company will also increase. The market risk premium will push up the return on equity, therefore debt financing will be a cheaper source of capital compared to equity and the gearing of a company will rise.
The risk-free rate is the return an investor can expect to earn on a risk-free investment, i.e.
the interest rate of an investment with no risk. An increase in the risk-free rate will pressure the market risk premium to rise. Considering that investors are able to earn a higher return on their risk-free investment; riskier assets should perform better to meet the new standards of required return for the investor. These assets will have a relative higher risk compared to the risk-free rate. Therefore, investors will demand a higher return on their riskier assets to compensate for the higher risk. Additionally, the risk-free rate will also push up the cost of debt, debt financing will be riskier thus these costs will increase as well. The beta will also increase if the risk-free rate increases, as the risk of a specific company increases and thus the asset becomes more volatile to the market.
Beta (equity beta) is a measure of systematic risk which cannot be diversified. It comes from the exposure of an investment compared to the general movements of the markets and is expressed as the sensitivity of the return of investment to the entire market. If the beta is larger (smaller) than 1, the company is more (less) volatile than the market and gives a larger (smaller) return than the market. If the beta is equal the 1, the price of an asset moves linear
with the market. The quantification of beta is influenced through assumptions about the time period, the market portfolio and frequency of data (Angelopoulos, et al., 2016). A company will become more volatile and riskier if the beta of a company increases, therefore lenders of debt will be more cautious and reduce the proportion of debt financing. The debt premium will increase if the beta increases, as debt financing will be a riskier source thus the price paid for debt will rise as well.
3.3.5 Independency
In order to run a correct application of the Monte Carlo simulation, the input variables should be considered as independent and random quantities, i.e. the input variables need to be completely uncorrelated. Otherwise, it could lead to simulation errors and problems in the interpretation of the simulation (Arnold, Yildiz, 2015). Therefore, we assume in this analysis that the input variables of the WACC are independent and viewed as random quantities in order to run the Monte Carlo simulation.
3.4 Deterministic model
A sensitivity analysis from the deterministic model is used to rank the input variables according to their sensitivity to the LCOE. The input variables are: gearing, debt premium, tax rate, market risk premium and beta. The deterministic model uses point estimates, in this case average values, as values for the input variables. To determine the sensitivity to the LCOE from the input variables, the average value of one of the input variables is varied, using a
±10/20% decrease and increase, keeping the other input variables constant (Arnold and Yildiz, 2015). Figure 1 shows the sensitivity analysis of the LCOE by using a ±10/20% decrease and increase in average value of the deterministic variables. We find that all the input variables show a linear sensitivity to the LCOE. The lines of market risk premium and the beta overlie, since they have the same coefficient. Therefore, the sensitivity of market risk premium and beta is the same on the LCOE. Market risk premium and beta show the largest sensitivity to the LCOE, followed by debt premium and tax rate.
Figure 1 – Sensitivity analysis of the LCOE in a deterministic model
The figure shows the sensitivity of the input variables to the LCOE. To determine the sensitivity, the average value of one of the input variables is varied, using a ±10/20% decrease and increase, keeping the other input variables constant. The deterministic input variables are: gearing, debt premium, tax rate, market risk premium and beta. We find that all the input variables show a linear sensitivity to the LCOE. The lines of market risk premium and the beta overlie, since they have the same coefficient.
Therefore, the sensitivity of market risk premium and beta is the same on the LCOE. Market risk premium and beta show the largest sensitivity to the LCOE, followed by debt premium and tax rate.
Source: Brindley, 2020, Fernandez, Martinex and Acin (2019), OECD (2020), PWC (2020), Thomson One (2020), Yahoo Finance (2020), Macrotrends (2020)
3.5 Monte Carlo simulation
It is hard to predict the future values of the input variables with absolute accuracy and precision. Therefore, the Monte Carlo simulation can be used to address the uncertainty related to these variables and their influence on financial risk and cost of capital (Berry, Betterton, and Karagiannidis, 2014). It uses the distribution of probability to model stochastic variables based on their randomness and uncertainty. From the distribution different outcome paths can be generated based on various possible outcomes or probabilities of outcomes for a decision. Compared to a deterministic analysis, the Monte Carlo simulation gives a superior simulation of risk, can improve decision making, and it shows the probability of occurrence of a certain situation (The Economic times, 2020).The execution of the Monte Carlo simulation consists of different steps: determining the five stochastic variables and their distribution, determining the base line and calculating the five variants to compare with the base line.
3.5.1 The stochastic variables
The future values of the five input variables of the WACC are associated with uncertainty and volatility. Therefore, these input variables are stochastic variables, and in this analysis, the stochastic variables are: gearing, debt premium, tax rate, market risk premium and beta. The average and standard deviation of the stochastic variables is based on a range of values based
-6%
-3%
0%
3%
6%
-20% -10% 0% 10% 20%
Sensitivity to LCOE (%)
±10/20% decrease and increase deterministic variables (%)
Gearing Debt premium
Tax rate Market risk premium
Beta
on literature analysis or historical data. From this range of values, the distribution type can be determined from the histogram and the Skewness and Kurtosis normality test. This test tests of the value of kurtosis and skewness of the variable is consistent with the value for a normal distribution. The null hypothesis for this test is that the stochastic variable is normally distributed. If the P-value is below the significance level, the null hypothesis can be rejected and the variable is not normally distributed.
3.5.2 The base line and five variants
The average and standard deviation of the stochastic variables are used to calculated a random varying number used to calculate the WACC (2) and LCOE (1). This variation is related to the uncertainty and volatility of the stochastic variables in the future. With the five random values for the stochastic variables and therefore a varying WACC and LCOE, the Monte Carlo simulation is run 150,000 times. 150,000 runs are chosen since at this number of runs, the probability does not change any more when running the simulation again. The 150,000 outcomes of the LCOE create a range from where it can be determined how many times the LCOE is lower than the capture price. If the LCOE is below the capture price, then the investment in offshore wind is economic profitable. Therefore, the result of the Monte Carlo simulation gives the probability of economic feasibility of an investment in offshore wind. The probability based on average values used as input for the stochastic variable is the base line.
Five new variants are created by increasing for every variant the average of one of the five stochastic variables by 10%, keeping the others equal to their original average value. These new input values determine a new random number for the WACC and LCOE. For each variant of the stochastic variable, the Monte Carlo simulation is run again and gives a new probability of economic feasibility. By comparing the probabilities of different variants with the base line, it can be determined which increasing stochastic variable has the largest impact on the economic feasibility.
3.6 T-test
A one-sample T-test is done for every variant to check whether the mean from the range of 150,000 LCOE equals the capture price. Additionally, from the test it can be concluded if the mean is below or above the capture price. The null hypothesis is that the mean of the range of 150,000 LCOE’s equals the capture price. The two-sample T-test tests if there is significantly difference in the mean of the ranges of LCOE’s from different variants. This difference is tested between the base line and the variants of the stochastic variables and the mean of the variants with each other. Also, it can be concluded if the difference in means are positive or
negative. The null hypothesis is that the mean of the range of 150,000 LCOE’s from variant x is equal to the mean of the range of 150,000 of variant y.
3.7 Hypothesis and evaluation
The hypothesis from this thesis is that the economic feasibility of an investment in offshore wind is impacted through financial risks. The market risk premium and the beta have the same and largest impact on the probability that an investment in offshore wind is economic profitable followed by the debt premium and tax rate.
In order to evaluate the impact of financial risks on economic feasibility of investment in offshore wind, two criteria are used:
1. The probability of economic feasibility of an investment in offshore wind farm. This is determined on how many times of the 150,000 runs of the Monte Carlo simulation, the LCOE is below the capture price.
2. The change in probability of economic feasibility. By comparing the probability of the base line with the variants from the stochastic variables, the impact on economic feasibility can be determined.
4. Data and descriptive statistics
4.1 Offshore wind farm: case illustration
To illustrate the application of the probability of economic feasibility of an investment in offshore wind, data on the offshore wind farm Hollandse Kust Zuid III&IV is used as case study.
The characteristics and costs only apply for this offshore wind farm, as these are different for every offshore wind farm depending on location, developer and timing of building. Hollandse Kust Zuid III&IV are located in the North Sea in front of the coast of South-Holland, The Netherlands. In 2019, the zero-subsidy tender for the sites was won by Vattenfall. In 2023 the wind park should be operational and deliver 760 MW of power. This should equal the supply of energy for one million households and should supply more than 2.5% of the electricity demand in The Netherlands (Rijksoverheid, 2019). Table 2 shows the characteristics and costs of offshore wind farm Hollandse Kust Zuid III&IV. It shows the capital expenditures (Capex) (EUR/kW), operational expenditures (Opex) (EUR/kW/year), Full Load Hours (FLH) (hours/year), Power (MW), Annual Energy Production (AEP) (GW/year), operational lifetime (years) and inflation rate. The annual energy production per offshore wind farm is equal to the power of the wind farm times the Full Load Hours. The operational life time used in this calculation is 25 years (Lensink and Pisca, 2019). The inflation rate is 2%, this is based on the Harmonised Index of Consumer Prices (HCIP) of the European Central Bank (2020). The operational expenditures are increased by the 2% inflation rate each year.
Table 2 – Offshore wind farm in The Netherlands: Hollandse Kust Zuid III&IV characteristics and costs, case illustration
The table shows the costs and characteristics of the offshore wind park Hollandse Kust Zuid III&IV used in the calculation of the LCOE. This offshore wind farm is used as case study to illustrate the application of the probability of economic feasibility of an investment in offshore wind. The table shows the capital expenditures (Capex) (EUR/kW), operational expenditures (Opex) (EUR/kW/year), Full Load Hours (FLH) (hours/year), Power (MW), Annual Energy Production (AEP) (GW/year), operational lifetime (years) and inflation rate. The annual energy production per offshore wind farm is equal to the power of the wind farm times the Full Load Hours. The operational life time used in this calculation is 25 years. The operational expenditures are increased by the 2% inflation rate each year.
Source: Lensink, Pisca, (2019), Rijksoverheid (2019), European Central Bank (2020) Hollandse Kust zuid III&IV
Capital expenditures (EUR/MW) 1,600,000 Operational expenditures (EUR/MW/year) 41,000 Full load hours (hours/year) 4,400
Power (MW) 760
AEP (GW/year) 3,344
Operational lifetime (years) 25
Inflation rate (%) 0.02
4.2 WACC
The WACC is calculated with equation two. Below, for each of the five stochastic input variables it is explained what determines the range of each variable based on literature analysis or historical data. Table 3 shows the descriptive statistics for the five stochastic input variables and their value if the average value is increased by 10%. The descriptive statistics are based on the range of values for every stochastic variable shown in appendix 2.
Furthermore, the distribution type of the stochastic variables is based on the histogram (appendix 3) and the Skewness and Kurtosis normality test (appendix 4).
4.2.1 Gearing
It can be expected that the gearing of offshore wind farm is between 70% and 80%, according to Brindley (2020). Determining the gearing based on historical data could be misleading, as gearing for offshore wind farm was 47% in 2010 (Brindley, 2019). Over the past decade, gearing has increased enormously through certainty of investors about the technology.
Therefore, it is not likely that in the near future the level of 2010 for gearing will be reached again. The optimal gearing level is 100%, because of the tax shield and the cheaper source of funding. However, these values are not likely to be reached because a debt provider always wants a reasonable amount of equity being invested to create confidence in the investment (Ioannou, Agnus, and Brennan, 2018). Gearing is an uncertain variable and cannot be predicted exactly for the future, but it can be expected that the gearing will fall within a certain range in the future. Therefore, in this analysis the range of gearing between 70% and 80% is considered. The number of observations is 11, with an average value of 75% and a standard deviation of 0.033, a random number to determine the WACC and to run the Monte Carlo simulation with. The histogram for gearing shows a normal distribution. Based on the Skewness and Kurtosis normality test, the p-value of skewness, kurtosis and combined are larger than the 0.05 significance level. Thus, the null hypothesis cannot be rejected with a 5%
significance level and can be concluded that gearing is normally distributed.
4.2.2 Return on debt
The debt premium is based on an analysis of Brindley (2020) which shows the debt premium between 2011 and 2019 in Europe. There are nine years between 2011 and 2019, therefore the number of observations is 9. The range for the debt premium is between 0.0325 above LIBOR in 2011 and 0.015 above LIBOR in 2019, with an average of 0.023 and standard deviation of 0.006. The average and standard deviation are used to calculate the random term. It can be noticed that the debt premium, expressed in bp above LIBOR, is decreasing over the past decade. The histogram for debt premium shows a normal distribution. This is
also concluded from the Skewness and Kurtosis normality test. This test shows that for debt premium the p-value of skewness, kurtosis and combined are larger than the 0.05 significance level. The null hypothesis cannot be rejected at a 5% significance level. Thus, debt premium is normally distributed. The average value of the debt premium is added to the average LIBOR yield to get the return on debt used in the WACC calculation. The average LIBOR yield in 2019 was 2.37% (Macrotrends, 2020).
4.2.3 Tax rate
The tax rate is based on the corporate income tax rate of the European OECD countries (OECD, 2020) and the tax rate in The Netherlands in 2021 (PWC, 2020). In the past decade, the corporate tax rate in The Netherlands is 25%. However, in 2021 the tax rate will be lowered to 21.7%. There are 25 European OECD countries and together with the rate in 2021, the number of observations is 26. The range of the tax rate is between 9% in Hungary and 32% in France. The average tax rate is 22% with a standard deviation of 0.05. These values are used to calculate the random value used to calculate the WACC. Based on the histogram and Skewness and Kurtosis normality test, tax rate has a normal distribution. The test shows that for the tax rate the p-value of skewness, kurtosis and combined are larger than the 0.05 significance level. The null hypothesis cannot be rejected at a 5% significance level. Thus, tax rate is normally distributed.
4.2.4 Return on equity
The return on equity is calculated with the CAPM equation three. It consists of the market risk premium, beta and the risk-free rate. The market risk premium times the beta plus the risk-free rate, result in the return on equity. The market risk premium is equal to the expected market return minus the risk-free rate. It is based on a study from Fernandez, Martinex, and Acin (2019) for the same European OECD countries used for the tax rate. This study determines the average market risk premium used in these countries in 2019. There are 25 European OECD countries used for the tax rate, however Iceland and Lithuania are not shown in the study for the market risk premium. Therefore, the number of observations is 23. This results in range for the market risk premium between 5.7% for Germany and 15.4% for Greece. The average value is 6.9% and a standard deviation of 0.02. These values are used to calculate the random value used to calculate the WACC. The histogram does not show a normal distribution, but a distribution that is skewed to the left. Also, based on the Skewness and Kurtosis normality test the p-value of skewness, kurtosis and combined are smaller than the 0.1, 0.05 and 0.01 significance level. The null hypothesis can be rejected at any significance level. Thus, market risk premium is not normally distributed.
The beta is based on the beta from three types of listed companies active in the offshore wind industry (Thomson One, Yahoo Finance, 2020). The first type of companies owns a part of an offshore wind farm in The Netherlands: Shell, Ørsted, Northland Power and Mitsubishi Corporation. The average beta of these companies has the smallest beta equal to 0.82. The second type are turbine suppliers for offshore wind: Vestas and Siemens. They have the second smallest beta of 0.93. The last category are companies who are active in the offshore industry and delivers for example monopiles for wind turbines. These are SIF, Boskalis and Fugro. These companies have the highest average beta of 1.31. The range of betas is between the minimum beta of 0.52 from Northland Power and the maximum beta of 2.07 for Fugro.
The range resulted in an average of 1.01 and a standard deviation of 0.48. These values are used to calculate the random value used to calculate the WACC. Nine companies are used as comparable companies for the beta analysis, the number of observations is nine. Based on the histogram and the Skewness and Kurtosis normality test, the beta is normally distributed.
The test shows that for beta the p-value of skewness, kurtosis and combined are larger than the 0.05 significance level. The null hypothesis cannot be rejected at a 5% significance level.
Thus, beta is normally distributed.
The risk-free rate is based on a study from Fernandez, Martinex, and Acin (2019), who concluded that the average risk-free rate used in The Netherlands in 2019 was 1.3%.
Therefore, the risk-free rate used in the calculation for the return on equity is 1.3%.
Table 3 – Descriptive statistics input variables WACC
The table shows the descriptive statistics of the stochastic variables and the value if the mean is increased by 10%. There are four main elements necessary to calculate the WACC, namely: gearing, return on debt, tax and return on equity. The debt premium together with the LIBOR yield (2.37%) determine the return on debt. The market premium, beta and the risk-free rate (1.3%) are necessary to calculate the return on equity. For every stochastic variable a range is determined based on historical data or literature analysis. This range has as boundary the minimum (min) and maximum (max) value. From where the mean and standard deviation for each variable are determined. These are used to calculate the random parameter used in the WACC calculation. Furthermore, the number of observations used in the range are shown.
Source: Brindley (2020), Fernandez, Martinex and Acin (2019), OECD (2020), PWC (2020), Thomson One (2020), Yahoo Finance (2020), Macrotrends (2020)
Input variables WACC Obs Mean Std Min Max Mean increased 10%
Gearing (%) 11 0.750 0.033 0.700 0.800 0.825
Debt premium (bp) 9 0.023 0.006 0.015 0.033 0.026
Tax rate (%) 26 0.221 0.053 0.090 0.320 0.243
Market Risk Premium (%) 23 0.069 0.020 0.057 0.154 0.076
Beta 9 1.008 0.481 0.520 2.070 1.109
4.3 Capture price
The capture price is considered to be an exogenous variable. Although, there is also uncertainty around the capture price, the value of this variable is determined outside the LCOE model and cannot be explained by the LCOE model. Therefore, the value is not influenced by the stochastic variables and assumed to be given in this analysis. We assume a capture price of EUR 39.5 per MWh based on analysis from Stead, Tarasewicz, and Husband (2020).
5. Results
A Monte Carlo simulation is run to calculate the probability of economic feasibility of an investment in offshore wind. Furthermore, the simulation determined which stochastic variable has the largest impact on the probability of economic feasibility. A one sample T-test is done to check if the mean of the variant is equal to the capture price. And a two sample T- test is done to check whether the mean of the base line is different from the mean of the variants from the stochastic variables.
5.1 Monte Carlo simulation
The probability that an investment in offshore wind is economic profitable is determined by the Monte Carlo simulation. This simulation is run 150,000 times to determine the probability that the LCOE is below the capture price. This probability is based on how many times of the runs the LCOE is lower than the capture price. It shows what the chance is that an investment in offshore wind is economic profitable, as the investment in offshore wind is economic profitable if the LCOE is lower than the capture price. The WACC is equal to the cost of capital and is the discount factor in the LCOE calculation. This WACC is influenced through stochastic input variables, namely: gearing, debt premium, tax rate, market risk premium, and beta, which future values are uncertain and volatile. Based on the average value for every stochastic variable the base line is determined. We find that in the base line, the probability that an investment in offshore wind is economic profitable is 84.8%. This means that out of 150,000 runs, the LCOE is below the capture price in 84.8% of the runs. In the other 15.2% of the cases the investment in offshore wind will not be financed by investors, as it is not profitable to invest because a return on capital cannot be obtained. Appendix 5 shows, based on the range of 150,000 LCOE’s from the Monte Carlo simulation, the probability that the LCOE is below the capture price. Also, it shows the average, minimum and maximum value created from this range, determining the histogram in appendix 6. The histogram for the base line shows a normal distribution.
Figure 2 shows for different capture prices the probability that an investment in offshore wind is economic feasible. The curve of the graph has a S-shape. This means that the probability increases slowly at the beginning if the capture price increases, then rises rapidly if the capture price falls between certain capture prices, where after it declines until it reaches 100% probability of economic feasibility. This rapid rise in probability are caused by the cumulative probabilities and due to the fact that limitation of the maximum probability (100%) stops the increasing factor after a certain capture price.
Figure 2 – Probability of economic feasibility for different capture prices
The figure shows for different capture prices (EUR/MWh) the probability that the offshore wind farm is economic feasible. It is economic feasible if the LCOE is below the capture price. In the cases that it is not profitable, investors do not want to invest since they cannot achieve their return on capital. The curve of the graph shows a S-shape. This means that the probability increases slowly at the beginning if the capture price increases, then rises rapidly if the capture price falls between certain capture prices, where after it declines until it reaches 100% probability of economic feasibility. This rapid rise in probability are caused by the cumulative probabilities and due to the fact that limitation of the maximum probability (100%) stops the increasing factor after a certain capture price.
Source: Brindley (2020), Fernandez, Martinex and Acin (2019), OECD (2020), PWC (2020), Thomson One (2020), Yahoo Finance (2020), Macrotrends (2020), Lensink and Pisca, (2019), European Central Bank (2020)
5.2 Influence on probability of economic feasibility
By comparing probabilities of different variants, it can be determined which variable has the largest impact on the economic feasibility. The variants are created by increasing the average value of one of the stochastic variables by 10%, keeping the others equal to their original average value, and running the Monte Carlo simulation again to get the new probability of economic feasibility. Comparing the probability from the base line with the variants from the increased stochastic variables, the influence of changing a stochastic variable on the probability of economic feasibility of an investment in offshore wind can be determined.
Figure 3 shows the impact on the probability that an investment in offshore wind is economic profitable expressed in basis points (bp), if one of the stochastic variables is increased by 10%.
Appendix 5 shows, based on the range of 150,000 LCOE’s from the Monte Carlo simulation, the probability that the LCOE is below the capture price. Also, it shows the average, minimum and maximum value created from this range, determining the histogram in appendix 6. The histograms for the different variants show a normal distribution.
0%
25%
50%
75%
100%
29.5 32 34.5 37 39.5 42 44.5 47 49.5
Probability
Capture price (EUR/MWh)
Figure 3 – Influence on probability by increasing stochastic variables by 10%
The figure shows the impact on the probability that an investment in offshore wind is economic feasible, when increasing the average of one of the stochastic variables by 10%. The probability from the base line is compared with the probabilities from the variants of the stochastic variables, to determine the impact on the probability of economic feasibility expressed in basis points (bp). The Monte Carlo simulation is run 150,000 times to determine the probability that an investment in offshore wind is economic feasible. This probability is based on how many times of the runs the LCOE is lower than the capture price. It shows what the chance is that an investment in offshore wind is economic profitable. First, the simulation is run with the base (average) values of the stochastic variables, to determine the base line. In the base line the probability an investment is economic feasible is 84.8%. The results for the probabilities of the variants are gearing: 94.5%; debt premium:
82.2%; tax rate: 86.2%; market risk premium; 80.2% and beta: 80.9%. In the other cases (100% minus the probability) the investment in offshore wind will not be financed by investors, as it is not profitable to invest because a return on capital cannot be obtained.
Source: Brindley (2020), Fernandez, Martinex and Acin (2019), OECD (2020), PWC (2020), Thomson One (2020), Yahoo Finance (2020), Macrotrends (2020), Lensink and Pisca, (2019), European Central Bank (2020)
Based on figure 3, we find that increasing the market risk premium has the largest impact on the probability that an investment in offshore wind is economic feasible followed by beta, debt premium and tax rate. Moreover, increasing the tax rate has a positive influence on the probability of economic feasibility and increasing the market risk premium, beta and debt premium have a negative influence. Increasing the tax rate by 10% results in the smallest positive influence on the probability that an investment in offshore wind is economic feasible.
The probability of economic feasibility increases with 145 bp to 86.2%. It changes the probability with 145 bp to a probability of the LCOE being below the capture price of 86.2%.
The smallest change of the tax rate can be devoted to the indirect effect taxes have on the probability of economic feasibility, as increasing the tax rate lowers the WACC through the tax shield resulting from having debt. Ioannou, Agnus and Brennan (2018) showed in their research that changing the tax rate had the smallest effect on the NPV. Increasing the market risk premium and the debt premium by 10% have a negative effect on the probability of economic feasibility of an investment in offshore wind. It decreases the probability of
970
-257
145
-452 -383
-600 0 600 1200
Gearing Debt premium Tax rate Market risk
premium Beta
Change in probability (bp)
Stochastic variables
