The behavior of base m etals prices So Paulo Novem ber 5 , 2 0 0 9 - PowerPoint PPT Presentation

The behavior of base m etals prices São Paulo Novem ber 5 , 2 0 0 9 1

Agenda  Unique features  The price behavior  Facts and fantasies  Minerals and metals for the long run 2

Unique features 3

Unique features  Commodity investing  Commodities and the weather  Commodities and inflation  Price elasticities  Cyclical drivers 4

The price behavior 5

Price performance  Key features  Asymmetric volatility  Price co-movement 6

Fat tails: large numbers of extreme values Aluminum Nickel Copper Freight Kurtosis 1 15.01 129.30 8.94 17.50 1 - The kurtosis coefficient measures the magnitude of the extreme values of the distributions. If returns are normally distributed, then the kurtosis should be three. 7

Unpredictability and volatility clustering Alum inum prices 7,000 12% Daily returns 6% 6,000 0% - 6% 5,000 - 12% 4,000 - 18% - 24% US$/ m etric ton 3,000 - 30% - 36% 2,000 - 42% 1,000 - 48% 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 8 8 9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 9 9 9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 Source: Vale and LME 8

Unpredictability and volatility clustering Copper prices 20,000 20% 18,000 10% 16,000 Daily returns 14,000 0% 12,000 10,000 - 10% 8,000 US$/ m etric ton - 20% 6,000 4,000 - 30% 2,000 0 - 40% 19 88 19 89 19 90 19 91 19 92 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 Source: Vale and LME 9

Unpredictability and volatility clustering Nickel prices 20% 90,000 10% Daily returns 75,000 0% 60,000 - 10% US$/ m etric ton 45,000 - 20% 30,000 - 30% 15,000 - 40% 0 - 50% 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 8 8 9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 9 9 9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 Source: Vale and LME 10

Brief summary of econometric tests  Autocorrelation in returns are not significant  unpredictability  Autocorrelation in square returns are significant  volatility clustering 11

Stock prices are inversely related to volatility S&P 5 0 0 x VI X 1,600 90 S& P 500 1,500 80 VI X 1,400 70 1,300 60 1,200 1,100 50 1,000 40 900 30 800 20 700 600 10 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Source: Bloomberg 12

Volatility asymmetry: base metals prices are positively related to volatility Volatility asym m etry¹ 1987 - 2009 Up cycle Dow n cycle S&P 5 0 0 0.86% 1.73% Alum inum 2.06% 1.64% 2.08% 1.90% Copper Nickel 2.97% 2.40% ¹ One month moving average of the standard deviation of daily returns Source: Vale 13

Prices co-movement in the short-term influenced by common factors Co-m ovem ent betw een returns: sim ple correlation Jan/ 04 - Oct/ 09 Iron Thermal Copper Aluminum Oil Corn Soybeans Wheat ore coal Nickel 0.54 0.53 0.33 0.18 0.34 0.22 0.22 0.15 0.28 0.37 0.23 0.27 0.27 Copper 0.72 0.63 Aluminum 0.26 0.45 0.56 0.29 0.29 0.20 Iron ore 0.19 0.29 0.18 0.20 0.32 Thermal Coal 0.50 0.24 0.36 0.24 Oil 0.29 0.31 0.21 Corn 0.76 0.53 Soybeans 0.57 Source: Vale 14

Price co-movement in the long term: specific market fundamentals prevail Co-integration analysis March 1 9 8 7 – October 2 0 0 9 Energy and All Food and Metals metals commodities metals prices prices prices High degree of co-movement Low degree of X X X X co-m ovem ent 1 1- We can say that the degree of the co-movement is low because there is only one co-integration vector. To find a strong co-movement between n variables it would be necessary to have n-1 co-integration vectors. Source: Vale 15

The relationship between iron ore and maritime freight prices  There is a high correlation betw een m aritim e freight and iron ore prices.  They also behave in a sim ilar m anner over the long term .  There is a com m on trend w hich influences the behavior of both prices in the sam e direction. 16

The relationship between iron ore and maritime freight prices Analysis of com m on trend influences % of freight price % of iron ore price Horizon variations explained variations explained by com m on trend by com m on trend 1-year 91.7% 33.7% 2-year 93.5% 69.8% 5-year 93.1% 85.8% 1 0-year 94.6% 87.9% 17

Price co-movement is likely to increase during financial crises Co-m ovem ent¹ Base metals and oil Base metals and food 1 .0 0 .8 0 .6 0 .4 0 .2 0 .0 -0 .2 -0 .4 -0 .6 -0 .8 1 9 8 3 1 9 8 5 1 9 8 7 1 9 8 9 1 9 9 1 1 9 9 3 1 9 9 5 1 9 9 7 1 9 9 9 2 0 0 1 2 0 0 3 2 0 0 5 2 0 0 7 2 0 0 9 Base metals  LMEX 1- Measured by 1-year correlation of price returns Food  CRB food 18 Source: Vale

Facts and fantasies 19

Facts and fantasies about minerals and metals prices  Mean reversal  Financial investments  The Chinese iron ore stockpiling 20

Iron ore prices are not doomed to revert to the mean Real iron ore prices¹ US$ cents/ fe-dmt 140 120 100 80 60 40 20 0 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 ¹ Nominal prices adjusted by the September 2009 US PPI Source: Vale 21

Base metals prices are not doomed to revert to the mean Alum inum price¹ Copper price¹ Nickel price¹ US$/ ton US$/ ton US$/ ton 60,000 6,000 9,000 5,500 8,000 50,000 5,000 7,000 4,500 40,000 6,000 4,000 3,500 5,000 30,000 3,000 4,000 20,000 2,500 3,000 2,000 10,000 2,000 1,500 1,000 1,000 0 1987 1992 1997 2002 2007 1986 1991 1996 2001 2006 1987 1992 1997 2002 2007 ¹ Monthly nominal prices adjusted by the September 2009 US PPI 22 Sources: Vale and LME

Brief summary of econometric tests  Unit root tests do not support the mean reversal hypothesis Unit root test: Augm ented Dickey-Fuller 1 9 8 7-2 0 0 8 Null hypothesis: price has a unit root Prices 1 I ron ore Nickel Copper Alum inum Augmented Dickey- 0.28 0.66 0.93 0.14 Fuller test P-value ¹ Annual real prices adjusted by US PPI Sources: Vale 23

However, metals price volatility reverts to the mean Nickel price volatility¹ 18 % 16 % 14 % 12 % 10 % 8 % 6 % 4 % 2 % 0 % 198 7 1 989 19 91 1993 1 995 19 97 1999 2 001 20 03 2005 2 007 200 9 ¹ Standard deviation of 20-trading day moving average of daily returns. 24 Sources: Vale and LME

Metal prices are determined by fundamentals Copper: prices and long positions 60 10,000 Net long positions Spot price 50 9,000 40 8,000 30 000 num ber of contracts 7,000 20 6,000 US$/ ton 10 5,000 0 4,000 - 10 3,000 - 20 2,000 - 30 - 40 1,000 1995 1997 1998 2000 2002 2004 2006 2008 Source: Vale, NYMEX and CFTC 25

Metals prices are determined by fundamentals  Financial investments are caused by prices, it does not cause prices. Granger causality test March 1995 – October 2009 Null hypothesis Obs F-Statistic Probability Long positions does not 380 0.41271 0.66215 Granger causes spot price Spot price does not Granger 380 9.90226 0.00064 causes long positions 26

There is no evidence of stockpiling Days of iron ore consum ption Days of iron ore im ports Mt Days Mt Days 140 40 140 60 35 120 120 50 30 100 100 40 25 80 80 20 30 60 60 15 20 40 40 10 Stocks at the ports Stocks at the ports 10 20 20 5 0 0 0 0 Sep-08 Nov-08 Jan-09 Mar-09 May-09 Jul-09 Aug-09 Oct-09 Sep-08 Nov-08 Jan-09 Mar-09 May-09 Jul-09 Aug-09 Oct-09 27

Minerals and m etals for the long run 28

Is there a bubble? US$/ ton Δ % All-time high Current price¹ Iron ore 205 92 -55.1 Aluminum 4,290 1,903 -55.6 Copper 8,985 6,575 -26.8 54,200 18,465 -65.9 Nickel 2,251 1,326 -41.1 Platinum² 145 77 -47.0 Oil³ Thermal coal 184 72 -60.8 Soybeans 4 1,649 978 -40.7 Wheat 4 1,283 494 -61.5 ¹ October 30, 2009 ² US$ per oz ³ US$ per barrel 29 4 US$ per bushel

This time metal prices were much more volatile than in the recessions of the last 40 years Commodity prices in global recessions and recoveries¹ -46% -26% Aluminum 38% 5% -56% -30% Copper 102% 5% -65% -32% Nickel 80% 13% Current recession Average of last 5 recessions Current recovery² Average of last 5 recoveries ¹ Recessions are measured as the percentage change of metal prices from the beginning of each recession to trough; recoveries are measured as the percentage change of metal prices from the trough to the end of each recession. Global recessions: 1974, 1981-82, 1990-91, 1998 and 2001. ² Percentage change from last trough until September 2009. 30 Source: Vale and IMF

The sharp drop of metal prices was caused by the steep decline of manufacturing output … Global industrial production grow th % 3mma, saar¹ 20 10 0 - 10 - 20 - 30 - 40 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 ¹ 3-month moving average, seasonally adjusted annualized rate 31 Source: Vale and JP Morgan