Welcome everyone to our presentation of Hydro's third quarter results, welcome also to all of you following us on the webcast today. The results will be presented by Hilde Aasheim, our CEO and CFO, Pål Kildemo. After that, we will have time for Q&As, both from the people that are in Oslo and also from webcast. With that, I'll give the word to you, Hilde.
Thank you, Inger. Good morning to all of you here in Oslo. Good morning to you that is following us on webcast. Let me say that I'm very happy to stand here when all the embargoes have been lifted in Brazil, which means that we can come back to normal operation of our Brazilian assets. Our third quarter are positively influenced by the fact that we have started the ramping up, as well as that cost is coming down in the upstream part of Hydro. The quarter is also impacted by indeed a volatile business environment, with increased uncertainty and low visibility, resulting in reduced demand for our products and solutions in the quarter, but also reduced expectations going forward in terms of demand.
That is why we forcefully now work on the improvement program that was launched on the Investor Day, as well as the forceful restructuring and optimization that we're doing, both in Rolled Products as well as in Extruded Solutions. Let me go to the highlights for the third quarter. The underlying EBIT for third quarter was NOK 1,366,000 ,000 down from NOK 2.7 billion in third quarter last year, but up from NOK 875 million in Q2 this year. As I said, I'm very happy that the last embargo was lifted on September 26 on the DRS2, the new deposit.
After 19 months of embargo, we can now start on activities on the DRS2, which is, as we have said many times, is the only long-term solution for Alunorte using the new press filter and using the new deposit and close the old one. When it comes to the result compared to last year, the results are down on lower aluminum prices, 20% lower aluminum prices and 30% lower alumina prices. The positive side is that, as I've said, we have started to ramp up in Brazil, and we are also seeing that cost is coming down in the upstream part of Hydro.
On our Investor Day on September 23rd, I announced an improvement program of NOK 6.4 billion to be realized by 2023, and I'm happy to stand here today and say that we are on track to deliver the first NOK 500 million by the end of this year. When it comes to the market, we see increased uncertainty, and we also see demand expectation going forward to be lower than what we had expected in the previous quarters. We are now estimating a flat global primary aluminum demand compared to 2018 to 2019 towards 2018, down from an estimation of 1%-2% as we guided on second quarter, but more in line with what we said at the Investor Day.
We are also now seeing reduced demand expectation in the downstream area, in particular in the extrusion, where we now expect a negative demand growth 2019 compared to 2018. I will come back to that. Before I go to the market, let's dwell a little bit again with Alunorte and Brazil. As I said, on the 26th of September, we got the last embargo, the criminal embargo on the DRS2. We have now started the activities when it comes to installation and commissioning of the DRS2 which is an important milestone in terms of coming back to normal in Brazil. I'm also happy to say that we have had a successful third quarter when it comes to ramping up in Alunorte.
We have been producing at a level of 83% in terms of the capacity utilization, which is showing that we are progressing when it comes to using the eight press filters. I've said many times, it's a new technology. It was stopped while we were in commissioning phase. We are learning how to operate the eight press filters, but we still guide on 75%-85% for the fourth quarter because we are working on the cycling time. We are working on what kind of intensity we need in terms of maintenance. Third quarter was a good sign that we are progressing in a good way. As mentioned before, we are waiting now for the ninth press filter. I was in Brazil two weeks ago, and I saw that the press filter was inside the big building together with the eight press filters.
We are commissioning the ninth press filter now in November and should be up at full speed at the end of the year, adding 10% in terms of capacity utilization. If you sum up the numbers, we are saying that at the end of the year or the beginning of the year, we should be in the range of 85%-95%, and then further optimization should make us come to 100% capacity utilization during 2020, at full capacity during the whole year of 2021. What is as important with the ramp-up is that we continue the good dialogue with the local communities. When I was in Brazil, I was meeting several of the community representatives.
To see myself the active dialogue, the constructive dialogue that we now have with the local community was very good for me to see that we are also progressing in this area. I also told them that we are fully committed to deliver on the technical as well as the social agreements that we have put forward to them. In terms of Brazil, I hope that we can continue to talk positively about the development going forward. Brazil is extremely important to us. To the improvement program. As many of you hopefully listened to me on the Investor Day on September the 23rd, we announced a full-fledged improvement program throughout the whole company, in all business areas, in all staff and support areas, within procurement, throughout the whole company. NOK 6.4 billion to be delivered by 2023.
A front-loaded program, NOK 5.5 to be delivered by 2021, and NOK 500 to be delivered by this year. The improvement program is about operational excellence, it's about fixed costs, it's about commercial excellence, it's also about restructuring. I have to say that I'm quite impressed about how the whole organization has mobilized through these few months in terms of working on what we can influence ourselves in a volatile environment. I'm happy to stand here to confirm that we will deliver the NOK 500 million by the end of the year. We do expect that the contribution from Extruded Solutions and the improvement program for Extruded Solutions, which is an EBIT improvement program, will be somewhat behind plan for 2019, reflecting the market headwind that we see now, even though we do have forceful restructuring activities, which I will come back to, that should mitigate the risk going forward.
In 2019, they might be somewhat behind. The good thing is that we have been ramping up faster in Brazil, so that sort of make up for the difference, and that is why I stand here today saying that we are on track to deliver the NOK 500 million by the end of the year. Let me go to the market and let me start by the third quarter. In third quarter we saw a flat development in terms of primary demand, Q3 versus Q3 last year. 2% growth in China, but negative 2% growth in world ex-China. On the supply side, we saw a decrease of 1% because we have seen production disruption in China. Hongqiao experienced a flood in the Shandong area, and there was explosions in two smelters, one in northwest, the Xinfa plant, and also one in Inner Mongolia, taking down the supply.
Which ends up with a deficit of around 500,000 tons in Q3. The major part in China, 400,000 tons in China, and more or less in balance in the world ex-China. To the right, you see the 12 months rolling, which is still in deficit in the tune of 1.3 million tons. When we look at the global demand for the whole year, as I said, we are expecting a flat development, 0.5 negative to 0.5 positive, which is down from the 1%-2% growth number we guided on in Q2, but it's more in line with what is said at the Investor Day. Let's look at the supply-demand balance for 2019. We still believe in a deficit for the whole year of around 1 million-1.5 million tons.
If we look at the composition of that, if we look to the world ex-China, we see now that we are in the negative territory, 1%-2% negative growth in the Western World ex-China. We see continued weaknesses in Europe, and we also see now weaknesses in the U.S. The growth is negative in the 1%-2%, while we see the supply growth of 2%-3%, more or less the same as we said in Q2. The reason for the supply growth is Alba coming up to full speed, and then also the restarts in the U.S. The deficit that we have seen in the Western World is still there, but less than what we have seen in the past. When it comes to China, we expect a 1%-2% growth, slightly down from 1%-3% last quarter.
We see the disruption that I just talked about, the negative supply growth of 1%-2% because of the disruption in the flooding and the explosions in these three plants that I mentioned. China is also in deficit for 2019, making up for the total of 1 million-1.5 million tons in deficit for 2019. Let's look into 2020. Here we see a slightly different picture. We see more or less a balanced or even a small surplus in 2020. We believe that the world ex-China will have a flat development, -1 to +1, flat development going into 2020. While we still see supply growth coming from Alba, coming from Brazil, coming from Canada, which makes the world like China more or less in balance. We see China expecting a slight tick-up on growth, not much, 1%-3%.
We see a supply growth of 4%-6% in China. Here we expect that the plants that was disrupted are coming back. We also hear about new greenfield plants in China, in several regions. Here we should be a little bit cautious because we believe that if demand is at a low level as we see here, we believe that also the growth numbers will come down because we don't believe that China will build up a big surplus in the environment we see now. As I said, the visibility is low, but that's what we expect based on our intelligence. We see that the market is more in balance or a slight surplus than what we have seen in the past with deficits. We normally follow the inventories very closely, and we do that also for this quarter.
We see that the inventories are now slightly below 11 million tons, approaching 60 days of inventories. We saw in the financial crisis, we were down to 50, which we considered to be a tight market. Now we are at 60. Inventories have also come down in Q3, but in China, because China was the one that was in deficit, while more in balance in the Western world. If we look to the right, we see the export from China, which is extremely important to follow, and we have seen growth year-on-year of export from China. If you compare third quarter 2019 to third quarter 2018, export has come down with 10%.
If you look at the export for the year of 2019, we have seen an increase of 3%, primarily in the first part of the year because since May, the arbitrage that we have observed has not incentivized that much export. The arbitrage it's more in balance, the Shanghai price and the LME, so we haven't seen that much export since May when the arbitrage is more in balance. Obviously for the aluminum industry to follow both the growth rate and the export in China is extremely important to us. When it comes to LME, we have seen a quite flat development when it comes to the LME development from Q2 to Q3. The average LME in Q2 was 1,818, while in Q3 it was 1,879.
Throughout the quarter, we have seen LME moving from $1,800 in the beginning of the quarter down to $1,720 at the end of the quarter. Yesterday it traded at $1,735. The fact that the Norwegian krone has depreciated to the U.S. dollar, we see improved LME in Norwegian krone, which is obviously a positive sign for part of the primary portfolio. We have also included on to the right here the development when it comes to the ingot premium. We see the black curve is the standard ingot duty paid, which is quite flat. We see the extrusion ingot premium, which has come down during the last year. The green is the spread between the standard ingot premium and the product premium for extrusion. Here I would like to remind the audience that in Hydro, we have very little production of standard ingot.
We produce so-called value added. If you look at our premium development, you will see that the premium development has been reduced quite significantly compared to if you just look at the standard ingot premium. This has an effect on the primary metal result, but also when we look into the metal market result with the remelters, as well as the remelters in extrusion ingot, in Extruded Solutions. To the demand for downstream products. Here we guide towards a negative growth. The extrusion demand is very much correlated to the macroeconomic development. When we talk about weak demand in Europe, weak demand in the U.S., this hits also the demand for extrusion ingots. Compared to what we guided on in Q2, we have taken down the expectation for extrusion ingots quite significantly. We are now in the negative territory.
If we look to the left here, we show the expected growth rate for extruded ingots compared from 2019 to 2018. We look at the two regions, North America and Europe, and we look into the different market segments. If I start with North America, here we guided on a growth rate of 1% in Q2. Now we are more in the range of 3% negative. There are two main parts. In the building and construction area, we believe that demand would pick up after summer. It has remained flat and now it has turned into negative territory. When it comes to the transport and automotive, in particular the truck and trailers, has moved from positive growth to negative growth, which makes the main composition of the fact that we are now expecting a negative growth number of 3% for Extrusion North America.
When it comes to Europe, we guide on an expected growth rate of -3%, while we had flat and 0 in Q2. Here we see the transport sector. We have heard a lot about the automotive industry in Europe, in particular in Germany. We also see the industrial segment, in particular in Germany, is now in negative growth. That is the main composition of the fact that we are guiding on negative growth in Europe for the whole year, 2019 versus 2018. To the right, we see how the numbers look like at Q3 level, Q3 versus Q3 last year. We are already now in the negative of 2%-3%.
Even though the visibility is low in Extrusion because it's the nature of the business, it's quite short-term, we guide now on the quarter-by-quarter comparison to 2018, a negative development in Europe of 8%, a negative growth rate of 8% compared to last year, while we guide on the 4% negative in North America for Q4. What should we expect for 2020? That's not easy to guide on. As I said, for the global demand growth, we expect a flat development in 2020 versus 2019. As I said, in the Extrusion business, it's low visibility. It's short-term in terms of the nature of the business. It's hard to guide how this would look like in the next quarter going forward. What we can do is to address this situation, and that is why we do rather forcefully.
We already in Q2, we saw a softer demand. That was why the management of Extruded Solutions started to look at their portfolio. We started with Europe, looking into how we could optimize the portfolio, how we could plan for a weaker market sentiment, but also to reduce costs. On Investor Day, we went through in more details what we have been doing in Europe. We have decided to fully close some plants. We have taken down warehouses in Europe. We have also divested a plant in Romania. When we saw that the market is deteriorating further, we also took a closer look in the U.S. We have now decided to fully close also two plants in the U.S.
These are our measures in order to be ahead of the situation in terms of optimizing our portfolio, in terms of reducing costs, and sort of compensate for a weak market. We have also now introduced and announced a full review of the SG&A costs throughout the whole Extruded Solutions in order also to take out costs in all parts of the organization, because the measures we take is not to be weaker in the market. We would still focus on being strong in the local market. This is a local business. To optimize and to reduce cost, to be even more robust, which also feeds into the improvement program that Extruded Solutions have as part of the 6.4 to reduce or to improve the EBIT by NOK 1 billion within the next three years. Doing this restructuring, there's also some costs following that.
In Q2, we have already taken NOK 228 million, we are booking another NOK 156 million in third quarter to cater for these restructuring, these closures, and these sales. Move to Rolled Products. In Rolled Products, transport and packaging are holding up, while we also see in Rolled Products that the demand is weakening in general. Here I use the same format as I had on Extrusion. To the left, we see the expected growth rates for North America and Europe in the different market segments compared to 2018. Here we see that we have taken down compared to second quarter, taking down the growth rate in North America from 4%-2%, primarily due to building and construction and the industrial segment, while transport is holding up. Automotive is holding up mainly due to the substitution in the automotive business in the U.S.
In Europe, we have taken down the growth rate from 1% to 0.5%. Here we see that packaging are holding up, while we see in Europe that the automotive is struggling, and also the industrial segment, particularly in Germany. More or less the same development in the Rolled Products as we saw in Extrusion, but not with the same negative growth rate. Yeah. To the right we see also the same development in the sense that here we have a strong Q3 versus Q3 last year, especially in North America, but more we see a flat development in the fourth quarter. The good spot in this market is the can market. The can market is growing on substitution. We've seen that the sustainability trend, looking for materials that can be recycled, is supporting can versus plastic.
We have heard several of the beverage companies and can producers now substituting the plastic towards using aluminum cans, which is good for our business. We expect that the can sheet demand in Europe will grow by 5%-6% in 2019, and 6%-7% in 2020. This is good news also for Rolled Products in Hydro. As you heard from the Investor Day, that we are moving our production from the low margin segments, the foil as well as the litho, towards automotive and can. Whenever we look into Rolled Products volume development year-on-year, it has been a rather flat development, but we have increased our production towards the can segment of 4%. We are targeting now to increase deliveries to can by 20% by 2023, which is part of the restructuring ongoing in our Rolled Product area.
Talking about the restructuring that is ongoing in Rolled Products. You heard from our Investor Day that we have taken quite a look at our Rolled Product business. We put it under strategic review as I announced on my first day as CEO. On the Investor Day, we announced an improvement program of NOK 900 million within Rolled Products to be delivered by 2023, and also release of operating capital of NOK 900 million by 2021. We are forcefully working on that improvement program. We have announced a shutdown of the mainline, and we are working on heavily manning reductions, especially in Grevenbroich as well as improvements in all parts of the business in terms of operational excellence, how we go to market, and also this shift from the more attractive segments, automotive and can, while leaving or reducing our exposure in the foil and litho business.
As part of this restructuring, we are taking down manning, we are taking down lines, we have booked for the third quarter restructuring and impairment costs of NOK 1,145 million of costs, which is not part of the underlying result. Very much in line with what we guided on at the Investor Day. While we are working on improving what we can influence within the Rolled Products business area, we are continuing the strategic review, looking for more value-creating strategic opportunities. This could mean also different scenarios for ownership. I know that you're all interested to hear more about that, we will revert to you with more information in due course. Moving to alumina. We have seen that the alumina prices has trended down in Q3 from in the range of $300 to $290.
On the backdrop of the fact that Alunorte is ramping up, and also the EMAL refinery in Abu Dhabi coming on stream. We see now that the alumina market is in surplus, and we have seen that the Chinese have started to import alumina, which is different from what they did when we had the peaks of alumina prices when China was exporting. That was an extraordinary situation rather than a normal situation. Alumina prices in China has gone up. Several alumina refineries has closed down, and that has opened the arbitrage or that has opened, let's say, the opportunity for the Chinese to import alumina. To the cost levels upstream. The upper level is the major cost elements, the raw material costs in the primary part, and to the lower end is the raw material cost for alumina.
We have seen cost reductions or price reductions in most of our raw materials costs since 2018. In petcoke, in pitch, and we now see also in alumina, prices has come down quarter by quarter, and we still see reductions coming down from second quarter to third quarter, and we even expect for the primary part to see also prices coming further down in Q4. On the alumina side, we have seen the same picture when it comes to caustic soda, while also fuel oil costs and steam coal has gone down, now more flattening out towards the fourth quarter.
Here I would like to remind the audience that there is a lag over one quarter before we see this in our numbers. We definitely see these costs coming down now in the upstream part, and we also expect that it will come further down in the Q4 for the primary part. More uncertain further reduction in bauxite alumina related to the fuel oil and to the steam coal. Obviously all of these input factors are globally traded, so this will impact the overall aluminum industry as well. If you look at the cost in the upstream part, cost has come down. To the left here you see the alumina cost coming down to 253, all-in implied, all-in alumina cost to $253 per ton, now approaching the fourth quarter 2017 level, which was 265 at the time.
This is very much influenced by the ramp-up of Alunorte, but also lower raw material costs as well as the lower sourcing cost of external alumina that we saw at very high levels during the embargo period. The margin though, in bauxite alumina has not gone up simply because the alumina price has gone down. If we look to the right of the all-in implied primary cost has definitely come down. We are now seeing an all implied cost of NOK 1,900, which at Q4 in 2017 was NOK 1,850. As I said, we still have lags to improve further in Q4. We see that the main contributor is the alumina cost, but also the fact that we have been ramping up and also the other raw material cost has improved as well as the continuous effort to reduce cost in general.
In the primary part, we see that the margin has ticked up, simply because the cost has come more down than the aluminum price has come down. A few comments on the Energy results, very much influenced by Nordic power prices coming down. It was quite at extreme levels in the Q3 last year. Now we see more normal levels simply because the hydrological balance is at normal levels. We also saw in Q3 that the fuel cost or fuel prices on the continent has come down. That explains sort of the main difference between Q3 this year versus Q3 last year. We saw that the hydrological balance in Q3 last year was seven terawatts lower than normal, while we see now for the Q3, we are more in a normal situation.
I would like to end my part of the presentation on a positive note relating to the demand for greener aluminum. We have talked about in this audience before that it's really the high-end players in the automotive and packaging, as well as in the building and construction industry that are leading the way in terms of demanding greener aluminum. Audi is a good example, it's a good partner for us. Now we deliver certified aluminum to the battery housing in the Audi e-tron, which is an exciting joint effort together with Audi. When we said certified, this is certified according to the Aluminum Stewardship Initiative, which is setting the standard when it comes to responsible and sustainable production of aluminum.
Here, we benefit from the fact that we are in the long value chain of aluminum, that we are able to produce in a responsible way, all the way from the mine to the fabricated product. In this sense, Hydro has the longest sort of unbroken chain of this stewardship certificate, for the fact that we are in that long value chain. Our Alunorte is certified, the primary smelters are certified, and the fabrication is also certified now at 19 different Hydro plants in eight countries. Several more are in the progress. This is what we talked about at the Investor Day, that our sustainability agenda is the platform for commercializing also our products in the market, meeting the demand of these advanced customers that are now looking for greener products in their products. That we find very exciting.
We see that these brands now are getting traction in the market. When we talk about negative growth in the market segment, it's good to have demands in these areas, which we can build on going forward. On that note, I leave the floor to our new CFO. He will do his first quarterly presentation. I leave the floor to you, Pål.
Thank you, Hilde. Good morning, everyone. I will now take you through Hydro's financial results in a bit more detail. Let me start with the high-level result development compared to the same quarter last year. In the third quarter of 2019, we delivered an underlying result before financial income items and tax of NOK 1.4 billion. This is down from NOK 2.7 billion in the Q3 of 2018. Most of the reduction in the result is explained by the decline in both the realized aluminum and alumina prices.
We have 18% lower realized all-in aluminum prices in dollars, which took the results down by NOK 2 billion. The alumina price is 33% lower, which reduces the results by a further NOK 1.6 billion. On the positive side, volumes in our upstream divisions increased, reflecting the ongoing ramp-up in Brazil, contributing positively with NOK 1.2 billion, mostly due to higher bauxite and alumina production, but some small effects from Albras also. This ramp-up also has positive scale effect on fixed costs of around NOK 0.2 billion. At the same time, falling market prices resulted in lower raw material costs, mainly for alumina, but also for caustic soda and carbon through pitch and pet coke. Overall, lower upstream costs added NOK 1.5 billion to our results compared to last period.
Stronger dollars also had a positive effect of 0.5 billion NOK given our cost base in NOK and BRL. Finally, we also had several other effects, which when compared to the positive effects last year, resulted in a negative deviation of 900 million NOK. This includes lower results in energy of 400 million NOK, lower earnings on sales of excess power in Brazil of 400 million NOK, as well as negative eliminations of another 400 million NOK delta quarter-on-quarter. If we compare the results to the previous quarter, which is more relevant from an upstream part due to our downstream divisions being driven by seasonality and upstream NOK, then this result improved by around NOK 500 million, from NOK 0.9 billion in the second quarter to NOK 1.4 billion in the third quarter of 2019. The main drivers are similar to what we saw on a year-on-year basis.
We have 4% lower aluminum prices. We have 15% lower alumina prices, which together impacts result negatively by NOK 0.5 billion and NOK 0.7 billion consecutively. The ongoing ramp-up in Alunorte, Paragominas, and Albras continued to contribute positively to the results with around NOK 800 million due to higher volumes, with around half or around NOK 0.5 billion in B&A and around NOK 0.3 billion in primary metal. The upstream results were further supported by the continued relief in raw material costs and some reduction in fixed costs, which gave a total effect of NOK 700 million. Finally, we have a stronger dollar, which supported the results also from Q2 to Q3 with around NOK 300 million. Other effects such as seasonally lower downstream results, negative eliminations, and a positive deviation on Albras power sales compared to a loss of around NOK 200 million in the second quarter have largely netted each other out.
If we then take a look at the key financials for the quarter, then as expected, revenues were around NOK 2.2 billion down compared to Q3, as we have declining prices, which are partly offset by the positive currency effects as well as somewhat higher upstream volumes. This quarter, we excluded from reported EBIT of NOK 222 million, a loss of NOK 1.1 billion, which I will get back to a bit later. With that, the underlying EBIT ended up at NOK 1.366 billion. We have high financial expenses in the quarter of NOK 1.6 billion, and that includes an interest expense of NOK 0.2 billion. The majority is the NOK net foreign exchange loss, mainly unrealized, of NOK 1.4 billion.
This reflects the embedded derivative in power contracts in euro with Norwegian counterparties, which is impacted by the strengthening of the euro versus the NOK or the weakening of the NOK versus the euro. As a result, income before tax was a negative NOK 1.4 billion, significantly down from the NOK 1.6 positive in Q3 last year. We have a marginal tax income of NOK 16 million in the third quarter compared to a large negative income before tax. The marginal tax income of NOK 16 million is due to the high power surtax, as well as certain tax considerations related to tax losses carried forward in Germany. This gives a negative income of around NOK 1.4 billion, down from a positive NOK 0.9 in the same quarter of the last year.
If we also here adjust for the net foreign exchange loss, then underlying net income was positive with NOK 606 million, also significantly down from NOK 1.7 billion of last year. Consequently, underlying EPS was reduced to NOK 0.33 per share from NOK 0.74 per share in last year. If we then get back to the items excluded of around NOK 1.1 billion, then as usual, we have several timing effects that net out to around NOK 43 million, but we also have several one-off effects. The major effect, which is in line with our earlier guidance, is the rationalization and closure cost of NOK 1.1 billion, which is related to the previously announced restructuring efforts in Rolled Products.
This is split between redundancy payments and impairments and cleanup and restructuring efforts, approximately NOK 900 on redundancy and NOK 200 on cleanup and restructuring. In addition, we excluded NOK 156 million in rationalization and impairment charges related to the ongoing restructuring in Extruded Solutions. This comes on top of NOK 228 million, which we took in Q2. We also expect a similar level as we have in Q3 into Q4. For the year as a total, we should end up at around NOK 500 million on restructuring costs in extrusion. For the third quarter, a large part of this is impairment charges of around NOK 95 million. We made a small adjustment to the provisions on the TAC/TC agreements in Brazil of NOK 30 million. We also reduced a provision related to time barring on a historical customs case.
Finally, we had the potential settlement in the U.S., which reduced historical liability to some extent. If we then move over to the detailed business areas, I will start with Bauxite & Alumina. Underlying EBIT for Bauxite & Alumina decreased from NOK 685 million in Q3 to NOK 481 million in Q3 2019. By far, the main reason behind lower results is significantly 30% lower realized alumina prices, driven by both lower PAX and lower LME. The negative price effect in Bauxite & Alumina is around NOK 1.6 billion. At the same time, as the ramp-up is progressing well, there was a positive volume effect due to higher alumina production of around NOK 1 billion. At the same time, higher bauxite volumes and lower production costs in Paragominas also contributed with around NOK 200 million positively.
Production costs per ton at Alunorte also decreased significantly, driven by positive scale effects as well as declining raw material prices. Caustic soda, fuel oil, as well as coal prices all came down in the third quarter. The positive effect from cost relief was around NOK 300 million. If we look into the fourth quarter, we will continue to ramp up production at Alunorte and Paragominas, with a positive effect on both volumes and costs. As Hilde has explained earlier, we expect to run in the range of 75%-85% of the nameplate capacity in the fourth quarter. Overall, we expect a fairly flat cost per ton development versus Q3, while cost levels should come down significantly compared to last year, as we have taken out a bit of the raw material costs as you saw on the chart earlier.
I should remind you that if the current PAX price remains at around the $280 level, then there will be a price-driven downside to our results compared to what we see in the third quarter of 2019. If we move over to primary, it's a bit of the same story. Underlying EBIT for primary decreased by around NOK 900 million, from NOK 861 positive in Q3 '18, to a small negative of NOK 39 in Q3 '19. By far, the main reasons behind these lower results was an 18% decrease in realized all-in prices, taking the results down by around NOK 2 billion. The third quarter last year was positively affected by a large gain of around NOK 450 million on excess power sales in Brazil.
This quarter, we had a marginal effect from Albras power sales, both due to the fact that power prices in Brazil are much lower than what they were in 2018, but also because we have lower power available for sale as we've ramped up Albras. The negative effects that we saw were partly offset, as in B&A, by lower raw material costs and somewhat lower fixed costs. In total, 1 billion NOK for raw material and fixed costs. The main part of this is the significantly lower alumina cost, which contributes with around NOK 0.9 billion, or NOK 900 million of the NOK 1 billion. Finally, a stronger U.S. dollar against the main cost currencies also had a net positive effect of around 300 million NOK.
If we then look into the fourth quarter, now that we have all the electrolysis cells at Albras restarted, we are increasing amperage to lift the production to full capacity. This means that we should have higher production in the fourth quarter than in the third quarter for primary metal. On the price side, we have by the end of September, sold around 55% of our primary aluminum production for Q4 at a level of around $1,765 per ton, which is indicating a somewhat lower level for prices in the fourth quarter. As you know, market prices have been trading in the range between 1,700 and 1,750 over the latter months. On the premium side, we have booked around 60% of our volume at around $325. As a consequence of this, we expect the realized premium in Q4 to come further down, ranging between $250 and $300.
We also look at the decline in prices for raw materials, we continue to expect a cost relief for primary into the fourth quarter, not only quarter-over-quarter, but also year-over-year. Compared to the fourth quarter, we are estimating around NOK 400 million in lower alumina costs. An additional NOK 100 million from lower carbon and energy, given the current market prices. I should also mention that we typically see higher fixed costs into the fourth quarter in primary metal, driven by seasonality. If we then move over to metal markets, metal markets delivered an underlying EBIT of NOK 362 million, significantly up from a marginal result of minus NOK 3 million in Q3 2018.
Remelters had another quarter with a strong performance, mostly driven by higher margins in the U.S., but in addition, metal markets also received an insurance compensation of around NOK 45 million for the fire that we had at the Henderson Remelter earlier in the year. We saw a strong contribution from sourcing and trading activities in this quarter. This is mainly related to the strong contango levels that we've seen on the forward curve, which has a positive effect on our risk mitigation strategy in periods of strong contango. In 2018, we saw this with an opposite effect due to a backwardation in that period. In addition, we had positive currency impacts of NOK 24 million due to a weaker NOK versus the dollar. Last year, the currency effect was negative NOK 81 million.
Excluding the currency effects, the result was NOK 338 million compared to NOK 78 million Q3 last year. When we look into the quarter of Q4 for metal markets, then we are seeing a tightening spread between the extrusion ingot and standard ingot premiums, as Hilde mentioned earlier. Also a tightening spread between the extrusion ingot premium and the scrap prices, which impacts our remelters going forward. Therefore, we do not, at current market prices, expect to see the same strong remelting results into the fourth quarter. I would also like to remind you that our quarterly guiding for metal markets stands at NOK 125 million per quarter, or averagely NOK 500 million per year, but that can be higher and lower, depending on sourcing and trading effects and also currency.
If we move to Rolled Products, results in rolled doubled from NOK 82 million in Q3 '18 to NOK 166 million in Q3 '19. However, if you look at rolling results from the rolling operation, these were stable between the two periods, as higher volumes and positive currency effects were largely offset by inflationary cost pressures in Rolled Products. The improvement we see year-on-year is mainly driven by the Neuss smelter results, which have increased despite lower aluminum prices, mainly due to lower alumina and power costs. Here, as in metal markets, we have also received an insurance compensation of around EUR 2.5 million related to the exhaust duct incident at Neuss earlier this year.
If we look into the fourth quarter for Rolled Products, while demand growth is still positive, we are seeing some softening growth rates in several markets, in addition to the continued margin pressure within some of our key segments, which may impact results for the fourth quarter in Rolled Products negatively. When it comes to the Neuss smelter, remember that, as always, the results are driven by metal prices and raw material price development. Similarly to primary metal, we expect further relief from lower raw material costs, mainly alumina, in the fourth quarter. While the market price for LME so far in the quarter has been trading somewhat below the level that Neuss realized in Q3. For Extruded Solutions, underlying EBIT improved from NOK 497 million in Q3 '18 to NOK 559 million in Q3 '19.
The main reason behind these better results is continued positive margin development in all business units, despite somewhat higher variable costs. We also saw somewhat higher fixed costs due to higher maintenance and labor costs. At the same time, volumes were 8% down, mainly due to softening markets. As guided, the cyber effect for Extrusion in the third quarter is very limited. If we look at the different business units, we see different performance. Results in Extrusion North America continue to improve, driven by higher margins, while results in Extrusion Europe were lower, driven by a decline in most market segments. For Building Systems and Precision Tubing, we see improvements in results compared to what we saw last year.
If we look into the next quarter, then Hilde talked quite a bit about the market for extrusion, and we do expect to see challenging markets with shrinking demand across segments and regions. This will impact the results of Q4 for Extruded Solutions. In addition, Extruded Solutions has remelting capacity, which will be impacted similar to metal markets and primary by the falling margin between extrusion ingot and scrap and standard ingot prices. The final business area is energy, and the underlying EBIT here decreased by 60% compared to Q3 last year, where we had 652 million NOK, and we're now recording 254 million NOK for the third quarter of 2019. The decline in result of around 400 million NOK is explained almost entirely by 20% lower production and 30% lower prices.
Remember also that Q3 last year was a record quarter for energy, with a combination of strong production and prices. As opposed to what we guided earlier, production in Q3 ended up a bit higher than in Q2 as a result of the improved hydrological balance which we have experienced over the Norwegian summer. If we look into Q4, we will continue to optimize production based on price signals, and we expect to move more volumes into the first quarter of next year. This indicates continued low production levels. However, as always, keep in mind that prices and production can change fairly quickly in response to hydrological developments. So far, in the fourth quarter, NO2 prices has increased somewhat from the Q3 levels of 328 and are now averaging around 360 NOK per megawatt hour.
For other eliminations, this netted to -417 in the third quarter, down from -97 in the third quarter last year and -258 in the second quarter of 2019. The other line consists mainly of corporate costs in addition to costs in holding companies and earnings from our industrial insurance. This quarter, we had NOK 160 million, slightly down from NOK 190 million last year and slightly below our usual guidance of NOK 175 million-200 million per quarter. As you remember, in Q2 2019, the level of NOK 253 million was much impacted by the higher costs related to the cyber attack. In Q3, however, we have an opposite effect as we have now received the first insurance compensation or recorded the first insurance compensation for the cyber attack of around NOK 33 million. As we've said before, we have a robust cyber insurance in place.
However, at this time, we are not able to provide more information as to the final compensation amounts and timings. Compensation amounts will be recognized when they are deemed virtually certain after discussions with the insurance companies. Finally, we have quite high eliminations of NOK 257 million in Q3. As you know, this is eliminations of internal profits and losses, and this quarter we have sold more alumina to primary from bauxite and alumina based on own production instead of external sourcing, and thereby we eliminate that profit until the profit is realized out in the market. If we look at the net development since the last quarter, then we have a small decrease in the net debt position, NOK 600 million lower than what we had in the second quarter. We started the third quarter with NOK 15.1 billion in net debt.
We generated underlying EBITDA of NOK 3.5 billion. We had a release of operating capital of around NOK 1 billion in the period. The release here is partly due to declining prices. It also reflects the efforts we are making on reducing safety raw material stocks that we accumulated in 2018 on the back of the Alunorte situation and also the Rusal sanctions. Reducing working capital levels will remain very high on our agenda going forward. We expect a further release as guided on into the fourth quarter due to the seasonal destocking, also very much due to our own improvement efforts. For taxes and other adjustments, they are NOK 900 million in the period. That includes NOK 300 million in tax payments, as well as the reversal of some non-cash effects from EBITDA.
We have cash effects here on provisions and net interest payments, which we should expect to be higher in the periods to come, reflecting the provisions we are taking for restructuring in Rolled. We generated net cash flow from operations at a positive 3.5 billion NOK in the third quarter. We had around 2 billion NOK in investments, we had 1 billion NOK in currency translation effects on cash and debt of around 0.8 billion NOK of the 1 billion NOK. With that, we ended up the third quarter with 14.5 billion NOK in net debt. If we move to adjusted net debt, it increased by 1.7 billion NOK despite the reduction in net debt. The main reason here is the same reason you see across many other companies.
Its increase in net pension liability driven by a significant reduction in the discount rates, mainly in Germany, but also in Norway, giving us a total effect of NOK 1.5 billion increase in net pension liability. We also have an increase in other adjustments with NOK 0.8 billion, and that is reflecting the downstream restructuring provisions after tax, so the Rolled effects that we talked about earlier. With that, the total adjusted net debt, including equity accounting investments at the end of Q3 2019 amounted to NOK 35 billion. We introduced the capital returns dashboard at our Investor Day in September to emphasize our increased focus on cash and returns. We have a somewhat reduction in capital employed in the third quarter of 2019. For the business areas, we have a total level of around NOK 99 billion. You know well the earning story over the last 12 months.
Our rolling ROACE figure is at 2.4%, so we still have a way to go to get up to the 10% over the cycle. We will continue to work on the roadmaps to profitability for Hydro, and for each business area to improve our returns on capital. If we look at our balance sheet and the key ratio of funds from operations to adjusted net debt, then we have seen 24% over the last 12 months, which is below the level we aim to be at over the cycle of 40%, impacted by many of the same elements that impact our returns. If we look at the spot rate of funds from operations to net adjusted debt, given the earnings we're seeing in Q3, we're closer to the 40%, which we target to be at.
We also have a second balance sheet ratio of total debt to equity, this is still well within our guided target. Far in 2019, we've generated NOK 0.7 billion in free cash flow. If we then turn to the measures, then on the improvement program, as Hilde mentioned, we are well on track towards our 2019 target of NOK 0.5 billion. When it comes to net operating capital, we have released NOK 2.3 billion so far in 2019, compared to our target of just NOK 4 billion by the end of 2019, and we expect to see a bigger release into the fourth quarter. On the CapEx side, we have so far spent NOK 6.5 billion compared to our guidance of NOK 10.5 billion for the year as a whole, and this indicates that you have NOK 4 billion to come in Q4.
Remember that we usually do see a higher share of CapEx into the fourth quarter, around 40% at the end of the year. As such, our guidance on CapEx remains unchanged. On that note, I would like to give the word back to Hilde.
Thank you, Pål. Good start. I started my presentation by saying that I was very pleased to the fact that all embargoes are lifted in Brazil. I'm happy that we see costs are coming down in the upstream part, that we in Hydro, we are mobilized for change. We are mobilized for improvement. Our earnings are simply not good enough, that is sort of what is the main theme going forward. That is why our priorities going forward is obviously safe and efficient operation. It's about to mobilize now, in terms of a weaker market, in terms of taking strong measures, compensating measures in order to maneuver in that situation, then to continue to deliver on the improvement program throughout the whole organization, as well as keeping a strict capital discipline in the situation that we're in. We're working hard. We released the NOK 3.8 billion.
That is the target for releasing net working capital at the end of the year. Then it's about differentiating ourselves through low carbon positions. As I said, I'm really impressed about how the organization has responded to this agenda, and we are mobilized for delivering better improvements, better earnings, better returns going forward. With that, we open for questions.
Yes, we do. Questions from the audience here in Oslo for Pål or Hilde. Hans-Erik Jacobsen, yes, please.
Hans- Erik Jacobsen, Nordea. Given your outlook for surplus in the primary aluminum market in 2020, there is certainly a need for a lower production. One of your competitors has announced that they will review as much as 1.5 million tons. Given your position on the cost curve, could you give us an insight to your own view on the necessity of capacity cuts? Are you seeing any other players in the market than the one I mentioned, considering taking out capacity?
I think I will comment, Hans- Erik, on the Hydro portfolio, saying that we have a good position on the cost curve with the primary portfolio as of today. Based on the market we see now, we have already taken down remelt capacity, which is the normal thing we do when we adjust for capacity utilization. At the moment, we don't have any plans to take out primary capacity.
You do agree that capacity on a global basis should be reduced given the outlook for demand?
It's about where you are positioned on the cost curve.
Thanks.
I think just to supplement on that, if you look at, for example, the margin between LME price, raw material prices in different currencies, you see for our Norwegian operations, this margin is at a level which is similar to what we've seen the last 10 years or so. In the third quarter, given spot prices in the market, it's actually quite a bit above, which would affect our Norwegian operations. If you look at the same margin in USD, you see that this is at the lowest level it's been for many years. Without naming names, you can see that the announcements which are coming now are coming from producers which are in a very different currency situation than what we are experiencing.
Thanks.
Thank you. Even Havning, DNB Markets. I have two questions, one upstream and one downstream. Firstly, on Extruded Solutions, is it possible to provide a bit more guidance or color on the outlook for Q4? On your market slide for Q4, you indicate that volumes in Europe and U.S. together are down 5%-6% there. Should we expect the same for your business? Where do you see margins going on the back of this? Will you take out high or low margin capacity? Also, how much did the remelting in Extruded Solutions contribute in Q3? Also on input costs upstream, I can see on your slide that now all input costs are below 2017 levels except for pitch. How should we look upon that compared to your cash cost guidance from the Capital Markets Day, where you say it should be similar to 2017 levels?
Should it be below? Thank you.
To start with the cost question first. I think, as we mentioned at Capital Markets Day, there are no big structural changes to our fixed cost base. There's no big changes to consumption factors of raw material once we get the alumina qualities from Alunorte back into the primary operations. Given that, and if the raw material prices are lower than what you saw in 2017, or taking into account cost inflation, then there's no reason why we shouldn't also move to the lower levels. I guess you see already for bauxite and alumina that we are trending below these levels. For primary, we are not there yet. As I guided on, we expect alumina costs to come quite a bit down in the fourth quarter together with carbon.
On the first one, on extrusion, we are a large player in the markets, both in Europe and North America. We have a significant position. If the market moves a lot, you should expect our volumes also to move with the market. I think what we are doing, and which you have seen already announced through the last couple of quarters, is that we are adjusting capacity and improving how much is produced at each of our facilities to try and react in advance of some of these movements. When we do that, we of course focus on lower margin versus higher margin products.
Okay. Thank you. Any more questions from Oslo? We'll turn to the questions from webcast. Stian?
Question from Daniel Major, UBS. CapEx is tracking below guidance, and you need to spend NOK 4 billion in Q4 to reach it. How realistic is this, and do you see downside risk to full year 2019 guidance?
Well, normally, we see a higher uptick of CapEx coming in fourth quarter. It's sort of the normal sequence. Obviously in the situation where we are in, and also the fact that we have discussed a lot about the capital allocation framework and how to spend our capital going forward, we will have a strict capital discipline. We will have to see at the end of the year how that turns out.
On the cyber claim, can you give any indication on the timing of the expected receipt of the remaining insurance proceeds?
I think I covered that already in the presentation. We are working through it case by case. It depends on the dialogue with insurance companies, et cetera. We will communicate at the time when this is made available to us.
Finally, on Extruded. Given the weak market conditions for Extruded, do you still expect to deliver earnings growth in 2020? Or is a flat outlook more realistic in the current macro environment?
Well, that is what we are working on in terms of initiating now even more restructuring and improvement initiatives in order to avoid not to meet the targets of 2020.
Okay. Any more questions this morning? No? We will say thank you everyone for joining us, and have a great day. Thank you.
Thank you.