asia's imprint on global commodity markets

8/22/2019 Asia's Imprint on Global Commodity Markets

1/58

December 2006

ERDECONOMICS AND RESEARCH DEPARTMENT

Working PapeSERIESNo.

90

Cyn-Young Park and Fan Zhai

Asias Imprint on GlobalCommodity Markets

Asias Imprint on GlobalCommodity Markets


2/58

ERD Wrin Paper N. 90

AsiAs imprinton GlobAl CommoditymArkets

Cyn-younG pArkAnd FAn ZhAi

deCember 2006

Cyn-Young Park is senior economist and Fan Zhai is economist at the Macroeconomics and Finance Research Division,

Economics and Research Department, Asian Development Bank. The authors are grateul or valuable assistance rom

Lea Sumulong and wish to acknowledge helpul comments rom Frank Harrigan and the Division sta. The authors

are solely responsible or any remaining errors.


3/58

Asian Development Bank6 ADB Avenue, Mandaluyong City1550 Metro Manila, Philippines

www.adb.org/economics

2006 by Asian Development BankDecember 2006

ISSN 1655-5252

The views expressed in this paper

are those o the author(s) and do notnecessarily reect the views or policies

o the Asian Development Bank.


4/58

FoREWoRD

The ERD Working Paper Series is a orum or ongoing and recently completedresearch and policy studies undertaken in the Asian Development Bank or onits behal. The Series is a quick-disseminating, inormal publication meant tostimulate discussion and elicit eedback. Papers published under this Series

could subsequently be revised or publication as articles in proessional journalsor chapters in books.


5/58

CoNtENts

Abstract vii

I. IntroductionI. Introduction 1

II. Commodity Prices ong-term Trends and Recent DevelopmentsII. Commodity Prices ong-term Trends and Recent Developments 2

A. ong-run TrendsA. ong-run Trends 2 B. Recent Developments 4

III. Developing Asias Imprint on lobal Commodity Markets 1III. Developing Asias Imprint on lobal Commodity Markets 14

A. rowth, Structural Change, and Commodity Demand 1A. rowth, Structural Change, and Commodity Demand 14 B. Evolving Patterns o Commodity Demand and Trade 18

I. ooking Ahead Prospects o lobal Commodity Markets into 2015 2I. ooking Ahead Prospects o lobal Commodity Markets into 2015 25

A. Baseline Scenario 2A. Baseline Scenario 25 B. Robustness Check 29

. Conclusions . Conclusions 2

Appendix EMAT A lobal eneral Equilibrium Model Appendix EMAT A lobal eneral Equilibrium Model 5

Selected Reerences Reerences 8


6/58

AbstRACt

Dynamic growth patterns o developing Asia will continue to make strongimpressions in world commodity markets. Driven by rapid income growth andeconomic development, developing Asia has suraced as a major demand orcebehind the price dynamics o primary commodities. The regions economic

growth and development has been tightly associated with rapid industrialization,urbanization, and massive inrastructure investments, all o which are highlyresource-intensive. These trends are set to intensiy as Asias mammoth economies

emerge. Real incomes in the Peoples Republic o China are now reaching a levelat which demand or energy and resource-intensive consumer durable goodsusually takes o. India may not be too ar behind in its catch-up process. Thepaper provides an overview o Asian inuence in world commodity markets and

examines its changing patterns. It also attempts to quantiy the impact o therapidly growing Asian economy on long-term resource utilization, using a eneralEquilibrium Model or Asian Trade (EMAT) to project regional growth scenarios or20052015. The model captures long-run equilibrium tendencies in product and

actor markets or natural resources. The estimated results point to undamentalchanges in market dynamics or a broad range o primary commodities.


7/58

I. INtRoDuCtIoN

Since the beginning o 2002, primary commodity prices have surged. The overall index o

primary commodity prices has more than doubled, driven by strong energy and metal prices. Energyprices rose by about 2% on average each year during 20022005 on signifcant increases in oiland natural gas prices. Prices o nonenergy commodities also rose by 11% annually in the sameperiod as base metal prices accelerated. Although metal prices have eased rom this years May

peak, the broad picture remains intact. Soaring prices can be traced to robust global demand andrestraints on supply stemming largely rom chronic underinvestment in earlier years, particularlyin energy and metal sectors.

Increasingly, global commodity price dynamics have been inuenced by the role o developingAsia as a consumer rather than producer o primary commodities. For example, the commodity priceupswing o the 1990s and the all in commodity prices in the late 1990s correlates with a period

o ast growth in Asia, subsequently punctured by Asias fnancial crisis. Again, more recently, theresumption o ast growth in developing Asia has helped drive global commodity prices up.

Allied to rapid growth are proound structural changes that are eeding Asias craving or

commodities. Industrialization, urbanization, and massive inrastructure investment have all increasedthe energy- and resource-intensiveness o output. The impact o an emerging middle class and risingauence on consumer preerences is also to exert leverage on demand or ood commodities. Thesetrends are set to intensiy as Asias mammoth economies emerge. Real incomes in the Peoples

Republic o China (PRC) are now reaching a level at which demand or energy- and resource-intensiveconsumer durable goods, such as automobiles, usually takes o. India, while lagging behind, maysoon begin to catch up (Asian Development Bank 2006). Asias rapid transormation will do more

than drive price. It is also likely to spur competition in trade and investment, which will reshape

commodity market dynamics and change global commodity prices in a more undamentally dierentway.

Changing dynamics o primary commodity prices have important implications or both developingAsia and the world. This paper assesses the impacts o emerging Asian economies on global commoditymarkets in the process o their rapid growth and development and examines challenges arising romthese trends. The ollowing section reviews commodity price dynamics in terms o both long-term

trends and recent developments. Ater a brie summary o overall commodity market dynamicsrom a long-term perspective, the actors that are responsible or the recent boom in commodityprices are examined in dierent sectors. The subsequent section will take a closer look at Asian

inuence behind the commodity price movements. Evidence presented here supports the view thatdeveloping Asia is playing an increasingly important role as a consumer in pushing up the prices

o primary commodities. Historical patterns also suggest that such evolving consumption patternsmay continue to shape the uture o world commodity trade. In the next section, a model-based

analysis helps quantiy the impact o developing Asia on global commodity demand and prices.The results provide the outlook or global commodity prices into 2015 based on the assumptionthat there are little changes in developing Asias economic and policy environments over the nextdecade. Simulations based on dierent sets o economic and policy assumptions or developing Asia

suggest an important role or policy in inuencing outcomes.


8/58

The paper concludes by discussing policy issues arising rom the longer-term prospects o

commodity prices. Challenges arise rom developing Asias thirst or a broad range o commodities andhow policies should address meeting the growing commodity needs eectively without precipitatingresource depletion and environmental degradation. These issues are tightly linked to other major

challenges that developing Asia aces regarding efcient resource management, environment protection,and poverty reduction, hence cutting across a broad array o development policies.

II. CommoDIty PRICEs: loNg-tERm tRENDs AND RECENt DEvEloPmENts

A. ln-rn trend

Figure 1 shows the path o the price indexes o both energy and nonenergy commodities since1980 (=100) (Box 1). Both plots show a sharp escalation in global commodity prices since 2002.

The recent surge in primary commodity prices has been driven by strong global recovery since 2002,combined with sustained high growth in developing Asia. However, a number o actors behindthe recent strength o commodity prices may reect long-term structural changes, which may have

enduring impact on market undamentals.

FIGURE 1

COMMODITY PRICE INDEXES

2001801601401201008060402001980 1983 1986 1989 1992 1995 1998 2001 2004

1980

=

100

Nonenergy Energy

Source: Primary Commodity Prices Database, downloaded 9 October 2006(International Monetary Fund 2006b).

box 1primAry Commodities

Primary commodities are defned as raw materials and industrial inputs that are close to the initialproduction stage. The price indexes o primary commodities employed throughout this paper are the

International Monetary Fund (IMF) commodity price indexes.The IMF overall commodity price index covers a total o 41 commodities. The overall price index

is divided into energy (47.8%) and nonenergy (52.2%) subindexes. The components o the nonenergysubindex are ood and beverages (ood [21.7%] and beverages [.1%]) and industrial inputs (agriculturalraw materials [11.%] and metals [16.1%]). The energy subindex is comprised o prices o crude oil

(9.9%), natural gas (4.5%), and coal (.4%). The weights used or the IMF commodity price indexesare 19951997 average world export earnings. Data or its nonenergy subindex starts in January 1980.

Since its energy subindex begins only in January 1992, crude oil price is used as a substitute or theenergy price subindex rom 1980 to 1991.

December 2006

AsiAs imprint on GlobAl commoDity mArkets

cyn-younG pArk AnD FAn ZhAi


9/58

Historically, uctuations in commodity prices have been highly associated with the global

business cycle (Figure 2). As the global economy enters into an expansionary phase, inventories alland a urther increase in demand pushes commodity prices sharply higher. However, as commodityprices become elevatedpartly responsible or heightened inationary pressures during the economic

boom, thus inviting contractionary policies and also slowing economic activityadditional productioncomes on stream and helps balance supply and demand. Once the business cycle begins to ebb,demand or primary commodities declines and prices begin to reverse. Cyclical adjustments couldbe steep i high prices in the expansion phase lead to a glut in inventory and supply.

FIGURE 2

GROWTH IN REAL GDP AND NONENERGY COMMODITY PRICES

6

5

4

3

2

1

0

15

10

5

0

-5

-10

-151980 1983 1986 1989 1992 1995 1998 2001 2004

Percent

Percent

GDP (left scale) Nonenergy commodity prices (right scale)

Sources: Primary Commodity Prices Database (International Monetary Fund2006b); World Economic Outlook Online Database, downloaded 16October 2006 (International Monetary Fund 2006c); CommodityPrice Indexes (World Bank Development Prospects Group 2006).

iven the generally high volatility o commodity prices, it is useul to put the recent evolution

in a longer-term perspective. Despite recent increases in commodity prices, real prices o nonenergy

commodities are deemed to be relatively low by historical standards ater a long-run decline (Box 2).Cashin and McDermott (2002) fnd that real commodity prices have declined by about 1.% per year

over the past 140 years. Technological innovations that have opened up new sources o supply andincreased total actor productivity; the development o synthetic substitutes; reduced transportationcosts; trade liberalization, economic reorms and changes in market structure; and fnancial innovationhave all contributed to a long-term trend o decline in commodity prices.

In sum, commodity prices are inuenced by both cyclical and structural orces. The recentsurge in commodity prices has been tightly linked to the strong global recovery that began in2002 and hence bears some o the traits o earlier episodes o commodity price ination during an

economic upswing. While there is much amiliarity about the current pattern o upswing in commodityprices, the past is not always a perect guide or the uture. There are two new elements behind

the recent upward movements as to suggest some ongoing structural changes in underlying marketundamentals have longer-term consequences. First, the pull o developing Asia, and particularly thePRC, on global commodity markets has a proound impact on world commodity market dynamics.To the extent that growth in developing Asia is driven by opportunities or economic catch uprather than the international business cycle, with the weight o developing Asia rising in the

global economy, there may now be greater persistence in the actors driving demand or primarycommodities. Second, there is now much greater dissonance about the energy supply outlook than

section ii

commoDity prices: lonG-term trenDs AnD recent Developments

erD WorkinG pAper series no. 90


10/58

ever beore. The inuence o geopolitical actors is certainly clouding the prospect o stable supply

in the short run, which could possibly drag on in the long run too. While these actors seem to mostdirectly aect crude oil markets, the generally high energy intensity o other commodity productionmay signal a more broad-based boom in commodity prices. Apart rom these two major elements,

structural actors are oten sector-specifc. The ollowing sections review the recent developmentsin various commodity markets and prices.

box 2lonG-run Commodity priCe dynAmiCs

Prebisch (1950) and Singer (1950) claim that the prices o primary commodities should decreaseover time relative to those o manuactured goods, due to (a) the lower income elasticity o demand orcommodities and (b) productivity and wage dierences between industrialized countries that produce

manuactures and developing countries that relatively lean toward commodity production. The Prebisch-Singer (P-S) hypothesis, which was named ater Prebisch and Singer or their seminal work, has been

one o the most intensely tested hypotheses in the economic literature.Empirical evidence on the P-S hypothesis has been largely mixed depending on the model

specifcation that identifes the sources o the downward trend. A school o studies that suggests

there has been a declining tendency in primary commodity prices relative to manuacturing prices,thus in support o the P-S hypothesis, include rilli and Yang (1988) and Ardeni and Wright (1992).

Spraos (1980) reports a declining trend o 7.% per year between 1900 and 198, but no statisticallysignifcant trend is detected when the data series is extended to 1970. Cashin and McDermott (2002)

confrm a trend decline in the series by 1.% per year, but claim that high variability o commodityprices renders a small long-term decline nearly meaningless. In contrast, Cuddington and Urza (1989),

Powell (1991), Bleaney and reenaway (199), and Cuddington et al. (2002) fnd structural breaks inthe series, but no deterministic trend.

Another salient eature in the commodity price movements is the large variability and generally

high persistency in their cycles. Cashin et al. (1999), in examining the cyclical behavior o commodityprices, fnd that a price slump tends to last longer and that the magnitude o the slump is oten greater

than that o the price boom. But they argue that the duration o the shocks cannot be determined by thetime already spent either in the slump or boom. Deaton and aroque (1992) also identiy an asymmetry

that commodity prices spend long periods in doldrums with occasional extreme spikes. Shocks to realcommodity prices are also ound to be quite persistent, lasting or several years or longer in each cycle.Cashin et al. (2000) also confrm that shocks to commodity prices are typically long-lasting.

The validity o the P-S hypothesis or the existence o a deterministic trend in real commodity pricesremains in question with many time-series properties o commodity prices let unanswered. However,

empirical fndings point to highly responsive commodity prices to shocks over the short to medium term.Over the long term, more persistent changes in the underlying supply and demand conditions seem to

dictate commodity price movements.

b. Recen Deepen

1. Ener Cdiie

Figure shows the trajectory o the price index o the energy basket along with the benchmarkBrent crude price rom 1980 until September 2006. Since 2002, prices have more than tripled. Amongall the commodity classes, recent price ination has been most pronounced or energy. Since energyis also an intermediate input into the production o many other primary commodities, movements

December 2006




11/58

in the price o energy reverberate through other markets. Crude oil prices, which are primarily

responsible or energy price movements, have risen steadily higher, leading to an impressive run-upwell into the frst hal o 2006.

FIGURE 3

ENERGY AND BRENTCRUDE PRICE INDEXES

200

150

100

50

01980 1983 1986 1989 1992 1995 1998 2001 2004

1980

=

100

Energy Brent crude


Oil prices are driven mainly by supply and demand actors like most other commodities.Crude oil, a still dominant resource or generating energy, has been particularly vulnerable to

supply disruptions, as its production is much more concentrated geographically than other naturalresources. The impact o geopolitical risks on global oil prices has been phenomenal as seen inthe drastic rises in oil prices during the frst and second oil shocks. Excluding the events o supplydisruptions, however, the crude oil price cycle, which may extend over several years each time, has

largely responded to changes in world demand.

The world demand or oil is dominated by industrialized countries, led by the roup o

Seven (7) accounting or roughly hal o the total demand over the last two decades. However,oil demand growth has been much aster in the developing world, apart rom the ormer SovietUnion (FSU), which experienced an economic collapse in much o the 1990s. Oil demand in totaldeveloping economies, including developing Asia, atin America, and Middle East (in descending

order according to share o total demand), grew by 6.2%, or 78.8% excluding FSU, compared to12.5% growth in the 7 over the 19902005 period. Developing Asia alone accounted or nearly60% o world oil demand growth during the same period. Figure 4 illustrates the rising importanceo Asian economies in oil price developments since 1990.

On the other hand, a rather long period o relatively low prices through the late 1980s to the1990s resulted in underinvestment in general on the supply side, creating signifcant bottlenecksalong the line o the supply chain, as global oil demand accelerated. Despite marginally improving

supply/demand conditions in the upstream oil market, generally low levels o spare capacity (Figure5) as well as binding constraints in the upstream production and downstream inrastructure continueto put upward pressures on global oil prices.

There are signs that indicate signifcant uncertainties surrounding supply/demand conditions orsome time ahead. The oil utures market continues to be volatile, reecting the heightened sense

section ii




12/58

o supply risks. The movement o utures prices has closely mirrored the uctuation o spot prices,

which tend to respond more strongly to news (Figure 6). A rise in utures contract prices in alltime horizons suggests a market assessment o an unlikely meaningul improvement in oil supplyand demand balance in the near term (Figure 7). Moreover, oil utures prices are again in contango

(distant utures prices exceed spot prices), encouraging a continuing buildup in inventories, whichitsel is usually a tell-tale sign o jittery market sentiment (Box ). Indeed, average commercialstock holdings in the Organisation or Economic Co-operation and Development (OECD) countrieshave steadily increased since 2004, despite rising prices.

FIGURE 5

GLOBAL SPARE OIL PRODUCTION CAPACITY

12

10

8

6

4

2

0

1980 1985 1990 1995 2000 2005

mb/d

OPEC-11 OPEC-10

Source: World Economic Outlook Online Database, Table 1.21 downloaded 23June 2006 (International Monetary Fund 2006c).

FIGURE 4A

OIL DEMAND BY REGION, 1990

FIGURE 4B

OIL DEMAND BY REGION, 2005

Developing Asia

13% Developing Asia

22%Middle East

5%

Latin America5%

Other

developingeconomies

16% Rest of the World

12%

Middle East

7%

Source: Statistical Review of World Energy 2006, downloaded 30 June 2006 (BP 2006).

G7

49% G744%Latin America

6%

Other

developing

economies8%

Rest of the World

13%

December 2006




13/58

$/bbl

80

70

60

50

40

30

20

10Jan-00

Source: Datastream, downloaded 24 August 2006.

Jan-01 Jan-02 Jan-03 Jan-04 Jan-05 Jan-06

9-11 terrorist attacksIraq invasion

Tight gasoline supplies in the US due tohurricane Ivan

Explosion at BPs Texas oil refineryHurricanes Katrina and Rita

Iran removes UN seals at its uranium enrichment facilitySupply disruptions in Nigeria

Fire at Venezuelas Amuay refinery

FIGURE 6

CHRONOLOGY OF BRENTCRUDE OIL PRICE SPIKES

FIGURE 7

BRENTSPOT AND FORWARD PRICES

78

76

74

72

70

68

66

64

62

60

$/bbl

Jan-06

Mar-06 Jun-06 Jul-06

Apr-06 Jul-06 Oct-06 Jan-07 Apr-07 July-07 Oct-07

Spot

Source: Datastream, downloaded 8 August 2006.

section ii




14/58

box 3oil priCes, reFineries, And Futures

Crude oil quality is important or refnery margins as it determines the level o processing andreprocessing required to obtain the optimal output portolio. Depending on its density and sulur

content, crude oil is classifed into light or heavy, sweet or sour. As lighter and sweeter crude isrelatively easy to refne and produces greater yields o high-quality products (also required by tightenedenvironmental regulations), world demand is increasingly driven by this crude grade. But recent additions

to production capacity have been rather concentrated in heavy and sour crude grades that require morecomplex refning acilities. iven the relative ease o supply in heavy and sour crude as well as greater

proftability o downstream processing, additional refnery capacity has leaned toward more complex andupgrading acilities since the 1980s.

As one o the worlds biggest oil consumers, the United States (US) accommodates the worldslargest refning acilities. It accounts or about a quarter o world crude distillation capacity, with theul Coast area hosting nearly hal o the countrys capacity. Downstream processing capacity is even

more concentrated in the US, especially in the ul Coast area. Thereore, when Hurricane Katrina hitthe region, it caused extensive damage to upgrading capacity, and the prices o some refned products,

particularly gasoline, shot up. Higher gasoline prices increased the dierence between the prices o

refned products and the prices o crude oilcalled crack spread. This stimulated refneries with lesssophisticated processing capacity to come on stream. However, these simpler refneries have relativelylower product yields (i.e., higher crude input demand), thus pushing crude oil demand and prices higher.In addition, many o them cannot even take heavy crude. As they sought light crude to maximize their

yields, the prices o light crude surged even urther.The rise in the crack spread also aected the utures markets. Refneries proft rom wide crack

spreads, but i the gasoline price alls at the time o sale or crude prices suddenly rise, the refnerieswill lose substantially. Thus, refneries have an incentive to hedge against price risks, by taking a short

position in gasoline utures (a legal obligation to sell gasoline at an agreed uture time at an agreedprice) and a long position in crude (a legal obligation to buy crude oil at an agreed uture time atan agreed price). In last years period o uncertainty, refneries started to buy crude utures, bidding

utures prices up, and to sell gasoline utures. But given the shortage in upgrading capacity, gasolineprices were unlikely to all signifcantly, thus the narrowing crack spreads came mostly rom rising crude

prices.The utures market situation in turn reinorced spot market conditions. iven tight refnery

conditions and unlikely improvement or the next couple o years, utures prices surged. Normally, aslonger-dated utures prices rise much higher than spot prices or near-month utures (contango), refnershave an incentive to hold larger inventories. This is an unusual situation generally, oil utures prices

are in backwardation (i.e., spot prices are higher than utures prices), reecting the convenienceyield, i.e., what refners will pay to hold stocks or ensuring smooth day-to-day operations (bearing in

mind that crude can be stored most efciently and at lowest cost with producers, not the refners). Thisconvenience yield is greatly discounted or distant utures, say 12 months ahead. However, when distant

utures prices are signifcantly higher than spot prices, thus creating a sufciently wide contango,refners are willing to hold the actual oil and to pay or the cost o carry. This leads to an increase inspot prices and, at the same time, a buildup in inventories.

Overall, strong market undamentals, i.e., robust demand and tight supply, have been the mainreason or currently high oil prices. Nevertheless, underlying market structures appear to play an important

role by reinorcing the crude oil/refnery products price dynamics through the utures markets.

December 2006




15/58

. Nnener Cdiie

Nonenergy commodities have seen the price rally that began in 2002 and stretched into May2006, largely on strong base metal prices (Figure 8). Despite some similarities in price movementsacross groups, individual commodities are requently subject to idiosyncratic inuences, both short-

term and structural (the supply o agricultural commodities, or example, is inuenced by weather).One common actor is the inuence o energy prices as many nonenergy commodities are energy-intensive in their production (Box 4).

. bae mea

Prices o base metals have made handsome gains since 2002 on the back o strong demand,low inventories, and high oil prices (Figure 9). Driven by strong demand or industrial productionin developing Asia, including the PRCthe PRC is currently the worlds biggest importer o iron andsteel, and second-biggest importer o metal ores and nonerrous metalsthe prices o aluminum,

copper, iron ore, lead, nickel, zinc, and other widely used metals have surged in recent years. On

average, base metal prices have grown at an annual rate o 24.9% rom 2002 to 2005. Copper, ironore, lead, nickel, and zinc (in descending order) have been the strongest movers. These metals are

generally used in steel production (a rapidly growing industry in the PRC and India), stainless steel,electrical wire, cable, and building inrastructure. Copper prices more than tripled in the period,whereas the prices o iron ore, lead, nickel, and zinc more than doubled.

The steady and steep increase in metal prices over the past several years accelerated evenurther in early 2006, largely on the ground o metals attractiveness to investors in search ohigher yields and driven by the prospect o capital gains. Investing in various commodities as apart o hedging against fnancial risks or portolio investment has a long history. The past ew years,

however, have witnessed a urry o hedge unds and institutional investors into commodity uturesmarkets in a hunt or better yields given the historically low interest rate environment (Box 5).This asset demand became more pronounced in early 2006, as the dollar precipitated the all amid

heightened global fnancial market and interest rate uncertainties. Changes in market expectationsor global liquidity conditions in mid-May brought about a large sell-o across a broad range ofnancial markets, including commodity markets. This underscores the underlying asset demand inthe latest rally in commodity prices.

section ii


1980

=

100

1980 1983 1986 1989 1992 1995

250

200

150

100

50

0

1998 2001 2004

Food and beverage Agricultural raw materials Metals

FIGURE 8

NONENERGY COMMODITY PRICE INDEXES


erD WorkinG pAper series no. 90 9


16/58

box 4enerGyAndthe CoststruCtureoF primAry Commodity produCtion

Higher energy prices generally translate into higher commodity prices. Typically, the production obasic commodities and semifnished manuactures (e.g., ertilizers) is highly energy-intensive compared

with that o manuactures. For example, most base metals including aluminum and steel are known ortheir energy-intensive production processes such as metallurgy and smelting. Paper and pulp, cement,and ertilizer industries tend to rank high in their use o energy. Highly energy-intensive production

o ertilizers in turn aects the prices o agricultural ood commodities in general. There are alsoindirect eects o high oil prices on demand and prices o nonenergy commodities. For instance, some

agricultural commodities are used to produce energy substitutes, such as sugar and soybeans or theproduction o ethanol and other biouels. Demand or crude rubber rises as energy prices increase given

the relatively high energy intensity o synthetic rubber production.A more ormal way o investigating the relationship between energy and nonenergy commodity

prices is to test ranger causality. The ranger causality test is simply to see i lagged values o one

variable (X) have any statistically signifcant inormation on uture values o the other variable (Y) giventhe lagged values o Y. I it does, X is said to ranger-cause Y. The box table provides test results or

ranger-causality between energy and nonenergy commodity prices. In this case, ranger causality is

tested or both directions between energy and nonenergy commodity prices using monthly data withdierent lags up to 12 months rom January 1980 until September 2006. Test results suggest thatchanges in energy prices ranger cause changes in nonenergy commodity prices but not t he otherway around. The results are statistically signifcant and indicate marginal signifcance o energy prices

in the equation or nonenergy commodity prices, and only in that direction.

numberoF lAGs(months)

null hypothesis

nonenerGy priCesdonotGrAnGer-CAuse enerGy priCes

enerGy priCesdonotGrAnGer-CAuse nonenerGy

priCes

1 1.58057 .76222*

2 0.50940 1.49629 1.11809 4.18756*

4 1.45697 .28997*

5 1.08427 4.7094*

6 1.1541 .641*

7 1.2077 .028*

8 1.04487 2.67009*

9 1.08210 2.70782*

10 0.96809 2.888*

11 0.75858 2.45541*

12 0.7875 2.02904*

* Indicates signifcance o F-statistics at the 5% level and rejection o the null.Note The results report F-statistics, which orm the basis or which the null hypothesis is accepted or rejected. In terms o the

actual numbers reported, a higher number represents greater statistical signifcance, thus leading to rejection o the null.For example, in column 2 with one month lag, the test statistic (F-stat) is 1.58, which is statistically insignifcant, thusleading to nonrejection o the null that nonenergy commodity prices do not ranger-cause energy prices. On the otherhand, in column with one month lag, the test statistic is .76, which is statistically signifcant, thus implying rejectiono the null that energy prices do not ranger-cause nonenergy commodity prices.

Source Sta calculations.

10 December 2006




17/58

FIGURE 9A

ALUMINUM PRICE INDEXES

FIGURE 9B

COPPER PRICE INDEXES

1980 1984 1988 1992 1996 2000 2004

1980 1984 1988 1992 1996 2000 2004

1980 1984 1988 1992 1996 2000 2004

250

200

150

100

50

0

1980

=

100

400

350

300250

200

150

100

50

0

1980

=

100

1980 1984 1988 1992 1996 2000 2004

1980 1984 1988 1992 1996 2000 2004

1980 1984 1988 1992 1996 2000 2004

160140

120

100

80

60

40

20

0

1980

=

100

300250

200

150

100

50

0

1980

=

100

500

450

400

350

300

250

200

150

100

50

0

1980

=

100

500450400350300250200150100500

1980

=

100

FIGURE 9C

IRON ORE PRICE INDEXES

FIGURE 9D

LEAD PRICE INDEXES

FIGURE 9E

NICKEL PRICE INDEXESFIGURE 9F

ZINC PRICE INDEXES

Nominal Real

Sources: Primary Commodity Prices Database, downloaded 9 October 2006 (International Monetary Fund 2006b); CommodityPrice Indexes (World Bank Development Prospects Group 2006).

section ii




18/58

box 5speCulAtionAnd Commodity priCes

The role o speculation in commodity price volatility is a highly contentious topic. There are two

major players in the commodity utures markets hedgers and speculators. Hedgers participate in uturestrading or insurance purposes against price movements. They are oten commercial traders who are actuallyinvolved in the economic activity related to the underlying commodity, and thus have intrinsic interest in

protecting themselves against adverse price movements. Unlike hedgers with commercial interests who wantto hedge against price movements, speculators seek trading profts coming rom price changes between the

selling and buying points in the utures market. That is, speculators assume price risks in the hope thatprices will move in their avor. By taking the price risks that hedgers want to transer, speculators act as acounter party to the utures contract, so that hedgers may shed unwanted risks.

Speculators participation in the utures markets contributes to market liquidity and diversity. Themarket liquidity literature explains that a large pool o speculators with diverse expectations and risk profles

help markets to unction more efciently by allowing the hedgers with specifc needs to unload the risks atlower costs. At the same time, however, an increase in trading volume due to their activities may lead to

higher price volatility in the short term. The utures market is a place where a seller meets a buyer through

clearing prices. ike any other market, the utures market is subject to asymmetric inormation between theseller and buyer. In order to fnd out the true value o the fnancial product involved, traders rely on trading

volume and prices to deduce the right inormation. This is called a price-discovery process. An arrival onew inormation triggers the process o price discovery, leading to an increase in trading volume and price

volatility. As traders flter out relevant inormation rom noise through vigorous trading activities, highertrading volume is generally accompanied by strong price reactions. Empirical studies also fnd a positive

relation between trading volume and price volatility (see Karpo 1987 or a survey). While an exogenousinormation shock along with an increase in utures trading could heighten the level and volatility o spotprices in the short run, the participation o a diverse group o investors in utures trading may improve the

overall inormation content in trading behavior and prices. Earlier literature suggests that better inormationcontent in increased utures trading helps stabilize spot prices by acilitating the price-discovery process

(see Anderson 1991 or a survey). Moreover, increased speculative activities in the utures markets provide

easier hedging and inventory-adjustment opportunities to help reduce fnancial risks. Empirical evidence onthe stabilizing eect o utures trading and speculation on spot price movements, however, remains largelymixed. Cox (1976) and Danthine (1978) ound that an introduction o utures trading helps stabilize spotprices thanks to improved inormation. However, Kawai (198) and Newbery (1987) provided evidence that

speculation could be destabilizing or storable commodities.The impact o speculation in utures markets on the level and volatility o spot prices is still under

debate. However, a diverse group o speculators enhance liquidity and broaden the scope o trading in thecommodity utures markets. Such benefts o market liquidity appear to be unambiguous over the long run

(see BIS 1999 or a survey). Increased market liquidity lowers trading costs and acilitates the price-discoveryprocess. Along this line, establishment o liquid utures markets with active participation o speculatorswould be a welcome progress, as more eective price discovery together with hedging opportunities through

the utures markets would help reduce undamental price volatility in the long run. In the meantime,however, increasing trade ows with heterogeneous belies accompanied by broader participation o fnancial

investors could continue to trigger sudden spikes in response to news in short-term price movements.

Nevertheless, the ate o dierent metal prices was critically inuenced by market undamentals.

iven frm demand driven by higher global growth and rapid industrialization in developing Asia,overall tight market conditions appeared to provide support or metal prices. For example, duringthe May sell-o, the price o nickel held its ground, given low inventory levels. ikewise, copper

1 December 2006




19/58

and zinc rebounded airly quickly, on the back o tight supply conditions. More generally, base

metal production has been constrained by tightened global mining and smelting capacity.1 Supplydisruptions and labor disputes, sometimes related to political instability in the producer countries,are also helping keep a tight supply/demand balance in recent years.

. Fd and oer Aricra Cdiie

Prices o ood and beverages have been on a long-term downward trend owing to increasedagricultural productivity and technological advances (Figure 10). Real prices o ood and beveragedeclined rather sharply during the 1980s and continued to be depressed until 2001. Food prices

started to rise partly in response to adverse weather conditions in 2002200. Ater reaching acyclical peak in early 2004, agricultural ood prices generally stabilized during late 2004 and 2005.The recent perormance o steady ood prices reects restored demand/supply balance based onboth frm demand and improved supply management over the past couple o decades. For some

ood commodities such as rice, soybeans, meat, ruits, and vegetables, Asian demand particularlyrom the PRC has been strong, largely reecting rapidly changing dietary patterns along with

rising income levels. A surge in soybean prices bears witness to growing consumption by the PRCo eedstock or live animals, as well as o vegetable oil. In the past ew years, high crude oil

prices have also lited demand or soybeans and sugar to produce ethanol, as a partial substituteor transportation uels.

Ater a sharp decline that coincided with Asias fnancial crisis, agricultural raw materials

prices are on a gradual recovery (Figure 11). In act, the economic perormance o the PRC sincethe early 1990s has exerted a growing inuence on movements o cotton, rubber, and timber prices.As the production o textile and clothing in the PRC rose rapidly in the 1990s, domestic demand

or cotton increased, pushing global prices higher. The PRC also emerged as the worlds majorconsumer o orest products since the 1980s, on the back o strong construction activities. In the1990s, the PRCs demand or timber was lited again by the takeo in the wood products industry

or urniture and interior decoration. Buoyant construction activities and a pickup in the domesticurniture industry in the PRC continue to be a major driver behind strong growth in global timberprices since the Asian crisis. Similarly, surging automobile production in the PRC has lited rubberprices strongly (due to demand or tires). The PRCs tire production has increased at about 20%

per annum since the early 1990s according to a report by the International Rubber Study roup(2004), which has pushed the PRC as the worlds largest consumer o natural rubber since 2001.Recently, high oil prices have provided additional upward momentum to rubber prices as syntheticrubber has become more expensive.

1 arious industry reports claim that a rather long period o relatively low prices and capacity overhang through the 1990s

led the mining industry to reduce excess capacity through mergers and restructuring (see Kitco Base Metals 2006 orexample). Investment in exploration and development o new mines has been also limited. Even when metal demand

started to rise ater the Asian crisis, several major producers continued to reduce their inventory overhang rather than

add new production capacity, a process that is now becoming more drawn-out as regulations on environmental concernstighten.

section ii




20/58

140

120

100

80

60

40

20

0

FIGURE 10A

FOOD AND BEVERAGE PRICE INDEXES

FIGURE 10B

CEREALS PRICE INDEXES

FIGURE 10C

VEGETABLE OILS AND PROTEIN MEALS PRICE INDEXES

FIGURE 10D

MEATPRICE INDEXES

120

100

80

60

40

20

0

1980

=

100

1980 1984 1988 1992 1996 2000 2004

1980 1984 1988 1992 1996 2000 2004

1980 1984 1988 1992 1996 2000 2004

1980 1984 1988 1992 1996 2000 2004

1980

=

100

1980

=

100

140

120

100

80

60

40

20

0

1980

=

100

140

120

100

80

60

40

20

0

Nominal Real

Sources: Primary Commodity Prices Database, downloaded 9 October 2006 (International Monetary Fund 2006b); Commodity

Price Indexes (World Bank Development Prospects Group 2006).

III. DEvEloPINg AsIAs ImPRINt oN globAl CommoDIty mARkEts

A. grw, srcra Cane, and Cdi Deand

In the past, global business cycles were dominated by the growth perormance o industrializedcountries, thus commodity price uctuations were dictated primarily by the business cycles in thesecountries. But with developing Asias growing presence in the global economy, the region exerts

increasingand now considerableinuence on international commodity markets and prices. Figure 12illustrates evolving correlations between gross domestic product (DP) growth rates o both 7

and developing Asian economies and commodity price uctuations. While the linkage between 7growth and the global commodity price cycle gradually weakened, that between developing Asiasgrowth and the global commodity price uctuations jumped since around the early 1990s. It canbe readily inerred that strong growth in developing Asias demand was a major driver behind anupswing in commodity markets through the mid-1990s prior to the Asian fnancial crisiswhile the

crisis itsel led to a slump in commodity markets during 19981999.

1 December 2006




21/58

FIGURE 11A

AGRICULTURAL RAW MATERIALS PRICE INDEXES

FIGURE 11B

COTTON PRICE INDEXES

FIGURE 11C

RUBBER PRICE INDEXES

FIGURE 11D

TIMBER PRICE INDEXES

200180

160140

120

10080

6040

20

0

1980

=

10

0

1980 1984 1988 1992 1996 2000 2004 1980 1984 1988 1992 1996 2000 2004

1980

=

10

0

1980

=

100

1980

=

100

140

120

100

80

60

40

20

0

1980 1984 1988 1992 1996 2000 2004 1980 1984 1988 1992 1996 2000 2004

250

200

150

100

50

0

300

250

200

150

100

50

0

Nominal Real

Sources: Primary Commodity Prices Database, downloaded 9 October 2006 (International Monetary Fund 2006b); CommodityPrice Indexes (World Bank Development Prospects Group 2006).

FIGURE 12

CORRELATION BETWEEN REAL COMMODITY PRICE INFLATION AND GROWTH

1.0

0.8

0.6

0.4

0.2

0.0

-0.2

-0.4-0.6

1980 1983 1986 1989 1992 1995 1998 2001 2004

G7 Developing Asia

Sources: World Development Indicators Online Database, downloaded 24 August2006 (World Bank 2006); Commodity Price Indexes (World BankDevelopment Prospects Group 2006).

section iii

DevelopinG AsiAs imprint on GlobAl commoDity mArkets



22/58

Although developing Asia as a whole exhibited remarkable dynamism in economic growthand development, the progress o economic achievement has not been uniorm across the region.At the subregional level, East Asia including the PRC grew most rapidly, ollowed by South Asia

including India. East Asia and South Asia grew at an annual average rate o 9.1% and 5.7% per

year, respectively, in the period 19912004. As o end-2004, East Asia and South Asia accountor 17.7% and 7.2% respectively o total world DP measured in terms o purchasing power parity(PPP), ollowed by Southeast Asia or 4.5% and Central Asia or 0.4% (Table 1). The relative weight

o these subregions in the world economy is in turn refected in their respective shares o worldcommodity consumption. For instance, the East Asian countries demand or crude oil rose by anaverage o 6.2% annually in the same period, nearly our times as ast as total world consumption,

contributing 6.0% to total world consumption growth. Steel use in the East Asian countries alsobecame intense, with consumption increasing by an estimated rate o 1.9% on average per yearbetween 1999 and 2004, about twice as ast as total world consumption.

tAble 1

Contributionto World Gdp And Commodity demAnd

shAreoF World ACtivity, 2004 Contributionto GroWth, 19912004

Gdp1 petroleum

demAnd

steel

demAnd

Gdp1,2 petroleum

demAnd2

steel

demAnd3

World 100.0 100.0 100.0 100.0 100.0 100.0

Developing Asia 29.9 21.6 4.1 41.2 6.0 75.

East Asia 17.7 12.1 4.9 26. 6.0 65.

South Asia 7.2 .6 .4 8.9 9.8 .7

Southeast Asia 4.5 5.0 4.8 5.1 12.8 6.

Central Asia 0.4 0.8 n.a. 0. 2. n.a.1 In purchasing power parity terms.2 19922004 or Central Asia. 19992005.Sources World Economic Outlook September 2006, downloaded 16 October 2006 (International Monetary Fund 2006e);International Energy

Annual 2004 (Energy Inormation Administration 2004); World Steel in Figures 2006, downloaded 1 August 2006 (InternationalIron and Steel Institute 2006); JP Morgan (2006).

Rapid population and income growth in East Asian economies have also been key driversbehind the rising demand or world ood commodities. Along with increasing ood demand, changing

dietary patterns play an important role. While the consumption o traditional coarse grains typicallydecreases with income growth, demand or meat, ruits and vegetables, and vegetable oil increases.

Table 2 illustrates considerable changes in the dietary composition in the economies o developingAsia in the last decade. Refecting its economic prowess, East Asia again leads the changes among

the subregions. In terms o total calorie intake, there has been a noticeable drop in the share ocereals, while the other ood products have gained considerable shares along with growing ooddiversity toward more protein and at consumption.

Overall, per capita income is an important variable or commodity demand. However, changesin technology and consumer preerences along with the long-term development path are oten

1 December 2006




23/58

tAble 2dietAryComposition (% oFtotAlenerGyConsumption)

1994 CereAls veGetAbleoils

suGArAnd

sWeeteners

meAtAnd

oFFAls

rootsAnd

tubers

milk,eGGs,

AndFish

FruitsAnd

veGetAbles

AnimAlFAts

pulses others

Developing Asia 6.2 9.11 5.74 6.4 .75 4.58 .88 0.68 2.09 0.52

East Asia 62.99 8.78 2.41 10.98 5.14 .22 4.50 1.14 0.54 0.29

South Asia 6.62 8.55 9.76 1.15 1.7 6.55 .12 0.14 4.5 0.86

Southeast Asia 6.99 11.50 6.4 4.6 4. .19 .86 0.56 1.4 0.4

Central Asia 61.17 7.84 5.0 7.78 .08 10.69 .46 0.55 0.07 0.

2004 CereAls veGetAbleoils

suGArAnd

sWeeteners

meAtAnd

oFFAls

rootsAnd

tubers

milk,eGGs,

And Fish

FruitsAnd

veGetAbles

AnimAlFAts

pulses others

Developing Asia 54.5 12.20 6.46 7.89 4.29 5.61 5.5 0.89 2.00 0.78

East Asia 47.01 12. .68 14.68 6.10 4.97 8.70 1.59 0.45 0.50

South Asia 59.96 12.27 9.17 1.27 2.12 7.02 2.71 0.15 4.2 1.09

Southeast Asia 60.87 12.16 7.9 4.97 4.4 .24 .72 0.76 1.19 0.84

Central Asia 57.06 8.25 4.2 7.66 5.07 12.07 4.19 0.5 0.12 0.7

Source Food and Agriculture Organization (2005), downloaded 1 July 2006.

as important or the consumption patterns. In general, economic development accompaniessignifcant changes in industrial structures as well as rising income levels, which in turn aectcommodity consumption. Earlier studies2 argue that resource use increases in the initial stage o

development but then tends to taper o ater income reaches a certain level. This creates an invertedU-curve or the intensity o resource use. For example, in the early stage o development, resourcerequirements particularly or metals are low as the economy oten relies on unmechanized subsistence

agriculture. Industrialization pushes resource demand to build urban, industrial, and transportationinrastructure. As the economy matures and the industrial structure shits toward services, resourcedemand or inrastructure building and industrial inputs would then decline. In light o resource-saving technological developments over time, Bernardini and alli (199) also claim that intensity

o use curves tend to shit downward over the long term.

Figure 1 graphically illustrates the intensity o resource use given per capita income. The dottedlines 15 represent the intensity o use curves o an individual country going through economic

transition in successive periods 15 or o a group o countries 15 at dierent stages o economicdevelopment (the higher number corresponds to the more advanced stage). Dierent intensitycurves either over time or cross sectionally reect the tendency o downward shits in intensity

o use curves as technology advances along dierent stages o economic development. The solid

line crossing the dierent intensity o use curves would indicate an actual intensity o use curve.

2 The relationship between the intensity o resource use and income has been established both in theory and empirics.

Earlier studies that laid the oundation or the inverted U-shaped intensity o use curves include Malenbaum (197),

arson et al. (1986), Bernardini and alli (199), and Jnicke et al. (1997). See Cleveland and Ruth (1999) or asurvey.

section iii




24/58

Intensity

of

resource

use

Per capita income

1

2

3

4

5

FIGURE 13

RELATIONSHIP BETWEEN INTENSITY OF RESOUCE USE AND PER CAPITA INCOME

raphical representation o the intensity o resource use curve as envisioned by Bernardini and alli (199). Thedotted lines 1-5 represent dierences in the intensity o use curve or countries at dierent stages o developmentover time. Countries that start the development process later in time enjoy lower intensity o resource use, as theycan avail o more efcient technology.

Figure 14 plots per capita consumption o mineral ores and petroleum o dierent countries

over time. Along the development path o selected countries emerges an early phase o the invertedU-shaped or S-shaped pattern. The general patterns suggest that resource consumption rises asincome grows. When income reaches about $5,00010,000 per capita along with industrialization,

measured in PPP terms, the consumption o metals and energy seems to grow rapidly. Ater a rapidexpansion period, the intensity o resource use appears to slow at incomes o about $15,00020,000

per capita in PPP terms. These patterns typically repeat among countries at dierent stages oeconomic development. Moreover, despite the rapid increases in the PRCs demand or metals and

energy, the fgure suggests that it may be only the beginning o the PRCs resource demand.

b. Ein Paern f Cdi Deand and trade

Rapid economic growth and industrialization accompany changes in the level and compositiono international trade. Although a countrys domestic resource endowment aects the extent o

the countrys reliance on external resources, in general rapid growth and industrialization will acedomestic resource constraints. Particularly, rapid economic development in relatively resource-poorcountries such as many developing Asian economies would likely signal signifcant changes in tradecomposition along with transition rom agricultural to manuacturing economies.

Developing Asias total trade with the world has increased dramatically over the 40-yearperiod, 19642004 (Figure 15). Changing trade patterns attest to the progress o substantiveeconomic transormation o developing Asia rom agricultural and largely resource-based economies

1 December 2006




25/58

0

500

1000

1500

2000

2500

3000

FIGURE 14A

ALUMINUM

FIGURE 14B

COPPER

FIGURE 14C

IRON ORE

FIGURE 14D

ZINC

FIGURE 14E

PETROLEUM

25

20

15

10

5

0Percapitaconsump

tion

(kg)

Percapitaconsump

tion

(kg)

Percapitaconsumption

(kg)


(kg)


(kg)

0 10 20 30

0 10 20 30 0 10 20 30

Real per capita GDP (PPP, 2000 international $000) Real per capita GDP (PPP, 2000 international $000)

Real per capita GDP (PPP, 2000 international $000)



0 5 10 15 20 25 30

201816

1412108420

12

10

8

6

4

2

0

1400

1200

1000

800

600

400

200

0

0 10 20 30

ASEAN4Linear (ASEAN4)

PRCLinear (PRC)

IndiaLinear (India)

JapanLinear (Japan)

KoreaLinear (Korea)

Sources: Statistical Review of World Energy 2006, downloaded 30 June 2006 (BP 2006); Commodity Yearbook 2003,downloaded 2 August 2006 (United Nations Conference on Trade and Development 2003); World Development

Indicators Online Database, downloaded 10 October 2006 (World Bank 2006).

section iii




26/58

to manuacturing and service-based economies. In general, the accumulation o both physical and

human capital along with economic development and industrialization accompanies changes in acountrys comparative advantage toward capital and skill-intensive products over time. Thereore, itis commonly observed that the share o manuactures in export composition increases in developing

countries with rapid industrialization. Faced with increasing domestic resource constraints, importso primary commodities also rise rapidly, not to mention capital equipment and intermediate inputsespecially at the early stage o development.

OthersElectronicsMedium- to high-skill manufacturesResource-intensive and low-skill manufacturesFuelsMetals, ores, and mineralsFood and agricultural commodities

Exports Imports Exports Imports Exports Imports Exports Imports Exports Imports

2000

1800

1600

1400

1200

1000

800600

400

200

0

1964 1974 1984 1994 2004

FIGURE 15

DEVELOPING ASIAS TRADE

Source: Commodity Trade Statistics Database, downloaded 16 October 2006 (United Nations 2006).

$

billion

While developing Asias total exports exploded rom $5. billion to $1,654.9 billion between1964 and 2004, the share o primary commodities in percent o total exports sharply dropped rom

74.0% to 1.2% over the same period. In particular, at the beginning o the period, exports romdeveloping Asia were dominated by agricultural and mining products. Propelled by industrialization,however, manuacturing exports gained an increasing weight, initially driven by labor-intensive (andresource) manuactures, such as textile and clothing in the early stage o development, and later by

more skill-intensive manuactures, such as machinery, industrial chemicals, and electronics. Alongwith the manuacturing export drive, Asian demand or primary commodities surged.

A close look at historical commodity import data also reveals developing Asias burgeoning

role as a consumer and importer o industrial raw materials and uels. Table summarizes theproiles and trends o primary commodity imports in PRC, India, Japan, and Korea, which isdesigned to provide an example o dierences and similarities in their changing import patterns

along with industrialization. The table highlights the growing share o total primary commodityimports o uels and metals during rapid economic growth and industrialization. It is clear rom UNCTAD (2005) reports that sustained rapid growth and rising incomes in Asia have been accompanied by a dramatic

shit in the pattern o international trade ows.

0 December 2006




27/58

the table that the weight o primary imports has shited rom agricultural ood and raw materials

to metals and uels in Japan, and more so in Korea over time. Fuels especially have continued toexplain a signifcant portion o their commodity needs, accounting or more than a hal o totalprimary commodity imports in recent decades. Not only accounting or a signifcant share in total

commodity imports, imports o uels and metals also rose sharply in real terms (deated by theimport prices o individual commodities) particularly during the period o rapid economic growthin Japan between 1964 and 1974 and in Korea in the latter decades (Figure 16). Real imports ouels have more or less stabilized in Japan since 1974, but continued to grow at an accelerated

rate in Korea along with its metal imports in the 1980s through the 1990s. On the other hand,real imports o agricultural raw materials have declined in Japan since 1984 and reached a plateaua decade later in Korea. This stands in marked contrast against a sharp pickup in the PRC andIndia between 1994 and 2004, highlighting a shit in the textile industry rom high-labor-cost to

low-labor-cost countries. This repeated pattern also underscores a successul shit in the industrialstructure rom labor-intensive to skill-intensive manuactures in these Asian economies. Moreover,the contemporary surge in the PRCs imports o uels and metals echoes experience o both Japan

and Korea during their early industrialization phase in the 1960s and 1970s. This pattern is most

likely associated with the increasing intensity o metal use during the early catch-up phase.

Interestingly, some agricultural ood commodities, particularly basic staples, have resisted the

general trend o rising import demand in Asia. As a consequence o signifcant improvements inagricultural productivity with the onset o the reen Revolution, and the subsequent protectiono the agricultural sector, import demand or cereals and cereal preparations actually ell in thePRC and India in the early stage o development (see Boestel et al. 1999). However, changing

dietary patterns appear to exert an obvious inuence on the level and variety o imports or otherood products including meats, vegetables, edible oil, and oil seeds. Rapid urbanization is anothercontributing actor to the changes in dietary profles, which oten generates additional demand or

higher-value processed ood and tropical beverages such as coee. It is clear that this general trendin the dietary pattern would have magniying impacts on ood imports by the PRC and India given

their population size. Table also illustrates that imports o meat, fsh, edible oil, and oil seedsrose sharply in Japan and Korea over the past our decades, while the share o cereals and cereal

preparations steadily declined in total ood imports. The PRC imports o meat, fsh, vegetable oil,and oil seeds started to take o in the mid-1990s. More recently, imports o vegetable oil and oilseeds in India have also shown ast growth.

With developing Asias growing presence in world trade, the region exercises increasing leverageover international commodity markets and prices. In particular, the remarkable economic expansioncombined with rapid industrialization in the PRC appears to have considerable repercussions or

world commodity demand and prices. Table 4 contrasts the PRCs top 10 primary commodity importsin 1984 and 2004, together with their global shares. In just two decades, ood products and inputsor primary and light manuacturing industries have given way to heavier industrial raw materials.

It also demonstrates how processes o ast growth and structural change in the PRC have made alarge impression on global commodity markets, even at comparatively low income levels. India maybe only one or two decades behind in its catch-up process given its recent growth perormance andoutlook. iven the size o the PRC and India economiesthey account or 7.4% o world populationand 21.4% o world DP in PPP termsworld commodity markets are expected to see hety changes.

The relatively low development stages o these countries also suggest that income elasticity o theircommodity demand would be higher compared with that o industrialized countries, which would

section iii




28/58

tAble 3shArein primAry Commodity import vAlue

JApAn prC

1964 1974 1984 1994 2004 1984 1994 2004

Primary products 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

All ood products 2. 16.9 16. 6.0 26.4 9.6 28.0 10.8

Meat and meat preparations 0.9 1.0 1.9 5.9 4.8 0.1 0.5 0.4

Dairy products and eggs 0. 0.4 0. 0.5 0.5 0.4 0.5 0.

Fish and fsh preparations 0.9 2.1 4.1 12.1 7.6 0.5 .2 1.8

Cereals and cereal preparations 10. 7.0 4.7 5.0 . 28.2 7.4 1.9

Sugar, sugar preparations, and honey 4.5 2.6 0.4 0.5 0. 4.1 2.5 0.

Coee, tea, cocoa, spices, andmanuactures

1.1 0.7 1.1 1. 1.0 0.8 0. 0.1

Fixed vegetable oils and ats 0.2 0. 0. 0. 0.4 0.9 9.7 .0

Agricultural raw materials .8 17.8 12.5 15.7 8.1 5. 29. 2.

Oil seeds, oil nuts, and oil kernels 4.6 2.6 2.2 1.7 1.5 0.0 0. 5.6

Crude rubber 2. 0.5 0.6 0.7 0.7 5.0 . 2.

Wood, lumber, and cork 7.2 7.5 .9 7. 2.8 11.7 . .

Textile fbers and waste 14.2 .8 2.5 1.4 0.4 10.9 16.7 5.2

Minerals, ores, and metals 19.9 14.9 11.2 11.7 1.1 22.9 20.0 29.0

Metallierous ores and metal scrap 15.8 10.8 6.5 5.7 6.6 5.0 10.6 17.9

Nonerrous metals 4.1 4.1 4.7 6.0 6.5 17.9 9.4 11.1

Fuels 22.9 50.4 60.0 6.5 52.5 2.2 22.8 6.9

Petroleum and petroleum products 19.2 42.8 45.2 24.9 5.7 0.8 20. 4.2

koreA indiA

1964 1974 1984 1994 2004 1984 1994 2004

Primary products 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

All ood products 5.8 26.9 1.4 16. 12.9 19.9 14.4 9.5

Meat and meat preparations 0.0 0.4 0.5 1.8 1. 0.0 0.0 0.0

Dairy products and eggs 0.9 0.0 0.1 0.2 0. 2.1 0.1 0.0

Fish and fsh preparations 0.0 0. 0.5 2.0 2.7 0.0 0.0 0.0

Cereals and cereal preparations 0.1 18.6 7.4 4.7 .0 2.8 0.2 0.1

Sugar, sugar preparations, and honey 1.9 4.5 1.6 1.4 0.6 1.1 6.2 0.6

Coee, tea, cocoa, spices, and

manuactures

0.1 0.2 0.4 0.8 0.4 0. 0.1 0.

Fixed vegetable oils and ats 0.1 0.1 0.5 0.6 0.5 10.9 1.8 5.

Agricultural raw materials 46.1 0.1 2.0 20. 7.7 10.8 14.0 6.9

Oil seeds, oil nuts, and oil kernels 0.6 0.6 1.8 1. 0.8 0.0 0.0 0.0

Crude rubber 2.9 2. 1.9 1.7 0.9 1.0 1.0 0.9

Wood, lumber, and cork 9.1 10.4 4.6 4.5 1.2 0.1 1.8 1.8

Textile fbers and waste 26.2 9.4 7.0 .8 1.0 2.7 5. 1.6

Minerals, ores, and metals 4.0 11.0 10.1 17.1 18.1 6.9 14.6 9.

Metallierous ores and metal scrap 1.9 7.8 6.1 7.7 9.0 2.1 6. 5.2

Nonerrous metals 2.1 .2 4.0 9.4 9.1 4.8 8.4 4.2

Fuels 14.1 2.0 5.5 46. 61. 62.4 57.0 74.

Petroleum and petroleum products 12.8 0.9 47.2 6. 45.7 61.7 50. 65.1

Source Commodity Trade Statistics Database, downloaded 29 September 2006 (United Nations 2006).

December 2006




29/58

Agricultural raw materialsMetals and minerals

Petroleum

Food products (right scale)

Food productsAgricultural raw materials

Petroleum

Metals and minerals (right scale)

FIGURE 16A JAPAN: INDEX OF REAL IMPORTS600

450

300

150

0

1964

=

100

1964 1974 1984 1994 2004

2000

1500

1000

500

0

1964

=

100

8000

6000

4000

2000

0

1964

=

100

1964 1974 1984 1994 2004

60000

45000

30000

15000

0

1964

=

100

FIGURE 16B: KOREA: INDEX OF REAL IMPORTS

FIGURE 16C PRC: INDEX OF REAL IMPORTS FIGURE 16D INDIA: INDEX OF REAL IMPORTS

Food products


Petroleum (right scale)

Food products


Petroleum

2000

1500

1000

500

0

80000

60000

40000

20000

0

1984

=

100

1984 1994 2004 1984 1994 2004

600

450

300

150

0

1984

=

100

Sources: Commodity Trade Statistics Database, downloaded 29 September 2006 (United

Nations 2006);International Financial Statistics Online Database, downloaded 29

September 2006 (International Monetary Fund 2006a).

1984

=

100

help sustain the strong demand for an extended period. Extrapolating from the experiences of Japan

and Korea with limited commodity endowments suggests that if the PRC and India continue to

grow quickly, this will result in enormous increases in their demand for energy, metals, other raw

materials, and later food commodities that can only be met through imports.

section iii




30/58

tAble 4top 10 primAry Commodity importsoFthe peoples republiCoF ChinA

19

perCentoFprCimports

perCentoFtotAlWorldimports

Cereals and cereal preparations 7.0 5.4

Nonerrous metals 4.4 2.7

Wood, lumber, and cork 2.9 .9

Textile fbers, not manuactured, and waste 2.7 .9

Metallierous ores and metal scrap 1.2 1.2

Crude rubber including synthetic and reclaimed 1.2 4.1

Pulp and paper 1.1 2.5

Sugar, sugar preparations, and honey 1.0 .

Tobacco and tobacco manuactures 0.4 1.5

Feed stu or animals excluding unmilled cereals 0.4 0.8

00

perCentoFprCimports

perCentoFtotAlWorldimports

Petroleum and petroleum products 7.9 5.5

Metallierous ores and metal scrap 4.1 21.9

Nonerrous metals 2.6 8.6

Oil seeds, oil nuts, and oil kernels 1. 27.9

Textile fbers, not manuactured, and waste 1.2 2.6

Pulp and paper 0.9 19.

Wood, lumber, and cork 0.8 8.7Fixed vegetable oils and ats 0.7 1.6

Crude rubber including synthetic and reclaimed 0.5 15.4

Cereals and cereal preparations 0.4 .4

Source Commodity Trade Statistics Database, downloaded 7 August 2006 (United Nations 2006).

December 2006




31/58

Iv. lookINg AhEAD: PRosPECts oF globAl CommoDIty mARkEts INto 01

Asias growth prospect in conjunction with rapid industrialization represents undamentalchanges in market dynamics or a broad range o primary commodities. Over the past hal century,developing Asia has exerted an increasingly strong pull on global commodity markets. Industrialization,

urbanization, and rising income levels have ed surging demand or primary commodities includingenergy, base metals, and sot commodities. In recent decades, the PRC has emerged as the worldsleading importer o petroleum, metals, oil seeds, textile fbers, and pulp and paper. The PRCs imports

o agricultural products, such as soy and meat products, have now also taken o as dietary patternschange and the scope or agricultural expansion is limited by constraints on land and water supply.Although India still lags behind the PRC in some way, it, too, has embarked on a process o ar-reaching economic transormation. I India ollows the pattern that has been seen in Japan, Korea,

and now the PRC, it will lead a third wave o explosive growth in commodity demands emanatingrom developing Asia. This section examines the outlook or global commodity markets to 2015 usingeconomic simulations based on a general equilibrium model (see Appendix or detailed inormationabout this eneral Equilibrium Model or Asian Trade [EMAT] model).

A. baeine scenari

The baseline scenario or global commodity markets and prices assumes that past trendsin economic growth and technological developments will continue into the uture. In particular,developing Asia is assumed to grow at the rate o the past decade (6.6%), and that the rest

o the world moves along a lower trend o growth at 2.8%. It is also assumed that the policyenvironment remains largely unchanged. The details o the baseline scenario are reported in Table5. Two alternatives (one in which circumstances conspire to soten commodity prices and anotherin which prices frm up) will be compared with this business-as-usual scenario. These experiments

help to defne a range within which the actual trajectory o commodity prices may lie. There aremany other actors that can potentially inuence commodity price trends and many ways in which

developing Asia might have an impact on them. Here attention is ocused on the impact o DPgrowth through our main channels (i) efciency o energy and mineral ores use, (ii) reserves oglobal ossil uels, (iii) productivity, and (iv) energy taxes and subsidies.

The interplay between demand and supply in global commodity markets is examined using

EMAT. This model captures links among dierent sectors o the global economy in terms o bothquantities and prices, and traces their impacts through to commodity markets and prices. Theappendix describes the models main eatures. An appealing aspect o the model is that it permitsquantitative analysis o the impacts o a variety o dierent policy and structural actors on

commodity prices and trade.

In the baseline scenario, the global economy grows at .% per annum over a 15-year period,which in calendar time starts in EMATs base year 2001, ending in 2015. Developing Asia4 grows at

6.6% over this interval and accounts or 27.2% o the expansion in global income. This economicgrowth creates additional demand or commodities, moderated somewhat by improvements in end-use efciency and income elasticities o demand that are generally less than 1 in value.

4 Throughout the modeling analysis in this section, developing Asia includes only East Asia, Southeast Asia, and SouthAsia. Central Asia and the Pacifc are excluded.

section iv

lookinG AheAD: prospects oF GlobAl commoDity mArkets into 2015



32/58

iven its aster growth rate, developing Asias share o global commodity demand rises. Figure

17 shows the increases in its demand shares or agriculture, ossil uels, and mineral ores comparedwith those o three major economic regions (European Union [EU], Japan, US) and the rest o theworld. Also shown in the fgure are developing Asias weights in the global economy in terms o PPP(which is held constant at the base year 2001 rate or 2015). Between 2001 and 2015, developing

Asias share in global income increases rom 26.2% to 8.2%. More striking still is developing Asias

contribution to the expansion o global commodity demand. For agricultural commodities, developingAsias share in global commodity demand rises rom 4.4% to 4.0%; or ossil uels rom 18.1% to

26.9%; and or mineral ores rom 6.0% to 50.1%. Over the 15-year period, the model estimatesthat developing Asia accounts or 61% o the growth in global commodity demand.

Developing Asias rapid growth and rising commodity need also leads to its increasing reliance

on external resources. Figure 18 illustrates changes in the import share o commodity demands overthe 20012015 period. By 2015, more than 70% o developing Asias (excluding the net oil exportersin Central Asia) oil needs will have to be met by imports. Faced with domestic and regional resourceconstraints, import shares in total demand or most other commodities are indeed projected to rise.

The projected increases in the regions import dependence or agricultural and orestry productsare associated with increasing losses o arable land and orestry to urban, industrial, and transport

inrastructure. These projections highlight the trend o depleting resources and the environmentaldegradation driven by commodity-intensive economic activities, changing liestyle with rising incomelevels in the regions major economies, and continuing high population growth.

tAble 5bAseline sCenArio Assumptions, 20012015

Assumption

DP and total actor productivity(TFP) growth

DP growth o each region is exogenous, and region-specifc nonagriculturalTFP growth is endogenous to match the exogenous DP growth. TFP growthis assumed to be identical across all non-agricultural sectors. Developing

Asia is set to grow at 6.6% and the rest o the world at 2.8%.

Efciency o energy and mineralores use

Improvement o 1.25% a year, leading to a 19.0% higher energy andmineral ores use efciency in 2015 relative to 2001.

Reserves o global ossil uels Determined by the speed o depletion, the size o unproven reserves,and the conversion rate rom unproven to proven reserves. See Table A1about the assumptions or reserves and conversion rates.

Productivity in agricultural andnatural resources sectors

rows by 1.02% per annum or the world average o crops and 0.72%or livestock, based on the projection by Hertel et al. (2006). There is

no TFP growth in orestry, fshing and mining (including ossil uels and

mineral ores) sectors.

Energy taxes and subsidies No changes.

Note The assumed 1.25% improvement per year in energy efciency lies in the high band o the range used in a number o otherstudies, but is smaller than the estimate or OECD in the period 19702000 (see Webster 2005).

December 2006




33/58

Table 6 (columns 2) summarizes model estimates o commodity price indexes in 2015.Table 7 (column 4) shows the implied price levels or energy products (in 2005 prices) according

to the model estimates. The commodity price indexes are measured relative to a base o 100 in2001 (or 2005). All price indexes are deated by the world consumer price index (CPI), which isheld constant throughout the simulation. It is important to understand that Tables 6 and 7 showestimates o equilibrium real commodity prices. These prices are derived to equate supply and

demand in commodity markets, given the other prices that clear all the other markets. As such,they do not easily translate to the market prices quoted in commodity exchanges, which are heavilyinuenced by inventory gluts and shortages. Moreover, market prices are susceptible to macroeconomicuctuations, weather conditions, political actors that may disrupt supplies, speculation, and price

setting behavior by oligopolies (such as the Organization o Petroleum Exporting Countries). Thus,even over a period o several years, market prices may diverge signifcantly rom these theoreticalequilibrium levels.

100

90

8070

60

50

40

30

20

10

0

FIGURE 17

ASIAS IMPRINT ON GLOBAL COMMODITY DEMAND

Percentoftotaldeman

d

2001 2015 2001 2015 2001 2015 2001 2015

GDP Agriculture Fossil fuels Mineral ores

Developing Asia US, EU, and Japan Rest of the world

Sources: ADB GEMAT model simulations, staff calculations.

FIGURE 18

DEVELOPING ASIAS DEPENDENCE ON COMMODITY IMPORTS

GrainOil seedCotton

Other cropsLivestockForestryFishery

CoalCrude oil

Natural gasMineral ores

Percent

2001 2015

10 20 30 40 50 60 70 800

Note: Including intraregional exports.Sources: ADB GEMAT model simulations, staff calculations.

section iv




34/58

A striking eature o the estimates is that they suggest a rising trend or (most) real commoditypricesnotwithstanding assumed improvements in productivity and efciency o resource use (see

details in the next section). On average, there is assumed to be about a 19% improvement inenergy and mineral ores use efciency by 2015. Even some agricultural commodity prices, whichhistorically have trended down, edge up. imited endowments and low productivity growth cause the

prices o fsheries and orestry commodities to rise sharply. The prices o other crops, includingruits, vegetables, and sugarcane and beet, increase marginally. O course, not all prices rise. Theprice o livestock and most crop products trend down; due to ast productivity growth in the caseo livestock and a particularly low income elasticity o demand in the case o crops. In miningsectors, global prices increase. Again, this is due to the presence o a tight resource constraint.

The model predicts the strongest increase in real prices or crude oil.

tAble 7reAl enerGy Commodity priCesinthe bAseline sCenArio (in 2005 priCes)

2001 (ACtuAl) 2005 (ACtuAl) 2015 (proJeCtion)

Coal ($/mt) 5.9 47.6 44.2Crude oil ($/barrel) 28.1 5.4 4.0

Natural gas ($/mmbtu) 6.1 8.9 6.5

Sources ADB EMAT model simulations, World Economic Outlook Online Database (IMF 2006c), sta calculations.

tAble 6Commodity priCes proJeCtioninthe bAseline sCenArio

proJeCtedpriCesin 2015

(2001 =100)

proJeCtedpriCesin

2015 (2005 =100)

ACtuAlpriCesin

2005 (2001=100)rain 9.9 88.5 106.1

Oil seed 94.8 77.9 121.7

Cotton 94.4 94.7 99.7

Other crops 102.4 86.4 118.6

ivestock 9.9 88.0 106.7

Forestry 116.7 111.5 104.7

Fishery 127. 126.4 100.7

Coal 12.2 92.9 12.5

Crude oil 15. 80.6 190.2

Natural gas 107.1 7.4 145.9

Mineral ores 110.6 67.9 162.8

Note All prices are real prices deated by the world CPI.Sources ADB EMAT model simulations; World Economic Outlook Online Database, downloaded 26 August 2006 (IMF 2006c); sta

calculations.

December 2006




35/58

O particular interest are the model estimates or the price trajectory o crude oil. Figure 19

shows the trajectory over 20012015. The trend calculated by the model suggests that the realprice o oil (measured in 2005 prices) rises by 5% rom about $28 per barrel to $4 per barrel in2015. This baseline orecast or the real price o oil in 2015 is in line with those o major energy

orecasting agencies, which predict oil prices will settle into a range o $550 per barrel (in 2004prices) by 2015, barring critical supply disruptions (IEA 2005 and EIA 2006).

FIGURE 19

BASELINE EQUILIBRIUM PRICE TRAJECTORY

AND CONVERGENCE OF MARKETPRICE FOR CRUDE OIL

65

60

55

50

45

40

35

30

25

20

$

/bbl(2005

prices)

Historic and projected nominal price

Historic and projected real price

Equilibrium price in the baseline

2001 2003 2005 2007 2009 2011 2013 2015

Sources: ADB GEMAT model simulations, staff calculations.

It is clear that the model estimates a gradual decline in the real price o crude oil by 2015rom current highs. The estimated real price o crude oil in 2015 is below the average price or2005. I or the sake o argument, global CPI ination is set at 2.5% (rather than 0% as assumedin the model calculations) over the period 20052015, this translates to $55 per barrel by 2015

in current prices, which is still lower compared to the peak reached in 2006. I, in the long-run,oil prices are driven by undamentals, these estimates suggest that the price o crude oil may belower in real terms by 2015 than it is today. Nevertheless, crude oil would still be considerablymore expensive than it was at the beginning o the millennium, or even a decade ago.

The baseline estimates suggest changes in terms o trade or developing Asia. At an aggregateregional level, developing Asia is a major exporter o agricultural and manuactured goods and asignifcant importer o energy and mineral commodities. Figure 20 summarizes model estimates o

terms-o-trade changes or major countries and subregions o developing Asia. It is notable thatthe projected terms o trade slide signifcantly in the PRC and India. Their hunger or commoditiesand their tendency or manuacturing-industry-ocused growth are likely to lower the prices o

manuactured and industrial output relative to global commodity prices, particularly o energy and

mineral ores.

b. Rne Cec

The baseline scenario makes a number o important assumptions that are subject to signifcant

uncertainty. To gauge how robust the associated price estimates might be, two alternative scenariosare considered. Table 8 describes two alternative sets o assumptions that may be compared with

section iv




36/58

those o the baseline scenario in Table 5. One scenario imagines that actors conspire to straincommodity markets and press urther price rises. The other scenario considers developments thatwould take the weight o o markets and allow commodity prices to soten. Figure 21 illustrates an

example o changes in the efciency o energy and mineral ores use under two alternative scenarioscompared to the baseline assumption.

tAble 8sCenArio Assumptions

mArketssoFten mArketstiGhten

DP and TFP growth The TFP growth in each region is

set at the baseline level and the

rate o DP growth o each regionis now endogenous, so that it canadjust to a shock through ollowingchannels.

The TFP growth in each region is set

at the baseline level and the rate o

DP growth o each region is nowendogenous, so that it can adjust toa shock through ollowing channels.

Efciency o energy andmineral ores use

Faster growth in all regions, leadingto a 0.9% improvement by 2015,

or 10% higher than the baselinelevel in 2015.

Slower growth in all regions, leadingto a 7.1% improvement by 2015, or

10% lower than the baseline level in2015.

Reserves o global oss

asia's imprint on global commodity markets

Documents