Commodities Comment. Chinese inventory overhang set to remain a problem over 3Q15 GLOBAL. Latest news

GLOBAL LME cash price % change US$/tonne day on day Aluminium 1,582-1.5 Copper 5,222 -.6 Lead 1,695 -.5 Nickel 1,994.1 Tin 16,39.8 Zinc 1,913-2.3 Cobalt 3,742. Molybdenum 13,53. Other prices % change day on day Gold (US$/oz) 1,98 1. Silver (US$/oz) 14.56 -.5 Platinum (US$/oz) 982 -.2 Palladium (US$/oz) 61-1.3 Oil WTI 47.71-2.7 USD:EUR exchange rate 1.13 1.1 AUD:USD exchange rate.734.8 LME/COMEX stocks Tonnes Change Aluminium 3,436,975-6,875 LME copper 345,475-65 Comex copper 33,823 Lead 218,575-1,525 Nickel 46,98-9 Tin 7,4 45 Zinc 434,8-2, Source: LME, Comex, Nymex, SHFE, Metal Bulletin, Reuters, LBMA, Macquarie Research, July 215 Articles of the Week Cutting deep into the cost curve Steel demand growth blunted, peak pig iron brought forward Trouble down pit: Copper mines begin to struggle Assessing the room for Chinese steel and aluminium export Analyst(s) Macquarie Capital Securities Limited Chen Shao +86 21 2412 941 chen.shao@macquarie.com Macquarie Capital (Europe) Limited Colin Hamilton +44 2 337 461 colin.hamilton@macquarie.com Jim Lennon Senior Commodities Consultant +44 2 337 4271 jim.lennon@macquarie.com Matthew Turner +44 2 337 434 matt.turner@macquarie.com Stefan Ljubisavljevic +44 2 337 4247 stefan.lj@macquarie.com Vivienne Lloyd +44 2 337 453 vivienne.lloyd@macquarie.com 3 August 215 Chinese inventory overhang set to remain a problem over 3Q15 Weak demand from China has been a headwind for commodities in 1H15, and manifested itself in the form of rising industrial product inventories as we highlighted in February (here) based on the NBS industrial enterprise database. With the NBS data for June coming out on Thursday, we review the situation in 2Q15. Industrial product inventory growth has slowed in April-June, but still ran much ahead of sales growth, suggesting downward pressure on industrial production and commodity demand from the ongoing inventory overhang will remain in place over 3Q15. Latest news After a week which included what was possibly the last FOMC meeting in nine years to keep interest rates unchanged, and US 2Q GDP with major revisions to previous quarters, it was a less well known US data series that caused the biggest impact on asset prices. Friday s Employment Cost Index (ECI) showed total labour costs, adjusted for changes in the composition of the workforce, rose just.2% QoQ in 2Q, the slowest pace since records began in 1982. On that the dollar fell 1%, causing gold and other metals to rally strongly, as traders queried whether the Fed, which sees the labour costs as important and the ECI as a good indicator of them, could really raise rates in September as expected given such a backdrop. Some of the moves were faded as economists queried some aspects of the report. Nevertheless it means next Friday s July Employment report, which provides information on wage growth (though not adjusted for compositional effects), will be more important than ever. Hot on the heels of First Quantum Minerals copper production downgrade, due to insufficient power supply from Zambian state utility Zesco, came the news on Friday that power distributor Copperbelt Energy Corp (which purchases power from Zesco) had informed miner customers including Vedanta, Glencore and Vale that supply would be reduced by 3% from midnight that day, local time. Later on Friday, however, it was reported that the planned supply cut had been delayed to allow talks to take place amid a government intervention. Total Zambian mine output is now (after FQM s downgrade) expected to be around 838kt of copper this year, down from our initial, pre-disruption expectations of 984kt. We await further details to determine whether another downgrade to the country will be necessary. Antofagasta Minerals and Barrick Gold announced on Friday their agreement for the former to acquire 5% of the latter s 12ktpa Zaldivar SXEW copper mine for $1bn. The acquisition will allow Antofagasta to maintain its position amongst the top ten copper miners to 22, and we think more importantly, sustain and grow its position in the copper cathode market following the closure of its Michilla SXEW mine this year and declining output at its Centinela SXEW mine over the next few years. ANTO will assume operational control of the asset as part of the deal. The latest data from the China Iron and Steel Association shows crude steel output from its member mills over July 11-2 to have been 622mtpa the lowest level since mid-may and 5.8% lower than the equivalent period in 214. Meanwhile, inventories held by these mills also rose by.26mt to 16.64mt over the 1-day period, reflecting a likely drop in trader purchasing activity amid weaker end demand. Please refer to page 26 for important disclosures and analyst certification, or on our website www.macquarie.com/research/disclosures.

Jan-11 Apr-11 Jul-11 Oct-11 Jan-12 Apr-12 Jul-12 Oct-12 Apr-13 Jul-13 Oct-13 Jan-14 Apr-14 Jul-14 Oct-14 Jan-15 Apr-15 Jul-12 Sep-12 Nov-12 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 Macquarie Research Inventory overhang to remain for China in 3Q15 Weak demand from China has been a headwind for commodities in 1H15, and manifested itself in the form of rising industrial product inventories as we highlighted in February (here) based on the NBS industrial enterprise database. With the NBS data for June coming out on Thursday, we review the situation in 2Q15. Industrial product inventory growth has slowed in April-June, but still ran much ahead of sales growth, suggesting downward pressure on industrial production and commodity demand from the ongoing inventory overhang will remain in place over 3Q15. The lack of reliable and comprehensive inventory data has been a major obstacle to analysts attempts to gauge China s industrial product inventory. We ve taken an innovative approach by looking at the difference between inventory growth and sales growth of industrial sectors (referred to as relative inventory growth in this note) with the straightforward theory that higher/lower growth of inventory than sales conceptually corresponds to restocking/destocking. The results have been supportive we do see a strong negative correlation between relative industrial inventory growth and Industrial Production (IP) growth (Fig 1); on a sectoral level, our indicator also seems to reflect reality well, e.g. the relative inventory growth for ferrous mining industry closely tracking a survey result of domestic iron ore mine inventory (Fig 2). Fig 1 Relative inventory growth correlates well with the overall economy Fig 2 and sectoral inventory indicators YoY % Industrial inventory relative growth % 15 Nominal IP, MoM - RHS 2. 1 1.5 5 1. YoY % 4 3 2 Ferrous mining relative inventory growth - LHS Domestic iron mine inventory - days of consumption by domestic mills Days 8 7 6 5.5 1 4-5 -1. -.5-1 3 2 1-15 -1. -2 Source: CEIC, Macquarie Research, July 215 Source: CEIC, MySteel, Macquarie Research, July 215 So as shown in Fig 1, relative industrial inventory growth remained positive in 2Q15, suggesting industrial companies were still building up inventory compared with their sales. This is not to say inventory growth per se was still rising; actually, both inventory and sales growth has been declining, but the former still grew faster (Fig 3). In other words, real demand for industrial products has been even weaker than the sluggish IP growth. While the general trends are largely the same across different industrial sectors for relative inventory growth, the further towards the upstream of industrial chains, the more volatile relative inventory growth tends to be. In recent months, we have also seen a sharper drop in our inventory indicator for raw materials than for other sectors (Fig 4), which is in line with the fact that the raw material producers have seen larger price drops than other industrial sectors owing to the whiplash effect from weaker downstream demand (and of course, lower prices). More specifically, besides the falling relative inventory growth at iron ore mines, relative inventory at ferrous smelting (which is largely steel-making) has remained low in 2Q15, and that at nonferrous smelting (i.e. base metal refining) hit a peak in January-February and slowed down since then. Coal inventory at miners has largely stopped rising in absolute terms, but given double-digit negative sales growth, the need for destocking remains (Fig 7). The conditions for a few other sectors are shown at the end of this note. 3 August 215 2

Jan-11 Apr-11 Jul-11 Oct-11 Jan-12 Apr-12 Jul-12 Oct-12 Apr-13 Jul-13 Oct-13 Jan-14 Apr-14 Jul-14 Oct-14 Jan-15 Apr-15 Jan-11 Apr-11 Jul-11 Oct-11 Jan-12 Apr-12 Jul-12 Oct-12 Apr-13 Jul-13 Oct-13 Jan-14 Apr-14 Jul-14 Oct-14 Jan-15 Apr-15 Jan-11 Apr-11 Jul-11 Oct-11 Jan-12 Apr-12 Jul-12 Oct-12 Apr-13 Jul-13 Oct-13 Jan-14 Apr-14 Jul-14 Oct-14 Jan-15 Apr-15 Jan-11 Apr-11 Jul-11 Oct-11 Jan-12 Apr-12 Jul-12 Oct-12 Apr-13 Jul-13 Oct-13 Jan-14 Apr-14 Jul-14 Oct-14 Jan-15 Apr-15 Macquarie Research Fig 3 Inventory still grew faster than sales for industrial products Fig 4 Inventory situation across different sectors YoY % 35 3 Industrial sales Industrial inventory YoY % 4 3 China industrial inventory situation Average Raw materials Average upstream Average mid-stream Average down-stream 25 2 2 15 1-1 1-2 5-3 -4 Source: CEIC, Macquarie Research, July 215 Source: CEIC, Macquarie Research, July 215 Fig 5 Inventory at steel-making industry Fig 6 and base metal refiners YoY % 6 Ferrous smelting relative inventory growth Crude steel inventory growth at mills YoY % 4 Non-ferrous mining relative inventory growth Non-ferrous smelting 5 3 4 2 3 1 2 1-1 -2-1 -3-2 -4-3 -5 Source: CEIC, MySteel, Macquarie Research, July 215 Source: CEIC, Macquarie Research, July 215 In the meantime, we have also gathered inventory data in volume terms for three sectors with particular relevance to commodity demand. The results are shown in Fig 8-1 for property, auto, and air conditioners respectively. Property inventory (measured on a month-of-sales basis) has come down significantly in May and June (Fig 8). Admittedly, the drop is partly seasonality as this measure has always come down towards the end of 1H and 2H each year, which is due to statistical features of the series. That said, the fall-off this year has been much sharper than in previous years, reflecting the genuine improvement in property sales in the past 2-3 months. By historical standards, property inventory could probably fall further, but we seem to be getting closer to a level that would demand a positive supply response. The auto market, however, looks to be in much bigger trouble, with inventory standing at the highest level in history and showing no signs of declining (Fig 9). This has probably cast doubts among investors about the resilience of household consumption in China. Inventory of air conditioners has instead been on a relatively low level throughout 2Q15 (Fig 1), showing the producers caution towards consumer demand. 3 August 215 3

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan-11 Apr-11 Jul-11 Oct-11 Jan-12 Apr-12 Jul-12 Oct-12 Apr-13 Jul-13 Oct-13 Jan-14 Apr-14 Jul-14 Oct-14 Jan-15 Apr-15 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Macquarie Research Fig 7 Coal inventory still too high on a relative basis Fig 8 Property inventory dropped significantly on stronger sales YoY % Coal ming relative inventory growth YoY % 5 Coal stock at key state-owned mines - RHS 12 4 1 3 8 2 6 1 4 9 8 7 6 5 Ratio between property inventory and monthly sales 212 213 214 215 2 4-1 -2-3 -4-2 -4-6 3 2 1 Source: CEIC, Sxcoal, Macquarie Research, July 215 Source: CEIC, Macquarie Research, July 215 Fig 9 Auto makers have too much inventory Fig 1 while A/C makers are much more cautious.9 Ratio between auto inventory and monthly sales 1.6 Ratio between A/C inventory and monthly sales.8 212 213 214 215 1.4 212 213 214 215.7 1.2.6.5.4.3 1..8.6.2.4.1.2.. Source: Wind, Macquarie Research, July 215 Source: Wind, Macquarie Research, July 215 Coming back to a more general perspective, inventory levels tend to be pro-cyclical weak/strong end-user demand doesn t only directly reduce/increase the need for finished product inventory as a buffer for fluctuations of sales, but also suppresses/boosts prices and the market value of inventory, which pushes companies to further lower/raise inventory levels. The implication is that major inventory cycles tend to coincide with general economic cycles and exacerbate the latter. In China s current case, industrial companies are facing the double whammy of weak demand and relatively high inventory, and such vicious cycles are usually not easy to break. Another implication is that inventory cycles could be indicative of the overall economic trends. From this angle, Fig 1 suggests the current level of relative inventory growth is still too high to be a positive factor for economic growth and commodity demand. As it takes time for relative inventory growth to be dragged towards/below zero, it seems likely that the inventory overhang will remain a negative factor for commodities outlook in 3Q15. 3 August 215 4

Macquarie Research Friday 31 July 215 Prices Closing price * Closing price * 31-Jul-15 31-Jul-15 3-Jul-15 3-Jul-15 % ch. day 215 YTD Ave 214 US$/tonne US /lb US$/tonne US /lb on day US$/tonne US$/tonne LME Cash Aluminium 1,582 72 1,67 73-1.5 1,761 1,867 Aluminium Alloy 1,696 77 1,696 77. 1,781 1,951 NAASAC 1,685 76 1,694 77 -.6 1,829 2,86 Copper 5,222 237 5,253 238 -.6 5,859 6,862 Lead 1,695 77 1,73 77 -.5 1,857 2,96 Nickel 1,994 499 1,978 498.1 13,341 16,867 Tin 16,39 743 16,265 738.8 16,693 21,893 Zinc 1,913 87 1,957 89-2.3 2,115 2,164 Cobalt 3,742 1,394 3,746 1,395. 29,894 31,251 Molybdenum 13,53 592 13,5 592. 16,998 25,548 LME 3 Month Aluminium 1,618 73 1,643 75-1.5 1,782 1,894 Aluminium Alloy 1,7 77 1,7 77. 1,794 1,974 NAASAC 1,71 78 1,72 78 -.6 1,855 2,117 Copper 5,23 237 5,26 239 -.6 5,851 6,828 Lead 1,72 77 1,713 78 -.6 1,867 2,113 Nickel 11,4 51 11,25 5.1 13,395 16,935 Tin 16,3 739 16,175 734.8 16,712 21,887 Zinc 1,916 87 1,951 88-1.8 2,122 2,167 Cobalt 3,5 1,383 3,5 1,383. 3,17 31,287 Molybdenum 13,25 61 13,25 61. 17,59 25,548 * LME 2nd ring price - 13 hrs London time. Year-to-date averages calculated from official fixes. Gold - London PM Fix (US$/oz) Silver - LBMA Silver Price (US$/oz) Platinum - London 3pm price (US$/oz) Palladium - London 3pm price (US$/oz) Oil WTI - NYMEX latest (US$/bbl) EUR : USD exchange rate - latest AUD : USD exchange rate - latest 1,98 1,88 1. 1,194 1,266 14.56 14.64 -.5 16.32 19.7 982 984 -.2 1,137 1,384 61 618-1.3 753 83 47.71 49.5-2.7 52.89 93.36 1.13 1.91 1.1 1.114 1.329.734.728.8.776.93 Exchange Stocks Change since last report Cancelled End-14 Ch. since (tonnes) 31-Jul-15 3-Jul-15 Volume Percent warrants stocks end-14 LME Aluminium 3,436,975 3,443,85-6,875 -.2% 1,431,45 4,25,225-768,25 Shanghai Aluminium 318,564 318,911-347 -.1% - 27,428 111,136 Total Aluminium 3,755,539 3,762,761-7,222 -.2% - 4,412,653-657,114 LME Copper 345,475 346,125-65 -.2% 26,85 177,25 168,45 Comex Copper 33,823 33,823.% - 24,15 9,673 Shanghai Copper 13,117 11,251 1,866 1.8% - 111,915-8,798 Total Copper 482,415 481,199 1,216.3% - 313,9 169,325 LME Zinc 434,8 436,8-2, -.5% 55, 69,825-256,25 Shanghai Zinc 177,214 176,763 451.3% - 83,471 93,743 Total Zinc 612,14 613,563-1,549 -.3% - 774,296-162,282 LME Lead 218,575 22,1-1,525 -.7% 57,5 221,975-3,4 Shanghai Lead 14,5 15,968-1,468-9.2% - 63,64-49,14 Total Lead 233,75 236,68-2,993-1.3% - 285,579-52,54 Aluminium Alloy 13,4 13,4.% 14 26,52-13,48 NASAAC 5,84 51,5-66 -1.3% 3,38 8,7-29,86 Nickel 46,98 46,998-9 -.2% 152,874 414,9 45,198 Tin 7,4 6,995 45.6% 1,58 12,135-5,95 Source: CME, LBMA, LME, Reuters, SHFE, Macquarie Research 3 August 215 6

Macquarie Research Summary of changes, week ended 31 July LME metal prices (%) Cash 3-Month Aluminium -1.3% -1.5% Aluminium Alloy -1.1% -.9% NAASAC -.1%.% Copper -.5% -.6% Lead -.5% -1.% Nickel -2.3% -2.3% Tin 6.1% 5.8% Zinc -1.7% -1.8% Cobalt -1.% -.8% Molybdenum.1%.% Other prices (%) Gold 1.6% Silver.5% Platinum -.3% Palladium -1.6% Oil WTI -.6% EUR : USD exchange rate.5% AUD : USD exchange rate.8% Exchange stocks tonnes % LME aluminium -43,525-1.3% Shanghai aluminium -347 -.1% Total aluminium -43,872-1.2% LME copper 2,225.6% Comex copper -338-1.% Shanghai copper 1,866 1.8% Total copper 3,753.8% LME zinc -1,7-2.4% Shanghai zinc 451.3% Total zinc -1,249-1.6% LME lead 6,85 3.2% Shanghai lead -1,468-9.2% Total lead 5,382 2.4% LME aluminium alloy.% LME NAASAC -2,98-5.5% LME nickel 6,768 1.5% LME tin -95-1.3% Source: CME, LBMA, LME, Reuters, SHFE, Macquarie Research, July 215 3 August 215 7

Macquarie Research Macquarie metals and bulk commodity price forecasts 214 215 215 215 215 215 216 217 218 219 22 Unit CY Q1 Q2 Q3 Q4 CY CY CY CY CY CY LT $215 Copper $/tonne 6,862 5,818 6,43 5,85 5,9 5,93 6,13 6,475 7,263 8,5 8,413 6,944 Aluminium $/tonne 1,867 1,8 1,765 1,66 1,69 1,729 1,7 1,75 1,8 1,8 1,9 1,95 Zinc $/tonne 2,164 2,8 2,19 2,7 2,15 2,122 2,413 3,2 3,25 3,315 3,265 2,6 Nickel $/tonne 16,867 14,338 13,8 12, 13,5 13,211 18,375 22,5 24, 25, 25, 24, Lead $/tonne 2,96 1,86 1,942 1,88 1,9 1,882 1,955 2,13 1,988 1,98 2,7 1,95 Tin $/tonne 21,893 18,392 15,586 15,25 15,25 16,12 15,75 16,5 17,75 19, 2, 19, Manganese ore $/mtu CIF 4.5 3.7 2.8 2.8 3. 3.1 3.3 3.5 4. 4.3 4.5 4.3 FeCr (EU contract) c/lb 119 18 18 18 112 19 122 129 141 142 142 126 Molybdenum oxide $/lb 11 9 8 8 9 8 9 11 12 13 13 13 Cobalt (99.8%) $/lb 14 14 14 14 14 14 15 16 16 16 16 13 Steel - Average HRC $/tonne 598 481 429 434 426 443 455 474 493 53 513 59 Steel Scrap - average #1HMS $/tonne 327 244 245 227 247 24 239 217 237 237 237 252 Iron ore - Australian fines c/mtu fob 142 93 85 69 77 81 8 9 11 17 15 113 Iron ore - Australian lump c/mtu fob 166 12 16 9 98 13 12 112 126 132 13 133 Spot 62% Fe iron ore China $/t cfr 97 63 58 48 54 56 56 63 7 75 75 8 Thermal coal - Australian Spot $/t fob 71 66 6 55 53 58 55 59 65 7 7 7 Thermal coal - S.African Spot $/t fob 72 62 61 56 54 58 57 6 72 72 72 7 Thermal coal - JFY contract $/t fob 82 82 68 68 68 68 63 68 73 78 78 75 Hard coking coal $/t fob 126 117 11 93 95 14 1 11 125 135 145 145 Semi-soft coking coal $/t fob 93 86 81 74 69 78 68 74 84 91 98 95 LV PCI coal $/t fob 14 99 93 73 81 86 8 94 14 111 118 11 Coke - China export spot $/t fob 195 174 152 16 16 161 17 17 17 17 17 21 Gold $/oz 1,266 1,218 1,192 1,75 1,125 1,152 1,163 1,256 1,344 1,4 1,4 1,25 Silver $/oz 19 17 16 15 16 16 17 19 22 23 23 18 Platinum $/oz 1,384 1,192 1,125 1,225 1,3 1,211 1,363 1,438 1,475 1,475 1,475 1,45 Palladium $/oz 83 786 758 835 875 814 913 994 994 95 95 8 Uranium spot $/lb 33 38 37 36 36 37 39 44 48 53 53 5 Source: LME, Platts, CRU, Metal Bulletin, Macquarie Research, July 215 3 August 215 8

Macquarie Research Selected Chinese commodity and macroeconomic data YoY % (unless specified) Jul-14 Aug-14 Sep-14 Oct-14 Nov-14 Dec-14 Jan-15 Feb-15 Mar-15 Apr-15 May-15 Jun-15 Macro Indicators GDP 7.4 7.4 7. Consumption 3.6 3.7 4.5 Investment 3. 3.6 1.2 Net Exports.8.1 1.3 Industrial Production in real terms 9. 6.9 8. 7.7 7.2 7.9 6.8 6.8 5.6 5.9 6.1 6.8 Industrial Production in nominal terms 8.1 5.6 6.1 5.3 4.3 4.3 2.2 1.7.8 1.1 1.2 1.7 IP, MoM %.6.2.8.5.5.7.5.5.3.6.5.6 Nominal IP, MoM %.5 -..5.1 -..1 -.6 -.3.2.4.4.2 Fixed Asset Investment 15.7 13.8 13.8 14.4 14.9 14.8 13.9 13.9 13.2 9.4 1. 11.4 Manufacturing 13.7 11.3 12. 11.3 13.5 13.5 1.6 1.6 1.3 9. 1.2 8.9 Infrastructure 21.6 16.3 17.7 24.4 16.6 16.2 2.8 2.8 24.5 16.1 15.1 2.5 Real Estate 13.5 1. 8. 1.9 1.5 Manufacturing PMI, % 51.7 51.1 51.1 5.8 5.3 5.1 49.8 49.9 5.1 5.1 5.2 5.2 New Orders, % 53.6 52.5 52.2 51.6 5.9 5.4 5.2 5.4 5.2 5.2 5.6 5.1 HSBC/Markit PMI, % 51.7 5.2 5.2 5.4 5. 49.6 49.7 5.7 49.6 48.9 49.2 49.4 New Orders, % 52.8 51.8 51.3 5.7 49.6 49.9 5.3 51.7 51.3 5. 49.3 49.7 Monthly Social Finance, Rmb bn 274 958 1,136 681 1,146 1,695 2,47 1,356 1,241 1,56 1,236 1,858 New Rmb Loans, Rmb bn 385 73 857 548 853 697 1,47 1,2 1,18 78 91 1,271 Outstanding Rmb Loans 13.4 13.3 13.2 13.2 13.4 13.6 13.9 14.3 14. 14.1 14. 13.4 Outstanding Social Finance 15.8 15. 14.6 14.3 14.1 14.3 13.5 13.8 12.8 12.3 12. 11.7 M2 13.5 12.8 12.9 12.6 12.3 12.2 1.8 12.5 11.6 1.1 1.8 11.8 CPI inflation 2.3 2. 1.6 1.6 1.4 1.5.8 1.4 1.4 1.5 1.2 1.4 CPI food 3.6 3. 2.3 2.5 2.3 2.9 1.1 2.4 2.3 2.7 1.6 1.9 CPI non-food 1.6 1.5 1.3 1.2 1..8.6.9.9.9 1. 1.2 PPI inflation -.9-1.2-1.8-2.2-2.7-3.3-4.3-4.8-4.6-4.6-4.6-4.8 Exports 14.4 9.4 15.2 11.6 4.7 9.7-3.4 48.3-15. -6.5-2.9 2.8 Imports -1.6-2.4 6.9 4.4-6.9-2.5-2. -2.6-12.7-16.3-17.7-6.1 NBS industrial enterprises ROA (EBIT), % 13.2 13.2 13.3 13.4 13.7 14.3 11.1 11.1 11.8 11.9 12.1 - NBS industrial enterprises Net Profit Margin, YTD % 5.5 5.5 5.5 5.6 5.7 5.9 4.9 4.9 5.2 5.3 5.4 - Retail Sales in nominal terms 12.2 11.9 11.6 11.5 11.7 11.9 1.7 1.7 1.2 1. 1.1 1.6 Retail Sales in real terms 1.5 1.6 1.8 1.8 11.2 11.5 11. 11. 1.2 9.9 1.2 1.6 Sector Indicators Industrial Production Cement 3.5 3. -2.2-1.1-4. -1.4 11.2 11.2-2.5-7.3-5.4-5.8 Crude Steel 1.5 1. - -.3 -.2 1.5-1.5-1.5-1.2 -.7-1.7 -.8 Ten Non-ferrous Metals 7.8 9.5 8.2 4.8 7.9 15.4 6.8 6.8 6.6 9.7 11.4 13.2 Automobiles 1.5 3.1 4.5 6.8 2.6 3.7 4.6 4.6 3.5 -.3-1.6.7 Power Generation 3.3-2.2 4.1 1.9.6 1.3 1.9 1.9-3.7 1. -..5 Fixed Asset Investment Coal Mining -3.2 6. -6.4-28.6-23. -19.2-16.3-16.3-23.8-16.3-5.7-9.1 Iron Ore 3.5-6.4 7.6-16.8-8.1-2.1-4.4-4.4-16.6-11.3-21.9-7. Non-ferrous Mining -4.1-8.3 1.6 6.5.9 Steel -7.8-1. 1.6-14.9 11. -18.4-4.1-4.1-6.4-17.5 4.4-22.4 Non-ferrous Smelting -2.4 12.5 3.8 5.4 8.2-18.1 1.7 1.7-8.8 14.5-5.6 1.6 Metal Product Manufacturing 27.4 24. 27.9 16. 17.9 11.6 16.2 16.2 15.2 9.1 9.3 5.1 Power and Thermal Supply 14.4 22.5 24.6 2.5 16.9 18.5 21. 21. 2.1 21.1 12.3 9.6 PPI Coal Mining -1. -9.6-1.6-11.4-11.6-12.2-13. -13.1-13. -13. -14. -15.4 Iron ore -8.2-1.4-12.6-14.7-16.6-19. -2.5-21. -21.4-22.9-22.5-2.1 Non-ferrous Mining.3 -.2-1.8-2.5-2.8-3.2-3. -4.2-5.8-4.9-4.6-4.7 Steel -4.2-5.7-7.9-8.5-8.9-9.7-11.5-13.3-13.7-14.5-15.7-16.7 Non-ferrous Smelting -1.2-1.3-1.6-2.8-2.9-3.4-5.5-6.3-4.4-3.6-3.9-5.9 Metal Product Manufacturing -1. -1. -1.2-1.2-1.2-1.4-1.6-1.9-2. -2.2-2.4-2.6 Heat and Thermal Supply.2.1. -.1.3.4.3 -.2 -.4 -.5-1.4-1.7 Imports of Commodities, tonnes Grain and Grain Power 1,559 1,87 1,963 1,253 1,492 2,374 2,418 2,353 2,136 2,729 2,864 3,787 Soybean 7,475 6,33 5,28 4,12 6,26 8,527 6,876 4,263 4,493 5,31 6,127 8,87 Iron Ore Concentrate 82,518 74,881 84,695 79,391 67,399 86,851 78,619 67,942 8,59 8,213 7,865 74,959 Copper Concentrate 91 96 1,29 956 1,154 1,167 928 758 1,323 1,39 984 988 Coal 23,29 18,864 21,159 2,134 21,27 27,219 16,78 15,263 17,26 19,952 14,249 16,63 Crude Oil 23,758 25,19 27,577 24,89 25,411 3,373 27,981 25,553 26,88 3,286 23,245 29,492 Refined Oil 1,865 2,53 2,472 2,281 2,371 3,196 2,347 2,679 2,874 2,471 2,32 3,99 Scrap Steel 248 26 24 184 195 223 185 165 194 31 22 28 Scrap Copper 36 343 392 332 322 367 36 187 288 39 285 317 Steel 1,223 1,174 1,364 1,87 1,127 1,212 1,155 871 1,28 1,198 1,53 1,166 Unwrought Copper 343 336 39 397 417 421 415 283 48 429 363 349 Property Investment 11.9 9.9 8.6 11.8 7.6-1.9 1.4 1.4 6.5.5 2.4 3.4 New Starts 8.2 6.2 -.2 42.9-31.2-26.1-17.7-17.7-19.5-14.9-12.8-15. Area under Construction 1.4 2.1 1. 51.4-35.3-22.1 7.6 7.6-2.2-12.3-18.7-18.4 Completion -15.7 24.5 11.4 1.5 11.4 1.2-12.9-12.9 1.3-18.4-23. -15.8 Sales by volume -16.3-12.4-1.3-1.6-11.1-4.1-16.3-16.3-1.6 7. 15. 16. Source: NBS, CEIC, Macquarie Research, July 215 3 August 215 9

Potash Gold Zinc Copper Thermal Coal Uranium PGM basket Iron Ore Aluminium Manganese ore Met Coal Nickel Potash Gold Zinc Copper Thermal Coal Uranium Iron Ore PGM basket Manganese ore Met Coal Nickel Aluminium Macquarie Research Articles of the Week Cutting deep into the cost curve Like a hot knife through butter, most commodity prices are now slicing with relative ease into industry cost curves. Given that for many metals and bulks there is an imminent need to take capacity offline amid clear oversupply, this is natural at the current point in the cycle. Concepts of cost support are thus less relevant in the short term where it is all about inflicting enough pain to force hard decisions on capacity. However, cost structures still matter for through the cycle equilibrium. Meanwhile, this reporting season has a common theme further cost outs at metals and mining companies. We review movements in cost curves and the factors driving the cost reductions many of which are unlikely to reverse anytime soon. Figure 1 below shows the extent of the challenge for metals and mining. Versus the 9 th percentile (a measure we still consider flawed for modern markets further details here) the vast majority have lower current spot prices significantly lower in certain cases. Or, looking at it another way in figure 2, decent proportions of global supply across metals and bulks are currently losing money on a cash plus sustaining capital basis - over a third in areas like met coal, nickel and aluminium. The only mined market where everyone remains cash positive is potash, and given the inherent overcapacity the spirit of producer discipline will be further tested in this market over the coming year we would expect a push into the cost curve sooner rather than later. Fig 1 The vast majority of mined commodities are now trading well into the cost curve Fig 2 with over a third of supply currently lossmaking in met coal, nickel and aluminium 2% 1% % -1% -2% -3% -4% Current spot price premium over 9th percentile 5% 4% 3% 2% 1% % -1% -2% -3% -4% Proportion of supply curve underwater Source: LME, Platts, Wood Mackenzie, Macquarie Research, July 215 Source: Platts, Wood Mackenzie, Macquarie Research, July 215 That a significant proportion of supply is under pressure should not be a surprise, given prices are at multi-year lows. Indeed, it could be said the surprise is that more isn t lossmaking, or that we haven t seen more aggressive supply responses. The main reason for this is that cost structures themselves are falling rapidly, and where potential for further cost gains can be shown, producers will continue to keep capacity in the market and try and improve their own competitive position. Unfortunately, when everyone is doing the same things, the fundamental imbalance is not solved. The current reporting season is showing further evidence of aggressive cost outs at mining companies, with Vale s $15.8/t FOB for iron ore reported on Thursday just the latest in a series of positive surprises. There are three core reasons why costs have been coming down: Currency: All currencies have been depreciating against the US dollar, but commodity currencies more than most (as would be expected given falling commodity prices). With decent parts of mining cost structures denominated in local currencies, the equivalent USD cost has of course fallen. Figure 3 shows YTD performance of 6 key producer currencies while the RUB may only be down 3% this year, on a 12-month view it has almost halved, while the BRL has certainly helped Vale s cost structure. It should be noted that, with many curves having Chinese or US producers at the top end, wider currency depreciation against the USD tends to steepen cost curves. 3 August 215 1

Coal Steel Iron ore mining Base metals Petroleum & Natural Gas Chemical Products Electronic components Fabricated Metal Product Plastic Products Ags + Food Paper Manufacturing Textile General Equipment Automobile Transport Equipment Consumer Good Printing & Recording Food Electricity Production Tobacco Beverages Clothing Pharmaceutical Food Services Water Supply PPI by sector, 214 Macquarie Research Fig 3 Commodity currency depreciation has been a key theme for 215 Fig 4 with currency alone moving industry average costs down by ~2% in some cases from 213 levels % YTD commodity currency depreciation versus USD 11 15 Production weighted USD index -5% -3% 1-1% -8% 95-15% -1% -11% -12% 9-2% 85-25% -3% -26% RUB ZAR CLP CAD AUD BRL DXY 75 21 211 212 213 214 215 Source: Bloomberg, Macquarie Research, July 215 Source: Wood Mackenzie, Macrobond, Macquarie Research, July 215 Energy: While the first oil price drop was almost a year ago now, these things take a while to feed into mining cost structures. But now the effects are being seen, and in many cases they are significant, particularly for those where diesel-heavy mining processes or freight rates as a significant part of the cost structure. The only miners not to have felt this benefit are those who previously has a subsidised rate, where the government has now reduced the subsidy for example smaller Indonesian thermal coal producers who were buying subsidised diesel illegally. While depreciating FX steepens cost curves, falling energy prices tend to make them shallower as inefficient producers gain most benefit. Producer self-help: This is not an area to be underestimated whether debottlenecking existing operations or stripping down contractor costs miners in the main have done much of what they can to help protect against further margin erosion. Moreover, many of these cuts are sustainable should discipline be maintained. The challenge now will be to eke out further gains given the easy hits have already been exhausted, particularly as sustaining capital may have to resume an upward trend. It is definitely fair to say the shift in Chinese cost structures was one of the key areas analysts (ourselves included) got wrong over recent years. While China was growing rapidly, core inflation was ~4% and mining cost inflation running at over 1% per annum on a like for like basis. Making assumptions that USD mining inflation would continue to rise at 5% per annum seemed fair, particularly as the RMB was still appreciating. 8 Gold Copper Fig 5 Lower energy costs are only now starting to push through into cost structures Fig 6 We have seen heavy disinflationary pressures running through Chinese domestic mining 16 14 12 1 8 6 4 Brent crude, US$/bbl 4% 2% Upstream Midstream % -2% -4% -6% Downstream -8% -1% -12% 2 27 29 211 213 215 Source: Reuters, Macquarie Research, July 215 Source: NBS, Macquarie Research, July 215 3 August 215 11

$/lb $/t CFR China, 62% basis 6 kcal NAR CFR China ($/t) Macquarie Research Instead, as China slowed, inflation bottlenecks gave way to overcapacity in all areas with disinflationary pressures running through the whole economy. As shown in figure 6, upstream sectors have indeed been worst hit with producer price inflation in 214 ranging from -4% for base metals to -11% for coal. As a result, Chinese cost structures have gone down, not up, adding further downward pressure to the sharp end of global cost curves. Of course, the nature of cost structures is very different across individual commodities. The drop in energy prices has had a notable effect on iron ore, given ~3% of costs are more or less oillinked. The same is true for thermal coal, where the effect has also seen a reasonable flattening of the curve, meaning everyone is feeling a bit of pain, but not really enough to invoke widespread cuts. While the higher premium certainly helped in 214, on a LME cash basis much of the curve has been losing money for a considerable period of time, just not enough. For copper, aside from a short period in 29 prices have consistently traded at the top end or even out of the cost curve, but not anymore we feel a decent proportion of copper mine output is struggling at these price levels. Unlike others however, we don t think there needs to be significant copper supply cuts. Fig 7 The move lower in the contestable iron ore cost curve has been dramatic Fig 8 while for thermal coal, the cost curve has also flattened considerably 18 Evolution of iron ore cost curve Seaborne thermal coal supply curve - CFR 16 12 South China 14 213 11 214 1 12 215 9 1 8 8 6 7 6 4 2 5 4 212 costs 214 costs 215 costs 2 4 6 8 1 25 5 75 1 125 million tonnes 15 175 2 Cumulative Volume (mt) Source: SMM, Company Reports, Macquarie Research, July 215 Source: Wood Mackenzie, Macquarie Research, July 215 Fig 9 On a LME cash basis, aluminium has traded into the cost curve for a while Fig 1 whereas copper has spent much of the past decade out the curve, but pressure has returned $/t 36 32 28 24 2 Aluminium costs and prices Price Ninth Decile Upper Quartile Median Lower Quartile 44 4 36 32 28 24 2 Copper costs and prices Price Ninth Decile 16 16 12 8 8 4 4 1984 1988 1992 1996 2 24 28 212 198 1984 1988 1992 1996 2 24 28 212 Source: Wood Mackenzie, Macquarie Research, July 215 Source: Wood Mackenzie, Macquarie Research, July 215 In the very short term, cost curves are not important for setting the price in many markets. Rather, they are important for determining how much balance sheet pain marginal producers can take before liquidity is drained and hard decisions on capacity cuts have to be made. The more cost structures fall, the more prices have to fall to force this point figure 11 looks at the drop in marginal costs across commodities over the past 2 years. 12 YTD 215 price 3 August 215 12

Zinc Copper Gold Nickel Aluminium Uranium Manganese ore Met Coal Thermal Coal Iron Ore Macquarie Research However, assuming suitable cuts are made, cost structures will again be important for through the cycle prices, particularly given general expectations of ongoing overcapacity. This may allow for some price recovery in selected commodities, but in a low inflation world this is barely anything to get excited about. Rather, raw material constraints (nickel, and eventually zinc and copper) still offer the better hope of price upside in the coming years. Fig 11 Marginal costs have fallen across all commodities over the past 24 months, but much more sharply for bulks % -1% -2% Change in marginal cost level over the past 24 months -2% -3% -4% -7% -8% -12% -15% -22% -32% -33% -33% -5% -6% -52% Source: LME, Platts, Wood Mackenzie, Macquarie Research, July 215 3 August 215 13

Jan-8 Jun-8 Nov-8 Apr-9 Sep-9 Feb-1 Jul-1 Dec-1 May-11 Oct-11 Mar-12 Aug-12 Jun-13 Nov-13 Apr-14 Sep-14 Feb-15 2 22 24 26 28 21 212 214 216F 218F 22F million tonnes Macquarie Research Steel demand growth blunted, peak pig iron brought forward If the Chinese economic slowdown has surprised in its severity, then the slowdown in steel consumption has been even more acute. And when you are talking about an economy which represents 5% of worldwide production and 45% of consumption then this becomes a global problem. With China now poised to be a much larger net exporter of steel than we had previously expected, we have taken the opportunity to update our global balance. The result is just.7% CAGR in global steel demand through the end of the decade, and China losing its position as the key consumption growth driver to India and other emerging Asian economies. With a relatively negligible growth rate in steel production, and the available scrap pool continuing to grow, 214 s global pig iron production now looks to have marked the peak for the foreseeable future. This has obvious implications for the raw materials, with required iron ore displacement now larger and more pressure on marginal met coal supply to exit permanently. China s construction sector has been important to the growth in global steel, and last year consumed over 3mt more than total consumption across the US and Western Europe combined. Its decline this year has been equally important in pulling global steel down in 215. The downturn in China s property sector has been well documented, and while property sales are now recovering strongly, early indicators for steel consumption remain weak developers seem content to run down inventory rather than advance new projects. YoY growth in property investment is now running at lower levels than in Q1 29, while prestressed concrete piles the first thing driven into the ground in medium-large building projects are seeing output down 1% YoY in 215 to date. We view this as a more tangible indicator than unaudited numbers such as new starts. That is not to say things won t improve into the second half after what has been an incredibly weak period indeed, we think they will as asset allocation towards property recovers. However, any aggressive uplift (such as that seen in H2 213) would have to be government directed, which currently looks unlikely, given they seem satisfied with the way the property market has trended in recent times. Given this, we see total construction demand for steel falling 6% this year to 286mt (crude steel equivalent), with the residential contribution down 1% YoY. Moreover, given that the base build for China is more like 1m residential units per annum than the ~12m units seen in recent years, we expect the general trend to be lower in the coming years. We may not yet have hit peak Chinese steel, but we likely have seen peak consumption in the construction sector. Fig 1 China s tangible construction market indicators continue to exhibit weakness... Fig 2 with peak steel in construction now past, and entering a period of steady decline YoY 5 % Property investment (3mma) Pre-stressed Concrete - RHS 6 35 Steel demand in Chinese construction 45 4 35 3 25 5 4 3 2 1 3 25 2 Industrial Commercial Residential 2 15 1-1 -2 15 1 5-3 -4 5 Source: NBS, Macquarie Research, July 215 Source: CISA, NBS, Macquarie Research, July 215 One area where we have changed our view is steel export volumes from China. While the overcapacity in the domestic steel industry has been apparent for the past couple of years, we thought a combination of rising demand and improving corporate behaviour (including capacity rationalisation and consolidation) would see export volumes drop over time towards ~3mtpa crude steel equivalent. Neither of those factors now look to be in place, with demand struggling and China s MIIT now encouraging exports as a solution to overcapacity, certainly delaying the hard decisions on capacity at least. As a result, we now model over 1mt of net exports on a crude steel equivalent basis for more details on where these will be targeted please refer to this recent report. 3 August 215 14

28 29 21 211 212 213 214e 215f 216f 217f 218f 219f 22f China Japan India Other Asia USA Latin America Western Europe CIS Other million tonnes million tonnes Jan-6 May-6 Sep-6 Jan-7 May-7 Sep-7 Jan-8 May-8 Sep-8 Jan-9 May-9 Sep-9 Jan-1 May-1 Sep-1 Jan-11 May-11 Sep-11 Jan-12 May-12 Sep-12 May-13 Sep-13 Jan-14 May-14 Sep-14 Jan-15 May-15 Chinese steel trade, crude basis, annualised, mt Macquarie Research Increasing China s production clearly has an impact on the global steel market essentially production expectations elsewhere have to drop, all other things being equal. Given the sustained period of low steel industry margins expected over the coming years, we have essentially removed any growth plans from the model from everywhere except India (where we have pared back production growth slightly), and moderated capacity creep expectations. Areas like Japan and the US will also have also seen production held at or below 215 levels the former due to China s impact on its key export markets and the latter due to a general lack of domestic industry competitiveness amid dollar strength. Fig 3 The growing gap between Chinese steel production and consumption... Fig 4 has led to net exports over 1mtpa on a crude steel basis, a level we now consider sustainable 25% 2% YoY change in Chinese steel production and consumption 16 14 12 Chinese steel exports Imports Exports Net Exports 15% 1 1% 8 5% 6 4 % 2-5% -1% Real Demand Steel production -15% Jan-14 Jan-15-2 -4 Source: NBS, Mysteel, CISA, Macquarie Research, July 215 Source: China Customs, Macquarie Research, July 215 While China is set to remain over 5% of global production, on our new modelling it never reaches 5% of steel consumption. Indeed, India and other Asian countries are set to become the new growth engines for the global steel economy. For 214-22, we now expect a global demand CAGR of.7%, but following the 2.3% drop in the current year, this means a 1.3% CAGR from this point forward. In terms of China, the equivalent 214-22 CAGR is zero, but post the 3.7% fall this year the CAGR is.7%, with autos and machinery gains offsetting construction losses. Meanwhile, we have India growing at a 7.5% CAGR far and away the best among major economies. Developed world consumption is essentially flat there is some upside risk here should we see a construction and infrastructure rebound, but given the fiscal pressures on many economies, these would likely have to be private sector-led rather than getting government kickers. Fig 5 China s steel consumption is in the peaking out phase... Fig 6 with India and other Asian countries now set to be the larger growth driver 1,9 1,8 1,7 1,6 1,5 1,4 1,3 1,2 1,1 1, 9 8 7 6 5 4 3 2 1-45 473 414 386 529 273 Crude steel consumption 53 496 497 514 531 548 565 583 483 494 451 655 675 729 756 728 735 739 747 752 755 67 344 366 36 36 381 377 381 381 378 374 37 4 35 3 25 2 15 1 5 - (5) (1) Change in steel consumption, 22 vs. 215 Developed China Emerging Markets Source: worldsteel, NBS, Macquarie Research, July 215 Source: worldsteel, NBS, Macquarie Research, July 215 3 August 215 15

Macquarie Research With a relatively negligible growth rate now projected for steel production, and the available scrap pool continuing to grow, global blast furnace iron output is in an even more challenging situation than steel. The leverage to China has been a benefit (and may well be again next year), but we are now looking at global production of 1,22mt in 214 marking the peak for the foreseeable future, and have global pig iron production 4mt lower than this by the end of the decade. Certainly, the scrap element does give some uncertainty to these forecasts, given that basic oxygen furnace steel mills have a certain degree of flexibility to substitute between produced pig iron and purchased scrap. The fixed asset element of the blast furnace and associated facilities (sinter plant, coke ovens) also add a layer of complexity. Presently, the equation is very much in favour of making pig iron using purchased raw materials to the greatest extent possible. This has the potential to delay the scrap cycle; however, we do tend to consider scrap inelastic in the long run such that any reprieve for pig iron output is likely to be a temporary one. Fig 7 Macquarie global steel production and consumption Steel Consumption - Crude Steel Basis (million tonnes) CAGR 213 214 215f 216f 217f 218f 219f 22f 214-22 China 729 756 728 735 739 747 752 755.7% Japan 69 72 69 7 7 7 7 7 -.5% India 78 8 86 93 1 17 115 124 7.5% Other Asia 152 158 155 161 166 169 173 176 1.8% USA 15 117 114 114 114 112 11 18-1.3% Latin America 81 86 82 84 86 88 9 91 1.% Western Europe 169 173 177 178 178 177 175 173.% CIS 78 7 66 68 69 7 72 73.8% Other 125 128 127 127 129 132 135 138 1.2% Total World 1585 164 162 163 1652 1672 1692 178 1.3% World, Ex-China 857 884 874 895 912 926 939 953 1.3% Crude Steel Production (m tonnes) CAGR 213 214 215f 216f 217f 218f 219f 22f 214-22 Japan 111 111 15 14 14 14 1 1-1.7% China 789 839 841 849 842 849 855 858.4% India 81 87 92 97 12 17 112 117 5.1% Other Asia 12 19 16 16 11 113 113 113.5% North America 119 121 111 113 113 112 112 111-1.4% Western Europe 179 18 177 179 18 18 181 181.1% Brazil 34 34 34 35 35 35 35 36.8% South/Central America 12 11 1 1 1 11 11 11 -.8% CIS 18 16 11 13 13 14 14 15 -.2% Other 81 83 84 84 84 85 85 86.5% Total World 1617 1681 1661 1679 1683 17 178 1716.3% World Ex-China 827 842 82 83 841 85 852 859.3% China share 48.8% 49.9% 5.6% 5.5% 5.% 5.% 5.1% 5.% Source: worldsteel, NBS, Macquarie Research, July 215 Peak pig iron output naturally means increased challenges for raw materials. Not least, the contestable iron ore market has hit saturation point, and is now in the peaking out phase which had previously been projected for nearer the end of the decade. With Chinese domestic iron ore, having already lost ~1mtpa over the past two years, now reaching the minimum expected volume, falling blast furnace output now means falling imports. This will be gradual rather than a collapse, but after over 2 years of consecutive growth including 28 and 29 is in itself a significant market event. 3 August 215 16

2 21 22 23 24 25 26 27 28 29 21 211 212 213 214 215f 216f 217f 218f 219f 1Q13 2Q13 3Q13 4Q13 1Q14 2Q14 3Q14 4Q14 1Q15 2Q15 3Q15F 4Q15F 1Q16F 2Q16F 3Q16F 4Q16F 1Q17F 2Q17F 3Q17F 4Q17F 1Q18F 2Q18F 3Q18F 4Q18F Macquarie Research Fig 8 Peak pig iron is set to mean falling Chinese imports... Fig 9 and with the majors still ramping, continued displacement of marginal capacity 18 Destination of contestable iron ore supply 16 China imports 14 China domestic 12 Ex-China imports 1 8 6 4 2 8 7 6 5 4 3 2 1 Supply required ex-majors, mtpa Source: Customs Statistics, NBS, Macquarie Research, July 215 Source: Customs Statistics, NBS, Macquarie Research, July 215 With the majors (BHP, Rio, Vale, FMG, Anglo and Roy Hill) continuing to grow output, the pressure will continue to bear on marginal players to make way. Under our previous forecasts, limited displacement was required from 217 onwards, with growth from the majors matching overall demand growth. With blast furnace output in a downtrend, even without growth from the majors we will need displacement, and with the growth between 6-8mt will be required each year through 219 not a pleasant environment for any iron ore producer. Fig 1 Macquarie iron ore balance more displacement to come million tonnes Pig iron + DRI production 213 214 215F 216F 217F 218F 219F China 716 749 737 743 725 717 77 Ex-China 464 471 466 469 475 48 48 Total 118 122 122 1213 12 1198 1187 Total ex-china seaborne iron ore demand 213 214 215F 216F 217F 218F 219F Europe 119 118 119 121 12 12 12 Japan 134 133 122 122 122 122 117 Korea 66 78 74 74 77 8 8 Other 69 13 114 16 15 16 16 Total 388 432 43 422 425 428 423 Total China iron ore demand 213 214 215F 216F 217F 218F 219F China demand 1144 128 1169 1192 1162 115 1133 Total iron ore demand (China + ex-china) 1531 164 1599 1613 1587 1578 1556 Seaborne iron ore supply 213 214 215F 216F 217F 218F 219F Total majors 867 111 116 1194 1245 1296 1336 YoY 144 95 88 5 51 4 Total variable supply annualised (after disruption) 665 629 492 419 342 282 22 YoY delta (displacement) 71 (36) (137) (73) (77) (6) (62) Source: worldsteel, Customs Statistics, Macquarie Research, July 215 3 August 215 17

Macquarie Research Our pig iron production downgrades also have bearish implications for our seaborne met coal supply-demand balance, which we have updated below. The key points are as follows: Ex-China met coal demand growth has been significantly reduced, from +43mt 214-22 to just +2mt 214-22. Europe aside, import projections for all major seaborne importers have been cut. India remains the only source of significant demand growth going forward. Chinese imports, now a function of protectionist trade policy, have been reduced by 15-2% from 215 to 218. This to reflect current import run-rates, which are below our previous expectations. In addition, we see China hitting peak metallurgical coal requirements in 216. The result of the former two points is that US export projections have been slashed. We now see steady state exports as being under 3mt, versus ~4mt previously. Fig 11 Metallurgical coal seaborne supply and demand Grow th Met coal demand (mt) 212 213 214 215F 216F 217F 218F 219F 22F 214-22 Europe 54 53 56 55 56 57 57 57 57 Japan 57 57 55 53 53 53 53 51 51-4 Korea 28 31 33 33 33 35 36 36 36 3 Taiw an 9 1 11 12 12 12 12 12 12 1 China 52 73 62 42 32 25 21 22 21-41 India 38 4 47 52 54 57 6 62 65 18 Brazil 15 15 19 2 2 2 2 2 2 1 Other 5 9 11 11 11 11 11 11 11 1 Total Demand 258 289 294 277 271 269 269 271 273-2 % change YoY 7.5% 12.% 1.7% -5.5% -2.2% -.7% -.6%.9% Grow th Met coal supply (mt) 212 213 214 215F 216F 217F 218F 219F 22F 214-22 Australia 145 17 186 188 187 185 183 183 183-4 Canada 3 34 3 27 27 27 27 27 28-3 USA 59 57 51 36 3 28 28 28 28-23 China 1 1 1 1 1 1 1 1 1 - Russia 1 14 13 13 13 13 13 13 14 1 Other 13 13 13 13 14 16 18 19 2 8 - Total Supply 258 289 294 277 271 269 269 271 273-2 % change YoY 7.% 12.% 1.7% -5.5% -2.2% -.7% -.6%.9% Source: Customs Data, Macquarie Research, July 215 Our ex-china demand revisions put into focus just how reliant the seaborne met coal market is going to be on Indian demand going forward. Unlike thermal coal, where concerns are building that domestic production could sort itself out to the detriment of imports, Indian domestic met coal qualities are poor and yields are low. India remains the market s structural demand growth driver, albeit our projections are more conservative than before. Meanwhile demand growth from all other importers is, on our forecasts, going to be more-or-less flat over the next five years. In previous years, this displaced volume may have been able to find a home in China, by pricing in. That is no longer possible with Chinese imports now effectively determined by protectionist policy rather than supply-push, which was the main driver in previous years. New coal quality standards introduced at the start of this year trace element inspections in particular have helped keep seaborne coal out and the risk is that this protectionism increases further. The implication is that we will need to see much larger supply cuts than we were previously assuming, almost all of which we have attributed to the US. There is no disputing that US met coal producers account for the marginal tonnes in the market and that if economics were to play out freely, these would be the cuts addressing market oversupply. 3 August 215 18