Long Term Price Forecast. Steam Coal and Freight

Size: px
Start display at page:

Download "Long Term Price Forecast. Steam Coal and Freight"

Transcription

1 Long Term Price Forecast Steam Coal and Freight December 2016

2 About this report Long Term Price Forecast Steam Coal And Freight is the fifth annual edition of the Long Term Price Forecast, published by Perret Associates. In the production of this report Perret Associates has drawn on its extensive knowledge of the coal and shipping market, accumulated since 2000, as well as its large network of independent analysts, specialising in particular countries and products. In each report and update, we examine those countries and products which we think are particularly relevant at that time. Each quarterly publication contains: Fully updated world seaborne steam coal supply/demand (imports and exports) forecasts for more than 50 countries up to 20 years ahead. Fully recalibrated top-down supply/demand models for China and India. Detailed steam coal demand / exports forecasts for up to 50 countries. Coking coal imports forecasts for China and India up to 20 years ahead. Fully updated cost of production curves for the 6 main exporting countries (Indonesia, Australia, South Africa, Russia, Colombia and the US), based on our comprehensive database of all the main mines in the world. Price forecasts for high quality steam coal FOB Newcastle, FOB Richards Bay, FOB Colombia and DES ARA, with sensitivity analysis. Price forecasts for medium and low quality steam coal FOB Newcastle 5,500, FOB Richards Bay 5,500, FOB Indonesia 5,900 and 4,900, and CFR China 5,500, with sensitivity analysis. Price forecasts for the main dry bulk freight routes: Capes 4TC, Cape C4, Cape C7 and Panamax 4TC, with sensitivity analysis. Section 1 includes all our main conclusions and assumptions, as well as the world supply/demand balance sheet and price forecasts for the various coal and freight products. Section 2 examines the main importing countries. Section 3 examines the main exporting countries. Section 4 examines the dry bulk freight market. About Perret Associates Perret Associates is a London-based consultancy, specialising in the coal, freight, iron ore and steel markets. The company offers a wide range of services in market analysis and forecasting, business development, strategy and risk management. Guillaume Perret, the founder and director of Perret Associates, has been influential in the development of today s coal trading market, having traded the first API2 swaps with a few other pioneers when developing the risk management desk of RWE Trading back in 2000 in London Perret Associates Page 2

3 This report reflects Perret s hands-on trading background, with an emphasis on facts and data. Other Perret Associates publications include: The International Coal Trading Market : An in-depth analysis of the physical and financial coal markets and how they could evolve Trading Iron Ore: An in-depth and independent analysis of the iron ore trading market and how it could evolve (2012, published by Metal Bulletin) The International Coal Trading Market Trading Iron Ore: A full feasibility study for a new derivatives contract (2008, published by Metal Bulletin) Steel futures: A turning point for the industry? (2008, published by Metal Bulletin). The International Coal Trading Market 2007 The International Dry Freight Market 2005 Table of Contents 1 Summary: supply/demand continues to tighten Imports China Slowing GDP growth Coal-fired generation finds some support Rebound in coal-fired generation, despite the cap in new capacity Renewed strength in cement production but for how long? Structural decline in coal production Chinese steam coal imports supported India Slight increase in GDP growth forecast Recent slowdown in total and coal-fired generation Further growth in Indian coal demand Robust demand from other sectors Further growth in domestic crude coal production caps imports Indian steam and coking coal imports forecast Japan South Korea Other countries Taiwan Perret Associates Page 3

4 2.5.2 Thailand Malaysia Vietnam Pakistan Philippines EU 15 imports forecast Germany UK Netherlands Italy Spain Turkey Egypt Other countries: Atlantic Israel Brazil Middle East Jordan 57 3 Exports Indonesia Further decline in exports Domestic demand picks up Structural erosion in Indonesian exports Indonesian cost of production curve Australia Exports erode Diversification and disinvestment in the Australian mining sector Average cost of production stable South Africa Exports stabilise South African cost of production Russia Russian cost of production Colombia Colombian cost of production US: Can a Trump presidency revive the coal industry? US cost curve of production Table of Exhibits Exhibit 1 Seaborne steam coal supply/demand, , m tonnes, November Exhibit 2 Coal inventories, main locations, November 2016, m tonnes Exhibit 3 Coal India ratio, stocks to production, m tonnes, % Exhibit 4 Main coal denominated currencies forecast, Exhibit 5 Brent spot price forecast, $/barrel Exhibit 6 Spot vs. cost of production, main origins*, December Exhibit 7 Base Case FOB price forecast, US$/t, nominal Exhibit 8 FOB NEWC., price forecast, sensitivity analysis, US$/t, nominal Perret Associates Page 4

5 Exhibit 9 FOB R.Bay 6,000 NAR price forecast, sensitivity analysis, $/t, nominal Exhibit 10 FOB P. Bolivar 6,000 NAR price forecast, sensitivity analysis, $/t, nominal Exhibit 11 FOB Russia Murmansk, 6,000 NAR, sensitivity analysis, $/t, nominal Exhibit 12 FOB NEWC., 5,500 NAR, price forecast, sensitivity analysis, $/t, nominal Exhibit 13 FOB R.Bay 5,500 NAR, price forecast, sensitivity analysis, $/t, nominal Exhibit 14 FOB Indo. 5,900 GAR, price forecast, sensitivity analysis, $/t, nominal Exhibit 15 FOB Indo 4,900 NAR, price forecast, sensitivity analysis, $/t, nominal Exhibit 16 CFR China, 5,500 NAR, price forecast, sensitivity analysis, $/t nominal Exhibit 17 Capes 5TC index, price forecast, Sensitivity Analysis, $/day, nominal Exhibit 18 Capes C4 index, price forecast, Sensitivity Analysis, $/t, nominal Exhibit 19 Capes C7 index, price forecast, Sensitivity Analysis, $/tonne, nominal Exhibit 20 Panamax 4TC index, price forecast, Sensitivity Analysis, $/day, nominal Exhibit 21 DES ARA price forecast, based on Colombia, $/tonne, nominal Exhibit 22 DES ARA forecast based on Russia, $/tonne, nominal Exhibit 23 Comparison BCase DES ARA based on Colombian and Russian coal, $/t Exhibit 24 China GDP growth forecast, % Exhibit 25 Coal s share of total Chinese electricity generation Exhibit 26 Chinese coal usage for electricity generation, forecast Exhibit 27 The cement bell curve relationship between GDP and cement usage Exhibit 28 Chinese steam and coking coal demand, forecast Exhibit 29 Chinese crude coal production forecast, steam and coking coal, m tonnes Exhibit 30 Chinese steam and coking coal imports forecast, m tonnes Exhibit 31 India GDP growth, historical and forecasts, % Exhibit 32 Coal and lignite, share in total electricity generation, % Exhibit 33 Indian steam coal demand for electricity generation, forecast, m tonnes Exhibit 34 India, monthly cement production, m tonnes Exhibit 35 Indian steam and coking coal demand, forecast, m tonnes Exhibit 36 Forecast Indian steam and coking coal production, m tonnes Exhibit 37 Indian steam and coking coal imports, m tonnes Exhibit 38 Japanese steam coal imports with split per origin, m tonnes Exhibit 39 South Korean electricity generation per source, TWh Exhibit 40 South Korean steam coal imports, m tonnes Exhibit 41 Thailand electricity generation, per source, TWh Exhibit 42 Thailand steam coal imports, m tonnes Exhibit 43 Pakistan coal and lignite power projects Exhibit 44 EU 15 seaborne imports steam coal imports, m tonnes Exhibit 45 German electricity generation per source, TWh Exhibit 46 Spanish electricity generation by fuel, TWh Exhibit 47 Turkey, electricity generation per source, TWh Exhibit 48 Indonesian steam coal exports, m tonnes Exhibit 49 Indonesian cost curve, FOB, US$/t, 6,000 kcal/kg NAR, December Exhibit 50 Australian steam coal exports with split per destination, m tonnes Exhibit 51 Australian cost curve, FOB, US$/tonne, 6000 kcal/kg, December Exhibit 52 India and Pakistan s share of South African exports, % Exhibit 53 South African exports, split by main destinations, m tonnes Exhibit 54 South African cost curve, FOB, 6000kcal/kg NAR, December Exhibit 55 Russia, total electricity generation, TWh Exhibit 56 Russian cost curve, FOB, US$/t, 6000 kcal/kg NAR, December Perret Associates Page 5

6 Exhibit 57 Colombian exports per main destination, m tonnes Exhibit 58 Colombian cost curve, FOB, 6,000 kcal/kg, December Exhibit 59 Capacity utilisation at US coal-fired power plants, % Exhibit 60 US coal- and gas-fired generation, by month, TWh Exhibit 61 US crude coal production, m tonnes Exhibit 62 US steam coal exports, m tonnes Exhibit 63 US cost curve, FOB, 6000 kcal/kg, $/tonne, December Copyright This is an individual user version of the Long Term Price Forecast Steam Coal and Freight for the exclusive use of the recipient. This document may not be copied, reproduced or redistributed in any form either for internal or external distribution. Disclaimer This document has been compiled by Perret Associates based on its own knowledge and experience. While every attempt has been made to ensure the accuracy of this document, Perret Associates is not liable for any consequences arising from following the suggestions or acting on any information contained herein. Provision of the data and information does not obviate the need to make further appropriate enquiries and inspections. The information is for the use of the recipient only and is not to be used in any document for the purposes of raising finance or financial investment without the written permission of Perret Associates Ltd. The data and graphics contained in this document which have been produced by Perret Associates are not to be used externally without our written agreement. Disclosure Perret Associates also provides investment advice regarding steam coal and iron ore financial instruments to a privately-owned UK-based trading company Perret Associates Page 6

7 1 Summary: supply/demand continues to tighten 1.1 Main findings Our seaborne supply-demand forecast is starting to rebalance, after the big deficit in our previous edition, as the impact of the strong price rally starts to make itself felt. We now expect 2016 to finish with a marginal surplus of 15m tonnes, compared with 7m tonnes in our last report. We also envisage a surplus of 23m tonnes in 2017, compared with a deficit of 68m tonnes in our last report. Nevertheless, we expect this rebalancing to prove short-lived and anticipate that the international steam coal seaborne market will tighten again from 2018 onwards. We still anticipate deficits of 55m tonnes in 2020 (compared with 103m tonnes in our previous edition), 29m tonnes in 2025 (78m tonnes in our previous edition) and 71m tonnes in 2030 (104m tonnes in our previous edition). One of our main assumptions is that the Chinese authorities will revert quickly (in H117 or maybe even in Q117) to their policy of cutting domestic crude coal production. Meanwhile, we forecast that Chinese steam coal demand will remain flat, with more risk on the upside than downside. As a result, after a slight dip in 2017 to 159m tonnes, we expect Chinese steam coal imports to remain in the region of m tonnes until We adjust slightly upwards our forecast for Indian imports in 2016, but lower them thereafter. After a period of stabilisation at just above 100m tonnes during , we expect imports to drop below this level in 2021 and decline further to 54m tonnes by We maintain our imports forecast broadly unchanged from the previous edition for Japan, South Korea, and Taiwan. Imports of emerging countries in India-Pacific, such as Vietnam, Thailand, the Philippines and Pakistan, are increasing strongly, in line with our previous forecast, which we maintain unchanged. However, due to the structural fall in Indian imports and the anticipated short term dip in Chinese imports in 2017, we decrease our 2017 imports forecast for India- Pacific by 62m tonnes compared with our previous forecast, and by 40-45m tonnes per year during the period We also decrease our imports forecast for EU15 countries by 5m tonnes in 2016, with all countries except for Germany and France (due to the nuclear outages) expected to import less Perret Associates Page 7

8 We expect the biggest declines in imports to take place in the UK, Italy, Spain and the Netherlands. Looking further ahead, we decrease further our imports forecast for EU 15 countries by 5-13m tonnes per year until This is partially offset by a further increase in imports by Turkey, Egypt, Morocco, which have so far risen in line with our expectations. Overall, we lower our 2017 imports forecast for EU15 and Mediterranean countries by 4.2m tonnes, 12.2m tonnes in 2025 and 0.5m tonnes in On the supply/exports side, we anticipate a relatively muted reaction from most large mining companies to the recent price rally. We think the situation is very different from 2011, when, after a strong rally, the international mining industry over- invested in greenfield projects, tipping the international market into a 5-year bear trend. For one thing, the recent rally has been too dependent on China s stance regarding managing its domestic supply. The Chinese government has the authority and capability to adjust its domestic supply to changing market conditions much faster than international mining companies. We suspect that many mining companies will be aware of this and remain wary of getting the timing badly wrong. More broadly, the war on coal that has been escalating since 2010 or so is having a significant impact, in particular on coal supply even more than on coal demand. Consequently, it will become increasingly difficult to find investors with sufficient appetite for Greenfield coal projects. Large mining companies are now generating cash again from their coal operations and they might use these funds to diversify their assets, rather than re-invest them all into coal assets. We increase our total exports forecast by 19m tonnes in 2016 and 26m tonnes in 2017, mainly due to a short term response from Indonesian suppliers. However, we expect this increase to stem mainly from a rise in capacity utilisation at existing mines rather than the opening of new mines. After this initial reaction, we forecast Indonesian exports to decline gradually to 310m tonnes by 2020 and 270m tonnes by 2030, due to the increase in domestic demand. The Australian mining market is going through a serious reshuffle in terms of ownership, but overall we forecast a limited increase in exports Perret Associates Page 8

9 We also anticipate a limited increase in South African exports due to increasing domestic demand. Colombian exporters have some spare capacity but the experience of recent years shows that anything more than a 10% rise in yearly in exports is difficult to achieve. The Russian market has been turned upside down in This was initially due to the strong recovery in oil prices, which also triggered the strengthening of the Rouble. Since October 2016, various political developments, such as the US presidential election and the events in Syria are also reinforcing Russia s position on the international scene. We anticipate that the sanctions imposed on Russia due to the Ukrainian crisis might be lifted within the next 24 months, partly because most of the leaders who pushed for them are no longer in power. This could further support the Rouble and therefore the cost of production of Russian coal. We think this is significant, as Russia is becoming the marginal cost supplier for the DES ARA market. We foresee some downside for international coal prices in 2017, although expect this to be relatively limited, at least during the first half. In any event, we are not anticipating the beginning of a long bear trend as was the case in On the contrary, we think that a bottom could be reached within the next 6-9 months, after which international coal prices could resume their long term improvement. Once again, this would be mainly the result of insufficient supply, due to a lack of investment, rather than an increase in demand. There will be a high level of political uncertainty in 2017 but overall we see more risk on the upside than downside. The election of Mr Trump in the US could provide a safety line to the US mining industry with a potential bearish impact on the international market. On the other hand, it could facilitate the development of coal-fired power plants in emerging countries. We forecast dry freight prices to remain subdued until March 2017, after which we expect a gradual recovery to take place Perret Associates Page 9

10 Exhibit 1 Seaborne steam coal supply/demand, , m tonnes, November 2016 Base Case Imports EU Turkey Israel Morocco Egypt Oman Dubai Jordan Yemen Kuwait Syria Total Europe + Med China (net) India Japan South Korea Taiwan Thailand Malaysia Hong Kong Philippines (gross) Vietnam (gross steam coal) Sri Lanka Pakistan Myanmar Other Asia/Indian Ocean Total India-Pacific US Chile Brazil Mexico Others Total Americas Others Total imports Base Case Exports Indonesia Australia South Africa Russia Colombia Vietnam (gross anthracite) Venezuela USA Mozambique Mongolia Madagascar Philippines Canada Norway Store Norkse Others Total exports Supply - demand before stocks Source: Perret Associates 1.2 Other developments Inventories down in China and Europe Our reference for coal inventories worldwide, based on all the locations we monitor, has continued to fall, standing at 417m tonnes as of 31 October This is the lowest level since at least December Our reference for global stock levels has fallen by 32.4m tonnes or 7.2% since our July 2016 update and by 68.5m tonnes or 14.1% since October Chinese stocks bottom out Interestingly, and maybe somewhat counter-intuitively, total Chinese stocks have bottomed out and in fact have increased slightly (by 4.4m tonnes) since our last update to 127.1m tonnes. As we highlighted in previous reports, stocks had been running at historically low levels during the first half of the year and were not sustainable. The sharp increase in 2016 Perret Associates Page 10

11 imports since June and backpedalling by the NDRC regarding its restrictions on domestic crude coal production have helped to stabilise stock levels since the summer. We expect total Chinese inventories to keep increasing, although only at a moderate rate, until the end of Exhibit 2 Coal inventories, main locations, November 2016, m tonnes 31-Oct Jun Mar Dec Oct Jul Apl Dec Dec.13 US producers and distributors (est.) US end users (est.) US electricity sector (est.) Total US Chinese power plants Chinese mines Chinese ports Total China Indian power plants like for like Coal India at pits Indian ports Total India Japanese power plants Australia Newcastle Richards Bay Terminal est UK power plants ARA Poland mines and power plants Ukrainian mines French power plants Finnish power plants Danish power plants Total Europe Total inventories Tentative numbers Source: Various Industry Sources, Perret Associates Indian stocks fall without impact on prices yet On the other hand, total stocks in India have been falling sharply since their peak at 95.5m tonnes in March 2016, down to 60.1m tonnes in October This represents a 37% fall. They have also fallen by 20.3m tonnes or 25% since our previous edition in July This decline was spread equally between a 9.1m tonnes fall at Indian power plants (like-for-like for the same 74 power plants) and a 11.05m tonnes fall at Coal India s pits (Coal India is the largest mining company in the country, accounting for 82% of total production in 2015). So far, we haven t seen any impact on prices, as like-for-like stocks at power plants are in fact just returning to normal levels for this time of the year. In fact, their long term average is 14.1m tonnes. We don t have enough historical data for Coal India to compare its current stocks levels with the historical average. Stock levels have been boosted by the fact that Coal India has increased production significantly in recent years, for instance by 78m tonnes or 14% between 2013 and our production forecast for 2016 of 641.4m tonnes. Still, the ratio of stocks to production has decreased significantly over the period, due to the overall decline in stock levels and the increase in production Perret Associates Page 11

12 Based on our forecasts for Coal India s 2016 production (at 541.3m tonnes) and pit-head stock levels as at 31 December 2016 (36m tonnes), the ratio of stocks to production has decreased as follows: 2014: 9.5%. 2015: 8.1%. 2016f: 6.7%. A stock to production ratio of 5% is generally considered as relatively low. Exhibit 3 Coal India ratio, stocks to production, m tonnes, % Coal India monthly stocks Coal India production Ratio stocks/produciton % 9.0% 8.0% 7.0% 6.0% 5.0% 4.0% 3.0% 2.0% 1.0% 0.0% Source: Company data, Perret Associates US market rebalancing Another potentially significant development in the medium- to long-term, which has larged passed unnoticed so far, is the significant fall in total US stocks. These stood at 189.3m tonnes as at 31 October 2016, down by 36.2m tonnes or 16% from their December 2015 peak of 225.5m tonnes. They are now back to a similar level to December 2014 (184.3m tonnes) and on a par with their historical average since January 2015 of 186.9m tonnes. The US market is rebalancing (as is the Chinese market). This also explains why, despite the significant rally in DES ARA prices, we have not yet seen a strong increase in US exports to Europe. US coal-fired generation has also been bottoming out in recent months after years of heavy decline, mainly due to the fact that the old, inefficient coal-fired capacity has now been dismantled. Meanwhile, US crude coal production has continued to fall, with 2016 output potentially falling to its lowest level since Consequently, we expect a further decrease in total US stockpiles Perret Associates Page 12

13 1.2.5 Stabilisation of coal currencies After a period of strong appreciation of the currencies of major coal-producing countries during H12016, a stabilisation has been taking place since July 2016, with a few currencies even losing some ground since then. The Russian Rouble has weakened by 2% to RUB65 against the US dollar in December 2016 vs. RUB63.85 in July. This follows a 7% appreciation between August 2015 and July In a similar vein, the US dollar gained 8% against the Colombian Peso between July and December 2016, with the latter standing at COP3,165 to the US dollar at the time of writing. Still, this remains below the recent peak of COP3,416 as at 15 February Given the strong performance of these two currencies during the first half of 2016, a period of consolidation was perhaps to be expected. In our Base Case scenario we continue to forecast a general appreciation of all major coal currencies again the US dollar. The nadir that most currencies reached in Jan.-February 2016 coincided with the low point of the Brent contract, which we also expect to recover gradually in the long term. The South African Rand has continued to strengthen to ZAR13.85 against the US dollar vs. ZAR14.37 in our previous edition. It has appreciated by 18% since its recent peak of ZAR in January We expect this trend to continue in the medium- to long-term, due to the overall development of the South African economy. The same is true for the Indonesian Rupiah. We also expect a general appreciation of the Australian dollar, supported by firmer commodity prices. Exhibit 4 Main coal denominated currencies forecast, Unit Dec.16 Dec.17 Dec.18 Dec.19 Dec.20 Dec.21 Dec.22 Dec.23 Dec.24 Dec.25 Dec.26 Dec.27 Dec.28 Dec.29 Dec.30 Russian Rouble vs. US$ Colombian Peso vs. US$ 3,050 2,988 2,925 2,863 2,800 2,760 2,720 2,680 2,640 2,600 2,580 2,560 2,540 2,520 2,500 S.African Rand vs. US$ Australian dollar vs. US$ Indonesian Rupiah vs. US$ 13,500 13,400 13,300 13,200 13,100 13,040 12,980 12,920 12,860 12,800 12,740 12,680 12,620 12,560 12,500 Source: Perret Associates Oil prices stabilise, for now, long term price rise forecast Having been capped at the $50/bbl. Resistance several times during the summer, the Brent contract finally breached this level in November 2016, after the conclusion of an agreement between OPEC suppliers to cut production and stabilise prices. In signing up to this agreement, Saudi Arabia appears to be signalling that the price war in the oil market which took place during is over. The main goal of the cheapest producing countries to grow their share of the international market at the expense of higher cost producers, such as US shale oil producers, seems to have been achieved Perret Associates Page 13

14 Still, OPEC is treading on thin ice as a sudden surge in oil prices could boost production by those companies who had to shut down in Hence, we expect OPEC will try to manage the rise in oil prices to be as gradual as possible and not to exceed the level at which marginal oil producers would be too tempted to come back. To some extent, in the years ahead we anticipate the same scenario of under-investment in oil assets as in the coal market. We anticipate that the prolonged and very painful bear market in both oil and coal will linger in the memory for years to come. Exhibit 5 Brent spot price forecast, $/barrel Unit Dec.16 Dec.17 Dec.18 Dec.19 Dec.20 Dec.21 Dec.22 Dec.23 Dec.24 Dec.25 Dec.26 Dec.27 Dec.28 Dec.29 Dec.30 Brent spot crude oil $/bbl Source: Perret Associates Coal prices well above the cost of production As a result of the continuation of the rally in coal prices between July and November 2016, the premium of the main spot contracts over their respective average cost of production has increased significantly, as illustrated by the table below. Exhibit 6 Spot vs. cost of production, main origins*, December 2016 Unit Colombia South Africa Russia Indonesia Australia US Average Cost of Production $/t Spot physical $/t Difference $/t Difference % 119.1% 117.0% 67.9% 76.3% 79.3% 53.2% *All prices based on 6,000kcal/kg NAR Source: Perret Associates To put things in perspectives, we estimate that the respective premiums were as follows in July 2016: Colombia: $15.15/t or 40%. South-Africa: $27.8/t or 82%. Russia: $8.9/t or 20%. Indonesia: $6.2/t or 14.2%. Australia: $15.8/t or 34%. US: $3.8/t or 7.6%. Understandably, Colombian and South African producers, which have the lowest cost of production, have the biggest profit margins. However, even Russian, Indonesian and Australian producers are now operating at a significant margin, according to our database. We also note that even US producers are now operating at a significant premium to their average cost of production. In normal circumstances, such price differences would prompt the full utilisation of existing capacity and, if required, investment in additional supply, until the point at which the market 2016 Perret Associates Page 14

15 becomes over-supplied, which would precipitate prices downward. This is exactly what happened in However, this time around we think things are different. The impact of the war on coal during the period has been such that we don t think that such investment in new coal supply will be made, or at least not enough to tip the market back into over-supply. Of course, there is a limit to how large the premium between spot prices and the cost of production can go. Overall, though, we don t expect to see the same kind of reaction from existing suppliers and potential suppliers as we have seen in the past. 1.3 FOB high quality: prices to decline during H117 before recovering Spot overshoots Since our previous edition the rally continued until early November, when the FOB Newcastle spot index peaked briefly at $109.70/t, API4 at $100.90/t and API2 at $91.26/t. Since then, the main indices have had a significant correction of about $20/t, before finding support at the time of writing in the low $80s/t for API4 and API2 and mid-$80s/t for FOB Newcastle. Based on the latest update of our world supply-demand, we expect the seaborne market to finish 2016 in balance and enter marginal surplus in This is mainly the consequence of the NDRC s temporary relaxation of production caps in China to enable an increase in production. We are already seeing the first impact of this, with some restocking taking place at Chinese ports and power plants. Our Base Case scenario is that once stocks are back at historical levels and domestic Chinese prices return to levels acceptable to the government, the mining industry and the end users (in the range CNY /t FOB Qinhuangdao), the NDRC will gradually revert to its strategy of structural cuts in domestic production capacity. Indeed, we estimate that the long term objective of the Chinese authorities remains to diversify away from coal as much as possible, both on the supply side and for electricity generation, even if the volumes of coal used will remain very significant at least until Within the next 6-12 months, the international seaborne market could swing back into oversupply, or at least perceived over-supply. The combination of rising Chinese stock piles and declining Chinese prices could encourage some end users in China and elsewhere to become complacent once again, in apparently confirming their view that the 2016 rally represented the swan song of the coal market. In fact, we think that a potential market dip in H117 might just be a short-lived affair before supply tightness starts to return to the market, as the Chinese authorities resume their efforts to rein in domestic production. Meanwhile, the structural lack of investment in new coal mines will gradually start to take its toll, as international coal demand remains supported Perret Associates Page 15

16 We expect the combination of all the factors above to push the market into structural undersupply as we move into 2018 and beyond. We therefore foresee a broadly bullish outlook for spot coal prices, especially as, in parallel, we anticipate an increase in oil prices and the strengthening of major coal-related currencies, which would also support the cost of production. As spot prices keep increasing, the natural floor will increase as well, limiting somewhat the downside potential. Of course, the best cure for high prices has always been high prices. Another significant coal price rally could push some investors to take the plunge and invest in greenfield mining projects at some point. However, we think it would require a significant price increase (say spot prices above $100/t) and / or a technological breakthrough in the usage of coal (for example carbon recycling rather than storage) to reverse the damage inflicted on the coal industry since It is way too early to speculate in further detail as to whether and when such events might occur but forward paper prices are too low, indicating potential contango In these circumstances, we struggle to understand why paper forward contracts are so discounted to spot: either we are wrong or the market is wrong. At the time of writing, for instance, the Cal.18 API2 contract is trading at $59.15/t, a discount of $31.05/t vs. the spot index at $90.20/t. True, spot prices have some downside in the next 6 months, as illustrated in our table below, but we don t expect them to fall as much as anticipated by the market, judging by the Cal.17 contract, for instance. We suspect that market participants may be using far lower consumption and import forecasts than ours and anticipate that the Chinese market will remain in over-supply for years to come. In this context, and given the downside we are expecting over the next 6-9 months before a potential tightening emerges, we are providing more granularity to our forecast by providing a quarterly forecast for the coming year, before switching to a yearly forecast thereafter. In fact, we anticipate that once the spot market has bottomed out, the forward curve could gradually rise above the spot values. Exhibit 7 Base Case FOB price forecast, US$/t, nominal 2017 Dec Dec Mar Jun Sep Dec-17 Dec.18 Dec.19 Dec.20 Dec.21 Dec.22 Dec.23 Dec.24 Dec.25 Dec.26 Dec.27 Dec.28 Dec.29 Dec.30 FOB NEWC FOB Richards Bay FOB Puerto Bolivar FOB Russia Murmansk Source: Perret Associates Sensitivity analysis The La Nina weather phenomenon has now replaced the El Nino in most parts of the world. However, according to the latest weather information and forecasts, especially those 2016 Perret Associates Page 16

17 published by Australian agencies, the risks of severe weather disruptions due to La Nina have decreased significantly. This means the risk of severe flooding at Australian mines over the Dec.2016-April 2017 period have decreased. Nevertheless, adverse weather conditions could continue to impact prices over the next 6 months, in particular in Russia, where low temperatures are affecting the domestic supply chain, and Indonesia, where persistent flooding at mines is also affecting supply. Low Case: 20% probability Our Low Case scenario is based on the assumption that the Chinese government overshoots in the current readjustment, i.e. that Chinese coal production starts increasing too quickly compared with the stagnation of domestic demand. In this scenario, Chinese imports will fall more than we forecast in 2017, piling further pressure on international prices. In this Low Case scenario we might also see the initial impacts of a Trump presidency in the US, which we think might push up US domestic crude coal production rather than demand, potentially boosting exports from the US, in particular to the European ARA zone. In this scenario, the $50/bbl. Support level for the Brent crude oil contract would also be short-lived, as neither OPEC nor non-opec countries stick to the agreed production cuts. The cost of production forecast is based on the spot cost of production to which is added an inflation ratio. Given the level of spot prices, which are well above their respective cost of production, we anticipate that even in our Low Case scenario spot values will take 6 to 12 months before falling back to their cost of production levels. High Case: 30% probability In our High Case scenario, the Chinese authorities do a good job and manage a soft landing by increasing domestic crude coal production just enough to stabilise the market, as well as domestic and therefore international prices. They revert to the 276 working days policy relatively soon, say around March 2017, and maintain their overall policy of cutting crude coal production as well as diversifying their electricity mix away from coal. In this scenario, international mining companies continue to operate at significant profit margins and generate hefty cash flow. However, they use this to diversify away from coal mining and don t invest in large Greenfield coal projects. A few mining operations are re-opened in Indonesia but this is not sufficient to swing the international coal market back into over-supply Perret Associates Page 17

18 Crude oil prices continue to rise strongly towards, approaching the $70/bbl mark by December 2017, as a Trump presidency boosts the US economy. The easing of tensions between Russia and Western powers also provides some support to the Russian economy. Exhibit 8 FOB NEWC., price forecast, sensitivity analysis, US$/t, nominal 2017 Dec Dec Mar Jun Sep Dec-17 Dec.18 Dec.19 Dec.20 Dec.21 Dec.22 Dec.23 Dec.24 Dec.25 Dec.26 Dec.27 Dec.28 Dec.29 Dec.30 Cost of production Low case Base case High case High case Base case Low case Cost of production 40 Source: Perret Associates Exhibit 9 FOB R.Bay 6,000 NAR price forecast, sensitivity analysis, $/t, nominal 2017 Dec Dec Mar Jun Sep Dec-17 Dec.18 Dec.19 Dec.20 Dec.21 Dec.22 Dec.23 Dec.24 Dec.25 Dec.26 Dec.27 Dec.28 Dec.29 Dec.30 Cost of production Low case Base case High case High case 100 Base case Low case Cost of production 20 Source: Perret Associates 2016 Perret Associates Page 18

19 Exhibit 10 FOB P. Bolivar 6,000 NAR price forecast, sensitivity analysis, $/t, nominal 2017 Dec Dec Mar Jun Sep Dec-17 Dec.18 Dec.19 Dec.20 Dec.21 Dec.22 Dec.23 Dec.24 Dec.25 Dec.26 Dec.27 Dec.28 Dec.29 Dec.30 Cost of production Low case Base case High case High case 100 Base case Low case 40 Cost of production 20 Source: Perret Associates Exhibit 11 FOB Russia Murmansk, 6,000 NAR, sensitivity analysis, $/t, nominal 2017 Dec Dec Mar Jun Sep Dec-17 Dec.18 Dec.19 Dec.20 Dec.21 Dec.22 Dec.23 Dec.24 Dec.25 Dec.26 Dec.27 Dec.28 Dec.29 Dec.30 Cost of production Low case Base case High case High case Base case Low case Cost of production 30 Source: Perret Associates 1.4 FOB medium and low qualities We derive FOB Newcastle 5,500 NAR from the FOB Newcastle 6,000 NAR quality, based on the historical spread. According to our data, the 5,500 quality stood at an average discount of $12.97/t below the 6,000 quality during the period , but it was lower during the period at $11.6/t. We are using a standard discount of $12.5/t for our forecast Perret Associates Page 19

20 Exhibit 12 FOB NEWC., 5,500 NAR, price forecast, sensitivity analysis, $/t, nominal 2017 Dec Dec Mar Jun Sep Dec-17 Dec.18 Dec.19 Dec.20 Dec.21 Dec.22 Dec.23 Dec.24 Dec.25 Dec.26 Dec.27 Dec.28 Dec.29 Dec.30 Cost of production Low case Base case High case Source: Perret Associates We use the same methodology for our FOB Richards Bay 5,500 NAR forecast with an average discount of $11.60/t. Exhibit 13 FOB R.Bay 5,500 NAR, price forecast, sensitivity analysis, $/t, nominal 2017 Dec Dec Mar Jun Sep Dec-17 Dec.18 Dec.19 Dec.20 Dec.21 Dec.22 Dec.23 Dec.24 Dec.25 Dec.26 Dec.27 Dec.28 Dec.29 Dec.30 Cost of production Low case Base case High case Source: Perret Associates For the FOB Indo. 5,900 GAR contract, we assume the same percentage price variation as our forecast for FOB NEWC. 6,000 NAR. The cost of production is based on our estimate of the cost of production for the 6,000 NAR material, adjusted for calorific value. The Low Case is based on the cost of production. The High Case assumes a premium of 20% above the Base Case. Exhibit 14 FOB Indo. 5,900 GAR, price forecast, sensitivity analysis, $/t, nominal 2017 Dec Dec Mar Jun Sep Dec-17 Dec.18 Dec.19 Dec.20 Dec.21 Dec.22 Dec.23 Dec.24 Dec.25 Dec.26 Dec.27 Dec.28 Dec.29 Dec.30 Low case Base case High case Source: Perret Associates We use a similar methodology for FOB Indo. 4,900 NAR. Exhibit 15 FOB Indo 4,900 NAR, price forecast, sensitivity analysis, $/t, nominal 2017 Dec Dec Mar Jun Sep Dec-17 Dec.18 Dec.19 Dec.20 Dec.21 Dec.22 Dec.23 Dec.24 Dec.25 Dec.26 Dec.27 Dec.28 Dec.29 Dec.30 Low case Base case High case Source: Perret Associates Our forecast for the CFR China 5,500 contract is based on FOB Indo. 5,900 GAR + the Panamax voyage rate from Indonesia (South East Kalimantan) to S.East China (Guangzhou), adjusted for calorific value. We have also adjusted the discount applied to FOB + Freight values from the 20% used in the recent years of over-supply to 11.2%, in order to better reflect the tightening of the market. We are seeing the same phenomenon in ARA, i.e. a reduction of the traditional discount of delivered price vs. the sum of FOB + freight Perret Associates Page 20

21 Exhibit 16 CFR China, 5,500 NAR, price forecast, sensitivity analysis, $/t nominal 2017 Dec-15 Dec Mar Jun Sep Dec-17 Dec-18 Dec-19 Dec-20 Dec-21 Dec-22 Dec-23 Dec-24 Dec-25 Dec-26 Dec-27 Dec-28 Dec-29 Dec-30 Low Case Base Case High Case Source: Perret Associates 1.5 Freight forecast The freight market enjoyed a brief rally during Q416, with the Capes 5TC index peaking at $19,500/day in mid-november while the Panamax 4TC index kept on increasing to $12,500/day by early December. As envisaged in our last report, based on our Replacement cost analysis, the surge in Chinese steam coal imports has been mainly beneficial to Indonesian coal suppliers, where a significant proportion of exports are still loaded on Panamax vessels. (Also, see the section on Chinese imports and Indonesian exports). Our previous 31 December 2016 forecast for Capes 5TC was $9,622/day, which is well in line with current spot values at the time of writing of $9,300/day. Still, we envisage further downside until the end of 2016 and in January 2017, due to a slowdown in Chinese coal imports and potentially iron ore. We decrease our price forecast until March 2017 vs. the previous edition and envisage a gradual recovery from that point. Exhibit 17 Capes 5TC index, price forecast, Sensitivity Analysis, $/day, nominal Dec Mar Jun-17 30Sept.17 31stDec.17 Dec.18 Dec.19 Dec.20 Dec.21 Dec.22 Dec.23 Dec.24 Dec.25 Dec.26 Dec.27 Dec.28 Dec.29 Dec.30 4,811 4,125 2,750 2,778 2,833 3,541 3,825 4,475 4,922 5,349 5,741 6,086 6,370 6,497 6,616 6,726 6,827 6,918 6,987 Low Case Base Case 4,811 5,500 5,500 5,555 5,666 7,083 7,649 8,950 9,845 10,698 11,482 12,171 12,739 12,994 13,232 13,453 13,655 13,837 13,975 High Case 4,811 6,875 8,250 8,333 8,499 10,624 11,474 13,424 14,767 16,047 17,223 18,257 19,109 19,491 19,848 20,179 20,482 20,755 20,962 Source: Perret Associates Exhibit 18 Capes C4 index, price forecast, Sensitivity Analysis, $/t, nominal Dec Mar Jun Sept Dec.17 Dec.18 Dec.19 Dec.20 Dec.21 Dec.22 Dec.23 Dec.24 Dec.25 Dec.26 Dec.27 Dec.28 Dec.29 Dec Low Case Base Case High Case Source: Perret Associates Exhibit 19 Capes C7 index, price forecast, Sensitivity Analysis, $/tonne, nominal Dec Mar Jun Sept Dec.17 Dec.18 Dec.19 Dec.20 Dec.21 Dec.22 Dec.23 Dec.24 Dec.25 Dec.26 Dec.27 Dec.28 Dec.29 Dec Low Case Base Case High Case Source: Perret Associates Exhibit 20 Panamax 4TC index, price forecast, Sensitivity Analysis, $/day, nominal Dec Mar Jun Sept Dec.17 Dec.18 Dec.19 Dec.20 Dec.21 Dec.22 Dec.23 Dec.24 Dec.25 Dec.26 Dec.27 Dec.28 Dec.29 Dec.30 3,706 6,000 6,060 6,121 6,194 7,123 7,465 8,227 8,720 9,174 9,577 9,922 10,200 10,322 10,436 10,540 10,635 10,720 10,784 Low Case Base Case 3,706 8,000 8,080 8,161 8,259 9,498 9,953 10,969 11,627 12,231 12,770 13,229 13,600 13,763 13,914 14,053 14,180 14,293 14,379 High Case 3,706 10,000 10,100 10,201 10,323 11,872 12,442 13,711 14,533 15,289 15,962 16,537 17,000 17,204 17,393 17,567 17,725 17,867 17,974 Source: Perret Associates 2016 Perret Associates Page 21

22 1.6 DES ARA price forecast Based on FOB Colombia We base our DES ARA price on FOB Colombia + C7 Capesize freight rate. We used to apply a discount of 10% to the sum [FOB + freight], which corresponded to the general pattern of a necessary discount on the replacement cost in the European market. Interestingly, and in another sign of the tightness of the European market, for this current forecast we have had to apply a premium of 3% to the [FOB + Freight] components. We reduce this premium progressively during 2017 before switching back to a discount from Jan.2018 onwards. Exhibit 21 DES ARA price forecast, based on Colombia, $/tonne, nominal 2017 Dec-15 Dec Mar Jun Sep Dec-17 Dec-18 Dec-19 Dec-20 Dec-21 Dec-22 Dec-23 Dec-24 Dec-25 Dec-26 Dec-27 Dec-28 Dec-29 Dec-30 Low Case Base case High Case Source: Perret Associates FOB Russia to become the marginal cost to ARA We continue to differentiate our DES ARA price forecast based on Colombian and Russian coal. Indeed, despite the continuous decline in EU28 coal imports, Colombian supply will not be sufficient to cover all European demand, in particular with additional imports from Turkey, Morocco and at some point probably the Middle East saw a turnaround in the Russian economy. The rebound in oil prices has triggered a prompt recovery in the Russian Rouble, which is one of the currencies that has appreciated most against the US dollar over the last two years. Recent global political developments are also reinforcing Russia s position on the international scene. The vacuum left by the West in Syria has opened the door to Russian intervention in the country reinforcing its position on the international scene. Donald Trump election as US President and his nomination of a Secretary of State with close links to the Putin administration should reinforce the dialogue between the US and Russia. The sanctions against Russia adopted by many Western countries over the Ukraine crisis have largely proven to be ineffective. We expect them to be dropped or significantly relaxed by 2018, partly because most of the country leaders who pushed for such sanctions are no longer in power. We therefore expect the Russian Rouble to keep appreciating against the US dollar, and faster than its direct counterparts for coal exports, such as the Colombian Peso. The Rouble will also be helped by our anticipated further rise in oil prices Perret Associates Page 22

23 A potential recovery in the Russian economy could also boost domestic coal consumption and potentially cannibalise exports. Consequently, we expect spot Russian coal prices, as well as the Russian cost of production, to rise more strongly, compared with other hubs. This is potentially another bullish element for the European market as we anticipate that Russian coal will become the marginal supplier to the European market. Exhibit 22 DES ARA forecast based on Russia, $/tonne, nominal 2017 Dec-15 Dec Mar Jun Sep Dec-17 Dec-18 Dec-19 Dec-20 Dec-21 Dec-22 Dec-23 Dec-24 Dec-25 Dec-26 Dec-27 Dec-28 Dec-29 Dec-30 Low Case Base case High Case Source: Perret Associates Exhibit 23 Comparison BCase DES ARA based on Colombian and Russian coal, $/t 2017 Dec-15 Dec Mar Jun Sep Dec-17 Dec-18 Dec-19 Dec-20 Dec-21 Dec-22 Dec-23 Dec-24 Dec-25 Dec-26 Dec-27 Dec-28 Dec-29 Dec-30 Base Case Colombia Base Case Russia Base Case Russia Base Case Colombia Source: Perret Associates Dec Jun-17 Dec-18 Dec-21 Dec-24 Dec-27 Dec Perret Associates Page 23

24 2 Imports 2.1 China Slowing GDP growth The IMF has marginally increased its GDP growth estimate for 2016 to 6.6% from 6.5% in our last edition. Their forecast is otherwise broadly unchanged until We are continuing to use the IMF forecast until 2021 and our own forecast thereafter. Exhibit 24 China GDP growth forecast, % China GDP growth forecast % 6.90% 6.59% 6.17% 6.03% 6.00% 5.90% 5.80% 5.71% 5.62% 5.53% 5.44% 5.36% 5.27% 5.18% 5.09% 5.00% Source: IMF until 2021, Perret Associates 2022 onwards Coal-fired generation finds some support Overall, Chinese electricity generation has gone from being flat, as we mentioned in our previous report, to robust growth in the second half of Indeed, whereas total generation improved only marginally (by 0.6%) y-o-y during Jan.-May 2016, the pace increased during the summer so that total generation reached 4,350.4TWh during Jan.-Sept. 2016, up 3.1% y-o-y. To put things in perspective, total generation eroded by 0.7% in Whereas thermal and coal-fired generation used to benefit the most from an increase in total generation, the impact is now more muted. Thermal generation increased by only 0.5% y-o-y during Jan.-September to 3,211.3TWh. By way of comparison, in 2013 total generation increased by 7.9% and thermal by 11.4%. This trend started to change in 2014 already, due to the increasing diversification of the Chinese electricity mix. In 2014, total generation increased by 9.5% and thermal by just 4.7%. In other words, the recovery in total generation, at best, spared thermal generation a more brutal decline. Having said that, thermal generation still increased by 7.5% y-o-y in August and 12.2% in September. The growth rates of alternative electricity sources was more pronounced. Hydro generation reached 810.7TWh during Jan.-Sept. 2016, up 8.7% y-o-y. This was mainly the result of strong precipitation, as hydro capacity increased by just 4GW so far this year at 284GW. Nuclear generation keeps increasing strongly at 154.4TWh, up 21.1% y-o-y, due to the addition of 4.5GW capacity so far this year. The growth in wind generation is roughly of the same order of magnitude, having reached 163.3TWh during Jan.-Sept. 2016, up 19.3% y-o-y. Wind generation capacity increased by 14GW during Jan.-April 2016 alone Perret Associates Page 24

25 As a result of all this, we expect coal s share of the generation mix to fall to 70% in 2016, from 74% in This would be its lowest level since our records started in 2007, when coal s share stood at 81.5%. Exhibit 25 Coal s share of total Chinese electricity generation 85% 80% 75% 70% 65% 60% Source: NBS, Perret Associates Rebound in coal-fired generation, despite the cap in new capacity We are revising upwards our forecast for full-year total electricity generation to 5,760TWh in 2016, up from 5,440TWh in our previous edition. This would represent a 2.5% yearly growth vs. the 0.5% decline we envisaged in our last report. We also increase our forecast for 2020 to 5,990TWh vs. 5,900TWh in our last edition. We keep our forecast from then until 2030 unchanged, representing an annual yearly growth of 1% over the period. We also revise upwards our forecast for 2016 thermal generation to 4,200TWh vs. 4,030TWh in our previous edition. This would represent a 0.6% increase vs. a 3.6% decline. Thermal capacity is continuing to grow, increasing by 68.4GW during Jan.-September to a total of 1,028.5GW. We have speculated in recent reports as to how many ghost coal power plants the Chinese government was prepared to build, in the face of lower growth in total electricity demand and the rapid diversification of the power mix. The government now seems to have acknowledged the problem, as in October 2016 the National Energy Planning Agency announced the halt of construction work on new coal power plants in 28 provinces. As always with such big announcements, the devil is in the details and it is difficult to quantify precisely how many power plants are directly affected. What seems clear for now is that: Work should not start on new coal-fired power plants, even if a permission had already been granted. In those cases where construction only started in 2016, construction work is to be stopped Perret Associates Page 25

26 Where construction work is on-going but started before 2016, there are to be adjustments to the construction schedules. The grey area lies mainly in the second and third scenarios above. As far as the adjustments are concerned in the third scenario, our understanding that construction is to be delayed as long as possible. Still, electricity generation and capacity are two different things, especially in countries, such as China, which are now building excess capacity in power generation, as Germany has also done. The capacity utilisation at Chinese coal power plants has declined sharply in recent years to less than 50% in 2015, according to our calculations. There is therefore already a significant cushion of spare coal-fired capacity available, which can be fallen back on in the event of a sudden surge in demand and/or fall in generation from alternative sources. The magnitude of the recent coal rally, which was driven primarily by sharp cuts in domestic crude coal production, serves to underline just what could happen if the surge in demand were to be combined with an unexpected fall in, say hydro, nuclear or wind generation. In any case, we are decreasing our forecast for coal-fired capacity as follows: 2016 (December): 900GW vs. 903GW in our previous edition. 2020: 929GW vs. 949GW. 2025: 954GW vs. 983GW. 2030: 979GW vs. 1,010GW. Our forecast was already conservative, for instance compared with the IEA forecast, which still stood at 982GW (for 2020) and 1,115GW (2030) at the time of writing. Our forecast is also lower than the latest target announced by the government in November 2016 to cap coal-fired capacity at 1,100GW until Still, even if China is going to build fewer coal power plants, it will still require a significant amount of electricity from its coal-fired sector. We have revised upwards our 2016 forecast for coal-fired generation to 4,032TWh, from 3,839TWh in our previous edition. This represents 1% yearly growth vs. a 3.9% fall, as per our last report. On the top of this, we expect 3GW less capacity to be commissioned this year, so we adjust the capacity utilisation from 48.75% in our previous edition to 51.2%. We think such levels are actually perfectly normal for a country like China, which is still in a developing phase. This compares well with a country like Germany, where the capacity utilisation has been in the range 50%-52% in recent years. A capacity utilisation level below 50% would be considered as an anomaly in China. We also estimate that there is some upside in our forecast for coal-fired generation in 2016, which could easily materialise if capacity utilisation at power plants increased further Perret Associates Page 26

27 We forecast a straight line decline in capacity utilisation between 2020 and 2030 to 50.5% although this remains well above the 48% assumption in our previous edition. Again, such a low capacity utilisation seemed unreasonably low to us and we feel more comfortable with the revised higher numbers. We have increased our estimate for the average coal plant efficiency at 36% in 2016 due to the rapid modernisation of the plants. As a result of this, and despite our downwards adjustment in coal-fired capacity, we increase our forecast for coal-fired generation between 2017 and 2030 by 2%. This equates to an increase in coal usage, compared with our previous edition, of 44m tonnes in 2020, 32m tonnes in 2025 and 35m tonnes in Exhibit 26 Chinese coal usage for electricity generation, forecast Steam coal demand - Electricity generation Coal-fired generation capacity GW Coal-fired electricity generation TWh 3,995 4,032 4,099 4,131 4,153 4,150 4,169 4,188 4,206 4,225 4,243 4,262 4,280 4,298 4,317 4,331 Steam coal demand m tonnes 1,868 1,817 1,837 1,841 1,841 1,830 1,828 1,826 1,825 1,823 1,812 1,815 1,819 1,823 1,827 1,825 Source: Perret Associates Renewed strength in cement production but for how long? March 2016 (after the Chinese New Year) marked a turning point in Chinese cement production. Having fallen on a yearly basis without interruption between March 2015 and February 2016 (at an average rate of 5.5%), cement production started to pick up in March It has been increasing without interruption since, averaging yearly growth of +2.4% in the period March-Sept As a result, Chinese cement production increased by 14.5m tonnes or 0.8% during Jan.- September. We are therefore adjusting upwards our 2016 cement production forecast to 2.37bn tonnes from 2.32bn tonnes in our previous forecast. This would represent a 0.9% yearly increase vs. the 1.2% decline forecast in our previous edition. Inevitably, this begs the question as to whether this increase in production will be sustainable. The Chinese property market has been picking up again, to the point where many local economists are again talking of the risks of a potential bubble. We have included again below the bell curve chart we featured in previous reports to highlight the specificities of the current Chinese cement market. In any event, we increase our cement production from the previous edition as follows: 2017: +37m tonnes 2020: +66m tonnes 2030: +50m tonnes. Given the relatively low ratio of coal usage vs. cement production, though, the impact of this revision on forecast coal usage is relatively muted, increasing it by: 8m tonnes in Perret Associates Page 27

28 16-18m tonnes during the period In a similar vein, Chinese steel production, which had a disastrous 2015, when it declined by 1.9%, started to bounce back in March The yearly growth has been uninterrupted since, averaging +1.8% during Jan.-Sept Exhibit 27 The cement bell curve relationship between GDP and cement usage Cement production per capita (kg) Iran Egypt China South Korea Turkey Poland Italy Spain France Germany 200 Brazil USA Source: World Bank, Various industry sources, Perret Associates Even so, due to the very slow start to the year, Chinese crude steel production reached 604.6m tonnes during Jan.-Sept. 2016, down by a marginal 1m tonnes y-o-y. As a result, we have increased our 2016 forecast to 801m tonnes, up by just 4m tonnes from our previous edition and flat y-o-y. We also increase our steel production forecast by 5-10m tonnes per year over the period. As a result the demand for coking coal increases from our previous edition as follows: 2016: +3m tonnes. 2020: +8m tonnes. 2025: +8m tonnes. 2030: +6m tonnes. Exhibit 28 Chinese steam and coking coal demand, forecast Source: Perret Associates India Thailand Mexico Japan 0 $0 $10,000 $20,000 $30,000 $40,000 $50,000 $60,000 $70,000 GDP per capita Summary demand Steam coal for electricity generation m tonnes 1,868 1,817 1,837 1,841 1,841 1,830 1,828 1,826 1,825 1,823 1,812 1,815 1,819 1,823 1,827 1,825 As % of total steam coal % 62.2% 64.8% 65.1% 65.4% 65.6% 65.8% 66.0% 66.3% 66.5% 66.7% 67.2% 67.3% 67.5% 67.7% 67.9% 68.2% Steam coal for heating, residential + other energym tonnes As % of total steam coal % 11.5% 11.5% 11.4% 11.2% 10.9% 10.7% 10.4% 10.2% 9.9% 9.7% 9.1% 9.0% 8.8% 8.7% 8.6% 8.4% Steam coal for cement and building material m tonnes As % of total steam coal % 13.0% 13.9% 13.8% 13.7% 13.7% 13.8% 13.8% 13.9% 13.9% 13.9% 14.1% 14.2% 14.3% 14.4% 14.5% 14.7% Steam coal for other industries m tonnes As % of total steam coal % 13.3% 9.8% 9.7% 9.7% 9.7% 9.7% 9.7% 9.7% 9.7% 9.7% 9.6% 9.5% 9.4% 9.2% 9.0% 8.7% Total steam coal m tonnes 3,001 2,803 2,821 2,815 2,804 2,780 2,768 2,756 2,745 2,733 2,696 2,696 2,695 2,694 2,692 2,675 As % of total % 78.7% 78.8% 78.8% 78.6% 78.4% 78.1% 77.8% 77.5% 77.3% 77.0% 76.6% 76.6% 76.5% 76.6% 76.7% 76.8% Total coking coal m tonnes As % of total % 21.3% 21.2% 21.2% 21.4% 21.6% 21.9% 22.2% 22.5% 22.7% 23.0% 23.4% 23.4% 23.5% 23.4% 23.3% 23.2% Total steam + coking coal m tonnes 3,812 3,556 3,580 3,579 3,577 3,562 3,558 3,555 3,552 3,549 3,519 3,522 3,521 3,517 3,510 3, Perret Associates Page 28

29 2.1.5 Structural decline in coal production Chinese crude coal production plummeted by 320.9m tonnes or 10.5% y-o-y to 2.72bn tonnes during Jan.-October As highlighted in previous reports, the trend started two years ago, with a fall of 80m tonnes in crude coal production in 2014 and 134m tonnes in The pace of this decline accelerated in April 2016, when China s top economic planning body, the Chinese National Development and Reform Commission, imposed on mining companies a limit of 276 working days per year in order to avoid a supply glut. At the same time, it accelerated its campaign to close of small and inefficient mines. It also suspended the approval of any new coal mines during the period As part of its plans to address over-capacity in a number of industrial sectors, the government intends to progressively shut down up to 500m tonnes of surplus coal production capacity through mine closures by Another m tonnes of production capacity is expected to be cut by way of mergers/industry consolidation. In our last report, we took this ambition with a pinch of salt but it is now looking more credible. In fact, the production restrictions were applied with such fervour that the marginal surplus in the domestic Chinese coal market disappeared very quickly over the summer of This was also due to the historically low inventories levels in China, which we have been highlighting since January Consequently, domestic prices surged by 85% or CNY326/t ($47.4/t) between April and November This also translated into higher imports. Since September, the NDRC has been back-peddling, realising that its reforms had gone too far, too fast. It gradually increased the number of working days back to 330 per year, initially until 31 December 2016 (on an annual basis) and at the time of writing until at least 31 March, marking the end of the winter heating season. We expect more short term adjustment measures from the NDRC in the coming weeks and months. Taking a step back from these short term measures, it seems to us (and these are our Base Case assumptions) that the NDRC s objective is to stabilise the domestic market and control price increases, rather than to revert to a structurally over-supplied Chinese coal market. Indeed, so far the output restrictions have been suspended only in the short term and for the largest mines of the country, not the small mines. The objective of closing down these small mines remains in force for the time being. This should result in greater consolidation in the Chinese mining sector, which is also a key government objective. The significant cut in crude coal production is also in line with the government s move to halt the construction of any new coal-fired power plants at least until 2018 (see above) Perret Associates Page 29

30 It therefore seems logical to anticipate further cuts in the medium/long term in crude coal production, as part of the global consolidation process of the coal industry. Despite the slowing in production cuts that should start filtering through from October 2016, we still anticipate a yearly fall in crude coal production of an average -8.1% during November-December 2016, vs. an average of -12.6% during April-September We therefore forecast a fall in crude coal production of 363m tonnes or 9.9% y-o-y to 3.32bn tonnes in We then anticipate a slight increase of 50m tonnes in 2017 from 2016 s levels. Indeed, there is typically a 2-3 month time-lag in the mining industry and we anticipate the measures recently taken by the NDRC will only have their full impact in Q117 or even H117. However, once the Chinese domestic market has stabilised, we forecast the structural decline in crude coal production to resume. We expect it to fall to 3.3bn tonnes in Exhibit 29 Chinese crude coal production forecast, steam and coking coal, m tonnes China coal production Crude coal production Mill. tonnes 3,638 3,330 3,370 3,347 3,323 3,300 3,291 3,282 3,273 3,264 3,255 3,245 3,236 3,227 3,218 3,200 Of which coking coal Mill. tonnes Of which steam coal Mill. tonnes 2,874 2,631 2,662 2,644 2,625 2,607 2,600 2,593 2,585 2,578 2,571 2,564 2,557 2,550 2,542 2,528 Total Mill. tonnes 3,638 3,330 3,370 3,347 3,323 3,300 3,291 3,282 3,273 3,264 3,255 3,245 3,236 3,227 3,218 3,200 Source: Perret Associates Chinese steam coal imports supported We now forecast Chinese net steam coal imports will reach 172m tonnes in We forecast a decline in 2017 to 159m tonnes, due to the temporary rise in domestic production. We then forecast a consolidation in imports in the range m tonnes during , before a gradual decline towards m tonnes by Exhibit 30 Chinese steam and coking coal imports forecast, m tonnes China net imports BC Net steam coal imports m tonnes Net coking coal imports m tonnes Total m tonnes Coking coal Steam coal Source: Perret Associates 2016 Perret Associates Page 30

31 2.2 India Slight increase in GDP growth forecast We are using the IMF forecast for the period until The IMF has increased its GDP growth forecast for India by 0.1%-0.3% from our previous edition. We have slightly increased our own long term GDP forecast for the 2030 anchor point from 6.3% to 6.5%. Exhibit 31 India GDP growth, historical and forecasts, % India GDP growth % 7.56% 7.62% 7.90% 7.67% 7.83% 7.95% 8.12% 7.96% 7.80% 7.63% 7.47% 7.31% 7.15% 6.99% 6.82% 6.50% Source: IMF until 2021, Perret Associates 2022 onwards Recent slowdown in total and coal-fired generation Indian power generation totalled 961TWh during Jan.-October 2016, up 5.8% y-o-y, despite a strong disparity between the first half of the year (+9.2% y-o-y) and the second half so far (up 1.2% y-o-y). Having been almost flat on a yearly basis over the summer, total generation picked up slightly in October. Overall, we have slightly decreased our 2016 full-year growth forecast from 5.8% in our previous edition to 5.2%, or 1,140TWh. We also decrease by 50TWh our generation forecast for 2020 to 1,350TWh but keep our forecast unchanged for 2030 at 1,800TWh. Thermal generation followed the same pattern as total generation, increasing by 13.1% y-o-y on average during H116 but levelling off (down 0.1% y-o-y) during July-October. Overall, it is still up 7.6% y-o-y on a year-to-date basis, although this overall growth should decline as the weight of the disappointing H216 filters through. Predictably enough, given that it accounts for 80.5% of total generation, hard coal-fired generation followed the same pattern, up 12.9% y-o-y during H116 but down 0.5% during July-October 16. Coal-fired generation has increased by 7.6% y-o-y on a year-to-date basis but we now forecast full-year annual growth at 6.4%, down from our previous forecast at 7.9%. We estimate that hard coal-fired generation capacity stood at 190GW on 31 October We are increasing our capacity forecast to 193GW by 31 December from 191GW in our previous edition. We estimate that lignite capacity is unchanged at 7GW. Given the continuous increase in hard coal-fired capacity and lower than expected generation growth, we adjust the capacity utilisation from 54.25% to 53%. Coal-fired generation has suffered from the disappointing growth in total electricity demand but also from higher generation from other sources. Hydro generation, which was 13% lower y-o-y during H116, has picked up since the summer and was 7.9% higher y-o-y during July-October 16. Nuclear generation, which was also 2016 Perret Associates Page 31

32 10.8% lower y-o-y during H116, has also risen significantly, increasing by 12.6% during July- October 16. Diesel generation continues to rise strongly, increasing by 18.5% during Jan.-October 16. Still, coal and lignite s combined share of the generation mix has averaged 80.6% so far this year, little changed from 2015 s average. Exhibit 32 Coal and lignite, share in total electricity generation, % 90% 85% 80% 75% 70% 65% 60% 55% 50% Source: CEA, Perret Associates Further growth in Indian coal demand We maintain our forecast for additional coal-fired capacity unchanged from the previous edition. We decrease our long term forecast for capacity utilisation from 54% to 53.5%, to reflect the developments in We also expect a gradual diversification of the electricity mix to reduce the dependency on coal. As a result, we decrease our forecast for coal usage for electricity generation as follows: 2016: -8m tonnes from our last report. 2020: -1m tonnes 2030: -3m tonnes Exhibit 33 Indian steam coal demand for electricity generation, forecast, m tonnes Steam coal demand - Electricity generation Coal and lignite capacity GW Coal and lignite-fired electricity generation TWh ,044 1,088 1,134 1,172 1,209 1,247 1,284 1,317 1,350 1,383 1,415 1,448 1,481 Steam coal and lignite demand m tonnes Source: Perret Associates Robust demand from other sectors Indian cement production reached 222.1m tonnes during Jan.-September 2016, up by 14.5m tonnes or 7% y-o-y. Although the yearly growth has eroded slightly from the 11.4% in Q116, this remains a very good performance, in particular compared with the disappointing 2015 at +1.5% y-o-y Perret Associates Page 32

33 We expect strong growth in cement production until 2030, although we have decreased our forecast for average growth by 1-2% from our previous forecast. We still expect average yearly growth of 5.8% during As a result, our forecast for coal usage for cement production has decreased as follows: 2016: down 3m tonnes from our previous forecast, at 43m tonnes. 2020: down 9m tonnes at 51m tonnes. 2030: down 27m tonnes at 90m tonnes. Exhibit 34 India, monthly cement production, m tonnes Source: Cement Manufacturers Association, Perret Associates Confirming the robustness of the Indian economy, steel production reached 70.15m tonnes during Jan.-September 2016, up 2.8m tonnes or 4.1% y-o-y. Unlike cement production, which was much firmer during the first half of the year than so far in the second, steel production surged by an average of 8.7% y-o-y during Q316 vs. average growth of 1.9% during H116. We increase slightly our 2016 full-year steel production forecast by 1m tonnes or +3.9% to 93m tonnes. On the other hand, we decrease our growth forecast by between 1.5% and 3% during the period vs. our previous forecast. We are still expecting average growth of +4.4% during the period, declining from 4.9% in 2017 to 3.5% in Exhibit 35 Indian steam and coking coal demand, forecast, m tonnes Summary demand Steam coal for electricity generation m tonnes as % of total steam coal % 76.6% 80.4% 80.3% 79.8% 79.5% 79.2% 78.9% 78.6% 78.2% 77.8% 77.4% 77.1% 76.8% 76.6% 76.3% 75.8% Steam coal for heating, residential + other en m tonnes as % of total steam coal % 4.8% 5.1% 5.2% 5.2% 5.3% 5.3% 5.4% 5.4% 5.5% 5.5% 5.5% 5.5% 5.4% 5.4% 5.3% 5.3% Steam coal for cement production m tonnes as % of total steam coal % 5.3% 5.6% 5.5% 5.6% 5.7% 5.9% 6.1% 6.3% 6.5% 6.8% 7.1% 7.3% 7.6% 7.9% 8.2% 8.6% Steam coal for other industries m tonnes as % of total steam coal % 13.3% 8.9% 8.9% 9.4% 9.5% 9.6% 9.7% 9.8% 9.8% 9.9% 10.0% 10.0% 10.1% 10.1% 10.2% 10.3% Total steam coal m tonnes ,009 1,027 1,045 1,055 as % of total % 87.3% 87.3% 87.1% 87.0% 86.8% 86.4% 86.2% 86.1% 85.9% 85.8% 85.6% 85.4% 85.2% 85.1% 84.9% 84.6% Total coking coal m tonnes as % of total % 12.7% 12.7% 12.9% 13.0% 13.2% 13.6% 13.8% 13.9% 14.1% 14.2% 14.4% 14.6% 14.8% 14.9% 15.1% 15.4% Total steam + coking coal m tonnes ,005 1,031 1,058 1,085 1,112 1,136 1,160 1,184 1,207 1,231 1,247 Source: Perret Associates 2016 Perret Associates Page 33

34 2.2.5 Further growth in domestic crude coal production caps imports Indian crude coal production reached 522.6m tonnes during Jan.-October 2016, up 8.7m tonnes or 1.7% y-o-y. However, the year so far has been split into two very different periods, mirroring the differences in the evolution of power generation, mentioned above. Indeed, crude coal production increased by 4.6% y-o-y on average during Jan.-July 2016, before falling by an average of 5.5% during August-October. The slowdown was not the consequence of logistics issues in the supply chain, but rather of a slowdown in total and coal-fired generation, as examined above. By the same token, we are assuming that Coal India, the largest mining company in the country, should be able to ramp up production again relatively quickly, if necessary. Nevertheless the current growth rate in crude coal production is much lower than our latest forecast of 5.9%, which we adjust downwards to 2%. We anticipate crude coal production will increase by only 11.4m tonnes in We maintain our crude coal production forecast unchanged from 2017 onwards. We still don t think that the government s very ambitious total crude coal production target of 1.5bn tonnes by 2020, of which 1bn tonnes from Coal India and 500m tonnes from other producers, will be achieved. The main reason for this is that we don t think India will actually require such quantities of coal. The cancellation of 16GW of potential additional coal-fired capacity is a good illustration of this. Exhibit 36 Forecast Indian steam and coking coal production, m tonnes Indian coal production Steam coal production m tonnes ,002 Coking coal production m tonnes Total crude coal production m tonnes ,000 1,020 1,040 1,060 1,080 1,100 Source: Perret Associates Indian steam and coking coal imports forecast According to preliminary customs and shipping data, we estimate that Indian steam coal imports fell by 13m tonnes or 9.8% during Jan.-October 2016 to 119m tonnes. Imports have declined steadily during the year, as they were already down 10m tonnes during Jan.-May We now forecast imports will fall to 135.9m tonnes in 2016, which would be 24.5m tonnes or 15% lower y-o-y. This would also confirm our view that Indian steam coal imports peaked at 170m tonnes in We expect this decline in imports to continue after The government announced in November 2016 that state-owned power generators have agreed to stop all steam coal imports from 1 April 2017, marking the beginning of the new fiscal year. We estimate that 2016 Perret Associates Page 34

35 state-owned coal power plants will account for 15m tonnes of hard coal imports during the fiscal year , compared with 20m tonnes during the fiscal. Our forecast for Indian imports has decreased further to just above 100m tonnes in and then below this level from 2021 onwards. This forecast assumes a slower rise in crude coal production than the government s target. There is potential downside of our coal imports forecast if domestic production were to increase faster than our forecast. Exhibit 37 Indian steam and coking coal imports, m tonnes Indian coal imports Steam coal imports m tonnes Coking coal imports m tonnes Total coal imports m tonnes Coking coal Steam coal Source: Perret Associates 2.3 Japan Japanese steam coal imports fell by 4.5m tonnes or 5.1% during Jan.-September 2016 to 84.5m tonnes. The fall accelerated during June-August, with an average yearly fall of 14.4% over the period. This is in line with our latest edition as we anticipated a yearly fall of 4m tonnes in imports in Looking at the origins, the fall was systematic across the board, although we saw some disparities. Japanese imports of Australian coal decreased by the same proportion as the total, which is logical as Australian imports have accounted for 72% of the total so far this year, in line with the average of 73.5% during the period They fell by 3.4m tonnes or 5.4% during Jan.-September 2016 to 60.8m tonnes. Indonesian coal was more severely hit, at least in percentage terms, as its exports to Japan fell by 8.7% or 0.9m tonnes y-o-y to 9.1m tonnes. Consequently, Indonesian coal s share of Japanese imports fell to an average of 11.4% so far this year and even reached an historical low of 6.1% in September This confirms the structural decline in Indonesia s share of Japanese imports, which peaked at 20.5% in Perret Associates Page 35

36 This development was exacerbated this year by the renewed strength in Chinese imports, which benefited mostly to Indonesian coal as well as the lack of availability, due to increasing Indonesian domestic demand. Russian coal, on the other hand, has been confirming its place as structural exporter to Japan and North East Asia. Japanese imports of Russian coal eroded by only 2.4% or 0.2m tonnes during Jan.-September 2016, at 7.5m tonnes. Consequently, Russian coal s share of Japanese imports has remained stable from 2015 at an average of 9.3% during Jan.- September 2016 and 11.6% during June-September only. Exhibit 38 Japanese steam coal imports with split per origin, m tonnes 12 Australia Indonesia Canada Russia China USA South Africa Source: Ministry of Finance, Perret Associates In the absence of updated data on the Japanese electricity mix, it is difficult to draw definitive conclusions. At this stage, we are assuming that the long term trend of declining total electricity generation is due to the persistently poor economic growth. Thermal and coal-fired generation have simply been following the trend. As a reminder, coal usage by Japan s ten largest power utilities declined by 0.4m tonnes or 2.2% y-o-y during Q116 to 16.05m tonnes. In this decline, coal tracked the overall negative trend in total power generation, which fell to 230.5TWh in Q116, down by 2.4% y-o-y. In fact, Japanese power generation has fallen almost every month on a yearly basis since Jan.2014, in a reflection of the sluggish state of the Japanese economy. Still, we anticipate Japanese steam coal imports will receive some support. Firstly, at the time of writing, there were still only 2 nuclear power plants in operation. This should remain the case during the winter due to the increasing level of local opposition to the restarting of additional plants. For instance, on 9 March 2016 the Otsu District Court ordered the utility Kansai Electric to shut down the two reactors it had started just a few weeks earlier. This was due to a petition from local residents who were concerned about the lack of safety measures. The decision was confirmed by another court on 20 June Perret Associates Page 36

37 In November 2016, another earthquake in the Fukushima area (7.4 magnitude) acted as a reminder of the vulnerability of the Japanese nuclear capacity. As things stand, of the 54 nuclear reactors in activity before Fukushima: 12 have already been decommissioned. 2 have restarted and are in operation. 6 to 10 reactors are considered as likely to restart. However, the rest of the fleet might never be restarted, if the recent example of Kansai Electric is anything to go by. In view of the uncertainty surrounding the Japanese nuclear sector and the risk that LNG imports might remain expensive, coal usage is likely to increase further in the next decade. Japan is planning full deregulation of its electricity market by 2020, which will trigger the commissioning of small new coal-fired power plants, in particular for captive supply to heavy industry. The environmental constraints for such small power plants are less strict than for large ones, which supply electricity to the grid, and this makes their financing and construction more likely. At this stage, 4 such coal-fired power plants, with a combined capacity of 1.9GW, are already under construction. In February 2016, in an interesting twist which potentially confirms the long term support for coal power generation in Japan, the Environment Minister Tamayo Marukawa, who had initially vetoed three of these power plants, due to concerns about CO2 emissions, announced that he would support them after all. Japan intends to keep its leading position in coal-fired power plant technology. In February 2016, Tohoku Electric submitted a construction plan for the 600MW coal-fired plant Noshiro No. 3, with an expected commissioning date in June The plant has been designed to use primarily sub-bituminous coal and could still achieve efficiencies of 44.8% by using so-called Ultra Super Critical Technology. An additional 49 plants with a combined capacity of 28GW are at different stages of planning. For instance, the construction of the 112MW co-combustion Soma Core Industrial project plant in Fukushima is expected to start in the coming months. In October 2016 Mitsubishi announced its intention to build two coal power plants of 540MW each using the Integrated Gasification Combined Cycle technology. Although nothing is certain, the government is assuming that at least 10GW of new coal-fired capacity will be commissioned over the next 5 years. We are assuming a more conservative 5GW in our Base Case scenario. We maintain our imports forecast unchanged from the previous edition Perret Associates Page 37

38 Jan-14 Mar-14 May-14 Jul-14 Sep-14 Nov-14 Jan-15 Mar-15 May-15 Jul-15 Sep-15 Nov-15 Jan-16 Mar-16 May-16 Jul South Korea South Korean electricity generation increased only marginally (by 0.7%) during Jan.-July 2016, at 303TWh. Still, this is an improvement from 2015, when it declined by 0.7 % y-o-y. However, coal and more generally thermal generation didn t benefit from this growth. Indeed, steam coal usage for electricity generation eroded by 0.9m tonnes or 1.9% over the period to 46.5m tonnes. This is equivalent to 2.6TWh fall in coal-fired generation at 135.3TWh. This also marks a reversal of the recent trend, as coal usage for power generation increased by 1.3m tonnes or 1.6% in Gas generation fell by about the same amount in absolute terms (2.4TWh) during Jan.-July to 53.9TWh, although the fall was even more significant in percentage terms (4.3%). Exhibit 39 South Korean electricity generation per source, TWh Renewables Hydro Nuclear Gas Coal 0 Source: KESIS, Perret Associates The fall in thermal generation was matched almost exactly by 5.6TWh or 6% increase in nuclear generation to 100TWh over the same period. Still, the second half of the year might well be very different as we are expecting a fall of up to 30% in nuclear generation between September and December Indeed, a planned maintenance programme was already under way in mid-september and on top of this an additional 4 units were taken off-line after a severe earthquake affected the country. We estimate that the loss in nuclear generation during September-December could reach 4-6TWh per month. On that basis, coal usage could increase significantly as 1TWh of coalfired generation requires 2.9m tonnes of coal in South Korea, according to our calculations. That said, looking back, the highest ever coal usage per month was 7.7m tonnes (January 2015). Coal-fired power plants have typically run at very high or even full capacity historically and net coal-fired generation capacity has increased only marginally recently to reach 31.7GW at the end of August Even full capacity utilisation equates to a maximum monthly usage of m tonnes Perret Associates Page 38

39 In other words, there should not be a direct benefit for coal from the expected slowdown in nuclear generation, due to capacity constraints. It should therefore fall to gas and LNG-fired power generation to fill the gap, at least to some extent. South Korean steam coal imports declined by 3.1m tonnes or 3.1% during Jan.-October 2016, at 88.6m tonnes. Exhibit 40 South Korean steam coal imports, m tonnes Australia Indonesia Russia Canada SA China US Colombia Others Source: Customs, Perret Associates We have highlighted in previous reports the trend that started more than 4 years ago of rising imports of Australian coal and falling imports from Indonesia. Just to put things in perspective, imports of Australian coal have increased from 29.6m tonnes in 2011 (25% of total imports) to 37.6m tonnes in 2015 (34% of total). Meanwhile, imports of Indonesian coal fell from 40.2m tonnes in 2011 (38% of total) to 33.9m tonnes in 2015 (31%). This trend stabilised during the first 10 months of 2016 as Australian coal has been the main victim of the decline in total imports. Imports of Australian coal decreased by 1.8m tonnes or 6.1% y-o-y during Jan.-October. Meanwhile, imports of Indonesian coal remained flat y-o-y at 29.2m tonnes. The resurgence in imports of Indonesian coal took place in particular during August-September 2016 (+37.5% y-o-y) while the fall in Australian imports has been so far focused during June- October 2016 (average fall 20.5%). We don t have a clear explanation for this, as according to our previous research the introduction of the three-tier tax system (vs. two previously) seemed to favour imports of medium calorific value Australian coal and not low calorific value Indonesian coal. One explanation might be the lack of availability of Australian coal during the summer period, which coincided with the sharp increase in Chinese imports. (Chinese imports of Australian coal increased by 28.6% y-o-y during July-August). Meanwhile, Russia is confirming its position as number 3 exporter to South Korea. Imports of Russian coal were flat during Jan.-October 2016, at 13.7m tonnes Perret Associates Page 39

40 Taking a longer term view, the country s coal-fired capacity will go through a serious shakeup over the next decade, with the simultaneous closure of old plants and opening of new ones. The South Korean government has signalled its determination to combat air pollution by declaring its intent to tackle inefficient coal-fired power plants. It plans to invest $8.6Bn to improve or shut down power plants in an aim to reduce air pollution by 24% by Ten coal power plants which are over 40 years old, with a combined capacity of 3.3GW, are scheduled to be shut down by On the other hand, the country could commission up to 20 new small coal power plants, with a combined capacity of 10GW by The government also plans to retrofit 8 coal power plants which are over 20 years old, to limit dust particle emissions. Overall, we decrease our imports forecast for 2016 by 2.5m tonnes from the previous edition to 107.5m tonnes (also down 2.5m tonnes vs. 2015). We maintain our imports forecast unchanged from 2017 onwards. 2.5 Other countries Taiwan Taiwanese steam coal imports have remained on a downwards track so far in 2016, falling by 3.1m tonnes or 6% y-o-y during Jan.-September In fact, the fall is accelerating, as it follows a decline of 0.3% in 2014 and Indonesian supplies have been hardest hit, most with imports from this origin falling by 3.9m tonnes or 19.9% over the period to 15.5m tonnes. Increased Chinese demand for Indonesian coal, as well as a lack of availability, might be the main reason for this. Imports of Australian coal were flat during Jan.-September 2016 at 24.3m tonnes. Imports of Russian coal increased marginally by 0.3m tonnes or 5.7% to 5.8m tonnes. We are revising our imports forecast further down as follows: 2016: 62.5m tonnes vs. 65.5m tonnes in our previous edition. This would represent a net fall of 4.4m tonnes y-o-y. 2020: 66m tonnes, down from 68m tonnes in our previous edition. 2025: 72m tonnes, down from 74m tonnes in our previous edition. 2030: 72m tonnes, down from 74m tonnes in our previous edition Perret Associates Page 40

41 2.5.2 Thailand Thailand s total electricity kept rising strongly during Jan.-September 2016 at 151TWh, up 4.5% y-o-y. Coal-fired generation increased even faster, by 9.5% y-o-y, to 28TWh. This is a noticeable trend reversal from 2015 when total power generation increased by 3.3% whereas coal-fired decreased by 8%. Coal benefited from a small decline in gas-fired generation to 96TWh, down 1.4% y-o-y. Gas remains the main source of electricity in Thailand but given its dominant market share even a small decline in gas-fired generation, in percentage terms, equates to significant additional coal demand. Exhibit 41 Thailand electricity generation, per source, TWh Coal Gas Oil and Diesel Hydro Renewable Source: Ministry of Energy, Perret Associates The recovery in coal usage is primarily the consequence of the gradual return on-line of 4 of the 13 units at the country s biggest state-controlled utility, EGAT s, 2.4GW Mae Moh plant. As part of the upgrade the efficiency will increase to 37%. The 1.4GW coal power plant operated by BLCP Power, as well as the commissioning of smaller coal power plants across the country, also contributed to the surge in coal-fired generation. Still, a significant part of this increase was fuelled by domestic lignite. Consequently, imports of hard steam coal increased but by a smaller percentage than the overall increase in coal-fired generation, at 5.4% y-o-y during Jan.-September 2016, taking them to 19m tonnes over the period. Indonesia remains the country s leading origin by some margin, due to its proximity, with imports from the country accounting for 68% of total imports so far this year. Still, this has fallen from a stable average of 85.6% during Digging further into the detail, we note that Indonesian sub-bituminous coal is outstripping bituminous for the second year in a row Perret Associates Page 41

42 Imports from Australia increased to 4m tonnes during Jan.-September 2016, up 1m tonnes or 31.8% y-o-y. The recent upgrades of the power plants might explain this recent development, as new boilers are typically designed to burn a wider range of coal qualities, providing more possibilities for blending. Utilities might be able to optimise their blending mix by combining low quality Indonesian coal with higher quality Australian coal. Exhibit 42 Thailand steam coal imports, m tonnes Indonesia Australia Source: Customs, Perret Associates Going forward, the prospects for steam coal imports are positive. EGAT, the largest power producer in the country, is currently only operating its coal power plants on domestic lignite. However, the company plans to commission 5.8GW of coal-fired power plants that will operate on imported coal. This reminds us of the evolution of some Turkish utilities. Of this 5.8 GW, the 800MW Krabi plant should be commissioned by December 2019 and the 2GW Therpha plant should be commissioned by December We adjust our imports forecast as follows: 2016: 25m tonnes vs. 27.5m tonnes in our previous edition (up 2m tonnes y-o-y). 2020: 29m tonnes vs. 32m tonnes. 2025: 32m tonnes vs. 35m tonnes. 2030: 35m tonnes vs. 37m tonnes Malaysia Malaysian steam coal imports increased by 0.6m tonnes or 6.3% during Jan.-August 2016 to 10.2m tonnes. This follows an already strong increase of 8.5% in The 1GW Tanjung Bin Energy (T4) has been coming on-line gradually over the period, with a potential annual requirement of 3m tonnes per year when fully operational. In addition, we expect 5GW of new coal-fired capacity to be commissioned within the next 5 years Perret Associates Page 42

43 Nevertheless, we decrease by 600KT our imports forecast for 2016 to 25m tonnes. We maintain our forecast unchanged thereafter Vietnam Vietnamese gross steam coal imports increased by a staggering 6.1m tonnes during Jan.- September, at 10.6m tonnes. At this pace, full-year imports could reach 14.5m tonnes, well in line with our forecast. The additional 3.4GW of coal-fired capacity commissioned by the state-owned utility EVN is now in full operation, with an equivalent coal usage of 10m tonnes per year. Despite its proximity, Indonesia, which used to be the largest exporter to Vietnam, has dropped back into third position this year, with 2.1m tonnes. Australia has taken the lead, with 3.4m tonnes, followed closely by Russia at 3.2m tonnes. Looking ahead, more companies will increase their coal imports. In November 2016 the state-owned oil and gas company PetroVietnam announced the signing of a 12m tonnes imports contract for three coal-fired power plants: 1.2GW Long Phu 1, due to start in GW Long Phu 3, due to start in GW Song Hau 1, due to start in We maintain our imports forecast unchanged from the previous edition Pakistan Steam coal imports from Pakistan increased by 22% or 0.6m tonnes during Jan.-July 2016 to 3.3m tonnes. This is the consequence of the gradual commissioning of the coal-fired power plants we mentioned in our previous edition. South Africa s share of imports averaged 78.2% over the period, in line with recent years (average 76.2% during ), but still much higher than the 37.2% during the period Of the potential 11GW of coal and lignite-fired capacity to be built by 2020, we estimate that 7.4GW could be based on imports. In September 2016, the Water and Power Ministry announced it was reconsidering the solid fuel sourcing for some of these projects, with the aim being to favour domestic coal instead of imports. This applies in particular to the Port Qasim 350MW Siddiqsons and 660MW Lucky Electric projects, for which the government has asked the developers to consider the option of switching entirely to domestic production. These plants are located close to the domestic Thar coal reserves, which in theory could supply them. Various feasibility studies are under way to assess the exact potential of the Thar basin in terms of quality and long term availability. The quality of the coal is in general 2016 Perret Associates Page 43

44 poor, closer to lignite, which might not be acceptable for brand new boilers, which were initially designed for higher quality imported coal. Exhibit 43 Pakistan coal and lignite power projects Project developer Location Capacity MW Coal source Commissioning date Huaneng Shandong Ruyi Energy Sahwal 1,320 Imported Dec-17 Engro Powergen Thar Limited Thar, Sindh 660 Domestic Oct-18 Sinohydro Resources Limited, Al Mirqab Capital Port Kasim, Karachi 1,320 Imported Jun-18 Shanghai Electric Thar, Sindh 1,320 Domestic Q119 HUBCO/CPIH China PowerHub Generation Balochistan 1,320 Imported 2019 Lucky Electric Power Company Port Kasim, Karachi 660 Imported 2020 Siddiqsons Port Kasim, Karachi 350 Imported 2020 Grange Power Arifwala, Punjab 163 Imported Jan-19 Oracle Coal Fields PLC England Thar, Sindh 1,320 Domestic Q219 Kot Addu Power Kot Addu, Muzaffargarh 660 Imported Late % imported Jamshoro Power Jamshoro 1, % Domestic 80% Imported Lakhra Power Lakhra 660 NA 20% Domestic Total 11,073 Source: Ministry of Planning and Development, Perret Associates The government has also asked the developer of the 650MW Lakhra project to consider using 100% domestic coal, as opposed to the initially proposed 80%/20% split between imported and domestic coal. Pakistan s imports are on track to reach our forecast of 6m tonnes in We maintain our imports forecast unchanged from the previous edition, until we get more clarity on this potential switch towards higher usage of domestic coal Philippines The Philippines gross steam coal imports (the country is also a coal exporter), keep rising steadily. They reached 9.8m tonnes during Jan.-July 2016, up by 1.3m tonnes or 16% y-o-y. Our imports figures are only based on Indonesian coal at this stage, which typically account for the vast majority of imports. We might start seeing some Australian coal being delivered to the Philippines, as the government announced in May 2016 its intention to cut the import tax on Australian coal from the current 5% to 0% by This is in accordance with the ASEAN Trade Agreement signed by the Philippines in Imports have been growing without interruption since our records started in 2006, when they stood at 4m tonnes. The outlook for coal imports is positive, as additional coal-fired capacity will be commissioned in the coming years. In the meantime, the domestic coal supply is facing some uncertainty Perret Associates Page 44

45 The government has reiterated its positon that, in view of the strong growth in electricity demand expected over the next decade and the fact that domestic power prices are amongst the highest of developing countries in South East Asia, the Philippines could not afford not to have coal. Renewables and coal-fired capacity are currently at about the same level of 6GW. Of these renewables, wind and solar account for 2.9GW, the rest being provided by geothermal and hydro. There is a significant pipeline of planned coal-fired capacity additions across the country, which accounts for 70% of the new capacity under discussion. The new coal units will be mainly small ones, as they have to supply electricity in all parts of the vast archipelago. For instance, in August 2016 the government approved the commissioning of a 135MW coal-fired unit that will supply the Visayas grid. The grid has a total capacity of 1.7GW and supplies 6 islands via underwater cables. This was the first new coal unit commissioned since A total of 6.8GW of new coal-fired capacity could be commissioned over the next 10 years. If it is achieved, coal s share of total generation could reach 44% from the current 34%, according to the government. On the coal supply side, the government made a volte-face in August Initially, it approved the largest domestic mining company, Semira Mining and Power (SMPC) s, expansion of the Molave open pit mine. The new pit was supposed to produce an additional 2mtpa of coal, which would have helped Semira to meet growing domestic demand. The company produced 8.1m tonnes in 2015, short of its target of 8.5m tonnes, due to a twomonth suspension after a fatal accident. The government s U-turn is part of a tougher line on the environmental consequences of mining in general, including nickel. We mentioned in our previous report that Semirara had increased its output target to 9.5m tonnes for It won approval in February 2016 to increase its capacity to 12m tonnes per annum in the coming years, which now looks in question. This might also have a negative impact on gross coal exports, mainly of lignite to China. We increase our imports forecast for 2016 from 16m tonnes to 17.5m tonnes. We maintain our imports forecast unchanged thereafter. 2.6 EU 15 imports forecast We keep decreasing our import forecast for EU15 countries at 105m tonnes in 2016 vs. 111m tonnes in our previous edition. The decline is general across the board for most countries in particular: The UK, down 1.5m tonnes from our previous edition. Italy and Spain down each 2m tonnes. The Netherlands down 1m tonnes Perret Associates Page 45

46 Denmark down 0.8m tonnes. We maintain our forecast for German imports unchanged and increase French imports by 1.6m tonnes due to the nuclear outages. Going forward we maintain our forecast of a structural decline in EU15 imports. We decrease our import forecast from our previous edition by: 13.2m tonnes in 2020 at 88.8m tonnes. 5m tonnes in 2030 at 69m tonnes. Exhibit 44 EU 15 seaborne imports steam coal imports, m tonnes Germany (only seaborne) UK Italy Spain Netherlands net France Portugal Finland Denmark Belgium Total Source: Perret Associates Germany After a robust first half of the year, when German gross steam coal imports (including by rail from Poland and Central Europe) were flat y-o-y, they fell quite sharply during the summer. As a result, imports fell by 0.7m tonnes or 2.2% y-o-y during Jan.-September 2016 to 29.9m tonnes. We also note the increasing volatility of German steam coal imports (up 37.6% y-o-y in February 2016, down 20.1% y-o-y in September), which also reflects the increasing volatility of coal usage. Indeed, German coal-fired generation has been yo-yoing between +28.2% y-o-y in January 2016 and -20.7% y-o-y in October. This was due to direct competition from equally volatile renewables generation, to which we have now become accustomed, but also an increase in gas-fired generation since July, which itself stemmed from the sharp increase in coal prices delivered Europe. German hard-coal fired generation fell by 11.8% y-o-y on average during July-September. This cancelled out the strong start to the year, leaving total generation during January- September flat y-o-y at 86.7TWh. Gas-fired generation followed the exact opposite path, as it fell by 10.5% y-o-y during Q116 and increased by 85.2% y-o-y during August-October. As a result, gas-fired generation increased by 3.4TWh or 14.8% y-o-y during Jan.- September 2016 to 26.6TWh. Given the additional supply of LNG to be delivered in Europe from 2017 onwards, gas could continue to encroach on coal s market share, if coal prices remain at current levels Perret Associates Page 46

47 Looking at other fuels, lignite generation continues its inexorable decline, down 7.3% y-o-y during the same Jan.-September period, at 108.6TWh. In fact, the decline is speeding up, and compares with -3% in 2014 and -0.9% in Nuclear generation keeps falling, for the sixth year in a row, down 8% y-o-y during Jan.- September at 66.05TWh. At this pace, nuclear generation might fall to 80TWh in 2016, which would be about half its 2006 level of 158.7TWh, before the global financial crisis, Fukushima disaster and subsequent EnergieWende. Wind generation increased by 9.1% over the period to 64.45TWh, in particular due to strong winds during Jan.-May 2016, when generation rose 23% y-o-y). Exhibit 45 German electricity generation per source, TWh Hard coal Lignite Gas Nuclear Wind Solar Biomass Hydro Source: Destatis, Fraunhofer ISI, Perret Associates In further evidence of the rebalancing of the German electricity mix, wind generation during Jan.-September was almost equal to nuclear at 66.05TWh and might even outstrip it for the entire year, should there be strong winds in Q416. This is not far off coal-fired generation of 86.7TWh and looks set to match lignite within a few years, given the opposite trend of the two sources. The increase in solar generation remains more muted, despite a strong performance during the summer. Overall, solar generation increased by 2.3% y-o-y to 34.4TWh. Looking at longer term developments, in November 2016 the German power producer, STEAG, announced its decision to shut down 5 coal-fired power units with a total capacity of 2.3GW: Voerde West 1 (1971, 322MW) and 2 (1971, 318MW), to be closed by September Herne 3 (1966, 280MW) Bexbach (1983, 721MW) Weiher (1976, 655MW). This would represent 32% of the company s current 8.2GW of coal-fired capacity and 10% of Germany s 29.6GW total coal-fired capacity Perret Associates Page 47

48 The disinvestment is not really a surprise as the plants were more than 40 years old on average and relatively inefficient, with an efficiency of 32-35%, according to our estimates. The more difficult economic situation since 2013, with the sharp fall in power prices, has put such power plants under increasing financial pressure, not to mention the additional CO2 cost and the increased competitiveness of gas-fired power plants. By comparison, the latest generation of coal-fired power plants commissioned by STEAG in 2013, such as the 725MW Walsum 10 unit, has an efficiency of 45%. This confirms our view that, as technology stands, there will be no new coal-fired power plants built in EU 15 countries. We are not even sure that potential investors would have appetite to build new coal-fired power plants with proven Carbon Capture and Storage technology. In fact, we estimate that an additional 650MW of STEAG s coal-fired capacity, which was commissioned in the 1960s and 1970s, might be closed in the coming years. In a sign of things to come, STEAG announced it was investing in 450MW of new onshore wind capacity, in which the company has already invested 500m. It is also continuing to invest in its battery project (already 100m invested), with 15MW of storage capacity. As reported in our previous edition, RWE is planning to close its 1.4GW Voerde coal-fired power plant by 1 April The plants two units were built in the early 1980s. After the EnergieWende, the Climate Protection Plan 2050 Taking a longer term view, just as the market is adapting to the already very ambitious EnergieWende announced in 2011, after the Fukushima disaster, an even tougher challenge, in the shape of the so-called Climate Protection Plan 2050 (CPP 2050), is now under review. The CPP 2050 has been produced as a consequence of the COP21 Paris conference in December Chancellor Merkel s government initially dragged its feet regarding this new proposal, but finally approved draft legislation in November Crucially, the introduction of the legislation is subject to an impact assessment in 2018, after the general election in the second half of This means that a potentially new coalition government will have to make the final decision on the CPP Mrs Merkel has announced she will stand for a fourth consecutive mandate in As things stand, the objectives of the CPP 2050 are as follows: Greenhouse gases emissions from the energy sector should fall to m tonnes of CO2 equivalent by 2030, or 61-62% of 1990 s level. More crucially, the benchmark year reference is brought forward to The cuts in GHG should represent a 48%-51% reduction from 2014 s levels. These 50% cuts vs would be a severe blow for the energy sector, and possibly spell the end of lignite and potentially hard coal-fired generation Perret Associates Page 48

49 Indeed, according to the German environmental office UBA, the power sector emitted 315m tonnes of CO2 in 2014, accounting for 88% of total energy sector emissions (358m tonnes). Of this: Lignite accounted for 50.5% or 159m tonnes. Hard coal accounted for 30.5% or 96m tonnes. Gas accounted for just 6.3% or 19.8m tones. The estimated emissions from these three sources declined only marginally in 2015 to 312m tonnes. In order to achieve the target cuts, UBA estimates that lignite generation would have to be cut by at least 70% by 2030 from 2014 s levels, and potentially altogether. The future of hard coal would depend to a large extent on what happens with lignite generation. However, if the CCP 2050 plan were to be adopted as it stands, gross hard coalfired generation would have to be cut to 45-50TWh in This would represent just 43-45% of the 105TWh hard coal-fired generation in German seaborne steam coal imports (excl. Poland) combined with domestic hard coal production, were as follows during : 2013: Imports 32.9m tonnes + production 7.6m tonnes = 40.5m tonnes. 2014: Imports 32.6m tonnes + production 7.6m tonnes = 40.2m tonnes. 2015: Imports 36.7m tonnes + production 6.2m tonnes = 42.9m tonnes. This yields average annual steam coal usage over the period of 41.2m tonnes. The remaining German hard coal mines will be shut down in the coming years and definitely by Thus, all steam coal usage will be based on imports. Based on current assumptions, German steam coal imports could fall to 17-19m tonnes by There are still a lot of uncertainties concerning the application of the CPP Still, we are reducing our imports forecast as follows: 2016: 36m tonne vs. 36.5m tonnes in our previous edition. 2020: 33m tonnes vs. 36m tonnes. 2030: 20m tonnes vs. 28m tonnes UK UK coal usage for electricity generation continued to plummet during the first 9 months of 2016, totalling just 8.6m tonnes, down by a massive 13.8m tonnes or 61.5% y-o-y. This is the consequence of the double whammy impact of the halving of coal-fired capacity, from 23.1GW in December 2010 to 12.5GW in December 2015, combined with the increasing competitiveness of gas-fired generation vs. coal, itself a consequence of the strong rally in coal prices. We now forecast coal usage for power generation will fall to 10.1m tonnes in 2016, down 19.1m tonnes or 65% y-o-y. This would be lowest level by far since our records started in 2016 Perret Associates Page 49

50 1995. To put things in perspective, this compares with a yearly average of 47.8m tonnes during and a peak of 57.4m tonnes in Predictably, steam coal imports have also fallen to levels unseen in recent history and sometimes too small to be even considered, such as 110KT in April Imports averaged 380KT per month during Jan.-September 2016, with a total of 3.4m tonnes, down 10m tonnes or 75% y-o-y. The UK s decision to leave the European Union will not bring massive changes as far as the UK coal market is concerned. The carbon tax at per tonne of CO2 emitted shall remain and if anything might even be increased. In November 2016 the government confirmed its intention to close all coal-fired power plants by 2025, although a few could remain in operation until 2030, if required. We are revising our imports forecast sharply lower from our previous edition as follows: 2016: 5.5m tonnes vs. 7m tonnes (down 12m tonnes y-o-y). 2020: 4m tonne vs. 6m tonnes. 2030: unchanged at 3m tonnes. On a cautionary note, there might be some upside to UK monthly imports in the coming months, in particular in the event of a severe winter. This would be mainly due to the historically low current levels Netherlands Dutch coal usage for power generation declined during Jan.-August 2016 at 10.6m tonnes, down 1.3m tonnes or 10.6% y-o-y. This contrasts with the sharp growth in coal usage for power generation during the period 2010 (12m tonnes) until 2015 (17.95m tonnes). The growth period coincided with the addition of 3.4GW of gross coal-fired capacity, which peaked at 9.3GW in December However, as the new coal-fired capacity is now fully operational, old power plants have been gradually taken off-line, as part of the replacement programme, including: Amer 8 (Essent), 645MW. Borssele (Delta), 406MW. Gelderland-13 (GdF), 602MW. The Dutch and German coal markets, which had been moving more or less in tandem in recent years, are now disconnecting. In addition, very low European gas prices, in particular in the Netherlands, which is supplied by different origins, displaced some coal consumption. The trend of decreasing coal consumption and imports is set to continue as in September 2016 the Dutch parliament voted to cut CO2 emission levels by 55% before 2030 and by 25% before According to various studies this would require the definitive closure of the 5 remaining coalfired plants in the country Perret Associates Page 50

51 In any case, two E.ON coal-fired power plants, Maasvlakte I and II, each with a capacity of 520 MW, are to be closed by July However, in the longer term the situation is not as clear-cut, as closing all coal-fired plants would mean shutting down the three that were only just commissioned in They are amongst the most efficient in the world with an average efficiency of 46% so their closure might falter on the grounds of economics. A change in the government at the general election taking place in March 2017 could also change the outcome. Still, we decrease further our net steam coal imports forecast (Gross minus transhipments) as follows: 2016: 14m tonnes vs. 17m tonnes in our previous edition. 2020: 10m tonnes vs. 16m tonnes. 2030: 8m tonnes vs. 13m tonnes Italy Italian coal-fired generation also declined in As in most other Western European countries, the reasons for this included the temporary or permanent closure of coal capacity and strong competition from gas-fired generation. This confirms a structural trend, as Italian coal-fired generation and imports have been falling systematically since According to the national operator Terna, coal-fired generation fell by 3.8TWh or 10% y-o-y during H116. The 605MW Brinidisi 2 plant was taken off-line between February and August The 270MW Fiumesanto 4 was taken off-line between April and June The 500MW Brinidisi Nord 3 and 4 units, operated by A2A, have remained off-line since March 2015 and are not due to come back at least until December The Vado Ligure 3 and 4 units, with a combined capacity of 296MW, have been off-line since 2014 will be permanently closed. More closures are on their way as Enel, which has 5.6GW of coal-fired capacity, has announced the closure of the following plants: 133MW Genoa 3 by MW Bastardo plant by MW La Spezia 3 by Italy s total coal-fired capacity of 6.7GW could therefore fall to 5.5GW by Not surprisingly, Italian steam coal imports fell by 1.5m tonnes or 14.8% y-o-y during Jan.- August 2016 at 9m tonnes. We decrease our imports forecast vs. our previous edition as follows: 2016: 13m tonnes vs. 15m tonnes (down 3.5m tonnes from 2015) Perret Associates Page 51

52 2020: 12m tonnes vs. 13m tonnes. 2030: unchanged at 10m tonnes Spain After a very poor start to the year, Spanish total electricity generation has started to recover since the summer. Indeed, total electricity fell by 3% on average during the first 6 months of the year, but only at an average of 1.8% during July-October Still, total electricity generation fell by 5.9TWh or 2.8% y-o-y during Jan.-October 2016 to 207.4TWh. Exhibit 46 Spanish electricity generation by fuel, TWh Coal Gas Nuclear Renewables Hydro Source: Red Electrica de España, Perret Associates Coal-fired generation has also had a very poor year, in particular during the first half, when it fell by an average of 56.5% y-o-y. It has stabilised since then, falling by 25.7% during July- October. Nevertheless, it still fell by a total of 19.2TWh or 43% y-o-y during Jan.-October to 25.5TWh. At this pace, Spanish coal-fired generation could decline to 34.7TWh in 2016, a level equivalent 2009, in the aftermath of the global financial crisis and the resulting collapse in electricity demand. Typically, low coal-fired generation also corresponds to high hydro generation. The latter has increased systematically on a yearly basis so far this year, reaching 35.2TWh during Jan.- October 2016, up 12.8TWh or 57.2% y-o-y. If this keeps up, hydro generation could reach 38.7TWh in This would be a new record, pipping the previous record of 38.65TWh in Nuclear generation has been flat during the period at 48TWh, up 0.3% y-o-y Perret Associates Page 52

53 Interestingly, renewables generation declined by 6% or 5TWh during Jan.-October 2016, in particular due to low solar generation at the beginning of the year and low wind generation in September and October. As a result, Spanish renewables generation might fall for the third year in a row. Gas-fired generation has also fallen by 2.5TWh or 11.7% over the same period, in particular due to very low generation during the first half of the year. Not surprisingly, Spanish steam coal imports fell by 2.3m tonnes or 28.7% y-o-y during H116 to 5.7m tonnes. We expect a slight recovery from this low level during the second half of the year, but we revise downwards our imports forecast for 2016 from 15m tonnes to 13m tonnes. We maintain our imports forecast unchanged thereafter, expecting a slow overall decline in Spanish imports. Still, we should bear in mind that they should remain highly volatile, depending on the inevitable swings in hydro generation. As is the case in Germany, the EU Commission has confirmed the end of state subsidies for the remaining Spanish coal and lignite mines and their definitive closure by Spanish coal production fell to 3m tonnes in 2015, down from 3.9m tonnes in Turkey Turkish power generation keeps rising strongly, confirming the trend identified in our previous report. It totalled 206.5TWh during Jan.-September 2016, up by 9TWh or 4.6% y-oy. Power output rose on a yearly basis every month except for September, which might be a consequence of the failed coup attempt during the summer. Hard coal-fired generation (based on imports) benefited most from the surge in total demand, as it increased by 23.8% or 7.3TWh during Jan.-September. We expect this growth rate to increase, as in June 16 the power producer Eren brought on-line the first 700MW unit of its 1.4GW Zetes 3 plant. The second unit should be fully operational in Q416. We estimate that the country s hard coal-fired generation capacity will reach 7.5GW by December Electricity generated by domestic lignite also continues to benefit from the overall growth in demand, totalling 27.3TWh during Jan.-September 2016, up by 4.9TWh or 21.7% y-o-y. This is a very different picture from 2015, when lignite-fired generation fell almost systematically on a yearly basis. In fact, the Turkish government is now favouring domestic lignite-fired generation in order to limit its reliance on energy imports. Gas-fired generation, which is highly dependent on Russian imports, fell by 5.5TWh or 7.7% y-o-y during Jan.-September could see the second yearly fall in a row in gasfired generation. In August 2016 the government tried to limit hard coal imports as well, by introducing a surprise $15/t tax on hard coal imports from non-eu countries. This was partly to promote the usage of domestic lignite instead. However, the timing of the tax was not optimal, as it 2016 Perret Associates Page 53

54 coincided with the strong rally in delivered steam coal prices and would have imposed significant additional costs for private Turkish electricity generators. Exhibit 47 Turkey, electricity generation per source, TWh Hard coal Lignite Gas Liquid fuels Geothermal Hydro Wind Source: Turkish Electricity Transmission Company, Perret Associates After a short but probably intense spell of lobbying from Turkish end users, the government promptly back-peddled and in October 2016 came up with a new tax system. This was based on a fixed price of $70/t, from which would be deducted the settlement price of the prompt month API2 contract on the Exchange ICE Futures for the previous week. Given the price levels since September 2016, the theoretical tax has been minimal or non-existent. Consequently, hard coal and lignite s combined share of total generation reached a new historical high at 37.7% in September It also averaged 32.1% during Jan.-September 2016 vs. a relatively stable average of 27.8% during We anticipate the strong growth in hard coal and lignite-fired generation will continue at least until the end of the year as hydro generation fell by almost 10% during May-September. However, due to the strong start to the year, it is still up marginally by 0.7% at 53.8TWh during Jan.-September Turkish steam coal imports have continued to rise in line with higher generation, totalling 21.5m tonnes during Jan.-September 2016, up 2.3m tonnes or 11.8% y-o-y. Colombia has confirmed its position as leading supplier, with a 50% average share of Turkish imports during Jan.-September, even hitting 62% at one point. Colombian exports to Turkey increased by 42.2% y-o-y over the period to 10.8m tonnes. Russia remains the second biggest origin, accounting for 8m tonnes of Turkish imports over the period, up 14.4% y-o-y. On the other hand, imports from South Africa fell by 50% y-o-y to just 1.8m tonnes. The supply of South African coal remains tight, due to the shift in shipments to other destinations. Going forward, Turkish steam coal usage should keep rising at a rapid pace, as 4GW of additional coal-fired capacity has received clearance from the Ministry of Environment: 1.2GW for Tosyali Holding, near Iskenderun, to be completed in Perret Associates Page 54

55 900MW for Bendis Energy, near Iskenderun, to be completed in MW for Atakas Energy, near Iskenderun, to be completed in GW for Filiz Enerji, in Canakkale, to be completed by Turkish steam coal imports seem to be on track to hit our forecast of 31m tonnes in We maintain our imports forecast unchanged from our last edition. 2.8 Egypt We estimate that Egyptian imports have reached at least 1.4m tonnes during Jan.- September 2016, up 500KT or 55% y-o-y. These figures are just for South African coal, as the data for Indonesian and Russian deliveries, which may also have taken place, has been delayed. On the other hand, and somewhat surprisingly, only one Panamax (65,000 tonnes) of Colombian coal has been delivered to Egypt. At this rate we expect Egyptian steam coal imports to reach 2.5m tonnes in 2016, broadly in line with our previous forecast of 3m tonnes, which we readjust to 2.5m tonnes. In November 2016, the Egyptian Central Bank unexpectedly announced the removal of currency controls, which automatically linked the Egyptian pound to the US dollar. This step was a pre-requisite for the IMF granting a $12bn loan for investment in heavy industry and infrastructure in the country. As a consequence, the local currency lost 35% of its value within days. The idea is that the free-floating currency will stimulate foreign investment in coal-fired power plants and cement plants, amongst others. On the flip-side, coal imports will become more expensive. Still, if the coal is procured by the same foreign companies who have invested in the assets, the impact could be less significant. The government has not announced any changes to the plan to build up to 30GW of coalfired capacity by 2030, although we continue to take a much more conservative approach. Still, three coal-fired power plants with a combined capacity of 8GW are at an advanced engineering development stage, with construction likely to start within 12 months: 3.2GW Hamrawain, to be developed by Shanghai Electric Group (China). 2.4GW also in Hamrawain, to be developed by Dongfang Electric (China). 2.4GW Altour, to be developed by Elnowais. Another two projects have been approved, although commissioning will take place at a later stage: 2.4GW in Kafr Elsheikh, to be developed by ACWA. 2GW in Matrouh, to be developed by Marubeni/Elsewedy. We maintain our imports forecast unchanged Perret Associates Page 55

56 2.9 Other countries: Atlantic Israel Based on tentative numbers, we estimate that Israel s steam coal imports fell by 2.5m tonnes or 28% during Jan.-October Coal consumption by state-owned Israel Electric fell by 19% y-o-y during H116 to 4.3m tonnes. This is the direct consequence of a directive from the government published in December 2015, fixing a target to reduce coal usage by power plants by 15% in 2016 and an additional 5% in The fall in coal usage should be offset by an increase in gas consumption, which should account for 60% of the country s total electricity generation in Taking a longer term view, the Environment Ministry has asked for the definitive closure of 4 units of the Hadera Orot Rabin coal power plant, accounting for 1.5GW of the 2.6GW capacity plant. The government had initially envisaged the possibility of upgrading the units, in particular by installing new scrubbers, but it finally decided to dismantle and replace them with more efficient gas units. The four units should be closed by December 2022 and will be used after that as strategic back-up. In fact. it seems Israel Electric will still have to equip its plants with new scrubbers until they are shut down. Our previous forecast for Israel s imports already anticipated a large fall in imports. We adjust marginally downwards our forecast for 2016 to 8m tonnes vs. 8.5m tonnes in our previous forecast. We maintain our imports forecast unchanged thereafter Brazil We estimate that Brazilian steam coal imports increased by 1.1m tonnes or 24% y-o-y during Jan.-September 2016 to 5.8m tonnes. There are only two coal-fired power plants in Brazil, Pecem and Itaqui, with a total capacity of 1.45GW. They were commissioned in 2012, initially only to supply the Brazilian grid during the dry season. Hydro generation has historically been the main source of electricity in the country but with growing domestic consumption this proved insufficient, in particular during the dry season. Brazil has faced several consecutive years of drought, meaning that most dams are typically operating below their theoretical capacity. For example, hydro generation in the northeast grid totalled just 1.6TWh in September 2016 vs. 3.4TWh in September Hydro generation in the northeast has fallen systematically on a yearly basis, except in February. Heavy industry, with captive smaller coal-fired power plants, is also importing more coal, in particular from Colombia and the US Perret Associates Page 56

57 We increase marginally our imports forecast for 2016 from 8.5m tonnes to 9.5m tonnes. We keep our forecast unchanged for thereafter Middle East Jordan Jordan imported 0.2m tonnes of coal for cement production in Imports could increase to 0.4m tonnes in 2016 with the commissioning of a 30MW captive coal power plant for cement production in Qatraneh, in the south of the country. We had been expecting Jordan to only start importing steam coal in We are bringing forward and increasing slightly our imports forecast. Dubai In July 2016 Dubai announced that construction of the 2.4GW Hassyan coal-fired power plant was due to start in the coming months. Chinese manufactuer Harbin Electric and the US General Electric will be in charge of the construction. The first 600MW is due to be commissioned in March 2020, with three more units of 600MW expected to be commissioned in March of 2021, 2022 and After this first phase, two additional units of 600MW each are also contemplated in a second phase, which could bring the total capacity to 3.6GW at some point between 2025 and Once fully operational, the plant could consume up to 10m tonnes of coal per year. It is likely to remain the sole coal-fired power plant in Dubai and supply 7% of the city-state s electricity requirements by The French company Louis Dreyfus Ports and Logistics is in charge of building and managing the port facilities, which should be able to discharge Capesize vessels Perret Associates Page 57

58 3 Exports 3.1 Indonesia Further decline in exports According to the Indonesian Ministry of Energy and Mines, Indonesian crude coal production fell by 10.7m tonnes or 3.5% y-o-y during Jan.-September 2016 to 290.8m tonnes. In fact, though, it seems that 2016 might turn out to be a year of two quite different halves. Production fell steadily on a yearly basis during H116, by an average of 8.1% y-o-y. This was a continuation of the negative trend that started in April 2014 and was the consequence of the prolonged international bear market. However, crude coal production bottomed out and started to pick up in July. In fact, it increased by 7.9% y-o-y on average during July-September This was due to the rapid increase in Chinese imports that started in June Exports fell heavily by an average of 11.3% y-o-y during Jan.-April 2016, when we were still broadly in a bear market, even if prices were already starting to pick up, in particular in Europe. Exports bottomed out in May and started a recovery since. Exhibit 48 Indonesian steam coal exports, m tonnes Source: Customs, Perret Associates Overall, Indonesian exports fell by just 13.2m tonnes or 6.1% y-o-y during Jan.-July 2016 and we expect them to have recovered some ground since then, even if we only have partial data. In fact, just considering Indonesian exports to China, which made up about 20% of Indonesian exports in recent years, these increased by 53% between H116 (monthly average of 6.65m tonnes) and August-October 2016 (monthly average of 10.2m tonnes). This recovery in Indonesian exports was predominantly a Chinese affair. Indeed, exports actually fell to most destinations between H116 and Q Perret Associates Page 58

59 Indian imports of Indonesian coal eroded by 7% between H116 (monthly average of 8.4m tonnes) and Q316 (monthly average of 7.8m tonnes). Japanese imports of Indonesian coal fell by 19% between H116 (monthly average of 1.1m tonnes) and Q116 (monthly average of 0.9m tonnes). On the other hand, South Korean imports of Indonesian coal increased by 6% between H116 (monthly average of 2.85m tonnes) and Q316 (monthly average of 3m tonnes). In other words, as could have been expected, Indonesian mining companies remain opportunistic and are able to cope to some extent with an unexpected increase in imports from one of its traditional destinations, such as China. We also note that China has reclaimed its position as primary destination for Indonesian coal exports. Based on our information on the ground, it seems that most of the additional exports since June are from existing mines, where there is some latent capacity, after production and exports were reduced in More crucially, though, we are not getting the sense that new investment in production capacity will be made on the basis of a spike in Chinese demand and imports, which might prove to be temporary Domestic demand picks up The Indonesian government s plans for a massive increase in overall and coal-fired generating capacity in the coming years are beginning to take shape, as two large coal-fired plants on the island of Java have been granted financing. China Development Bank has granted a $1.8bn loan for the 2GW Java 7 plant, on which construction started back in April The project is a joint venture between PLN, the country s largest power producer and distributor, and the Chinese firm Shenhua Energy. An additional 1GW unit has been agreed and financed for the Cilacap plant, also in Java. The current 1.3GW unit is operated by PLN. The new unit will have an efficiency of 45% and could use up to 5m tonnes of per year. This is on top of the 2.4GW of new coal-fired capacity delivered in March Still, this falls short of the ambitious plans announced a few months ago by President Widodo of 35GW of additional capacity (about half of it coal-fired) to be commissioned by Structural erosion in Indonesian exports We are revising slightly upwards our Indonesian exports forecast for 2016 and 2017, in line with the rise in Chinese imports. Longer term, we still forecast a structural decline in Indonesian exports. This is mainly due to the anticipated further increase in domestic consumption and also a lack of investment in new mining capacity, mainly on environmental grounds Perret Associates Page 59

60 There were some official announcements in November 2016, including a target to cut Indonesian exports by half to 165m tonnes by We are taking a more conservative approach and envisage a softer decline. We adjust our exports forecast vs. our previous edition as follows: 2016: up from 330m tonnes to 342m tonnes (but still down 23.7m tonnes y-o-y). 2017: up from 327.5m tonnes to 350m tonnes. 2020: down from 320m tonnes to 310m tonnes. 2025: down from 300m tonnes to 280m tonnes. 2030: down from 290m tonnes to 270m tonnes Indonesian cost of production curve The Indonesian Rupiah has depreciated slightly to IDR13,532 to the US dollar vs. IDR13,115 in our previous report. This has offset the moderate increase in the Brent reference contract to $51.8/bbl vs. $48.5/bbl in our previous edition. As a result, the average Indonesian cost of production remains flat at $43.76/t vs. $43.61/t in our last report. Exhibit 49 Indonesian cost curve, FOB, US$/t, 6,000 kcal/kg NAR, December Cash Costs ($/t) Weighted Average = $43.8/tonne Production Volume (m tonnes) Source: Perret Associates 2016 Perret Associates Page 60

61 3.2 Australia Exports erode Australian exports of steam coal and anthracite fell by 2m tonnes or 1.3% y-o-y during Jan.- September 2016 to 148.1m tonnes. This decline is slightly greater than the 0.7m tonnes decline we had forecast for full-year Still, as was the case with Indonesia, Australian exporters have been quick to react to the surge in Chinese imports since the summer. Indeed, the average yearly fall was -1.6% during H116, whereas exports increased by 0.4% y-o-y during Q316. Moreover, Australian exports to China decreased on average by 31.4% y-o-y during Jan.- May 2016, whereas they increased on average by 25.4% during June-September. Even so, Australian exports to China still fell by 2.7m tonnes or 9.6% over the period. Australian exports to Japan also dropped 3.7m tonnes or 5.9% y-o-y over the period, in line with the general decline in Japanese imports. Exports to South Korea also fell to 24.8m tonnes in Jan.-September 2016, down 3.1m tonnes or 11% y-o-y. There has also been a degree of diversification in Australian exports, with exports rising to other destinations on an annual basis during Jan.-September, such as: Taiwan: + 2.7m tonnes or +16.5% y-o-y to 18.7m tonnes. Thailand: + 0.1m tonnes or 2.8% y-o-y to 3m tonnes. India: +0.2m tonnes or +7.4% to 3.2m tonnes. Malaysia: +0.6m tonnes or +15.7% to 4.7m tonnes. Others: +5.9m tonnes or 131% to 10.4m tonnes, of which 1.4m tonnes to Vietnam, 1.2m tonnes to Chile, 0.8m tonne to Mexico, 0.4m tonnes to Turkey. Exhibit 50 Australian steam coal exports with split per destination, m tonnes Japan China South Korea Taiwan Thailand India Malaysia Others Source: Customs, Perret Associates 2016 Perret Associates Page 61

62 3.2.2 Diversification and disinvestment in the Australian mining sector Despite the recent improvement in the FOB Newcastle index, which briefly breached the $100/t mark in November 2016, we still anticipate very little appetite for investment in new greenfield projects. Indian conglomerate Adani cleared another legal hurdle in August 2016, when the Federal Court dismissed a case brought against its Carmichael mine by Australian conservationists. Still since then other appeals from various environmentalist groups have been raised. Aside from the legal challenges, the group s financial situation, especially its large debt position, gives it little room for manoeuvre in securing financing for this project. We still doubt that the company will be able to find external investors in the current environment. More broadly, the 5-year bear market continues to reverberate around board rooms everywhere, with large international mining companies still divesting their assets. Anglo American is continuing with its gradual exit from the steam coal market. In October 2016 the company concluded the sale of its Callide open pit thermal coal mine in Queensland to private company Batchfire Resources. The mine produced 7.9m tonnes in 2015 and 5.5m tonnes during Jan.-September Around 5m tonnes of the annual production is sold to two adjacent power plants. In August 2016, Anglo American sold its 70% stake in the Foxleigh PCI coking coal mine to the private equity firm Taurus Funds Management. Anglo American is still willing to sell its 51% stake in the Dawson coking and steam coal mine (total capacity 4.2mtpa) as well as its Grosvenor and Moranbah North coking coal mines. The company also announced in October the definitive closure of its Drayton steam coal mine in the Hunter Valley. The mine began operation in 1983 and was producing 5mtpa at its peak. It produced just 1.4m tonnes in H114 and H115, due to weather disruptions. Anglo American submitted a plan to start an open cast operation near the existing mine, but this application was rejected by the New South Wales government. Peabody, which is still under Chapter 11 bankruptcy protection in the US, sold its Metropolitan metallurgical coal mine to South 32 (ex BHP.Billiton) in November 2016 for $200m. In August 2016 Rio Tinto completed the sale of its Mount Pleasant mine to the Indonesian conglomerate Salim Group for $220.7m. The mine has a capacity of 10.5mtpa. Rio Tinto initially had plans to extend its Mount Pleasant operations but dropped these due to the depressed coal market and relatively high Australian dollar. In November 2016 Rio Tinto also announced the termination of its contract with the operator of the Abbot Point Terminal for 5m tonnes export per year or 18% of the port capacity. It is very rare for an Australian mining company to terminate a contract with a stevedore. The company already announced in August it would not renew the contract and took a $496m after tax charge Perret Associates Page 62

63 Adani, which is another operator of the Abbot Point Terminal, will need to find another exporter or use the capacity itself. It is estimated that Rio Tinto has divested a total of $4.7bn in assets worldwide since January 2013, of which $2bn in coal assets. In an interesting move, the New South Wales government announced in August 2016 that it will buy back the exploration licence for the Caroona thermal coal project it initially sold to BHP.Billiton. After further studies, the government concluded that the development of a coal mine could have a negative environmental impact on local farmland. Indeed, the area is characterised by highly fertile black soil, which could be used for the production of cereals and rearing of cattle. BHP.Billiton will be compensated for its 10 years of investment and research, as the government will buy back the license at a A$120m premium to the initial selling price (It paid A$220m for the license issued in 2006). The Caroona mine had an expected capacity of 10mtpa and could have operated over 30 years, meaning that some m tonnes of steam coal production over a 30-year period will not materialise. The announcement might also create a precedent for the nearby Shenhua steam coal mine project, for which a license has also been granted. The New South Wales government has confirmed it had started discussions with Shenhua to buy back the license. In fact, Shenhua has quietly put its 10mtpa Watermark project in New South Wales on the back burner. The company didn t apply for the mining lease in August 2016, which would have been the normal step, as all legal hurdles had been removed. It also did not submit the water management plan. The company has spent nearly 10 years on the project and invested an estimated A$700m. The additional investment required to bring the mine into operation has been estimated at A$1.5bn. The mineable reserves of the project were estimated at 290m tonnes over 25 years. In a similar vein, in November the Australian quoted company Wesfarmers announced its intention to sell its coal assets, comprising total output of 18mtpa of thermal, PCI and metallurgical coals. The company s two main assets are its 100% ownership of the Curragh coking coal mine in Queensland (8.5mtpa capacity) and a 40% stake in the Bengalla steam coal mine in New South Wales (9mtpa capacity). South 32 announced in November 2016 that the company has no plans to buy or develop thermal coal resources outside its existing assets because of investor concern about climate change. Glencore closed its Newlands Northern Underground and West Wallsend underground mines in June The group has announced its intention to close the Blakefield South underground mine in 2017 and Tahmoor underground mine in Perret Associates Page 63

64 To offset these closures, at least in part, the company is pursuing a joint venture with Peabody for a new 10mtpa run of mine thermal coal project in the Hunter Valley, called United Wambo. The potential saleable output would be 3mtpa for each company. In October, Glencore announced the restart in the coming months of its Collinsville steam coal operation, due to better market conditions. The mine was shut at the end of 2015 due to low coal prices. It could produce 1.7m tonnes in 2017 and 4.8m tonnes in 2018, which is its normal capacity, assuming demand remains supported, in particular in developing countries in South East Asia. In August 2016 Glencore s fully owned subsidiary, Carbon Transport and Storage (CTSCo), announced it has received A$8.8m for a carbon capture and storage project in the Surat basin (Queensland). This confirms the willingness of the Australian government to decarbonise its electricity mix. The international utility Engie announced the definitive closure in 2017 of the 1.5GW Hazelwood lignite power plant. The plant is over 50 years old. Total CO2 emissions from the plant were estimated at 15.3m tonnes during the financial year The bulk of the lost generation will be replaced by nearby hard coal-fired power plants, which might also undermine the availability of material for steam coal exports. More broadly, the New South Wales government has announced a plan for zero emissions by We maintain our exports forecast unchanged Average cost of production stable According to our database and model, the average cost curve of production for Australian coal has remained very stable at $46.86/t vs. $46.62/t in our previous edition. This was mainly due to a slight increase in our reference Brent price to $51.8/bbl from $48.5/bbl. The Australian Dollar has remained very stable at AUS$1.34 to the US dollar vs. AUS$1.32 in our previous edition Perret Associates Page 64

65 Exhibit 51 Australian cost curve, FOB, US$/tonne, 6000 kcal/kg, December Cash Costs ($/t) Weighted Average = $46.86/tonne Production Volume (m tonnes) Source: Perret Associates 3.3 South Africa Exports stabilise South African steam coal and anthracite exports continue to be very volatile on a month by month basis, swinging from a low of 4.8m tonnes in July to a high of 7.25m tonnes in September). Overall, though, exports fell to 54.2m tonnes during Jan.-Sept. 2016, down 1.7m tonnes or 3.1% y-o-y. Somewhat surprisingly, South African exports to India have increased by a significant 3.75m tonnes or 14.6% y-o-y to 29.4m tonnes. This is all the more telling as we are currently forecasting Indian steam coal imports will fall by a total of 26.90m tonnes during full-year South African exports to Pakistan have also increased by 0.7m tonnes or 22.3% y-o-y to 3.6m tonnes As a result, India and Pakistan s combined share of total South African exports has averaged out at 61% so far this year, vs. 50.3% in 2015 and just 25.7% in This confirms the structural trend of the increasing concentration of South African exports to these two destinations. Indian steam coal imports are being driven increasingly by the large, privately owned power plants on the coast, which in most cases require a minimum quantity of high quality coal that is blended with Indonesian material. Indian sponge iron importers also remain heavily dependent on South African coal Perret Associates Page 65

66 Exhibit 52 India and Pakistan s share of South African exports, % 80% 70% 60% 50% 40% 30% 20% 10% 0% Source: Customs, Perret Associates Elsewhere in the region, we note the 0.3m tonnes increase in South African exports to the UAE to 1.44m tonnes during Jan.-September On the other hand, South African exports to Mediterranean countries fell by a massive 4.5m tonnes or 48.1% during Jan.-September 2016 to 4.8m tonnes. The biggest falls were to: Turkey, down by 2.05m tonnes y-o-y to 1.2m tonnes. Morocco, down by 1.7m tonnes y-o-y to 1.8m tonnes. Israel, down by 1.2m tonnes to 0.5m tonnes. Only exports to Egypt increased by 0.5m tonnes y-o-y to 1.3m tonnes. Exports to EU28 countries also fell by a significant 2.4m tonnes or 29% y-o-y to 5.8m tonnes. Predictably, the biggest falls were to countries which reported a significant fall in imports, such as: Spain, to which S.African exports fell by 1.5m tonnes y-o-y to 0.7m tonnes. Italy, where they fell by 0.6m tonnes y-o-y at 1.9m tonnes. The Netherlands, where they fell by 0.5m tonnes to 1.7m tonnes. Exhibit 53 South African exports, split by main destinations, m tonnes India EU28 China Taiwan UAE Mozambique Pakistan MED AREA Others Source: Customs, Perret Associates 2016 Perret Associates Page 66

67 On the domestic front, projects for building additional coal-fired capacity, which have been on-going for years, are taking shape. In October 2016, the government announced that it had selected two coal-fired power projects. In phase 1, the Thabametsi coal power plant with a net capacity of 558MW is now scheduled to be commissioned in It will be the first Independent Power Producer project in South Africa, alongside the giant state-owned Eskom (largest coal-fired utility in the world). The plant will be developed and managed by a consortium made up of Japanese trading firm Marubeni, South Korean state-owned utility Korea Electric Power and South African mining company Exxaro. Exxaro is to supply the power plant from its Thabametsi mine, which is part of the larger Grootegeluk mining project. The average coal quality is expected to be 4,500-4,800 kcal/kg NAR and the power plant could consume 1.5-2m tonnes per year. Exxaro is expected to supply the plant for up to 30 years. The second project is the 305MW Khanyisa coal-fired power project, for which fewer details have been published. More broadly, the South African government intends to build 2.5GW of additional coal-fired power capacity through IPPs (Independent Power Producers). However, only the two projects above have received backing so far. On another note, the musical chairs in the mining sector that is happening Australia is also taking place in South Africa, even if it is less active. The Gupta family, which acquired the South African Optimum mine from Glencore in a controversial deal in 2015 for $135m, has decided to sell back the mine. Year-to-date exports are well in line with our 2016 exports forecast of 75m tonnes (down 2.9% y-o-y). We maintain our exports forecast unchanged for 2016 and beyond South African cost of production We estimate that the average cost of production for South African coal has increased by $2.18/t since our previous report to $36.87/t. This is due to the $3.30/bbl increase in the Brent reference contract over the period as well as the slight strengthening of the South African Rand to ZAR13.85 against the US dollar, vs. ZAR14.37 in our last report Perret Associates Page 67

68 Exhibit 54 South African cost curve, FOB, 6000kcal/kg NAR, December Cash Costs ($/t) Weighted Average = $36.87/tonne Production Volume (m tonnes) Source: Perret Associates 3.4 Russia Russian crude coal production has continued to increase, growing at a robust annual rate of 12.5m tonnes or 3.7% to 348.7m tonnes during Jan.-November In a potentially interesting signal for 2017, we note that the monthly coal production declined on a yearly basis during September-November 2016 to an average yearly rate of 3%. This was due to a major programme of open-top railcars capacity reduction, which forced most mining companies to lower their output levels. Thermal and coking coals are typically transported on open railcars, either for domestic consumption or export. In Q216 the national railways provider, RZD, stepped up the capacity reduction programme, having already scrapped 51,000 or 9% of an estimated total 565,000 units in An additional 72,700 open-top railcars (13% of initial capacity) were scrapped during Jan.- July 2016, bringing the total number down to 440,000 units. The company is supposed to take delivery of 4,000 new units between September 2016 and December 2017 but this might not be sufficient to offset the loss in capacity. In fact, a further 14,000 old units are expected to be scrapped in 2017, 7,000 units in 2018, 3,000 in 2019 and 4,000 in On the top of this, the early onset of winter has caused some coal stockpiles to freeze, delaying coal loading operations further Perret Associates Page 68

69 As a result, coal producers have not been able to obtain maximum benefit from stronger international demand. They have even had to reduce output, albeit temporarily, due to the increase in stock levels at pits, caused by the lack of railcars. Total pithead stocks in the Kemerevo area, which mainly supplies the export terminals on the East Coast, surged to an estimated 16m tonnes in October 2016, compared with a normal level of 9m tonnes. After a very slow start to the year (down on average by 3.4% y-o-y during Jan.-May), total electricity generation started to recover in June, achieving an average growth of 2% during June-November Overall, total generation increased by 7TWh or 0.7% during Jan.-November 2016 to 965.2TWh. This represents a slight improvement from the even more anaemic growth of 2014 and 2015 (average growth of 0.2%) and the 2013 decline of 1.2%. It seems that the economic sanctions imposed on the country by the US and European Union following the Ukraine crisis are gradually losing their impact on the Russian economy. Exhibit 55 Russia, total electricity generation, TWh 120 Output Av Source: Ministry of Fuel and Energy, Perret Associates According to the Ministry of Fuel and Energy, total Russian coal exports (including steam, coking, seaborne and railways) increased by 9.3m tonnes or 7.4% y-o-y during Jan.-October 2016 to 135.2m tonnes. Of this we estimate that seaborne steam coal exports totalled 87.9m tonnes, up by 6m tonnes or 7.4% y-o-y. Shipments increased to India-Pacific countries, in particular, such as: China, up by 3.5m tonnes y-o-y during Jan.-October 2016 to 14.2m tonnes. Vietnam, up by 2.7m tonnes during Jan.-October to 3.5m tonnes. Japan, up 1.9m tonnes during Jan.-September to 10m tonnes. Taiwan, up 2.3m tonnes during Jan.-September to 6.5m tonnes. South Korea, up 2.2m tonnes during Jan.-September to 12.8m tonnes Perret Associates Page 69

70 Turkey, up by 1.4m tonnes during Jan.-October to 9.3m tonnes. The shift in focus to the Far East was mainly due to a further collapse in Russian exports to the UK, which fell by 5.1m tonnes during Jan.-September to 8.6m tonnes. This follows a fall of 6.9m tonnes in Despite the fact that Russian exporters are not able to take the full advantage of the additional demand, due to the difficulties in domestic transportation, we increase our exports forecast as follows: 2016: up from 108m tonnes in our previous edition to 111m tonnes. 2020: up from 114m tonnes to 119m tonnes. 2025: up from 118m tonnes to 125m tonnes. 2030: up from 122m tonnes to 130m tonnes Russian cost of production We estimate that the average cost of production for Russian mines has increased to $46.47/t vs. $44.58/t in our previous edition. The moderate increase in oil prices has been offset by a slight weakening of the Russian Rouble to RUB65 against the US dollar vs. RUB63.86 in our last report. Moreover, with effect from 22 October 2016 the railway operator has scrapped the discount on thermal coal transportation it had introduced earlier to help exporters when prices were at their lows. We estimate that the cancellation of this discount will increase the transportation cost by $1-$1.5/t, depending on the location of the mine and the export terminal. Exhibit 56 Russian cost curve, FOB, US$/t, 6000 kcal/kg NAR, December Cash Costs ($/t) 50 Weighted Average = $46.47/tonne Production Volume (m tonnes) Source: Perret Associates 2016 Perret Associates Page 70

71 3.5 Colombia Colombian steam coal exports have continued to rise, increasing at a robust rate of 5.8m tonnes or 8.8% y-o-y to 72.4m tonnes during Jan.-October Digging into the detail, we notice an interesting development in the split by exporting companies. Drummond ramped up its exports to 23.5m tonnes during Jan.-September 2016, up 3.8m tonnes or 19.3% y-o-y. Indeed, after a few years of under-performance, due amongst other things to the non-compliance of its loading facilities, the US-based producer has been catching up fast, surfing the rally in coal prices and benefiting from the growth in India-Pacific demand for Colombian coals. Cerrejon, on the other hand has focused on its traditional markets, in particular Turkey. In fact, the producer s exports declined by 1.1m tonnes or 4.6% y-o-y over the same period to 23.9m tonnes. For the first time this year, the two companies are exporting similar volumes, whereas Cerrejon used to dominate the export market. Prodeco (Glencore) also increased its export volumes by a significant 3.1m tonnes or 27.2% y-o-y at 14.6m tonnes. The smaller producer, Colombian Natural Resources (now Murray Energy), increased its exports by 0.6m tonnes or 25.4% y-o-y at 2.9m tonnes. The bulk of the increase in Colombian exports during Jan.-October was to non-traditional destinations, such as: India, at 2.6m tonnes, from zero in South-Korea, at 1.1m tonnes, from zero in China, at 0.3m tonnes, from zero. We also note a strong increase to various other destinations, such as: Mexico, up 1.3m tonnes y-o-y to 1.6m tonnes. Turkey, up 4.3m tonnes y-o-y to 13.3m tonnes. Italy, up 0.9m tonnes y-o-y to 3m tonnes. And even France, up 0.4m tonnes y-o-y to 0.7m tonnes. Indeed, Colombian exporters continue to benefit from historically low international freight rates, in particular Capesize, enabling long haul transportation as far away as the Far East. Capesize vessels are using the traditional route (via the Atlantic Ocean and Cape of Good Hope) and not the recently enlarged Panama Canal. Interestingly, Colombia hasn t benefited directly from the surge in Chinese imports, with a relatively modest 300KT exported to the country so far this year. This is not for economic reasons, as our Replacement Cost Analysis, updated on a weekly basis in our Short Term Price Forecast, clearly indicated from January 2016 that Colombian coal was becoming competitive to China, North East Asia and India. In fact, the absence of Colombian exports to China is primarily due to quality issues Perret Associates Page 71

72 On the other hand, Colombian exporters have benefited indirectly. Chinese importers have sucked up all available exportable surplus, in particular from Indonesia but also Australia. As a result, traditional importers of such material, such as South Korean Gencos, had to turn to other origins to fill the supply gap. As they don t have the same quality constraints as their Chinese counterparts, they have apparently been satisfied with Colombian coal. The same is true for Indian end users. In fact, we estimate that Colombian exports to these destinations could have been even higher, given the origin s significant competitive advantage at that time. However, for many users Colombian coal was a completely new origin, untried and untested. Technicians at coal power generators are relatively conservative, and rightly so, about adopting a new blending mix. This explains the relatively small volumes of exports so far compared with what might have been expected, based solely on the product s competitive price advantage. This also means that if such a window of opportunity were to occur again, we might expect a faster reaction from users in India and the Far East. Exhibit 57 Colombian exports per main destination, m tonnes Denmark NL Germany France UK Spain Portugal Others Source: Customs, Perret Associates In fact, taking a longer term view, we anticipate that Colombian exports to such destinations could become more frequent, if not structural. For one thing, Capesize freight rates could remain relatively subdued for at least another 12 months, meaning that Colombian coal might well price into the Far East in Also, our Base Case scenario worldwide is that the seaborne market could remain undersupplied, due to a lack of investment in new mining projects. If this turns out to be so, this would provide a natural support for international prices at destinations such as India, China and South Korea. The price of Colombian coal delivered to these destinations could become the marginal cost, setting a price reference and potential ceiling for international destinations Perret Associates Page 72

73 Unfortunately, the turnaround in prices came too late for a smaller Colombian mining company, Holding Minero, which filed for bankruptcy protection in August The company had total liabilities of $54m at the end of This is the third Colombian coal mining operation to file for bankruptcy protection over the last two years, after CCC and Sanoha, although they were operating in the coking coal sector. In further illustration of the fact that a sharp increase in actual exports is not a smooth process, even if the theoretical capacity is there, the railway transportation company Fenoco announced in November 2016 that it will miss its target of 10% growth for 2016 at 58m tonnes. Still, the company is confident of achieving a 9% yearly increase from 2015 s level of 53m tonnes. A combination of severe weather conditions, especially torrential rain, along with popular protests regularly undermine Colombian railway transportation. Colombia experienced a shock in September 2016 when the population rejected a peace deal which could have ended a 50-year conflict between the insurgency group FARC and the government. In November 2016, FARC and the government signed a renegotiated peace deal which this time will not be subject to a referendum for approval. Although this is a long term prospect, if implemented the deal could free up some lands which are currently in the hands of the rebels and used for the production of coca, particularly in regions such as Catatumbo, Uraba and Antioquia. Preliminary mining studies indicate that such areas contain significant steam coal reserves with high calorific value (up to 8,000 kcal/kg NAR). We increase by 1m tonnes our exports forecast for 2016 to 87.5m tonnes and maintain our exports forecast unchanged thereafter Colombian cost of production We estimate that the average cost of production for Colombian coal has eroded marginally to $36.05/t vs. $37.87/t in our previous edition. The rise in the Brent contract was offset by the weakening of the Colombian Peso to COP3,159 against the US dollar vs. COP2,928 in our last report Perret Associates Page 73

74 Exhibit 58 Colombian cost curve, FOB, 6,000 kcal/kg, December Cash Costs ($/t) 40 Weighted Average = $36.05/tonne Production Volume (m tonnes) Source: Perret Associates 3.6 US: Can a Trump presidency revive the coal industry? 2016 will be an interesting year for the US coal market, after years of decline and economic challenges under the administration of President Obama. During his election campaign, President-elect Trump claimed that he wanted to help US heavy industry, including coal, which has historically been the main fuel for its development. Trump seems to be following through with this ambition as, at the time of writing, he has nominated Scott Pruitt, a so-called climate change sceptic, as head of the Environmental Protection Agency. The nomination comes at a time when the US coal market is rebalancing anyway, as most of the damage has been done. To start with, focusing on coal usage for electricity generation, we notice that although it is still falling, the rate of decline has diminished significantly from an average of -19.7% during H116 to just -3.1% y-o-y in July and -0.1% in August. Even so, coal usage was still down a massive 66.5m tonnes or 14.1% y-o-y during Jan.- August 2016 at 404.7m tonnes. This follows a fall of 101.8m tonnes in 2015 and 6.1m tonnes in The old and obsolete coal-fired power plants that are due to be closed in accordance with the numerous clean energy plans passed under the Obama administration now have to be removed from the grid. The country is left with relatively efficient coal-fired power plants with superior utilisation rates Perret Associates Page 74

75 Jan-73 Jan-75 Jan-77 Jan-79 Jan-81 Jan-83 Jan-85 Jan-87 Jan-89 Jan-91 Jan-93 Jan-95 Jan-97 Jan-99 Jan-01 Jan-03 Jan-05 Jan-07 Jan-09 Jan-11 Jan-13 Jan-15 During June-September 2016 capacity utilisation of coal-fired power plants averaged out at 64.5% vs. 63.2% during the same period in However, the collapse in coal usage in 2016 was mainly the consequence of a crash in capacity utilisation at coal-fired power plants to an average of 43.5% during Jan.-May This was an unprecedented low level for this time of the year, compared for instance with an average of 61.7% during Jan.-May The graph below illustrates how volatile coal-fired capacity utilisation has been during 2016, even taking into account the typical seasonality. The usual February peak didn t occur: on the contrary, capacity utilisation fell to an all-time low at 35.6% in March After that it recovered strongly to an average of 68.8% during July-August 2016, very close to historical highs for this time of the year. Exhibit 59 Capacity utilisation at US coal-fired power plants, % 80% 70% 60% 50% 40% 30% 20% Source: EIA, Perret Associates Jan-13 Mar-13 May-13 Jul-13 Sep-13 Nov-13 Jan-14 Mar-14 May-14 Jul-14 Sep-14 Nov-14 Jan-15 Mar-15 May-15 Jul-15 Sep-15 Nov-15 Jan-16 Mar-16 May-16 Jul-16 Sep-16 Nov-16 Meanwhile, gas-fired generation kept increasing but at a lower rate compared with previous years. It reached 1.09bnTWh during Jan.-September 2016, up 75.7TWh or 7.4% y-o-y. Although this is solid growth, it compares with +18.3% in As a result we are seeing some consolidation between coal- and gas-fired generation. Exhibit 60 US coal- and gas-fired generation, by month, TWh Coal Gas Source: EIA, Perret Associates 2016 Perret Associates Page 75

76 In parallel, US crude coal production keeps plummeting. Output fell to 488m (metric) tonnes during Jan.-September 2016, down by a massive 132m tonnes or 21.3% y-o-y. This follows massive declines in previous years, as follows: 87.5m (metric) tonnes in m tonnes in m tonne in m tonnes in At its current pace, we forecast that US crude coal production could fall to 660m (metric) tonnes in This would be the lowest level since 1978 and represent just 62% of the peak production reached in 2008 at 1.06bn tonnes. Exhibit 61 US crude coal production, m tonnes 1,100 1, Source: EIA, Perret Associates Indeed, the 8 years of President Obama s administration had a massive negative impact on the US coal mining industry, with almost all US mining companies (Arch Coal, Patriot, Alpha Natural and Peabody) seeking Chapter 11 bankruptcy protection, except for Murray Energy. Foresight, now controlled by the Murray Group, has also announced it is in serious financial difficulties. As a result and despite the fall in demand, US steam coal exports (excl. Canada) have also been falling sharply so far this year. They reached only 10.9m tonnes during Jan.-October 2016, down another massive 9.8m tonnes or 47% y-o-y. This follows falls of: 5.8m tonnes in m tonnes in m tonnes in The fall was relatively evenly distributed across destinations during Jan.-October 2016: Down 3.4m tonnes or 58.3% y-o-y to the Far East. Down 5.1m tonnes or 46.5% y-o-y to Europe. At this rate we now forecast US steam coal exports will fall to 13.7m tonnes in 2016, below our most recent forecast of 15m tonnes. This would be 10m tonnes less than in 2015 and 34.3m tonnes less than the record 48.1m tonnes exported in Perret Associates Page 76