Cost of capital

Page 69 of 162

Page 70 of 162 structure, as Ørsted’s capital structure should converge towards an industry average. Ørsted will likely have a high portion of equity compared to debt in their target capital structure because of their previous problems with their credit rating and use of HC. To ensure consistency, the capital structure used when computing the beta will also be used as Ørsted’s target capital structure.

5.2.1.2. Risk-free rate

Ørsted’s cash flows are denominated in DKK and to avoid issues such as inflation, the local Danish government bond is used as the risk-free rate (Plenborg & Petersen 2012). The duration of the Danish government needs to match the duration of the forecasted cash flows (Damodaran, 2012). For this reason, the current 10-year Danish government bond is chosen as the best proxy for the current risk-free rate. The current yield on the bond is 0.5% as of March 31^st, 2018, according to Bloomberg. As mentioned in the economical part of the PESTEL, there has been a general reduction in government bond yields across Europe due to QE. It is questionable whether the risk-free rate will remain at this low level going forward. The consequence of applying a low risk-free rate is that it will result in a lower WACC, which, all else being equal, creates a higher valuation (Damodaran, 2012). Ernst & Young (2015) advises using an average government yield over a selected period. Hence, it was decided to use the 10-year historical average of the 10-year Danish government bond as a proxy for a normalised risk-free rate. The 10-year historical average of the 10-year Danish government bond is calculated to 1.95%. In comparison, Fernandez et al. (2017) reports in his survey of 4,368 answers from professionals, that the average risk-free rate in Denmark is 1.6% and the median is 1.9%. Thus, 1.95% seems to be fair and will be the risk-free rate when calculating the cost of equity.

5.2.1.3. Beta

The estimation of the beta is one of the most critical parts in the process of risk adjusting the discount rate to market risk. Furthermore, it is the only factor in the CAPM formula that is company specific. If Ørsted’s current capital structure of 78% equity and 22% debt turns out to also be the target capital structure, then the WACC will be highly sensitive to the beta value (Damodaran, 2012).

To account for Ørsted’s divestments of oil & gas and its current mix of businesses, a bottom-up beta approach is chosen. To ensure consistency, all the data used below is found through the same external source:

Bloomberg. Bloomberg reports both an adjusted and a raw beta. The adjusted beta assumes that the beta of a company converges to the market average of one. The strategic analysis highlighted that the offshore wind industry is expanding rapidly and is not sensitive to the general economy. Therefore, the raw beta will be used instead of the adjusted beta. The raw beta is the slope of the regression of the company’s stock return against the market index. The raw beta reflects the company’s levered beta. To account for company-specific financial leverage, the levered beta must be unlevered (Damodaran, 2012). The average unlevered beta is estimated for each of Ørsted’s divisions. This is done by finding peers that are similar to Ørsted’s Wind Power, Distribution

Page 71 of 162 and Customer Solutions, and Bioenergy and Thermal division. Only two peers are identical to the Bioenergy and Thermal division: Drax and Enea. For the other two divisions, there are sufficient comparable companies.

With the peers defined, the average levered beta, corporate tax rate and median D/E ratio are computed for each business. Alternatively, the levered beta could be unlevered for each company and an average could be taken but, given the standard errors of the individual regression betas, it will give a noisy beta (Damodaran, 2012). According to Damodaran (2012) the unlevered betas need to be adjusted for cash because investment in cash and marketable securities have a beta that is close to zero:

Beta unlevered, corrected for Cash = Beta unlevered

1 − Cash

Enterprise Value

The total unlevered beta for Ørsted is calculated by taking a weighted average of the unlevered betas based on Ørsted’s EBITDA segmentation. Finally, the unlevered beta needs to be re-levered to reflect Ørsted’s market capital structure.

Beta levered = Beta unlevered ∗ (1 + (1 − Tax) ∗D E)

As mentioned previously, the capital structure used to re-lever the beta should also be the target capital structure for Ørsted. None of the median capital structures from the peers reflect Ørsted’s capital structure.

Ørsted has a much lower debt compared to its peers and will likely not issue more debt with their goal of having a 30% FFO to adjusted net debt. Moreover, Ørsted is planning to increase its dividend by a high single-digit rate compared to the dividends for the previous year up until 2020, reflecting that they are satisfied with their current capital structure (Ørsted, 2017a). As a result, Ørsted’s current capital structure is assumed to be their target capital structure. Hence, it is used to re-lever the unlevered beta and to calculate the WACC. The calculated bottom-up beta for Ørsted is 0.67.

Table 3 – Calculation of Beta

Source: Authors’ own creation from Ørsted’s annual reports from 2007-2017 and Bloomberg

Business Number of

Peers

Average Levered Beta

Median MVD/MVE ^Tax

Unlevered Beta

Cash/Firm Value

Unlevered Beta Corrected for Cash

EBITDA Weight

Wind Power 10 0,8 86% 24% 0,49 11% 0,55 90%

Distribution 7 0,9 91% 23% 0,51 4% 0,53 9%

Bioenergy & Thermal 2 0,5 108% 20% 0,26 19% 0,32 1%

Weighted Unlevered Beta 0,55 Ørsted's Capital Structure 29%

Ørsted's Relevered Beta 0,67

Page 72 of 162 The beta of 0.67 reflects that investing in Ørsted’s stock involves less systematic risk than investing in the market portfolio (Damodaran, 2012). Stocks with a beta that is less than one generally moves more independently than the broader market, confirming that the energy sector is not perfectly correlated with the general economy, as energy is almost always in demand. From a theoretical point of view, this is counterintuitive. The industries in which Ørsted operates in are all asset heavy and capital intensive.

Subsequently, Ørsted’s investment in fixed costs is large compared to operational costs, which implies high operating leverage. High operating leverage corresponds to high betas under normal circumstances (Ibid.).

Ørsted’s operating leverage can be measured by the EBIT variability measure (Ibid.). The measure takes each year’s change in EBIT and divides it by the change in revenue. Hence, the measure shows how quickly EBIT changes with revenue. The higher the number, the greater the operating leverage. Figure 36 shows that Ørsted has operating leverage at the lower levels compared to the peers. Coupled with Ørsted’s low financial leverage, Ørsted’s beta should be in the lower end compared to its peers. This is the case with the found beta of 0.67.

The average EBIT variability measure among all the peers is 45.84; as a reference point, the average across entertainment companies is 1.35 (Damodaran, 2012).

Figure 36 – Ørsted and competitors’ EBIT Variability Measure

Source: Authors’ own creation from Ørsted’s annual reports from 2007-2017 and Bloomberg

Ørsted’s beta of 0.67 must be a reflection of the subsidising governments bearing a large portion of the risk related to the investments. With the risk of no subsidies, Ørsted will bear all the risk associated with building offshore wind farms. In relation to this, Martin Neubert, the newly appointed head of Ørsted’s Wind department, commented:

“It is an unfair distribution of risk that no one can control or assess … Ørsted may opt to build projects if the risk becomes too large … it is good for the politicians who then can say that it is not them who are taking the risk” (Børsen, 2018b. l. 3-15).

His comment reflects the riskiness of offshore wind and that the found beta of 0.67 is a function of government support. The question is whether the beta should be based on theoretical correctness or beliefs about the development in subsidies. This stressed the importance of using market values when computing a beta. The

13 8

42

63

35

76

0 9 20 40 60 260

Centrica Iberdrola

EDPR EDF ENEL E.ON Fortum Engie RWE SSE Ørsted

1 0 0

257 EBIT Variability Measure

Page 73 of 162 market value of equity, a function of the share price times number of shares outstanding, reflects the market’s expectations to Ørsted’s future cash flows. The share price is up around 45% year-to-date (see figure 1) reflecting that investors are not nervous about the zero subsidies. Therefore, the WACC calculation will be based on a beta of 0.67.

5.2.1.4. Equity Risk Premium

The equity risk premium is the spread between historical returns and returns on the market portfolio and risk-free investments (Damodaran, 2012). As mentioned in the theoretical review, the implied equity risk premium will be used. The implied valuation approach requires estimating the equity and earnings in future periods, solving backwards for the implied cost of equity (Ibid.). The drawback of this method is that it relies on significant assumptions about future growth and return on capital and, thus is very sensitive to these inputs.

However, it is the method that best reflects the equity risk premium investors are actually paying. Figure 37 shows how the measure has changed since 1960. According to Damodaran (2017) the current ERP is 4.95%.

In comparison, Fernandez et al. (2017), using the survey approach, reports an average ERP for Denmark as 4.5% for 2017. Thus, 4.95% will be the ERP when calculating cost of equity. Denmark is assumed to reflect the total geographical equity risk premium as the credit ratings for the countries that Ørsted operates within are identical (Damodaran, 2018). However, if Ørsted seeks more businesses in, e.g., Taiwan, it could have a slightly negative impact on their ERP, making WACC higher.

Figure 37 – Equity Risk Premium

Source: Authors’ own creation from (Damodaran, 2017)

5.2.1.5. Cost of Equity

Plenborg & Petersen (2012) argues that for smaller and less liquid stocks, an additional risk premium should be added to compensate for smaller stocks being more volatile. Due to the size of Ørsted and the volume in their stock, no company-specific liquidity or risk premiums are added. With the risk-free rate of 1.95%, beta of 0.67, and the market risk premium of 4.95%, the cost of equity is equal to 5.33%. As a rule of thumb, the cost of equity is normally 3-4% above the risk-free rate, which is also the case here (Vibig et al., 2008). In comparison, figure 38 shows the cost of equity for Ørsted’s peers. Due to the low beta, the cost of equity for Ørsted is the lowest among the companies.

6 5 0 7

1990

1960 1970 1980 2000 2010 2017

-1,3% Equity Risk Premium

Page 74 of 162 Figure 38 – Ørsted and competitors’ cost of equity

Source: Authors’ own creation from Ørsted’s annual reports from 2007-2017 and Bloomberg

5.2.1.6. Cost of Debt

From the 2017 annual report, it can be found that Ørsted’s weighted average effective interest rate for general borrowing was 5.3% in 2017 (Ørsted, 2017a, p. 127). This represents the current cost of borrowing for Ørsted.

However, looking at the historical borrowing rate, the 2017 rate is higher than previous years. For this reason, the rate will be challenged with more theoretical correct approaches to calculate the cost of debt (Damodaran, 2012).

Ørsted’s yield to maturity on its outstanding bonds can be used to determine the cost of debt (Ibid.). According to Bloomberg, the current yield to maturity for a bond maturing in 2032 is 3.97%. However, according to Koller et al. (2010), when the credit rating is low, the yield to maturity is a poor proxy for the cost of debt.

Another approach is to look at Ørsted’s credit rating and default spread. Bloomberg reports that the Moody’s credit rating on Ørsted is Baa1, while Standard & Poor’s has a rating of BBB+. This is at the bottom of the investment grade, reflecting that there is a default risk in investing in bonds issued by Ørsted. Standard &

Poor’s define the rating as:

“... exhibits adequate protection parameters. However, adverse economic conditions or changing circumstances are more likely to weaken the obligor's capacity to meet its financial commitments on the obligation” (S&P, 2018, table 1).

The credit rating translates into a credit spread of 1.27%, which gives a pre-tax cost of debt of 3.07%

(Damodaran, 2012). This is at the lower end of the spectrum compared to Ørsted’s historical borrowing rates and the yield to maturity. When discounting operating leases, the discount rate used is the pre-tax cost of debt (Ibid.). Ørsted reports that it used 3.5% in 2017 and 4.5% in earlier years to discount its lease payments (Ørsted, 2017a). These values are in line with the other calculated values. Lastly, a synthetic rating can be estimated from Ørsted’s interest coverage ratio (Damodaran, 2012). However, according to Damodaran (2012) the formula needs to be adjusted to include Ørsted’s use of operating leases:

EDF 11,81

EDPR

7,08

Engie

Centrica ENEL Iberdrola

10,84

E.ON Fortum RWE SSE Ørsted

10,24

Ø 9,50

8,02 9,15 9,05

11,62 11,41

9,94

5,33 Cost of Equity

Page 75 of 162 Modified Interest Coverage Ratio = EBIT + Operating Lease Expense_t0

Interest Expenses + Operating Lease Expense_t0

Table 4 – Modified Interest Coverage Ratio

Source: Authors’ own creation from Ørsted’s annual reports from 2007-2017

Table 4 illustrates Ørsted’s problem with its credit rating and why it needed to issue hybrid capital in order to borrow debt at a reasonable rate. Particularly in the years from 2012-2015, the credit rating is characterised as extremely speculative with very high credit risk. However, the ratio did become better in 2016-2017 and it is currently at investment grade. The value for 2017 is 3.22% and corresponds to a BBB rating, which is close to the official rating. The table below summarises the findings and an average of these values is used as the pre-tax cost of debt.

Table 5 – Cost of Debt

Source: Authors’ own creation

The final input needed to estimate the cost of debt is the tax rate. The Danish corporate tax rate is 22% and preferred over the effective tax rate since it can fluctuate over time (Koller et al, 2010; KPMG, 2018). The Danish tax rate is also close to the EU average of 21.29%, which is important due to Ørsted’s European operations (KPMG, 2018). This results in an after-tax cost of debt of 3%.

5.2.1.7. WACC

After estimating all the inputs, the last step is to calculate the WACC. The WACC is equal to 4.75%, which reflects the low beta. According to Morgan Stanley, the most troublesome aspect of the calculation is holding the WACC constant when the leverage ratio changes throughout the year (Vibig et al., 2008). Therefore, they prefer to use a single measure that represents the average of all the individual annual WACCs (Ibid.). To test whether the found WACC is a realistic long-term WACC, a Monte Carlo simulation is done with 100.000 simulations. The input variables are seen in table 6. With rising interest rates, the risk-free rate will not be lower anytime soon, hence the minimum and most likely are the same. The beta range is based on the bottom-up calculation. Given the confidence of the beta calculation, the beta will not be much lower. The maximum beta is inspired by Ørsted’s peers and with the use of the accounting beta approach, it is adjusted downwards

Year 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

Modified interest coverage ratio 2,14 1,98 0,94 1,60 1,00 -0,65 0,30 -0,13 -0,61 1,49 2,98

Corresponding Credit Rating Ba2/BB B1/B+ Caa/CCC B2/B Caa/CCC D2/D C2/C D2/D D2/D B3/B- Baa2/BBB

Spread 2,38% 2,98% 8,64% 3,57% 8,64% 18,60% 13,95% 18,60% 18,60% 4,37% 1,27%

Risk Free rate 1,95% 1,95% 1,95% 1,95% 1,95% 1,95% 1,95% 1,95% 1,95% 1,95% 1,95%

Pre-tax cost of debt 4,33% 4,93% 10,59% 5,52% 10,59% 20,55% 15,90% 20,55% 20,55% 6,32% 3,22%

Historical rate Bond YTM Default Spread Lease (kd) Modified ICR Average

5,30% 3,97% 3,07% 3,50% 3,22% 3,81%

Page 76 of 162 to reflect Ørsted’s lower operating and financial leverage (Damodaran, 2012). The range for cost of debt reflects the previously defined range.

Table 6 – WACC assumptions

Source: Authors’ own creation

The distribution in figure 39 shows that 4.75% is at the lower end of the distribution. This a function of not believing that the beta will be much lower than 0.6 and with a max of 0.9, skewing the WACC to the right.

The most extreme events are a minimum WACC of 4.3% and a maximum WACC of 6.4%, which are realistic scenarios due to the nature of the triangular distribution. Having stored the variables for each simulation, the sensitivity for each input can be measured. Each variable regressed against WACC clearly shows the beta’s influence on the WACC (Appendix 14). Hence, having a carefully researched beta is important.

It can be discussed whether WACC should be higher. In a recent conference call regarding Ørsted’s zero subsidy win at a German auction, Ørsted commented that their WACC for a zero-subsidy project is 2.5%

higher than their normal WACC for wind power projects (Ørsted, 2017f, p. 5). If Ørsted should win more zero-subsidy auctions, the 4.75% WACC is arguably too low. However, Ørsted withdrew from the auction in the Netherlands, reflecting they carefully assessed the earnings spread over WACC (Reuters, 2017b). Furthermore, Ørsted’s farm-down model allows them to diversify faster into a larger number of projects, which reduces the relative exposure of Ørsted’s cash flows to the contribution of one single project. With this advantage, Ørsted’s WACC should be at the lower end compared to their peers. Ørsted’s WACC of 4.75% seems reasonable compared to the peers with a median WACC of 5.83%. Therefore, a WACC of 4.75% (beta of 0.67) is used with the acknowledgement that it could be in the lower spectrum of Ørsted’s real WACC. The potential weakness and consequence of this choice will be accounted for when performing Monte Carlo simulations of the DCF.

Figure 40 – Ørsted and competitors’ WACC

Source: Authors’ own creation from Ørsted’s annual reports from 2007-2017 and Bloomberg

WACC Assumptions Min Most likely Max

Risk free rate 1.95% 1.95% 2.50%

Beta 0.6 0.67 0.9

MRP 4.50% 4.95% 5.50%

Cost of debt 3.70% 3.81% 5.30%

EDPR Centrica EDF ENEL E.ON Fortum Engie Iberdrola

4,62

RWE SSE Ørsted

6,03

Ø 5,71 2,73

5,83 6,69

4,75

6,53 6,53

5,67

8,70

4,75 WACC

Figure 39 – WACC Monte Carlo

Source: Authors’ own creation with Python

Page 77 of 162 5.2.1.8. Historical WACC

The historical WACC calculation is slightly modified to the WACC used for discounting future cash flows.

The historical WACC reflects Ørsted’s historical business mix, where the division Exploration and Production of oil & gas was included. Furthermore, Ørsted had businesses in Norway, also reflecting different geographical risks (Ørsted, 2010a). However, all of the countries Ørsted has historically operated in have the same credit ratings, reflecting that the implied historical equity risk premium from Denmark is covering the geographical risk (Damodaran, 2018). Ørsted was a private company until the IPO in 2016, so the market values for equity and debt had to be estimated from peers’ capital structure (Damodaran, 2012).

The historical WACC shown in Appendix 15, is higher than the current, which is a result of a higher risk-free rate but also a higher unlevered beta. In the early years, the Power division, accounting for Wind Power and Thermal Power, had a significantly higher beta, which corresponds to higher operational risk. This is a reflection of offshore wind not being a truly global mainstream generation source due to its high LCoE compared to fossil fuels. Investors at this time required a higher compensation when investing in Ørsted. The lower beta in the later years is a product of renewable energy becoming an important energy source in many of the European countries and governments starting to support the renewable companies through subsidies, which lowers the risk for investors. The current WACC, representing the risk for investors going forward, of 4.75% is a natural extension of the trend seen over the years. Figure 41 shows how the beta has changed historically with an increasingly higher portion of EBITDA stemming from Wind Power. From 2007 the beta has decreased c. 56%.

Figure 41 – Historical WACC – EBITDA for each business with the relevered beta

Source: Authors’ own creation from Ørsted’s annual reports from 2007-2017

In document Executive Summary (Sider 73-81)