Orsted A/S

CSE:ORSTED

Earnings Call Transcript 2019-Q3

[Operator]

Welcome to this Ørsted AS Interim Report for the first 9 months of 2019. [Operator Instructions] Today's speakers are CEO, Henrik Poulsen; and CFO, Marianne Wiinholt. Speakers, please begin.

[Henrik Poulsen, President & CEO]

Thank you, and good morning, everyone. I will start out summarizing third quarter results, and then I'll cover the update on our long-term targets before ending up with an update on our projects and markets. Our company continued its strong financial and strategic performance in the third quarter, where EBITDA amounted to DKK 4.1 billion. This was an increase of DKK 1.9 billion or 85% compared to same quarter last year. The increase in EBITDA was mainly driven by our offshore wind farms in operation, where we saw a year-on-year increase of 35%, driven by ramp-up in generation from Hornsea 1 and Borkum Riffgrund 2 as well as higher wind speeds. We increased our full year EBITDA guidance with DKK 0.5 billion during the quarter as a consequence of the reversal of a provision related to the Elsam competition case and strong wind conditions, especially in the month of August. Our financial performance is in line with our expectations and keeps us well on track to reach our full year guidance of DKK 16 billion to DKK 17 billion. The continued buildout within our offshore and onshore businesses brings our [ green ] share of heat and power generation to 87% for the third quarter compared to 71% in third quarter last year. In September, as you know, we signed an agreement to divest our Danish power distribution residential customer and city light businesses to SEAS-NVE at a price of DKK 21.3 billion on a cash and debt-free basis. We expect the transaction to close during the first half of next year, subject to a regulatory approval by the competition authorities in the Danish Energy Agency. At closing of the transaction, approximately 750 employees will be transferred to SEAS-NVE and continue to serve the divested businesses. The outages and curtailments we experienced across our portfolio during the first half persisted into the third quarter, where we saw a profit impact of approximately DKK 150 million. The operational issues at London Array, Borkum Riffgrund 2 and Race Bank have largely been resolved during third quarter, and going forward, we don't expect further production losses related to these issues. In early October, we installed the last turbine at Hornsea 1. The wind farm is now undergoing a period of extensive testing and commissioning and is expected to be commercially operational later this year. When commissioned, Hornsea 1 will, as you know, become the world's largest offshore wind farm with its 1,218 megawatt, and it will be able to supply more than 1 million U.K. households with green power. As you likely noticed, we experienced a rapid reduction of power at Hornsea 1 in August during an unusual set of circumstances affecting the transmission grid. This was caused by an unexpected control system response only revealed during these unusual circumstances, and it has now been resolved. Hornsea 1 is progressing through the necessary compliance tests with National Grid. And by mid-November, we would expect to lift the maximum export capacity from the current 800 megawatt to the full 1,218 megawatt. During the turbine commissioning ramp-up period, we have seen high turbine availability at Hornsea 1. In September, we selected GE as preferred turbine supplier for our U.S. mid-Atlantic cluster covering Ocean Wind and Skipjack. These projects will pioneer the deployment of GEs, Haliade-X 12-megawatt turbine, continuing our track record as a first mover on new technology. We have signed contracts with Siemens Gamesa to supply turbines for our greater Changhua 1 & 2a project and the Northeast cluster in the U.S., consisting of the Sunrise Wind, Revolution Wind and South Fork Offshore Wind Farms. In August and September, respectively, we submitted bids in the Massachusetts and Connecticut offshore wind solicitations. Earlier this year, our 880-megawatt Sunrise Wind project off the coast of New York was selected as the preferred bidder. Since the award, we have been negotiating an offshore wind renewable energy certificate for the project. And in October, Sunrise Wind signed a 25-year power purchase agreement with the New York Energy Research and Development Authority. The project will receive a fixed all-in price of $110.37 per megawatt hour from 2024, corresponding to a levelized 2017 price of $79.6 per megawatt hour. Following a structured competitive process, we were recently selected by PGE to commence discussions regarding the sale of a 50% stake in 2 Polish offshore wind projects in the Baltic Sea, with a total capacity of up to 2.5 gigawatt. The subject of discussions will be a sale of 50% of the Baltica 3 development project with a planned capacity of approximately 1 gigawatt for construction by 2026, and 50% of the Baltica 2 development project with a planned capacity of approximately 1.5 gigawatt for construction by 2030. We're quite excited about developing a potential partnership with PGE. Poland has strong offshore wind conditions and could develop into a sizable new market. In onshore, we've taken final investment decision on our 230-megawatt onshore wind farm, Plum Creek in Nebraska. Plum Creek is expected to be commissioned during fourth quarter next year. Last year, the Danish Western High Court acquitted Elsam, now Ørsted, of the competition authorities' claim that Elsam abused a dominant position on the Danish wholesale power market back in 2005 and 2006. In light of this ruling, the parties have agreed on dismissing the competition authorities' similar claim for second half of 2003 and all of 2004. Consequently, the cases between Elsam and the competition authorities have now reached a conclusion in favor of Ørsted. Despite the fact that the competition authorities' claims against Elsam and Ørsted have now been dismissed, the claimants have chosen to maintain their claims for damages and to continue with their legal action. Finally, in October, we established a commodity trading unit in Chicago based on our experiences from our successful European trading setup. The role of the U.S. trading activities will be to manage and mitigate merchant risks derived from onshore wind farms in the U.S. Turning to Slide 4 and the update of the long-term targets shared at the Capital Markets Day a year ago. I'm sure we'll come back to the topic in the Q&A, but let me just walk you through the primary drivers of the adjustment of our targets. And before I dive into it, I should reiterate that these targets, obviously, are impacted by many factors, including CapEx and OpEx estimates, production forecasts, expected [ and ] long-term power prices, currency, et cetera. Therefore, our estimates will remain very dynamic, and today's update is as such just a snapshot that will continue to evolve. We have, as announced, 3 things that have added pressure to our long-term target. Starting out with the adjustment of our production forecast. We have been running a comprehensive project to upgrade the tools and processes we use to forecast the annual energy production from our offshore wind farms, exploiting our unique access to production, wind and turbine data from our large asset portfolio. Forecasting offshore wind is inherently a complex task. You need to forecast the wind speed, its direction, how it flows, how strong it is at each turbine position, how it interacts with the turbine, how the turbines impact each other, how the wind at each position translates into electricity production, how often and when the turbines will be available and not, et cetera, et cetera. Given the high number of variables and the use of advanced analytical models, crunching millions of data points, such an exercise comes with a fair amount of uncertainty. Now with that said, the project has made us conclude that our current production forecast underestimate the blockage and wake effects across our asset portfolio. Blockage arises from the wind slowing down as it approaches an object, in this case, our wind turbines. There is an individual blockage effect for every wind turbine position as well as a global effect for the whole wind farm, which is larger than the sum of all of the individual effects. When the wind hits the front row of a wind farm, it will slow down as it approaches the front wall, so to speak. Our new wind simulation models suggest that we historically have underestimated these blockage effects. This finding is also supported by a recent report on blockage from industry consultants, DNV-GL, which indicates that this effect is more broadly underestimated. The second effect is the wake within wind farms and between neighboring wind farms. This effect where the turbine shield and impact each other has been subject to extensive modeling by the industry for many years, and it is still a highly complex dynamic to model. We have now introduced a more advanced model for estimation of wake effects within a wind farm. It leverages data from our entire operational portfolio and benchmarks, [ to ] the predictions against production data from our SCADA systems. The results point to a higher negative effect on production [ than ] earlier models predicted. With respect to wake effects between neighboring wind farms, we are also here in the process of developing a new model, capable of more accurately predicting wake effects over longer distances. We have, among other things, leveraged a first-of-its-kind advanced radar system, collecting 3-dimensional data on the wind flow. This system has been used at the West, the most rough wind farm in the U.K., where it has given us a lot of new insight into the wind flows. We will also deploy it in Taiwan. The new model, albeit still being refined, suggests a slower wind speed recovery, and therefore, higher wind effects -- wake effects. At the same time, we have now factored in a more extensive offshore wind build-out in the different regions, which will increase the wake effect from neighboring wind farms. As the global offshore wind build-out accelerates the whole industry will see higher wake effects from neighboring wind farms. We have, over the years, benchmarked our internal production estimates against third-party views from industry experts and other developers. In comparison, most production estimates from third parties have been trending towards a more positive view than ours. Therefore, we do believe that underestimation of blockage and wake effects likely is an industry-wide issue. These higher-than-forecasted blockage and wake effects have also been embedded in our historical actual production numbers, but they have been captured in more broadly defined deviation [ buckets ] like wind content, availability, curtailments and ramp-up effects. We have until now not had the data and the advanced analytics models to do a more granular breakdown of the production deviation. These new tools, leveraging all our production data, including the last new assets build over the past couple of years, have given us a more detailed insight into the factors impacting our production. While the production deviation, we have discovered is negative, I'm firmly convinced that Ørsted's unparalleled access to production data and our advanced analytics capabilities will help drive our competitive advantage, getting smarter and gaining a more granular insight into our production dynamics is, in itself, a good thing. And it goes without saying that we will seek to leverage the recent findings to enhance the design of future wind farms. And while not immaterial, the forecast adjustment is not something that changes the competitiveness of offshore wind nor does it change Ørsted's ability to drive profitable growth. We remain very confident about both. Moving on to Slide 5 and the other negative and positive developments we have seen since we announced the CMD targets a year ago. As a second negative impact, and as you know, we ultimately had to accept a 6% reduction in our feed-in tariff and a cap on full-load production hours on our Changhua 1 & 2a project in Taiwan, back in Q1 this year. This is old news, but still part of the mix of factors that impact the targets communicated at the Capital Markets Day a year ago. And thirdly, in the U.S., we have raised the CapEx estimate for the deepwater development portfolio, primarily related to the transmission assets. We have, since the acquisition of deepwater wind, been working to [ mature ] the EPC case for the deepwater portfolio. And this work, while still very much being work in progress, has so far led to higher CapEx estimates. In terms of positive development since the Capital Markets Day, CapEx estimates for some of our construction projects have improved a bit. Low interest rates have led to lower return requirements on our OFTO assets, which translates into lower ongoing transmission charges in the U.K. And we have seen higher-than-budgeted availability on one of our newer turbine platforms, which positively impacts some of our assets.Furthermore, we will reduce our overhead cost base by DKK 500 million to DKK 600 million between 2020 and 2022, recognizing that tight cost control remains an imperative in a competitive market environment. Roughly half of the cost reductions will be fall-away costs relating to the simplification of our structure following the divestment of our Danish downstream assets. And the other half will come from reductions across our staff functions, both internal and external spend. When we combine these key impacts since the Capital Markets Day, we come to the status on the long-term financial targets released yesterday. Let me just reiterate. Average growth in site EBITDA, around 20% for the period 2017 through 2023, is unchanged, and we remain comfortable with the projected tripling of site EBITDA from 2017 to '23. Our growth outlook, in other words, remains very strong, and we continue to have strong visibility on our EBITDA growth for the coming many years. Average return on capital employed around 10% for the period 2019 through 2025 is unchanged. Our growth, in other words, remains healthy and profitable. The unlevered life cycle IRR capacity weighted for 7 offshore wind projects across Europe, U.S. and Asia, where we set the target at 7.5% to 8.5%. This target has been reduced to 7% to 8% due to the production forecast adjustment, the reduced feed-in tariff in Taiwan and the higher CapEx estimate for the revolution wind project in the U.S. I should remind you that this IRR metric is fully loaded with all development costs and corporate overhead included, and it does not include any uplift from potential farm downs in the 7 projects. Despite the adjustment of the target, we maintain a healthy value creation in this portfolio of 7 construction and development projects. Also in light of declining interest rates that lead to lower cost of capital. Some of you may wonder why we see an adjustment to the IRR target and not the return on capital employed and EBITDA growth targets. The main reason is that the effects from the Taiwanese tariff reduction and the higher transmission CapEx estimate for revolution wind, have a more concentrated impact on the IRR target as it only comprises 7 projects, including the Changhua and Revolution Wind projects. Contrary to this, the return on capital employed and EBITDA growth targets, both built on the entire offshore and onshore asset portfolio. When spreading the positive and negative developments, including the overhead cost reductions across the entire asset portfolio, our targets remain unchanged. The average share of EBITDA from regulated and contracted activities targeted to be around 90% for the period 2019 through 2025 is also unchanged. In other words, our portfolio merchant exposure remains very low. The estimated lifetime load factor of 48% to 50% for a defined European portfolio of 10 wind farms is reduced to around 48% due to the adjustment of the production forecast. The CapEx and OpEx multiples communicated at the Capital Markets Day, remain unchanged. Let me reiterate that our fundamental economics and value creation remain solid. And Ørsted remains uniquely positioned to tap into the vast growth opportunity offered by global renewables markets that only continues to expand at a still higher rate. This concludes the update on our long-term financial targets from last year's Capital Markets Day. Let's turn to Slide 6., where I'll give an update on the key offshore construction projects currently in progress. At Hornsea 1, we installed the last of 174 wind turbines in the beginning of October. With the final turbine installed, the project will undergo a period of testing and commissioning and will be commercially operational later this year. More than 8,000 people have worked offshore at the Hornsea 1 site, and we are truly proud to complete the work at this record-breaking project, becoming the world's largest offshore wind farm. At our Borssele 1 & 2 wind farm, the construction of the O&M building in Vlissingen is progressing and is expected to be completed in fourth quarter this year. Offshore installation works will start ahead of schedule with monopile installation starting in late December this year, and turbine installation expectedly in April 2020. We still expect the wind farm to be completed by the end of 2020 or in early 2021. The Virginia EPC demo project remains well on track. In October, the U.S. Bureau of Ocean Management -- Ocean Energy Management issued a no-objection determination on the facility design report and fabrication and installation report for the project. This is a significant milestone as we move forward of building the first-ever fully permitted offshore wind project in federal waters. The Coastal Virginia offshore wind project will be the first offshore wind farm in the U.S. Mid-Atlantic, and we expect to complete the 2-turbine, 12-megawatt project by the end of next year. At the Hornsea 2 project, the onshore construction work is progressing according to plan, and we continue to expect completion of the wind farm in the first half of 2022. At our Greater Changhua 1 & 2a project, we are now finalizing the signing of supplies and installation contracts. The onshore construction work is progressing according to plan, and we expect the wind farm to be fully commissioned by 2022. In September, we achieved first power at the Formosa 1 Phase 2 project, and the wind farm is scheduled to be officially inaugurated over the coming weeks. Taiwan's first-ever offshore wind farm is now fully up and running and supplying green power to the grid. Turning to Slide 7 and an update on construction projects outside offshore. In our onshore business, we continue to see good progress on our construction projects. In August, we secured tax equity funding commitment for Sage Draw. Construction commenced back in June with road and foundation installation well underway. The 338-megawatt onshore wind farm is expected to be completed by first quarter of next year. The construction of Willow Creek is also progressing according to plan with expected completion in fourth quarter next year. In September, the construction of Plum Creek in Nebraska began with foundation construction, and the project is on track to be completed in fourth quarter 2020 as well. In bioenergy, the bioconversion of the Asnæs powerplant is progressing according to plan, and we still expect final commissioning by the end of the year. The ramp-up of the waste throughput and production on our first full-scale Renescience plant in the U.K. is progressing, and we see small, but steady improvements. Further work and mechanical optimization is needed before we have visibility on a stable and coherent, technical and commercial formula for the Renescience technology. We still target final commissioning of the plant by the end of this year. In August, Radius reached an important milestone with the installation of Smart meter number 1 million. We have seen the main phase of the installation work now being successfully completed on time and budget. The project will be finalized as planned by the end of this year. To conclude on the ongoing construction projects, let me just say that I remain very happy and satisfied with the execution capacity and capability of the organization. Let's turn to Slides 8, 9 and 10 and take a look at the latest market development and offshore wind opportunities across the different regions. Starting in Massachusetts, where we have submitted bids in the 800-megawatt offshore wind solicitation with our Bay State Wind project in a joint venture with Eversource. The selection of projects for negotiation will expectedly soon be announced. In Connecticut, we've submitted bids in the up to 2-gigawatt offshore wind solicitation with the Constitution Wind project in a joint venture with Eversource. We expect an outcome from the auction in November following the award in Massachusetts. Yesterday, we announced that we will enter into exclusive negotiations with PSEG for them to potentially become an equity investor in our 1,100 megawatt Ocean Wind project in New Jersey. Subject to further negotiations towards a joint venture agreement, due diligence and any required regulatory approvals, PSEG would acquire 25% of Ocean Wind. In New York and New Jersey, we continue to see strong commitment to offshore wind, with auctions scheduled for second half of 2020 in both states. In Maryland, we expect the first auction to open in the first half of 2020, followed by sequential [ indiscernible ] auctions in 2021 and 2022. Finally, Virginia Governor Ralph Northam signed an executive order establishing a nonbinding 2.5-gigawatt offshore wind capacity target to be fully commissioned on an accelerated time line by 2026. In addition to the updated target, the Governor also announced that the 2.5 gigawatt of offshore wind will be deployed in federal waters east of Virginia Beach leased by Dominion Energy in 2013. Back in July 2017, Ørsted entered into a strategic partnership with Dominion Energy, where we are to build the 2 turbine pilot project off the coast of Virginia Beach with the capacity of 12 megawatt. In addition to this, we signed a memo of understanding, which gives us the exclusive rights to discuss a strategic partnership with Dominion Energy about developing their commercial site based on the successful deployment of the pilot wind farm. Turning to Slide 9 and the recent market developments in Europe. The determination date for Hornsea 3 development consent has been postponed from October 2019 to end of March 2020. The Secretary of State is seeking additional information to consider potential impacts on protected sites, including the associated effects of other wind farm projects. We are working with the relevant bodies to respond to the Secretary of State on these matters and still anticipate a positive decision in due course. Moving to Germany, where the offshore wind capacity target has been increased from 15 gigawatt to 20 gigawatt towards 2030. The first centralized 900-megawatt German tender is expected to take place in 2021. In the Netherlands, the draft tender conditions for Holland Coast North were released in October. The tender will have a capacity of up to 760 megawatt with bid deadline, April 16, 2020. In Denmark, the next tender of 800 to 1,000 megawatt, has been launched with the expected bid deadline in fourth quarter 2021. In France, we also expect the next tender to be issued during 2021. The capacity of the fourth round will expectedly amount to 1 gigawatt. As I mentioned earlier, we've been selected by PGE in Poland to commence discussions regarding the sale of a 50% stake in their 2 offshore wind projects in the Baltic Sea, with a total capacity of up to 2.5 gigawatt. The Polish government has agreed to have a 10.3 gigawatt target for offshore wind, commissioned by 2040. The government is working to support this ambition through a regulatory framework to be codified into law through the Offshore Wind Act. The sector expects to have the Offshore Wind Act in place by first quarter next year, which should cover all relevant regulatory areas for development of offshore wind. As I mentioned earlier, we are quite excited about the possibility to develop offshore wind in Poland through the agreement with PGE. Finally, turning to Slide 10 and the market development in Asia Pacific where I will focus on Japan. In Japan, we continue to develop our partnership with TEPCO, with a focus on the Choshi zone off the coast of Tokyo. The Japanese government has quantified that the announced designated 11 zones, potentially suitable for development of offshore wind, have a capacity of approximately 7 gigawatt. Four of these areas, including the Choshi zone have been selected as prospective areas and will work towards qualification during first quarter next year. The Ministry of Economy Trade and Industry is pursuing a targeted time line for a first auction round to take place in second half next year. Turning to Slide 11. I will not go through this slide in any detail. It is just for your reference. It seeks to provide an overview of the many upcoming offshore wind auctions and tenders in 2020 and 2021, highlighting the strong global demand for offshore wind. We have never been more optimistic about the competitiveness of offshore wind as a technology as also alluded to by the International Energy Agency in their report from last week. We are looking into a global offshore wind market that will further accelerate over the coming years. The technology has some very strong competitive characteristics, and what just a few years ago was seen as a niche technology, is now on track to become a cornerstone power source in many countries as we move towards 2030 and beyond. This concludes the offshore market development. Let's turn to Slide 12 and the progress of our U.S. onshore business. The U.S. onshore business continues to expand its portfolio of operating and development projects. With the final investment decision on the Plum Creek onshore wind farm in August, our total installed and [ decided ] onshore capacity now stands at 1.7 gigawatts. We expect to take the final investment decision on our first large-scale solar farm, the 400-megawatt Permian Solar project in Texas, later this year. With this, I will now pass on the word to Marianne.

[Marianne Wiinholt, Chief Financial Officer]

Thank you, Henrik, and good morning from me, too. Let's start on Slide 13, where I will go through the group financials for Q3 '19. In Q3 '19, we realized an EBITDA of DKK 4.1 billion, a year-on-year increase of DKK 1.9 billion, in line with our expectations. In offshore, earnings from our operating wind farms increased by 35% due to the ramp-up of generation from Hornsea 1 and Borkum Riffgrund 2, as well as higher wind speeds. Offshore also realized higher earnings from construction of offshore wind farms for partners, mainly driven by the construction of Hornsea 1. Onshore contributed with DKK 0.3 billion in the quarter, while bioenergy was above last year due to the reversal of a provision of DKK 0.3 billion following the acquittal in the Elsam competition case. In customer solutions, we saw lower earnings from LNG, mainly driven by extraordinary high earnings from LNG in Q3 '18. And furthermore, lower gas prices impacted the accounting value of our gas at storages and thus, led to a temporary negative impact in markets. Net profit totaled DKK 1.4 billion, an increase of DKK 1 billion year-on-year. The increase was driven by the higher EBITDA, partly offset by higher depreciation from more wind farms in operation as well as the implementation of the new IFRS 16 accounting standard regarding leases. The free cash flow from continuing operations was negative DKK 6.1 billion. In Q3 '19, cash flows from operating activities came in at DKK 0.9 billion, mainly driven by the EBITDA, a tax equity contribution from our partner at the Lockett onshore wind farm and lower gas inventories. This was partly offset by more funds tied up in work in progress. Our gross investments for the quarter totaled DKK 7.2 billion, which mainly related to the construction of Hornsea 1, the Greater Changhua 1 & 2a and Borssele 1 & 2 and last, our onshore projects. If we then turn to Slide 14 and our net interest-bearing debt and financial ratios. Our net debt at the end of Q3 '19 amounted to DKK 12.1 billion. The DKK 7.1 billion increase compared to 30th of June '19 primarily reflected the contribution from the free cash flow, as I just described, as well as a DKK 0.8 billion impact from exchange rate adjustments. Our key credit metric FFO to adjusted net debt stood at 47%, well above the target of around 30%. Our return on capital employed came in at 29%, a 6 percentage point increase compared to the same 12 month last year. The increase was significantly impacted by the farm-down gain from Hornsea 1, whereas the same period last year was impacted by the farm-down gains from Walney Extension and Borkum Riffgrund 2. If we then move to the results of the business units, starting with offshore on Slide 15. Power generation amounted to 2.8 terawatt hours, an increase of 0.9 terawatt hours compared to Q3 '18. This was primarily driven by the ramp-up of generation from Hornsea 1 and Borkum Riffgrund 2, which together accounted for 0.5 terawatt hours. Wind speeds for the quarter amounted to an average of 8.5 meter per second, up 0.8 meter per second compared to last year. This was also above the normal wind speed for the quarter of 7.9 meter per second across our portfolio. For the first 9 months of '19, the wind speeds were 9.0 meter per second, which was also higher than the normal wind speed of 8.8 meter per second for the portfolio. EBITDA for the quarter amounted to DKK 3.3 billion, up DKK 1.3 billion on Q3 '18. Earnings from wind farms in operation increased 35%, again, driven by ramp-up of Hornsea and Borkum Riffgrund 2 as well as higher wind speeds. As Henrik mentioned in the beginning of the call, the operational issues at London Array, Borkum Riffgrund 2 and Race Bank have been resolved during third quarter. And going forward, we don't expect further production losses related to these issues. Earnings from partnerships amounted to DKK 1.2 billion, an increase of DKK 0.5 billion compared to last year. The construction agreements in this quarter, primarily concern Hornsea 1. Finally, the project development costs amounted to DKK 0.6 billion, mainly related to development activities in U.S. and so on. Free cash flow came in negative at DKK 6.2 billion for the quarter and a decrease of DKK 3 billion compared to last year and mainly driven by more funds tied up in work in progress due to the construction of Hornsea 1, whereas we had a positive cash inflow in Q3 '18 as we received milestone payments related to Borkum Riffgrund 2. If we then turn to the results for onshore on Slide 16. In onshore, power generation amounted to 0.9 terawatt hours for the quarter, and the wind speed averaged 6.6 meters per second, which was slightly below the normal wind speeds of 6.7 meters per second in Texas for the quarter. We had a very high availability of 98% across the portfolio. The EBITDA came in at DKK 0.3 billion for the quarter, with sites contributing DKK 0.2 billion positively affected by high peak power prices in Texas in August. As the Lockett wind farm was completed a couple of months ahead of schedule and therefore, was fully exposed to merchant prices, it significantly benefited from these peak prices. Production tax credit added an additional DKK 0.1 billion, and this was partly offset by project development costs. The free cash flow amounted to a negative of DKK 0.3 billion and related to gross investments in Sage Draw, Plum Creek, Willow Creek and Permian Solar, partly offset by the tax equity contribution from our partner at the Lockett Wind farm and less funds tied up in other net working capital. Turning to Slide 17, covering the results in bioenergy. EBITDA came in at DKK 0.2 billion, up DKK 0.4 billion on Q3 '18. The increase was mainly due to the reversal of the provision of DKK 0.3 billion related to Elsam case. The free cash flow in Q3 '19 amounted to a negative of DKK 0.6 billion, a decrease of DKK 0.4 billion compared to last year. The decrease was mainly due to lower payables, due to higher fuel inventories at the beginning of the period. If we then turn to Slide 18, covering the results in customer solutions. The EBITDA for Q3 totaled DKK 0.2 billion, a decrease of DKK 0.3 billion on last year. The lower EBITDA was mainly driven by LNG and our gas portfolio within markets. LNG contributed with extraordinary high earnings in Q3 '18 due to the utilization of location spreads and optimization of physical assets. The lower gas prices we saw during Q3 '19, resulted in a temporary negative effect from revaluation of the LNG at storage. In addition, we have seen temporary negative effects from oil-indexed LNG purchase agreements that are hedged with a time lag. Consequently, we see a timing difference between the date when the market value of the hedging contract is recognized and the physical delivery date. The lower EBITDA in markets was driven by continued lower gas prices in Q3 '19, which has led to a decrease in the accounting value of the gas at storage and thus, a temporary negative impact for the quarter. This negative impact will be offset if the gas prices increase or when we sell the gas in '19 or '20, as we have hedged most of our gas margin. In distribution, EBITDA increased by DKK 0.1 billion, which was mainly due to timing of activities between years. The free cash flow from operating activities for the quarter amounted to DKK 1.1 billion, primarily from lower receivables and lower gas inventories due to the low gas prices and gas sales. Slide 19 show our 2019 guidance and our long-term financial estimates and policies. On the 25th of September, our 2019 EBITDA guidance for the group was increased by DKK 0.5 billion. We still expect EBITDA to be between DKK 16 billion and DKK 17 billion for the year. Our directional EBITDA guidance for each business unit is unchanged relative to the guidance in the interim report for the first half year. Our gross investment guidance is unchanged relative to our guidance in the annual report for '18. Gross investments are expected to amount to between DKK 21 billion and DKK 23 billion. And with that, we will now open for Q&A. Operator, please.

[Operator]

[Operator Instructions] And our first question comes from the line of Casper Blom from ABG Sundal Collier.

[Casper Blom, Lead Analyst]

Two questions from my side, please. First of all, now that you lowered the predicted output from the wind farms, does that in any way change the economics in the farms -- parks that have already been farm down. Or said in another way, do your partners at these farms now have the -- have you say possibility to come back to you and say we would like a refund because this is not producing as we were promised? That's my first question. Secondly, as you said, Henrik, this is a snapshot. And clearly, there is an uncertainty in the long-term economics of wind parks. Now that we see a negative revision of the outlook, does that in any way change your return requirements for such parks? Those are my 2 questions, please.

[Henrik Poulsen, President & CEO]

Thank you, Casper. To the first question, our farm-down agreements are done on a shared risk basis. So this type of risk is fully shared between us and our partners. So there is no basis for any claims against us. And in terms of the snapshot and our return requirements, this does not in and of itself change our return requirements, our return requirements would typically change in response to changes in the cost of debt and cost of capital -- cost of equity, obviously. So I don't see this in and of itself changing our return requirements.

[Operator]

Next question comes from the line of Deepa Venkateswaran from Bernstein.

[Deepa Venkateswaran, Senior Analyst]

I have a few questions. So I think on the news from yesterday, just wanted to understand how does having larger turbines change the impact of the Block and [indiscernible] make it better? And the second question is...

[Henrik Poulsen, President & CEO]

The line is not very good, Deepa. I can -- we can barely hear what you're saying?

[Deepa Venkateswaran, Senior Analyst]

Is it better? Hello?

[Henrik Poulsen, President & CEO]

Yes, it's better. Thank you.

[Deepa Venkateswaran, Senior Analyst]

Hello?

[Henrik Poulsen, President & CEO]

Yes. Can you hear us, Deepa.

[Operator]

Okay and we seem to have just lost Deepa's line. So our next question comes from the line of John Musk from RBC.

[John Musk, Analyst]

I'll just have 2 questions for now. Firstly, on the OpEx savings that are partly offsetting some of the negative impacts, within the 50 basis points move down in the IRR, how much of those OpEx savings have you put in there versus how much of those will be going against the existing asset base? And then secondly, longer term, if we -- as you indicated, the blockage and wake effects happen as we get more and more build-out, if that, sort of, accelerates, are we likely to see another leg down in load factors in a few years' time, as we get more assets built around your existing asset base?

[Henrik Poulsen, President & CEO]

Thanks, John. In terms of the OpEx reductions, those OpEx reductions would translate into improvements across the entire portfolio of offshore and onshore assets. And therefore, they will also benefit the 7 projects that are in the bucket that we use for the IRR guidance, but only with their, sort of, pro rata share of the total overhead reduction. We have added more neighboring wind farms into our simulations of the neighboring wake effects as part of this exercise. And on that basis, we do not expect a further reduction of our load factor due to continued build-out of offshore wind.

[John Musk, Analyst]

Okay. So just to come back on the OpEx, is there any split of that number into the construction portfolio and the existing portfolio?

[Henrik Poulsen, President & CEO]

The allocation key is the number of hours that is spent on the individual projects. So it's allocated with an hourly rate.

[Operator]

And our next question comes from the line of Deepa Venkateswaran from Bernstein.

[Deepa Venkateswaran, Senior Analyst]

I'm just checking. Can you hear my question?

[Henrik Poulsen, President & CEO]

Yes, we can, Deepa. Thank you.

[Deepa Venkateswaran, Senior Analyst]

Okay. All right. Apologies for that. So my question was, firstly, what's the impact of moving to larger turbines on the blockage and wake effect? Is it neutral? Or does it worsen it or just does it improve the output? Second question is the better availability on one of your turbine platform, is it safe to assume that you're really talking about your Hornsea 1 and Borssele 1 and 2, the platforms there? Or is this a wider and impacting more wind farms? And my third question is on the farm-down of Ocean Wind to PSEG, keeping in mind the, sort of, risk sharing you have, what kind of pricing should we expect? Should it be like at cost as you did to Eversource? Or should it be like the typical module you've been using in Europe for financial investors or somewhere between?

[Henrik Poulsen, President & CEO]

Thanks, Deepa. When we move to larger turbines, again, we would obviously need real data to fully understand the dynamic of moving to a larger turbine. The expectation right now would be that larger turbine may lead to marginally higher blockage effects. On the other hand, we would expect wake effects to go down on a relative basis to the capacity of the turbine.

[Deepa Venkateswaran, Senior Analyst]

Okay. So at best, it's indeterminate now, or should we just assume that it will marginally worsen?

[Henrik Poulsen, President & CEO]

I would, for now, expect this to be a net 0 as we move to larger turbines. And in terms of the better availability of one of our turbine platforms, I'm not going to be specific, obviously, you know our assets, you know our turbine portfolio extremely well. So there are only that many options, of course. But I'm not going to specify it by name. I'm looking to the third question was.

[Deepa Venkateswaran, Senior Analyst]

Farm-down to PSEG.

[Henrik Poulsen, President & CEO]

To the PSEG, yes. Well, it's an ongoing negotiation. So we're still negotiating with PSEG on the price for this equity stake. So I could not give you any details today.

[Operator]

Our next question comes from the line of Kristian Johansen from Danske Bank.

[Kristian Tornøe Johansen, Senior Analyst]

So 2 questions for me. First question, if we look at some of your most recent wins like Ocean Wind and Sunrise Wind, which are not included in the portfolio of wind farms where you lowered the return assumption by 0.5 percentage point, should we expect a similar adverse effect on expected returns on these projects? Or are you still at a stage where you should be able to mitigate this lower production forecast?

[Henrik Poulsen, President & CEO]

Well, the impact of the changes from this exercise, it varies by project, obviously. So -- and when you look at Ocean Wind and Sunrise Wind, there's also an impact on these 2 projects. But I cannot, obviously, start specifying these impacts at an individual project level, but there is an impact. We have also reduced our production forecast for these 2 projects. We obviously are doing everything we can to mitigate this small adjustment to the production forecast by mitigating through CapEx and OpEx levers. And if we can find ways of optimizing the production to bring some of it back, we obviously will. But that's all work in progress. So it would be too early for me to share anything specific. But the fundamental economics of the project are still healthy.

[Kristian Tornøe Johansen, Senior Analyst]

Okay. So there is no risk on the FID from this?

[Henrik Poulsen, President & CEO]

Risk of FID on Ocean Wind and Sunrise Wind? No, I do not see a risk on FID.

[Kristian Tornøe Johansen, Senior Analyst]

Okay, that's quite clear. Then my second question, how should we think about your competitive strengths as a result of this. I mean, in future auctions I assume you would have to take in a lower production forecast, which if you keep your return assumption would yield a higher bid price. And if competition doesn't do the same immediately, will that make you less competitive?

[Henrik Poulsen, President & CEO]

Well, all other things equal, the answer would be, yes. It's quite clear that anyone can win an offshore wind project, if they want to, sort of, just dial up their expectations on specific assumptions, including production forecast. So this all comes back to being a disciplined and prudent allocator of capital. And our challenge, obviously, is to make sure that even if we have, in our view, more prudent and better production forecast that we should still be able to win projects in future and have a more robust value creation in our wins. That is our challenge. But it goes without saying that if competition has much more optimistic production forecast than us, that will, obviously, in isolation impact our competitiveness. But when you look at it from a shareholders' perspective, Kristian, you also have to think that our primary task is to make sure that we only take on projects where we do indeed create value. It is not in the interest of any shareholder for us to operate with inflated production forecasts and winning projects on that basis. That's not going to bring anything good to anyone.

[Operator]

Our next question comes from the line of Peter Bisztyga from Bank of America.

[Peter Andrew Bisztyga, Head of Pan]

So my first question regards these higher costs relating to transmission assets for your U.S. portfolio, I'm just wondering if you can specify exactly where those cost pressures have come from? Is it the cables? Is it labor? Is it installation cost? And also, were those high costs included in your recent bids for Ocean Wind and Sunrise? Or is that another factor that is going to affect the economics of those projects? Second question, I think you've already answered this, but just to double check, is the plan to sell the Ocean Wind stake to PSEG, sort of, cost on a similar basis that you did with Eversource? Or is this, sort of, a farm-down where you could secure premium? And then finally, I'm just wondering, do you think there's any kind of political angle to the U.S. interior department requiring the BOEM to redo their environmental impact assessments? And do you worry the federal government could curtail some of the very ambitious state-level targets for offshore wind development?

[Henrik Poulsen, President & CEO]

Thank you. When it comes to the higher estimated transmission CapEx in the U.S., it comes from a number of sources across the Deepwater Wind portfolio. Some of it relates to export cable, some of it relates to substations, some of that relate to onshore grid upgrades, as well as the cable landing. So it's a relatively broad set of factors that impact these numbers. It is not necessarily a surprise that as we mature these projects, we sometimes tend to see CapEx estimates either moving up or down, compared to our original estimates. When it comes to Sunrise Wind and Ocean Wind, we have had visibility on these transmission CapEx estimates for Deepwater when we submitted our bids. So we have accounted for these higher transmission assets or transmission CapEx in our bids for Sunrise and Ocean Wind. When it comes to the negotiation with PSEG for an equity stake in Ocean Wind, it is a commercial negotiation, where we're obviously taking a number of factors into account, including what is still developing into a very good partnership with PSEG, who's obviously able to contribute a number of benefits to our position in New Jersey and the Mid-Atlantic market. But with that said, it is a full commercial negotiation. As it relates to the interior department's request for BOEM to do this so-called cumulative impact assessment, it will obviously be something that we are eagerly awaiting to see the outcome expectedly early next year. We are not concerned that this is going to stop the large-scale deployment of offshore wind along the U.S. East Coast, that potential is huge and the states along the East Coast they need renewable energy. They can see that offshore wind is a very competitive solution, and they are quite determined to deploy offshore wind. I see this more as a natural step for BOEM to gain a broader understanding of how this offshore wind build-out will impact different key stakeholders along the coast, including fishing communities, local communities, where we land the cables, et cetera. So there are many stakeholders that you need to consider. You need to have a robust and productive dialogue. And you need to make sure that all concerns are listened to and develop a comprehensive framework for how to build-out offshore wind as effectively as possible. This is the exercise they're going through. Frankly speaking, that makes sense. And it may cause some delays here and there, but I don't think it's going to stop the build-out or slow it down as such.

[Operator]

Our next question comes from the line of Sam Arie from UBS.

[Samuel James Hugo Arie, Managing Director and Research Analyst]

I think a lot of good questions asked already, so just one for me. And it's bit of a difficult one to formulate. But what I want to do is drill into this point about the return guidance that you've given now being lower, but the return on capital and EBITDA targets being unchanged. And I think we follow that logic. There is some positive at the group level, which offset the lower project returns. But the project returns are over 25, 30-year duration, I suppose, whereas the return on capital and EBITDA is just the next 4 or 5 years. So what you seem to be saying is that the net effect of everything you've talked about today and yesterday, there's no downside on numbers at group level for the next 4 or 5 years. And what I want to ask is, if you think there's any downside actually at all when you go further out. So I thought to be clear, I'm not asking you to share your own company valuation, but I think that you do run valuation estimates internally. And what I'm asking is just directionally whether your overall internal enterprise valuation has come down as a net result of everything you've announced in the last 24 hours, or as if it might have stayed about the same? Or maybe I suppose this possibly could have gone up? So is that something you could comment on, it would be very helpful.

[Henrik Poulsen, President & CEO]

Thanks, Sam. It's quite clear that in what we released yesterday, there are obviously a number of moving parts, some being negative, some being positive. I'm not going to give you our own estimate of what is the net present value impact of these -- those different moving parts. It goes without saying that the adjustment to production forecast is a negative number, and we've done our utmost to mitigate it as best possible. But I would rather not start, sort of, extending our view on what the impacts will be because I essentially just end up, sort of, start building new targets. So I think we have already given fairly detailed targets here, and we will stand by those targets. That's also why we came out yesterday to make sure that we give you absolutely full transparency and timely transparency. But I would rather not start digging further into the details. I think we've given you a number of data points, that should give you a pretty good idea as to what the impact [ is here ].

[Samuel James Hugo Arie, Managing Director and Research Analyst]

Yes, okay. I somehow thought that, that might be your answer but then it was worth to try anyway.

[Operator]

Our next question comes from the line of Timothy Ho from Morgan Stanley.

[Timothy Ho, Equity Analyst]

Just a couple of questions for me, building on, I think, some of them have already been asked. So the first one, so this is clearly a very comprehensive review that you've done. Could you provide some more color, please, on whether you intend to -- how often you intend to do these reviews from here on in. And you referred to this as a snapshot, a point in time. How do you think about the, kind of, updates on this snapshot as you go along as well? And the second thing is just on mitigation. I think you briefly alluded to some parts of mitigation. But is there anything that you can say on formation of wind farms in the future? Or can the turbine [ designers ] help you at all to, kind of, help reduce these effects. So yes, anything you can provide on the [ litigation ], that would be very helpful.

[Henrik Poulsen, President & CEO]

Yes. Thanks, Timothy. I mean we -- when we came out with the update on the targets yesterday, yes, it is, of course, a snapshot. I mean, we're operating a very big portfolio of assets and development projects. And we see movements in all of those assets and projects on a daily basis. So we have an almost endless number of variables that will make this a very dynamic picture. We obviously cannot come out updating you every time we see a change somewhere in that portfolio. So we're striking -- trying to strike a balance here. We filed the news yesterday, especially because we have found this -- these adjustments to our production forecast, we felt it was an important piece of news for our shareholders but also for the offshore wind industry more broadly. And we felt that it would be prudent for us to go out and be open and transparent about it. When it comes to the long-term financial targets, it would be, and I say that without making any firm commitments, but I would certainly expect that when we come to the Capital Markets Day in June next year, that is currently being planned, that we would give you an update on a like-for-like basis on the targets that we updated yesterday. We may, at some point, want to move away from those targets. Over time, they will obviously become less relevant as the portfolio progresses. So we'll take a look at giving you an update in June next year, I would expect, and then we may look for some new targets that give you more information. If we remain totally static in our targets, they will obviously become less relevant over time. In terms of mitigation, Timothy, what was the question again? The second question on...

[Marianne Wiinholt, Chief Financial Officer]

Layout.

[Henrik Poulsen, President & CEO]

On the layout, yes.

[Timothy Ho, Equity Analyst]

Yes, just on whether it's layouts or can turbine [ supplies ] helped you in this regard if we look more into the medium term, just a bit more on potential mitigation to try and -- of these load factor potentially in the future or keeping static anyway?

[Henrik Poulsen, President & CEO]

Yes. I mean, you also asked about how we're going to, sort of, use this exercise that would be going through here going forward. And we have now build these new advanced analytics models that will help us more accurately predict the number of the underlying dynamics in the production. And we'll obviously continue to use these models and refine them. And then we are also looking at if there is any learning from this more granular insight into our production numbers that would allow us to mitigate part of the impact by adopting different layouts of the wind farms. We are looking into all kinds of correlations to see if we can identify ways of mitigating some of the impact that we have seen. But that is a still work in progress and too early for me to say whether we can actually mitigate some of the impact by adopting a different design of our wind farms.

[Timothy Ho, Equity Analyst]

And could the turbine supplies help you in this regard at all?

[Henrik Poulsen, President & CEO]

The turbine supplies?

[Timothy Ho, Equity Analyst]

Just in regards to, does the actual design impact things at all here? Or is that kind of a side issue.

[Henrik Poulsen, President & CEO]

The physical design of the nacelle, for instance, again, I couldn't tell you, Timothy. For now, we actually did discuss it the other day. We don't think it's a real factor. If you think about, sort of, the size and the actual design of the nacelle, we think that's a relatively minor thing in these blockage and wake effects. But we'll obviously discuss this with our turbine suppliers.

[Operator]

Our next question comes from the line of Mark Freshney from Crédit Suisse.

[Mark Freshney, Research Analyst]

3 questions, please. Firstly, on the cost out or the 5 -- I think it's DKK 500 million to DKK 600 million cost out in central costs. I guess most of that will relate to offshore wind. My understanding is a lot of the items within there are things like leases, development costs where you haven't taken FID. Surely, reducing those costs by 20% to 25% would have a material impact on your ability to win projects on good returns at the end of [ next ] decade. Could you go through that, firstly? Secondly, on the curtailment issues, I think, particularly things like London Array in Q2. Could you talk about the impact that they've had in Q3? And thirdly, on wake and blockage effects, that's what's within your control which is where you guys have 2 wind farms together. But what if you get other wind farms close to yours, that's something outside of your control. So as the seabed gets more congested, is there a risk that somebody else's return partly comes at your expense? Those are my 3 questions.

[Henrik Poulsen, President & CEO]

Thanks, Mark. When it comes to the cost reductions, the cost reductions come from costs that fall away as we divest Radius and the residential customer business. There are no cost reductions that relate to offshore lease rights or anything like that. These are mostly costs coming out of our staff functions across the company. And as we allocate our corporate overhead into our projects, the projects will, of course, benefit from these corporate overhead reductions. And the majority of our overhead is allocated to the offshore division, given it's done on an hourly allocation [ key ] . So of course, the majority of these overhead cost reductions will ultimately benefit offshore wind. When it comes to the London Array cable repair campaign, it also had an impact in third quarter. It is included in the DKK 150 million impact from outages and curtailments that we referred to earlier today. I cannot give you the specific number for London Array in isolation. That would become a bit too granular, but we do not expect further losses from this cable campaign at London Array moving forward. When it comes to neighboring wake effects from wind farms being built by other developers, we have accounted for that in these models that we have been building. We have added in extensive build-out based on our knowledge of all projects currently in progress, not only our own projects, but also projects from other developers. So we have tried to account for a broader set of wind farms being built around our own sites. So we feel that we have diligently accounted for this.

[Operator]

Our next question comes from the line of Marcus Bellander from Nordea.

[Marcus Bellander, Senior Analyst]

Just one follow-up question on the wake and blockage effects. I believe you mentioned, Henrik, that other observers or players have more optimistic estimates on those effects? And I guess, I'm just curious. And I realize it's hard to quantify, but how certain are you that your model is right and everyone else's models are wrong. And also on the load factor -- your new load factor target, is that your best guess? Or is there a certain element of conservatism included in that?

[Henrik Poulsen, President & CEO]

Marcus, when it comes to these effects and the reason why we think there is a broader issue across the industry is basically not that we've seen any specific estimates for blockage and wake effects, but we've seen other production forecasts for some of our assets. And we've seen also indications of production forecasts on projects where we have been competing in a tender or an auction with some of our competitors. So -- and whenever we've done farm-downs, we have also seen a third-party expert contributing production forecast as part of those farm-down processes. So we do have a number of data points suggesting that traditionally, external experts, other developers have had a tendency to be slightly more optimistic on the total production outlook than us, which is why we believe that we may be looking at an industry-wide slight overestimation on the production forecast. And I, obviously, say that to the point you make, assuming that we had better visibility on this topic than the rest of the industry, and it's obviously, for anyone to judge whether they consider that credible or not. I would say we have more operation data to leverage than any other player in the industry, and we have spent very significant resources into this project to develop what we would expect to be probably the most advanced models for predicting different underlying variables in our production forecast. So on that basis, I would actually expect that we have better visibility and better prediction power than most other players or all other players in the industry. When it comes to the load factor, we now expect it to be around 48%. That is our best estimate at the moment for the 10 projects included in that target.

[Operator]

Our next question comes from the line of Elchin Mammadov from Bloomberg Intelligence.

[Elchin Mammadov, Utilities Analyst]

I have 3 questions, please. The first one is on your guidance. You confirmed you recently increased DKK 16 billion to DKK 17 billion EBITDA guidance. However, at the current run rate, you're set to achieve about DKK 19.5 billion. So my question is, what kind of headwinds are you expecting in 4Q? Or what fewer net positive one-offs are you expecting in 4Q that you haven't had in the first 9 months? The second question is on the farm-downs. I mean, besides the Ocean Wind, what other assets are you looking to include in your farm-down policy? If you can update on that, that will be great. And the final one is on Poland. I mean, it's early days, but assuming you do sign your partnership with PGE in Poland, given it's a 50-50 joint venture, what will be your role? And what will be PGE's? So for instance, will you be the one to build it and PGE to operate it? Or the other way around. So that will be really useful.

[Marianne Wiinholt, Chief Financial Officer]

Yes. Thank you for your questions. On the first question on the guidance. The reason why we have a front end-loaded result development, you can say, is that all our farm-down gains more or less are coming in the first 3 quarters. We are now very close to being complete with Hornsea 1. And therefore, you will not have this farm-down gain in Q4. So that's the reason for that. On the -- your second question is on farm-downs. Yes, you are right. We are working on Ocean Wind. And then it's also Taiwan, where we are also out, we are in a process right now. And we have said that we expect it to farm down probably next year in Taiwan. But aside from that, we don't have any active processes on the farm-down side.

[Henrik Poulsen, President & CEO]

And when it comes to Poland, the exact roll split between PGE and Ørsted is still part of the ongoing discussions, I cannot give you a specific answer. We are, obviously, bringing together quite complementary capabilities and experiences that we would tap into, but the exact detailed [ role ] split is still to be finally settled.

[Operator]

Our next question comes from the line of Klaus Kehl from Nykredit.

[Klaus Kehl, Chief Analyst]

Two questions from my side. First of all, if I look at your onshore business, you have had a very strong performance here in -- both in Q3 and in -- for the first 9 months. And especially, if I look at -- on the EBITDA line, then I guess you have had an EBITDA margin of above 100%. I guess, a big part of that is due to the PTC subsidies that you have received. But could you give us any kind of indication of what would be a reasonable long-term EBITDA margin for this business? That would be my first question. Secondly, you have had these curtailment problems throughout '19. And I guess, the full year figure is now at around DKK 350 million. Could you just confirm what the number is for the full year? And secondly, would it then be reasonable to expect that your EBITDA would go up by a similar number next year?

[Marianne Wiinholt, Chief Financial Officer]

Yes, on your first question on onshore results. Yes, you are right. We have had a very strong Q3, driven by these benefits we have had from the high [ peak ] prices in Texas in August. The 2 other quarters has been very much as expected. If you look at EBITDA margin in this business. It's not really a meaningful number because you are right, the reason why we have such a high margin is these PTC, the way we account for them. So just to say that it's not really the right way of -- right metric for that business. Then on curtailment, yes...

[Henrik Poulsen, President & CEO]

On curtailment class, I mean, at half year, I believe we indicated a DKK 400 million impact from outages and curtailments. And we have set DKK 150 million for third quarter so amounting to DKK 550 million impact for the first 9 months of the year. We've also said that we expect fourth quarter to be better given that we have resolved many of these issues at London Array, Race Bank and Borkum Riffgrund 2. So there will probably still be some impact. There always tend to be some impact from outages and curtailments in the portfolio, but we'd expect them to trend down in Q4. Looking into next year, there will, again, always be some impact from outages and curtailments. But we would hopefully be back to what we consider a more normal level, which hopefully also should be below what we have seen here in 2019.

[Operator]

Our next question comes from the line of Iain Turner from Exane.

[Iain Stewart Turner, Analyst of Utilities]

Can I just ask you a couple of questions? On the GE turbines that you've chosen for the U.S. Mid-Atlantic projects. Is that purely on the basis of cost? Or are you acting strategically and trying to bring on a third scale suppliers as, I think, you've done in the past, a strategic purchaser? And then secondly, on the Polish projects, could you just outline the attractiveness of the Baltic compared with the North Sea? I mean, is it windier? Is it shallower? How does it stack up?

[Henrik Poulsen, President & CEO]

Yes. Thanks, Iain. I mean, we have certainly picked the GE turbine on the basis of its cost competitiveness, which obviously both includes the cost of the turbine and the productivity of the turbine. And with that said, it has never been a secret that we do believe it would be beneficial for the whole industry, if we could introduce and ramp up a third significant supplier of offshore wind turbines. And with GE being the preferred supplier for our Mid-Atlantic cluster, and as I understand, also for the Dogger Bank cluster coming out of U.K., Round 3, it seems like they have built a meaningful ramp-up volume for the 12-megawatt turbine. When it comes to the Polish projects, they have very good site conditions. We are actually quite positive on these sites in the Baltic Sea in terms of distance to shore, seabed conditions, water depth and wind speeds. So overall, we consider them high-quality sites.

[Operator]

Our next question comes from the line of Peter Bisztyga from Bank of America.

[Peter Andrew Bisztyga, Head of Pan]

Just a couple of follow-up questions. You've mentioned that having more prudent load factor assumptions could be a disadvantage in auctions. I'm just wondering whether there actually could be any rationale for sharing your analytics capability with other developers, perhaps even as a service going forward to make sure that everybody has equally advantageous information?And then just another quick one. Do you know -- happen to know how close together, separate wind farms need to be before wake effects from one project start to impact another one?

[Henrik Poulsen, President & CEO]

Thanks, Peter. On the sharing of the insights that we have generated, it is certainly something that we will discuss and consider how we can best share them with the broader industry. I do believe that it would make sense for us to find ways of sharing this information. That is also why we came out yesterday to make sure that we are transparent and also that we make sure that whoever can benefit from these insights should be allowed to. In terms of neighboring wake effects, I can't give you a very specific sort of threshold for when you start seeing these wake effects from a neighboring wind farm, but you'll probably be surprised how far out, it actually has an effect. Let me put it this way. When you get sort of within 25 kilometers, it actually starts to have an impact. We can detect an impact further out even beyond 25 kilometers, but then it begins to be a very marginal and small, almost negligible impact. But as you move within that 25-kilometer boundary and down to sort of 10, 15, 20 kilometers, it begins to have a real impact.

[Operator]

[Operator Instructions] And our next question comes from the line of Alberto Gandolfi from Goldman Sachs.

[Alberto Gandolfi, Head of European Utilities Research]

Sorry I joined a bit later. So if questions have been asked, just please feel free just to tell me. I'm going to read the transcript. But I have 2 follow-ups. One is just trying to understand, on -- in light of what you said yesterday, what should be broadly the [ actual ] earnings, let's say, EBITDA impact. So you talk about -- you gave 48%, 50%. So let's assume this rate should have been in the middle at 49%. Now you say 48%. Am I right in saying that by 2024, that, in theory, is less than 1 terawatt hour impact, which is about DKK 700 million. And on top of that, you're talking about DKK 500 million, DKK 600 million cost savings. So on the back of that, I would say, your EBITDA will barely move. And so I'm a bit surprised that there was such a big deal in the communication about it, number one. And number two is, you really -- I felt yesterday, most investors really focused on not just on this, but what you said about the returns. The perception was the leading offshore developer is cutting the return assumption. Now when you get 7.5% to 8.5% originally, the interest rate environment was very different. So I mean, my point would be, when we think about WACC plus spread, has the spread changed at all, besides these few million Danish krone, I would say, on volumes or not? Because [ eventually ] , I suspect, otherwise, we are blowing out of proportion this issue. The second point is, now that you think you have superior analytics, does it mean that you need to put more space in between each single turbine? Does it mean that each seabed you've already leased will be able to host fewer turbines to maximize the volume? Does it just mean -- how can you basically circumvent this problem, not what's under construction? Because, I guess, the design is what it is, but maybe on the future ones.

[Henrik Poulsen, President & CEO]

Thank you, Alberto. When it comes to the EBITDA impact, we didn't change the guidance, obviously. We kept the target at an average EBITDA growth from our operating sites of around 20% over that period from '17 through '23, which obviously indicates that big picture, the negative and the positives offset each other to an extent where there was no reason for changing the EBITDA outlook. So our EBITDA expectation as such is unchanged. I haven't done the math on exactly what a 1 percentage point load factor drop would translate into in terms of EBITDA at a future point. But obviously, it all comes back to what I said earlier, which is that when you take the bigger picture of positives and negatives, we are not changing our EBITDA outlook, and we continue to have strong visibility on that EBITDA growth. When it comes to the returns, and we take it down by 50 basis points, as I said earlier, it is not -- least driven by the fact that within that bucket of 7 projects, you have a couple of very targeted negative impacts on the Changhua project and Revolution Wind, which is why that particular target is adjusted, which is also one of the reasons why we come out yesterday. And you can rightfully say, it's a lot of fuss for relatively marginal adjustment. I will leave that assessment to the market. But when we put a target into the market as we did at the Capital Markets Day, and we see it changing for one reason or the other, we feel it's incumbent upon us to come out and be forthright and transparent about it, and then we leave it for the market to assess it. If you look at the declining interest rates, you're absolutely right. They have come down probably by magnitude of 100 basis points since the Capital Markets Day. So if we had been guiding on our spread between returns and cost of capital, we probably wouldn't have changed anything. So in that respect, you're right, but that was not how we set the target back at the Capital Markets Day. In hindsight, maybe we should have, but we didn't. So that's why we're now...

[Alberto Gandolfi, Head of European Utilities Research]

Could you allow me?

[Henrik Poulsen, President & CEO]

Yes, please.

[Alberto Gandolfi, Head of European Utilities Research]

Forgive me, this is very clear. If you allow me to follow up. Are you saying that the main reason for cutting the return is down to interest rates?

[Henrik Poulsen, President & CEO]

No. No, no. We are counting the returns because the net impact of the positive and negative developments, including the production forecast adjustment leads us to actually lower our return expectation for that portfolio of 7 projects.

[Alberto Gandolfi, Head of European Utilities Research]

Okay. So it's Taiwan and Revolution Wind. But going forward -- going forward, besides those 2 projects, you would say the impact on any other developments, are we right in saying in terms of spreads is really negligible?

[Henrik Poulsen, President & CEO]

When you look at the spreads, and you take into account that interest rates have come down between the Capital Markets Day and today, the impact on the spread would be negligible. Yes, that would be a fair conclusion. When it comes to superior analytics, Alberto, I mean, again, it's too early for us to say exactly how this may impact the way we design future wind farms. That is going to be a key part of the effort as we move forward. So whether we will actually move to a different layout, whether we will use different spacing between the position, it all remains to be seen.

[Operator]

Thank you. And as we have no more questions registered, I now hand back to our speakers for any closing comments.

[Henrik Poulsen, President & CEO]

Thank you so much, everyone, for joining. Thank you for a lot of very good questions, and thank you for the interest in the company. Should you have any further questions, as always, please reach out to our investor relations. Have a continued good day.