《KAPSARC:2023世界石油和关键矿产研究报告(英文版)(72页).pdf》由会员分享,可在线阅读,更多相关《KAPSARC:2023世界石油和关键矿产研究报告(英文版)(72页).pdf(72页珍藏版)》请在三个皮匠报告上搜索。
1、1Global Crude Oil Storage Index:A New Benchmark for Energy Policy World Oil and Critical Mineral Study:A Global VAR Analysis Jennifer Considine,Abdullah Al Dayel,Philipp Galkin and Emre HatipogluSeptember 2023Doi:10.30573/KS-2023-MP032World Oil and Critical Mineral Study:A Global VAR Analysis:A Kaps
2、arc Methodology Paper About KAPSARCKAPSARC is an advisory think tank within global energy economics and sustainability providing advisory services to entities and authorities in the Saudi energy sector to advance Saudi Arabias energy sector and inform global policies through evidence-based advice an
3、d applied research.This publication is also available in Arabic.Legal Notice Copyright 2023 King Abdullah Petroleum Studies and Research Center(“KAPSARC”).This Document(and any information,data or materials contained therein)(the“Document”)shall not be used without the proper attribution to KAPSARC.
4、TheDocument shall not be reproduced,in whole or in part,without the written permissionof KAPSARC.KAPSARC makes no warranty,representation or undertaking whetherexpressed or implied,nor does it assume any legal liability,whether direct or indirect,or responsibility for the accuracy,completeness,or us
5、efulness of any information thatis contained in the Document.Nothing in the Document constitutes or shall be implied toconstitute advice,recommendation or option.The views and opinions expressed in thispublication are those of the authors and do not necessarily reflect the official views or position
6、 of KAPSARC.3World Oil and Critical Mineral Study:A Global VAR AnalysisExecutive SummaryCritical minerals(CMs),such as lithium,cobalt,nickel,and rare earth metals,are essential for the development of clean energy technologies across the whole value chain of wind and solar power,electricity networks,
7、and electric vehicles(EVs).Demand for these minerals is expected to grow quickly as energy transitions accelerate.Other sectors that depend on CMs include the electronics,defense and space industries.Unsurprisingly,CMs have been the focus of major industry participants,policy makers,and academia.A g
8、rowing body of literature explores the impact of CMs on macroeconomic indicators by applying sophisticated techniques,such as Markov switching and vector error correction models or a system dynamics approach.However,to the best of our knowledge,neither CM nor rare earth metal(REM)prices have been in
9、cluded in a global vector autoregressive(GVAR)system.In this study,we develop a model that is capable of examining the consequences of market disturbances from CM price shocks across a wide variety of locations,diverse economic and political systems,and market conditions.It is a stylized representat
10、ion of world oil markets,specifically a quarterly GVAR model,which breaks the global industry down into numerous country-and region-specific disaggregated models.The GVAR modeling approach has numerous clear advantages,including transparency and interdependence on the national and international leve
11、ls that can be empirically evaluated and the ability to model both long-run and short-run relationships that are consistent with the data.We expand the model scope and features by:Including a new variable:the CM price index.Increasing the geographical coverage and extending the estimation:We estimat
12、e the GVAR for 36 countries,including Russia and Venezuela.Expanding the coverage period from 1979Q2 to 2022Q1.The result of these modifications is a revised version of the global oil vector autoregressive(GOVAR)model,a theoretical framework for examining factor interdependencies and the internation
13、al co-movements of variables affecting the global macroeconomy,with an emphasis on the interplay between the crude oil industry and CMs.In the KAPSARC specification of the GVAR model,the CM price is affected by changes in variables such as the world oil price,GDP inflation and world oil inventories
14、with a lag.The CM price,in turn,has the potential to affect individual country-specific economies.This model represents a policy tool that will have the capacity to perform scenario and counterfactual analysis of market disturbances to the CM price and oil markets and make potential policy prescript
15、ions.In the future,the framework can be expanded to include the creation of a unique monthly and real-time database from a variety of sources and timeframes,including satellite inventory data,real-time GDP forecasts,and prices.We run numerous counterfactual simulations of a CM price shock.The result
16、s suggest that CMs are a rising new industry that is starting to have an impact on the macro level,in addition to their already crucial role in energy transitions and other industrial value chains.The industry and its impacts are not yet fully developed but appear to have diverse implications across
17、 countries similar in some respects to the 4World Oil and Critical Mineral Study:A Global VAR Analysiscurrent major commodityoil.A CM price shock has statistically significant implications for inflation in the United Kingdom(U.K.)and South Korea.At the same time,geopolitical shocks to crude oil pric
18、es have significant implications for CM prices.These scenarios indicate the suitability of the proposed GVAR model specification for an examination of the national and country-specific relationships between the oil and CM sectors,for policy prescriptions and for the development of CM scenario analys
19、is.World Oil and Critical Mineral Study 5World Oil and Critical Mineral Study:A Global VAR AnalysisThe role of critical mineralsCritical minerals(CMs)have been playing an increasingly important role in the global energy transition.These elements are critical for the whole value chain of solar and wi
20、nd power.In thin-film solar cells,elements such as gallium,indium,and tellurium are.Lithium,cobalt,cadmium,tellurium,and magnesium are central to battery storage,ranging from electric vehicle(EV)batteries to grid-scale storage.The emerging uses of hydrogen as an energy vector have further increased
21、the demand for CMs,especially in fuel cells such as lithium and graphene.CMs are also used in conventional fossil fuel applications;minerals such as chromium,nickel,manganese,and molybdenum play a central role in the manufacture of tubular goods for operation in high-pressure/highly corrosive enviro
22、nments.The industries where CMs play a key role extend beyond energy,spanning the production of consumer electronics,medical imaging equipment,auto parts,and many other high-tech industrial goods.CMs are also critical for the defense and space industries,for example,to produce sonar,night-vision gog
23、gles,laser range finders,sophisticated communications equipment,and advanced aviation systems.The definitions used to characterize this category of commodities,such as rare earths,critical earth minerals,and CMs,generally imply(1)strategic importance to the national economy or technological developm
24、ent and(2)high risks associated with supply disruption(IEA 2022;Australian Government 2022;Burton 2023).Thus,the set of CMs can vary over time and across countries.For example,in 2022,the U.S.Geological Survey added nickel and zinc while removing helium,potash,rhenium,and strontium from its list of
25、CMs(Burton 2022).For the purpose of this study,we represent the CMs in the model via the price index derived from the U.S.Bureau of Labor Statistics for inorganic chemicals and organic or inorganic compounds of precious metals,of rare earth metals,of radioactive elements or of isotopes(U.S.Bureau of
26、 Labor Statistics 2022).1 For a comprehensive overview of CMs and their fields of application,see Bazilian(2018).CMs tend to include the rare earth metals group,that is,17 metals that share electromagnetic chemical properties(IFPEN 2023).Although many minerals in this category are called critical an
27、d/or rare,they are actually widely available in the Earths crust(Eggert 2011).However,the minable concentrations of these minerals are generally lower than those of other minerals,making the extraction process challenging and capital intensive.Considerable investment is also needed to transform CMs
28、into usable intermediate goods or end products.These challenges,coupled with the unequal distribution of resources,have led to heavy market concentration,which is especially evident in certain CM market sectors.As a result,in 2021,China accounted for 60%of global rare earth production(USGS 2021),whi
29、le the Democratic Republic of Congo produced over 70%of global cobalt(Kitco 2022),and South Africa and Russia captured over 75%of total palladium production(Statista 2022).The disparities are exacerbated by strong demand projections.According to the United States administration,the next several deca
30、des will see an overall CM demand increase of 400%-600%.Specific segments,such as the minerals used in EV batteries,are projected to rise precipitously,with a 40-fold surge(The White House 2022;Foss 2023).Unsurprisingly,the increasing demand for CMs,coupled with the concentrated means of extraction
31、and processing,has created vulnerabilities in the global commodities market and economic development.In 2010,a maritime clash led China to stop exporting CMs,including rare earth oxides,Introduction6World Oil and Critical Mineral Study:A Global VAR Analysisrare earth salts and pure rare earth metals
32、,to Japan for two months,creating considerable difficulties for the Japanese automotive industry(Bradhser 2010).Most recently,the 2022 Russia-Ukraine conflict has illustrated how shocks to CM supplies and prices have the potential to influence the course of country-specific and perhaps even global e
33、conomic development.The ensuing significant upward price shocks to CMs such as nickel and lithium have raised the risk of CM supply shortages for the European EV industry(Shaikhmahmud 2022).There is a growing body of literature addressing the potential implications of a supply shock to critical rare
34、 earth minerals.Sophisticated econometric techniques such as Markov switching models have been used to measure price spillovers between rare earth stocks,financial markets,and oil prices(Reboredo and Ugolini 2020).Keilhacker and Minner(2017)use a system dynamics approach to study individual companie
35、s reactions to restrictions on exports from China.Vector error correction models have been used to disentangle the complex interactions between Chinas complex rare earth metal quotas,statecraft,and pricing policies(Vekasi 2018).To the best of our knowledge,rare earth metal prices have not been inclu
36、ded in a global vector autoregressive(GVAR)system.From the policy perspective,the majority of studies to date tend to focus on the micro level,particularly on industry development(Dou et al.2023),risk management(Keilhacker and Minner 2017),and impacts on downstream segments(Liu et al.2022).On the co
37、untry level,the main policy research agenda concentrates on related geopolitical(Fan et al.2022;Guliyev 2022)and supply security aspects(Bartekova and Kemp 2016).For many economies,the issue of identifying and classifying CMs remains relevant to this day(Galos et al.2021).In the macroeconomic policy
38、 domain,the most explored area is the CM trade-security nexus at both the country(Hau et al.2022;He 2018)and global(Yu et al.2022)levels.A few particular studies explore the macroeconomic characteristics of CMs(Proelss 2020)and their links with other indicators(Cerny 2021).However,there is an eviden
39、t research gap in applied macroeconomic analysis and policy support studies that focus on the role of CMs in the economy of countries and the global economy as well as on relevant potential scenarios.The dynamics of the CM markets and the need for more representation of this sector in global macroec
40、onomic models make understanding how CMs relate to global financial and economic markets a timely and relevant endeavor.We model the dynamic interactions between CMs and oil,which is arguably the most important globally traded commodity.In addition,we investigate into how CM prices,i.e.,the supply-d
41、emand balance,respond to changes in the GDP of the main economies in the world,especially the U.S.,China,and Russia.Understanding the basic VAR frameworkSince its origin in 2004,the use of GVAR models to study the importance of trade and financial links among countries has become well recognized.To
42、cite only a few examples,in the energy field,Dees et al.(2007)examine the international linkages of the Euro Area and for counterfactual analysis including evaluation of the entry of the U.K.into the Euro Area(Dees et al.2007;Pesaran et al.2007;Konstantakis 2015a,2015b).Mohaddes and Pesaran(2016a)de
43、velop a GVAR model for the world oil market and integrate it with a quarterly model of the global economy,a GVAR-Oil model for 27 countries,to investigate the effects of country-specific supply shocks on the global economy.The system is expanded to the GVAR-Oil model by adding a Introduction7World O
44、il and Critical Mineral Study:A Global VAR Analysissimple dynamic oil price equation and combining it with country-specific models(Mohaddes and Pesaran 2016b).Other studies include the use of GVAR models to explore trade linkages between the Caribbean and the United States as well as crude oil price
45、s(Vargas and Hess 2019)and the impact of exchange rate policy(Shah et al.2020;Maral et al.2018).More recently,the GVAR has been used to investigate the international implications of a monetary policy shock in the Euro Area using shadow interest rates as a proxy for monetary policy.Authors have propo
46、sed a new method of identifying a euro-specific shock by using a step procedure for individual and aggregate variables(Beneck et al.2018).Bettendorf(2017)studies the potential implications of shocks to key U.S.macroeconomic variables and the oil price for international trade balances using a GVAR ap
47、proach.The effects of a shock are quantified by means of a variance decomposition of the generalized forecast errors.2 McAdam et al.(2022)employ a structural Bayesian GVAR model to investigate trade imbalances between the South Euro Area(SEA)and the North Euro Area(NEA).Long-and short-run restrictio
48、ns are used to disentangle the structural shocks to the system.In addition,the authors use counterfactual analysis to show that if fiscal austerity or policies improving competitiveness had been employed prior to 2010,the EU debt crisis might have been averted.The GVAR modeling framework makes it po
49、ssible to analyze potential spillover effects from economic shocks and sanctions(Hoyn 2021;Kwok 2022).To cite only a few examples,Sznajderska(2018)employs a GVAR model to estimate the spillover effects of a negative demand shock in China on global GDP growth.Kempa and Khan(2017)analyze the spillover
50、 effects of public debt and economic growth in the Euro Area.They find that debt shocks do not impede growth trajectories but tend to raise debt levels across the Euro Area.Salisu et al.(2022)investigate the spillover effects of financial uncertainty in the United States using a GVAR framework in wh
51、ich uncertainty shocks to developed and emerging economies are conditional on the state of the global financial cycle.There are a growing number of GVAR models utilizing forecasting and counterfactual scenario analysis(Greenwood-Nimmo 2013;Zahedi 2022).The estimation of the trade weight matrix can p
52、lay an important role in both the construction of foreign variable vectors and the estimation of the GVAR system.On the one hand,misspecified weights have the potential to bias the GVAR estimation process,thereby distorting the system dynamics.On the other hand,accurate specification of the weight m
53、atrix has the potential to improve model accuracy and the system dynamics.Gross(2013)outlines a variety of different strategies for the construction of weight matrices,including the use of trade data when the application involves macroeconomic variables such as GDP,inflation,and monetary policy.For
54、financial applications,weights can be constructed by calculating asset exposures such as portfolio equity,investment,and debt.The analysis demonstrates that weight matrices can be estimated jointly with the parameters in cases where the appropriate trade weights cannot otherwise be constructed from
55、the data.A 2020 study by Maral examines the implications of a downturn in a major economy,such as Germany,China and the United States,for the growth paths of smaller economies.The author examines 4 scenarios,setting China,the U.S.and Germany as dominant countries,including the oil price as a global
56、variable,changing trade weights to a 10-and 20-year average,and using the Autometrics algorithm to search for level and impulse dummy variables to Introduction8World Oil and Critical Mineral Study:A Global VAR Analysiscontrol for structural breaks/misspecification in the underlying true data-generat
57、ing process(Maral 2020).The author studies the effects of changes in trade patterns by comparing trade weights calculated using a 20-year average with trade weights calculated using a slightly more contemporary 10-year average.The results show that the impact of shocks from the U.S.and China are rob
58、ust with respect to the change in weights.Dungey et al.(2018)use weight matrices calculated in 1998Q2 and 2015Q4 to examine the implications of transmission shocks for ASEAN-4 and NIE-4 economies.3 They find that the international propagation patterns change significantly across the two distinct tra
59、de regimes.Khan(2020)uses time-varying trade weights to study the implications of trade shocks in Central and Eastern European Baltic countries for economic growth and inflation in the region.The contribution of this studyThis paper contributes to understanding energy markets in three novel ways.Fir
60、st,the model we present in this paper is,to the best of our knowledge,the first of its kind to incorporate CMs in a global economic and energy model.To that end,we expand the GVAR model developed by Mohaddes and Pesaran(2016a)and KAPSARC(Considine et al.2020)to include a new global variable,the CM p
61、rice index.Second,this study features the updated version of the King Abdullah Petroleum Studies and Research Center(KAPSARC)global oil vector autoregressive(GOVAR)model including data for all variables to the first quarter of 2022.Finally,the updated model adds Russian long-term interest rates and
62、extends the temporal coverage of the trade weights and linking matrices to 2022Q1 and 2021,respectively.The original GOVAR model extended Mohaddes and Pesaran(2016)by adding Russia,Venezuela,and Iran as well as oil inventories as an additional variable(Considine et al.2021).The KAPSARC GOVAR model h
63、as been employed to evaluate the time sensitivity of oil shocks under tight and loose market conditions(Considine et al.2022)and to assess the extent of the regional spillover effects of trade and/or financial sanctions on an oil-producing country(Hatipoglu et al.2022).4The modifications to existing
64、 GVAR models include the following:1)The inclusion of a new variable,the CM price index.2)An increase in geographical coverage and an extension of the estimation.We estimate the GVAR model for 36 countries,including Russia and Venezuela,over the period from 1979Q2 to 2022Q1.3)Sufficient flexibility
65、to accommodate the future use of trade matrices to investigate the effects of trade sanctions.The result of these modifications is a revised or augmented version of the GOVAR model,a theoretical framework for examining factor interdependencies and the international co-movements of variables affectin
66、g the global macroeconomy,with an emphasis on the interplay between the crude oil industry and CMs(Considine et al.2022).The structure of the paper is as follows:Section 2 provides a description of the augmented GOVAR model,including a brief discussion of the new oil and CM model and the GVAR system
67、.Section 3 presents the empirical results of a base case scenario,including the estimation of country-specific models,and it illustrates the dynamic properties of the model,including persistence profiles.Section 4 presents the conclusion and suggestions for future research.Appendices A and B describ
68、e the data Introduction9World Oil and Critical Mineral Study:A Global VAR Analysissources and statistical properties of the country-specific models.Appendix B also presents the results of the weak exogeneity tests for the country-specific foreign variables and shows the ability of the model to accou
69、nt for interdependencies and international co-movements via the calculation of pairwise cross-sectional correlations for the endogenous variables and residuals.The main contribution of this study is the creation of a policy tool that will have the capacity to perform scenario and counterfactual anal
70、ysis of market disturbances to the CM price and oil markets and make potential policy prescriptions.The GVAR econometric model is uniquely suited to this analysis.We design a GVAR model that is able to capture the interdependencies between CM prices and the world oil market.In the KAPSARC specificat
71、ion of the GVAR model,the CM price is affected by changes in variables such as the world oil price,GDP inflation and world oil inventories with a lag.The CM price,in turn,has the potential to affect individual country-specific economies.In the future,the framework can be expanded to include the crea
72、tion of a unique monthly and real-time database from a variety of sources and timeframes,including satellite inventory data,real-time GDP forecasts,and prices.Introduction10World Oil and Critical Mineral Study:A Global VAR AnalysisThe framework for studying world oil and CMs builds on a model develo
73、ped by Dees et al.(2007)and Mohaddes and Pesaran(2016a).We first develop a GVAR model to examine the effects of oil and CM price shocks on global economies.The oil and CM prices are modeled separately and introduced into the GVAR model by adding the prices and their lagged values in individual vecto
74、r autoregressive with exogenous foreign variable(VARX*)models(Smith and Galesi 2014).5 In a departure from the literature at the time,we model the oil price equation separately and introduce the oil price variable as weakly exogenous in all countries,including the United States(Mohaddes et al.2020).
75、The data series utilized in the model and corresponding data sources are described in detail in Appendix B.The model for oil and critical mineral pricesThe KAPSARC oil price and CM model expands on the GOVAR model presented in Considine et al.(2020).The dynamics of the global oil and CM markets can
76、be described by the following equations for the dynamic aggregate demand for oil and the CM price.()()()=+aaL yaLaL pQdRptdyytRRtppt000()+a L IIItdt0(1a)()()()=+bbL ybLbL pRpRptdytRtpt000()()+bLbLDpQsDptQstdt00(1b)where:Qdt0 Lagged value of oil demand Qst0 Lagged value of oil supply Rpt0 Lagged valu
77、e of the rare earth metal price index Dpt0 Inflation first difference of the lagged value of the CPI Yt Lagged value of real seasonally adjusted GDP Pt Lagged value of the real price of oil It0 Lagged value of oil inventories!(),(),#(),$()aL a L aLa L(),(),(),()ypRIPolynomials in the lag operator,L,
78、whose coefficients sum to 1=+aLaa La L()yyyy0122=+aLaa La L()RRRR0122=+a Laa La L()pppp0122(2)=+a Laa La L()IIII0122=aLaLa La L()()()()1yRpI bLb Lb LbLbL(),(),(),(),()ypRDpQs Polynomials in the lag operator,L,whose coefficients sum to 1It can be shown that y,p,R and I are the long-run income,price,a
79、nd inventory elasticities of the demand for oil and R is the long-run cross elasticity of the demand for oil and the CM price.Oil prices respond to supply and demand imbalances to create equilibrium or balance on global oil markets(Considine et al.2020).!=#+(Qd!Qs!)+#!(3)where:The speed of adjustmen
80、t between oil supply and demand!A fixed constant representing the scarcity of oil!Speculative oil price changes that are not related to fundamental factorsModeling Oil and Critical Mineral Prices11World Oil and Critical Mineral Study:A Global VAR AnalysisSubstituting equation 1a into equation 3 and
81、solving for pt0 yields the following:=+paaL ya L paL()()()RptpyytpptRRt000+a L I()Qs)IIttpt00(4)where:+aaapsd+ptsdtSolving for pt0 yields a standard autoregressive distributed lag(ARDL)model of oil prices,rare earth metal prices,income,inventory,and world oil production.!=#11+#$#$)#+#1#$#$1+#$#$)!%&
82、,(!&)(!*!-#+!%+-.+,(#&)(!*!-/()!+#01+#$#$)0()!+#1+#$#$)Qs!+#11+#$#$)1()Rp!+#11+#$#$)#!(5)Following Considine et al.(2020),we estimate the ARDL model described by equation 6.Endogeneity problems are avoided by using lagged values of yt and Qst0.=+=pca pyQtplmlt llmlt llmlt l s01011,0pyqo0+=RpIlmlt ll
83、ml t lt10100RoI(6)where mp0,my,mqo,mRo and mI are allowed to vary across the different variables and will be selected using the Akaike information criteria(AIC)(Akaike 1981).It can be shown that the long-term price,income and inventory elasticities and the long-run cross elasticity of the demand for
84、 oil and the rare earth metal price are as follows:!=%#$!#%&(&(1#$#$#%&)(=%#$!#%&(&%#$%#%&()=%#$!#%&(&%#$%&(*=%#$!#%&(&%#$#%&(7)Expanding the system:An international perspective in the GVAR modelThe GVAR methodology is a two-step modeling procedure.In the first step,countries are estimated individ
85、ually by means of country-specific vector error correction models that include domestic and foreign variables and two global variables that are common across all countries,specifically oil and rare earth metal prices.All countries,except the United States,are treated as small open economies.In the s
86、econd step,the individual models are combined,and the GVAR model is solved for the world as a whole because all the variables are endogenous to the system as a whole.Modeling Oil and Critical Mineral Prices12World Oil and Critical Mineral Study:A Global VAR AnalysisStep 1:Estimating the country-spec
87、ific vector error correction modelsWe begin by estimating a single equation for each country-specific model:=+11,1,1,1*xaa t xxA xitioiii tipi tpii tii+,*A xuiqi t qiti(8)Or equivalently:()()=+L p xaa txuiiitioii titL q,ii1,*(8a)where:aio,ai1=K 1 vectors of fixed intercepts and coefficients on the d
88、eterministic time trends.=k1ixit vector of country specific domestic variables=k1ixit*vector of country specific weakly exogenous star(foreign)variables!,!,!#,!$!1 vectors or matrices of fixed coefficients that vary across countries=k1iuit vector of country-specific supply shocks.(,)the shocks are s
89、erially uncorrelated with zero means and the nonsingular singual covariance matrix.(,)=#the matrix lag polynomial of the domestic variable coefficients(,)=$the matrix lag polynomial of the foreign variable coefficientsThe variables=xYP IrrlDpep eq(Qs,)ititittitititititit00 are the country-specific d
90、omestic variables.Qsit 0 Lagged value of oil supply Yit Lagged value of real seasonally adjusted GDP Pt Lagged value of the real price of oil Iit 0 Lagged value of oil inventories rit=0.25*In(1+Rit/100)Rit Nominal short term interest rates rlit=0.25*In(1+Rlit/100)Rlit Nominal long term interest rate
91、s Dpit Inflation first difference of the lagged value of the CPI epit Equity prices,lagged value of nominal equity prices divided by the CPI eqit Exchange rates,lagged value of nominal exchange rates divided by the CPIThe variables!=#!,!$,!,!,!,!,!.are country-specific star foreign variables constru
92、cted using country-specific trade shares and defined by equation 8a.Note that ai1=0,as there is no time trend in this specification of the model.The variables Qsit 0,and Rpt 0 are excluded because they have already been included in the models for oil and rare earth metal prices.These variables are c
93、ommon factors present in all of the country-specific models.They can be modeled as global or dominant variables and have implications for both the world and individual countries(Pesaran 2015).Following Mohaddes and Raissi(2020),the real exchange rate is defined as the logarithm of the real exchange
94、rate(the nominal exchange rate divided by the CPI),and the U.S.dollar is the“reference currency”for the model.=xw xitjNijjt*1(8b)where wij,I,j=1,2,N,are bilateral trade weights with=0,=$.The trade weights,Modeling Oil and Critical Mineral Prices13World Oil and Critical Mineral Study:A Global VAR Ana
95、lysiswij,are computed as a three-year moving average to reduce the impact of extreme annual movements on the trade weights.Specifically,=+wTTTTTTijijijijiii,2019,2020,2021,2019,2020,2021(8c)where Tijt,I,is the bilateral trade of country I with country j during a given year t and is equal to the aver
96、age of exports and imports of country I with country j.=%(the total trade of country i)for t=2019,2020,2021 and j=1,2,N.The weights used for the world oil and rare earth metal study are presented in Table 1.In addition,the KAPSARC GVAR specification makes it possible to analyze regional responses to
97、 shocks.We define the following regions:Europe,the Euro Area,net oil exporters and importers,Latin America,Asia Pacific and the rest of the world(ROW).Following Dees et al.(2007),we use weights based on the purchasing power parity(PPP)valuation of the individual countries real GDP for regional aggre
98、gation and the derivation of aggregate impulse response functions.The PPP method has Modeling Oil and Critical Mineral PricesTable 1.Fixed trade weight matrix(2019-2021).ArgentinaAustraliaBrazilCanadaChinaChileEuroIndiaIndonesiaIranJapanKoreaMalaysiaMexicoNorwayNew ZealandPeruPhilippinesRussiaSouth
99、AfricaSaudi ArabiaSingaporeSwedenSwitzerlandThailandTurkeyUKUSAVenezuelaArgentina0.00000.00180.05200.00110.00500.02860.00420.00720.00560.01830.00120.00170.00290.00210.00040.00240.02180.00200.00220.00370.00390.00050.00110.00320.00420.00280.00140.00270.0101Australia0.00830.00000.00380.00390.05340.0053
100、0.01250.02730.02580.00140.04610.03090.02890.00300.00260.17730.00270.01390.00170.01100.00510.03950.00730.00800.03630.00510.01530.01100.0004Brazil0.22430.00310.00000.00750.03470.06490.01850.01580.01020.04750.00820.01010.01000.01050.01000.00210.04310.00630.01270.01150.01540.01110.00700.00760.00950.0130
101、0.00610.01740.0629Canada0.01000.00770.01770.00000.02550.01560.02000.01290.00890.00810.01900.01310.00660.03820.01780.01330.04540.00970.00420.00900.01280.00430.00870.01160.00920.01020.02380.16300.0065China0.16590.40220.33570.08170.00000.36040.17740.18470.27850.43990.28530.34170.28950.08150.06300.27820
102、.33800.35030.26340.26270.27200.20190.07480.06300.26320.09620.11050.18020.2773Chile0.04950.00130.02400.00230.01500.00000.00500.00370.00100.00010.00730.00640.00090.00390.00120.00260.03610.00140.00180.00120.00050.00030.00230.00200.00270.00260.00150.00710.0044Euro0.14870.07830.14720.06390.16830.11440.00
103、000.13520.06740.11490.09690.08880.07280.06690.39720.09420.11140.07100.36610.24190.15040.09460.57390.50890.07550.43560.45090.15290.1263India0.04070.03290.02140.00720.02770.01430.02510.00000.06040.09580.01480.02250.03280.00900.00830.01220.03280.01170.02150.07520.10850.05550.00790.04000.02830.02700.017
104、20.02650.2274Indonesia0.02180.01890.00880.00310.02510.00230.00730.03040.00000.01140.02470.01700.03490.00160.00250.01840.00360.04090.00480.01030.01640.05820.00270.00420.03490.00580.00280.00860.0019Iran0.00910.00330.01240.00450.02170.00110.01470.05520.01520.00000.02500.02540.00790.00110.00140.01030.00
105、110.00640.00200.02510.00000.01870.00520.00450.01600.01780.00560.00780.0001Japan0.01490.12440.03040.02320.09690.06150.03840.02960.08730.01190.00000.08450.06330.01840.01550.06830.03760.10500.04340.05440.10200.06130.01710.02390.13420.01290.02130.06010.0048Sourth Korea0.01670.06840.02420.01330.07960.038
106、50.02590.03690.04900.02260.06570.00000.03800.01990.01910.04010.04270.06140.04870.01790.08460.04620.01200.00720.03330.02630.01150.04180.0029Malaysia0.01590.02620.00790.00220.02600.00270.01050.02710.05490.01240.02730.02110.00000.00330.00210.02460.00310.03550.00400.00790.01690.12600.00340.00420.05490.0
107、1030.00460.01290.0270Mexico0.01750.00290.02830.02700.03010.02480.01900.01650.00570.00010.01810.02620.02520.00000.00170.00660.02470.01770.00440.00490.00090.00610.00460.00560.01640.00540.00570.16580.0220Norway0.00090.00100.00510.00360.00520.00230.02440.00250.00110.00040.00230.00320.00110.00070.00000.0
108、0130.00320.00050.00510.00250.00050.00210.07810.00250.00220.00860.03520.00290.0001New Zealand0.00210.02170.00050.00110.00630.00130.00230.00160.00450.00050.00410.00330.00380.00060.00060.00000.00110.00300.00090.00120.00250.00400.00120.00070.00590.00070.00230.00230.0000Peru0.02300.00050.00820.00420.0064
109、0.01880.00300.00360.00130.00010.00260.00370.00050.00270.00040.00190.00000.00120.00090.00100.00020.00020.00130.00390.00130.00120.00090.00450.0043Philippines0.00470.00380.00300.00130.01070.00060.00430.00460.02400.00120.01790.01230.01340.00100.00040.00770.00080.00000.00210.00100.00490.02330.00100.00160
110、.02350.00060.00130.00540.0001Russia0.01230.00190.01370.00180.03590.00680.05630.02060.00800.06560.01610.02700.00590.00370.01470.00570.00630.00530.00000.00640.00650.00500.01470.01200.00520.09720.02670.00820.0078South Africa0.00720.00390.00420.00080.00870.00150.01180.01540.00420.00020.00690.00290.00280
111、.00090.00170.00310.00070.00120.00210.00000.01430.00220.00430.00350.00790.00420.00910.00410.0001Saudi Arabia0.00910.00330.01240.00450.02170.00110.01470.05520.01520.00000.02500.02540.00790.00110.00140.01030.00110.00640.00200.02510.00000.01870.00520.00450.01600.01780.00560.00780.0001Singapore0.00220.03
112、400.01010.00290.03080.00170.01830.03060.11410.00050.03020.03460.16300.00430.00770.03510.00090.07560.00760.00580.02360.00000.00530.01850.05340.00510.01250.02120.0096Sweden0.00210.00440.00410.00190.00560.00360.04660.00370.00180.00120.00330.00310.00160.00100.19340.00320.00350.00120.01080.00590.00400.00
113、260.00000.00550.00260.01080.01770.00540.0090Switzerland0.02030.00890.00970.00750.01250.00870.08210.03970.00670.00590.01090.00520.00430.00310.00680.00880.02760.00500.00970.02220.00930.02390.01600.00000.01960.02140.04790.02280.0022Thailand0.01770.03160.01060.00290.02660.00690.01010.02220.04590.00750.0
114、4780.01500.04410.00410.00400.03150.00580.04970.00610.01900.02390.03670.00430.01640.00000.00540.00630.01460.0017Turkey0.00790.00350.01000.00290.00850.00520.03660.01370.00510.12840.00380.00810.00540.00180.00910.00210.00660.00130.06110.00960.01550.00220.01300.01130.00380.00000.02010.00670.0312U.K.0.015
115、20.02910.01390.02410.03110.01080.12970.02690.00800.00280.01570.01220.01010.00540.16360.02880.01060.00700.04840.06370.02050.02070.05590.06960.01260.07030.00000.03590.0047U.S.0.12970.08100.17830.69960.17990.19590.18070.17140.09010.00130.17360.15430.12150.70990.05340.11010.18680.10920.06220.09970.08490
116、.13360.07140.15590.12730.08450.13640.00000.1552Venezuela0.00210.00000.00230.00010.00110.00060.00050.00570.00010.00020.00010.00000.00060.00040.00000.00000.00080.00000.00010.00000.00000.00040.00050.00000.00000.00100.00010.00050.0000Source:Internal KAPSARC calculations,International Monetary Fund,Direc
117、tion of Trade Statistics,2022.14World Oil and Critical Mineral Study:A Global VAR Analysisbeen shown to be more reliable than weights based solely on U.S.dollar valuations(Dees et al.2007).Including the dominant and global variables:The oil price and critical mineralsThe global variables,oil and rar
118、e earth metal prices,are added to the systemthe conditional country modelsas global or dominant variables.The addition entails the following augmentation to equation 8a:(,)=+%(,%),+(,)+(9)where is a vector of global or dominant variables and its lagged values.The model can be augmented to allow for
119、feedback effects from the domestic variables as follows:=%,)+%,)+(10)where(,)=#is the matrix lagged polynomial of the global and dominant variable coefficients.In the new specification,is allowed to vary and can be selected by the AIC or Schwarz information criterion(SBC)methodologies(Mohaddes et al
120、,2020).Oil prices and rare earth metals are treated as dominant or global variables,and equation 10 is specified by the oil and rare earth metal price models specified in equations 1b and 6.While the common variables can be treated as a foreign variable for the purposes of modeling and share the sam
121、e lag order(q),this specification allows for different lag orders for the dominant and foreign variables!,!$,!,$*.The oil and rare earth price equations 1b and 6 are standard ARDL models of oil and CM prices.The fixed weights used to construct the feedback variables are(i)PPP for the financial econo
122、mic variables,real GDP,inflation,real equity prices,real exchange rates,and short-and long-term interest rates;(ii)the contribution to OECD inventories for the inventory variable;and(iii)the contribution to total oil production for the crude oil production variable.The weighting system is derived as
123、 follows:!=!)=!=!)=,!)(10a)where:jY is calculated as a three-year(2019,2020 and 2021)average of the PPP-GDP weights of country j,and=$jI is calculated as a three-year average from 2019Q1 to 2021Q1 of the quarterly weights of country j in terms of its contribution to OECD inventories,and=$jQ is calcu
124、lated as a three-year average from 2019Q1 to 2021Q1 of the quarterly weights of country j in terms of its contribution to total oil production from the producing countries listed in the GVAR model,and=$.Modeling Oil and Critical Mineral Prices15World Oil and Critical Mineral Study:A Global VAR Analy
125、sisThe weights used in the feedback equations for the oil and rare earth price model are given in Table 2.The combined GVAR model permits a two-way link between real GDP,oil supplies,oil inventories,the CM price and the world oil price.Changes in GDP,inventories,CM prices and oil supplies affect the
126、 world oil price with a lag.World oil prices,in turn,can affect the variables in country-specific economies.The CM price is affected by changes Modeling Oil and Critical Mineral PricesTable 2.Fixed feedback weight matrix(2019-2021).Country weights for real GDP,production and inventories.CountriesGDP
127、-PPPProductionInventoriesArgentina0.00896710.00000000.0000000Australia0.01208600.00000000.0000000Brazil0.02883820.00000000.0000000Canada0.01647100.09662480.0547898China0.21980870.09456370.0000000Chile0.00444850.00000000.0000000Euro0.13184720.00000000.1604838India0.08429340.00000000.0000000Indonesia0
128、.02992340.07481770.0000000Iran0.01148570.09131460.0000000Japan0.04718960.00000000.1523164South Korea0.02053720.00000000.0566884Malaysia0.00827860.00000000.0000000Mexico0.02215840.08576580.0000000Norway0.00328410.08477590.0000000New Zealand0.00203100.00000000.0000000Peru0.00378610.00000000.0000000Phi
129、lippines0.00861840.00000000.0000000Russia0.03982980.10689710.0000000South Africa0.00732000.00000000.0000000Saudi Arabia0.01484310.10509320.0000000Singapore0.00523740.00000000.0000000Sweden0.00513040.00000000.0000000Switzerland0.00560200.00000000.0000000Thailand0.01160200.00000000.0000000Turkey0.0208
130、9440.00000000.0000000U.K.0.02859650.07821430.0225983U.S.0.19140960.10767230.5531233Venezuela0.00548240.07426040.0000000Sources:EIA 2022,World Bank Development Indicator database 2022,KAPSARC internal calculations.16World Oil and Critical Mineral Study:A Global VAR Analysisin the world oil price,GDP
131、inflation and world oil inventories with a lag,and the CM price,in turn,has the potential to affect country-specific economies.The fact that aggregate global GDP,oil inventory,inflation,and oil production variables are excluded from the global price equations allows us to identify and evaluate count
132、ry-specific shocks to inventories,income,and oil supplies(Mohaddes and Pesaran 2016a).Step 2:Solving for the system as a wholeIn the second step of the estimation process,we combine the oil and CM equations and individual country-specific models to complete the GVAR model.The GVAR model is solved fo
133、r the world as a whole,and all variables are treated as endogenous to the system.The GVAR model and solution are described in detail in Appendix A.Modeling Oil and Critical Mineral Prices17World Oil and Critical Mineral Study:A Global VAR AnalysisThe scope of the modelThe world oil and CM GVAR model
134、 represents 36 countries,4 regions and subregions,and 8 country-specific variables(see Table 3).The data set includes quarterly data from 1979Q1 to 2022Q1 and was taken from a variety of industry sources,including the U.S.Energy Information Administration(EIA),the World Bank,the International Moneta
135、ry Fund(IMF),and Bloomberg.The data sources for the study are described in detail in Appendix B.The oil-producing countries include net oil exporters:Canada,Indonesia,Venezuela,Russia,Iran,Mexico,Norway,and Saudi Arabia.The new oil importers include Brazil,China,the U.K.,and the U.S.These 12 countri
136、es account for 67%of world crude oil production and hold approximately 67%of the worlds proven oil reserves.See Table 4(BP 2021,2022).Empirical Results:Estimation of Base Case Country-Specific ModelsTable 3.Countries and regions in the GVAR model.Countries utilized in the world oil and CM modelArgen
137、tinaIndonesiaRussiaAustraliaIranSouth AfricaAustriaItalySaudi ArabiaBelgiumJapanSingaporeBrazilSouth KoreaSpainCanadaMalaysiaSwedenChinaMexicoSwitzerlandChileNetherlandsThailandFinlandNorwayTurkeyFranceNew ZealandUnited KingdomGermanyPeruU.S.IndiaPhilippinesVenezuelaRegions and subregions accounted
138、for in the world oil and rare earth metal modelNet oil exportersEuropeLatin AmericaEuro AreaNet oil importersAsia PacificRest of worldGlobal variables in the world oil and rare earth metal modelWorld oil priceRare earth metal priceCountry-specific variables in the world oil and rare earth metal mode
139、lReal GDPOil inventoriesReal exchange ratesInflationCrude oil productionShort-term interest ratesReal equity pricesLong-term interest ratesSource:KAPSARC calculations,2022.18World Oil and Critical Mineral Study:A Global VAR AnalysisThe euro block of countries includes Austria,Belgium,Finland,France,
140、Germany,Italy,the Netherlands,and Spain,and the time series for the Euro Area are calculated using weighted averages of the eight Euro Area countries using PPP-GDP weights averaged over the 2019-2021 period.Specification of country-specific vector autoregression modelsGiven the quarterly estimates o
141、f the data for the domestic variables from 1979Q2 to 2022Q1,we estimate the 36 individual country-specific models.The modeling exercise assumes that the country-specific foreign variables are weakly exogenous variables and that the parameters are stable over time.Unit root tests performed on the var
142、iables utilized by the GVAR model show that the variables utilized in the model are integrated of order one.The unit root tests,weak exogeneity,and structural stability test results are reported in Appendix C.The weight matrix that is used to calculate the foreign-specific star variables and the sol
143、ution to the GVAR model,including the W or link matrices(see equation 8b)and bootstrapping,is shown in Table 1.The variables specified by the country-specific VARX models are illustrated in Table 5.The model Table 4.Crude oil production and reserves.2021 CountryOil production(thousand barrels daily)
144、Oil reserves(billion barrels)Net exportersCanada5429169.10Indonesia6922.40Iran3620157.80Mexico19286.10Norway20257.90Russia10944107.80Saudi Arabia10954297.50Venezuela2110303.18Net importersBrazil65411.90China399426.00United Kingdom8742.50United States1658568.80Total598091160.98Rest of the world300685
145、71.42World total898771732.40Model producers as a percentage of the world total67%67%Sources:Total proven reserves at end of 2020,thousand million barrels.BP(2021,2022).Empirical Results19World Oil and Critical Mineral Study:A Global VAR AnalysisEmpirical ResultsTable 5.List of variables included in
146、the country-specific VARX models.ModelsDomestic variablesForeign variablesGlobal variablesYitDpiteqitepitritrlitIit0Qsit0YitDpiteqitepitrlitIit0pit0Rpit0Argentina1Australia11Brazil111111111Canada11China1Chile1Euro111India1I
147、ndonesiaIranJapan11South Korea111Malaysia1111111111MexicoNorway111New Zealand1Peru1111111111PhilippinesRussia1South Africa1Saudi Arabia111111Singapore1Sweden11Switzerla
148、nd11Thailand1111111111TurkeyU.K.1111U.S.Venezuela111111111Note:The value of 1 is given if the variable is included in the analysis.Source:KAPSARC calculations,2022.20World Oil and Critical Mineral Study:A Global VAR Analysisfor the United States differs fr
149、om the specification for all other countries in one respect:U.S.dollar exchange rates are included as endogenous variables in all other countries,except the U.S.This reflects the importance of the U.S.financial system in the world economy and is supported by empirical evidence that the global financ
150、ial cycle in capital flows,asset prices,and credit growth is driven primarily by the monetary policy settings of the United States(Chudik et al.2013)(Mohaddes and Pesaran 2016b).Foreign,country-specific interest rates are omitted from the country-specific models.The selection of lag orders,cointegra
151、ting relationships,and persistence profilesThe lag orders for the domestic and foreign variables are selected by the AIC test statistic,which is applied to the underlying VARX models,with maximum lag orders set to 2.The results of the selection process are presented in Table 6(Akaike 1981).The coint
152、egrating relationships,also shown in Table 6,are chosen using the MacKinnon trace test statistics and 95%critical values(MacKinnon 1990).The dynamic properties of the model in response to a system-wide shock are described by the persistence profiles(PPs)and based on an infinite moving average of the
153、 GVAR model(see Appendix A).As shown in Figure 1,the PPs are normalized to a starting value of 1 for the impact of the system-wide shock.The rate at which they tend to zero provides information to the analyst on the speed with which the system tends to return to equilibrium after a shock.If the rela
154、tionship underlying the model is cointegrated,the PPs have the potential to overshoot but will converge to zero in a finite period(Esfahani,Mohaddes,and Pesaran 2012).All the variables return to their long-run equilibrium values after the initial system-wide shock.In most countries,the speed of conv
155、ergence was very fast,taking one to two years.In 50 out of 59 cointegrating relationships,the value of PP was less than 20%after two years.By the fourth year,all of the countries had returned to their equilibrium values.The countries reaching full convergence to equilibrium levels faster than two ye
156、ars include Argentina,Australia,Brazil,Canada,Chile,the Euro Area,India,Indonesia,Iran,South Korea,Malaysia,Mexico,Norway,New Zealand,South Africa,Singapore,Sweden Switzerland,Thailand,Turkey,the U.K.,the U.S.,and Venezuela.China,Japan,Peru,Russia,Saudi Arabia,and Singapore all took longer than 3 ye
157、ars to converge.The PPs for the net oil-exporting countries and their 95%bootstrapped error bands are shown in Figure 2.For all of the net oil exporters,the speed of convergence is very fast,and convergence is achieved in 3-4 years.Among these countries,Saudi Arabia and Russia are the slowest,at app
158、roximately 2 years.In the case of Saudi Arabia,this result might be attributed to the sovereign wealth fund,which can absorb shocks and lead to a more sluggish response to system-wide shocks(Esfahani,Mohaddes,and Pesaran 2012).Empirical Results21World Oil and Critical Mineral Study:A Global VAR Anal
159、ysisTable 6.The lag orders for the domestic and foreign variables.Lag orders for domestic variablesLag orders for foreign variablesCointegrating RelationshipsArgentina211Australia113Brazil211Canada123China112Chile213Euro212India211Indonesia212Iran111Japan223South Korea213Malaysia211Mexico222Norway21
160、3New Zealand222Peru222Philippines213Russia211South Africa212Saudi Arabia222Singapore212Sweden213Switzerland211Thailand213Turkey211U.K.221U.S.213Venezuela212Source:KAPSARC calculations,2022.Empirical Results22World Oil and Critical Mineral Study:A Global VAR AnalysisFigure 1.Persistence profiles of t
161、he effect of a system-wide shock to the cointegrating relationships,with bootstrap medians.Source:KAPSARC calculations,2022.Empirical Results23World Oil and Critical Mineral Study:A Global VAR AnalysisFigure 2.Persistence profiles for net oil-exporting countries.Source:KAPSARC calculations,2022.Empi
162、rical Results24World Oil and Critical Mineral Study:A Global VAR AnalysisThe PPs for the net oil-importing countries and their 95%bootstrapped error bands are shown in Figure 3.Interestingly,the speed of convergence for net oil importers is considerably more diverse,ranging from less than one year f
163、or Brazil to over 3 years for China.The faster convergence speeds in some of the underdeveloped major oil-exporting countries might be attributed to the relatively underdeveloped capital markets(Mohaddes and Pesaran 2016a).Estimating the contemporaneous effects of foreign variables on their domestic
164、 counterpartsThe impact elasticities of foreign variables on their domestic counterparts are shown in Tables 7a and b.6 The elasticities reflect the international linkages between domestic and foreign variables and show how an individual country is affected by changes in foreign variables such as eq
165、uity prices,inflation,interest rates,GDP,crude oil inventories,and crude oil production.As expected,the majority of the estimated coefficients are statistically significant.The income elasticity for the Euro Area is approximately 0.77%,suggesting that a 1%increase in foreign GDP results in a 0.77%ch
166、ange in real GDP in the Euro Area.The real income elasticities for developed countries are all similar or slightly lower,as is the case for the U.S.(0.54%),Canada(0.76%),and Norway(0.31%).China,Japan,and South Korea are considerably lower at 0.44%,0.05%and 0.18%,respectively.The“foreign”income elast
167、icities for the Middle Eastern nations included in the model,Saudi Arabia(0.30%)and Iran(0.02%),are considerably lower.Figure 3.Persistence profiles for net oil-importing countries.Source:KAPSARC calculations,2022.Empirical Results25World Oil and Critical Mineral Study:A Global VAR AnalysisTable 7a.
168、The contemporaneous effects of foreign variables on their domestic counterparts.YitDpiteqitrlitIit0ArgentinaCoefficient0.853*0.6841.104*Whites adjusted SE0.3670.5600.331 Newey-Wests adjusted SE0.4390.6200.344AustraliaCoefficient0.380*0.432*0.851*0.985*Whites adjusted SE0.0920.1100.1050.146 Newey-Wes
169、ts adjusted SE0.0840.1220.1190.148BrazilCoefficient0.212*Whites adjusted SE0.153 Newey-Wests adjusted SE0.118CanadaCoefficient0.768*0.902*0.731*Whites adjusted SE0.1380.0370.168 Newey-Wests adjusted SE0.1850.0330.173ChileCoefficient0.531*0.127*0.869*Whites adjusted SE0.1870.0670.153 Newey-Wests adju
170、sted SE0.1980.0780.160ChinaCoefficient0.442*0.135 Whites adjusted SE0.1560.171 Newey-Wests adjusted SE0.1360.159EuroCoefficient0.765*0.120*1.153*0.649*0.505*Whites adjusted SE0.1060.0430.0560.0760.078 Newey-Wests adjusted SE0.1320.0460.0740.0810.097IndiaCoefficient1.050*0.250-0.036 Whites adjusted S
171、E0.4630.1800.110 Newey-Wests adjusted SE0.5030.1840.114IndonesiaCoefficient0.659*0.688*Whites adjusted SE0.2630.324 Newey-Wests adjusted SE0.2660.358IranCoefficient0.0280.0740.285*Whites adjusted SE0.2810.1560.161 Newey-Wests adjusted SE0.2480.1490.166JapanCoefficient0.0500.756*0.449*0.392*Whites ad
172、justed SE0.0670.0810.1020.075 Newey-Wests adjusted SE0.0640.0850.1030.072South KoreaCoefficient0.1800.1550.775*0.0820.696 Whites adjusted SE0.1160.1440.1770.1880.697 Newey-Wests adjusted SE0.1130.1830.1810.2050.868MalyasiaCoefficient1.155*Whites adjusted SE0.145 Newey-Wests adjusted SE0.175MexicoCoe
173、fficient0.754*-0.225 Whites adjusted SE0.2000.421 Newey-Wests adjusted SE0.2320.306New ZealandCoefficient0.570*0.835*0.469*Whites adjusted SE0.1600.1020.197 Newey-Wests adjusted SE0.1930.0840.215Source:KAPSARC calculations,2022.Note:*,*,and*indicate that the estimated coefficient is statistically si
174、gnificant at the 10%,5%and 1%levels,respectively.Empirical Results26World Oil and Critical Mineral Study:A Global VAR AnalysisTable 7b.The contemporaneous effects of foreign variables on their domestic counterparts.YitDpiteqitrlitIit0NorwayCoefficient0.313*0.386*0.932*0.799*Whites adjusted SE0.1300.
175、1320.0960.137 Newey-Wests adjusted SE0.1030.1460.1090.134PeruCoefficient1.121 Whites adjusted SE1.254 Newey-Wests adjusted SE1.199PhilippinesCoefficient0.3400.0191.333*Whites adjusted SE0.2340.2170.146 Newey-Wests adjusted SE0.2690.2360.160RussiaCoefficient0.583*0.1681.718*Whites adjusted SE0.2271.1
176、000.20665 Newey-Wests adjusted SE0.1750.9440.23680Saudi ArabiaCoefficient0.306*Whites adjusted SE0.125 Newey-Wests adjusted SE0.138SingaporeCoefficient1.096*0.300*1.145*Whites adjusted SE0.2430.0810.117 Newey-Wests adjusted SE0.2740.0840.131South AfricaCoefficient0.713*0.936*0.582*Whites adjusted SE
177、0.1850.1370.232 Newey-Wests adjusted SE0.1860.1470.223SwedenCoefficient0.939*0.759*1.048*0.961*Whites adjusted SE0.1460.1440.0650.117 Newey-Wests adjusted SE0.1930.1400.0750.132SwitzerlandCoefficient0.599*0.258*0.835*0.486*Whites adjusted SE0.0550.0990.0630.068 Newey-Wests adjusted SE0.0470.1240.072
178、0.073ThailandCoefficient0.509*1.145*Whites adjusted SE0.1970.116 Newey-Wests adjusted SE0.1860.115TurkeyCoefficient1.154*-0.131 Whites adjusted SE0.2890.639 Newey-Wests adjusted SE0.3380.449U.K.Coefficient1.228*0.328*0.812*0.791*0.751*Whites adjusted SE0.2570.0980.0540.1090.107 Newey-Wests adjusted
179、SE0.2920.0960.0590.1000.114U.S.Coefficient0.545*0.450*Whites adjusted SE0.0830.086 Newey-Wests adjusted SE0.0760.131VenezuelaCoefficient0.470*-1.614 Whites adjusted SE0.1601.448 Newey-Wests adjusted SE0.1991.207Source:KAPSARC calculations,2022.Note:*,*,and*indicate that the estimated coefficient is
180、statistically significant at the 10%,5%and 1%levels,respectively.Empirical Results27World Oil and Critical Mineral Study:A Global VAR AnalysisThe elasticities for long-term interest rates are reasonably high and positive for all countries except India,which has a negative long-term interest rate ela
181、sticity of 0.04%.This result illustrates a strong relationship between long-term rates and monetary policy reactions across countries.The elasticities for oil inventories are strong and statistically significant at the 5%level for most variables across countries.Notably,a 1%increase in global invent
182、ories results in a 0.50%increase in inventories in the Euro Area,rising to 0.73%in Canada,0.75%in the United Kingdom,and 0.45%in the United States.Pairwise cross-sectional correlationsIn the GVAR framework,an“idiosyncratic”shock to the country-specific models is assumed to be weakly correlated acros
183、s countries,cov x uas N(,)0,ijit*.The country-specific models are conditioned on weakly exogenous foreign variables.These weakly exogenous foreign variables,in turn,serve as proxies for the common global factors that affect the system.Following Dees et al.(2014),the remaining shocks across countries
184、 should be modest and reflect only residual considerations,such as the spillover effects of monetary and fiscal policy and trade movements.We test the proposition that the foreign variables are successful in capturing common global factors and,as a result,reduce the cross-sectional correlation of th
185、e variables.To do so,we compute the average pairwise correlations for the levels and first differences of the endogenous variables and the associated residuals for the sample period,i.e.,from 1979 to 2022(Dees et al.2007).The pairwise cross-sectional correlation tests are reported in Tables 8a-c.As
186、expected,the correlations are high for the levels of the endogenous variables and fall when the first differences and residuals are considered.The level of real GDP show the highest levels of cross-sectional correlations,which range between 0.84 and 0.91 for all countries except Russia(0.38)and Vene
187、zuela(-0.47).The second highest cross-sectional correlations are reported for the level of equity prices and range from 0.80 for Norway to 0.19 for Argentina.These are followed closely by the cross-sectional correlations for the exchange rates,which range from 0.79 for Australia to-0.14 for Iran.Uns
188、urprisingly,Iran and Venezuela(0.29)display the lowest cross-sectional correlations for the exchange rate variable,as the other countries fall into a significantly higher(0.42 to 0.79)range.For the level of long-term interest rates,the cross-sectional correlations range from 0.78 to 0.45 in the Euro
189、 Area and fall to-0.75 in Iran.The large negative value for Iran reflects years of war and harsh economic sanctions.The short-term interest rate correlations are considerably lower,ranging from 0.64 to-0.63 for Iran.Unsurprisingly,for both Iran and Venezuela(-0.21),the correlations are negative,refl
190、ecting years of economic isolation.In the oil sector,the pairwise cross-sectional correlations are significant for the level of oil inventories,with an average correlation of 0.11,ranging from-0.53 to 0.32.The highest level is seen in South Korea(0.32),and the lowest is reported in the U.K.(-0.53).T
191、he effects taper off for the first differences(0.07-0.38)and fall significantly for the residuals,which range from-0.18 to 0.02.The results for crude oil production are considerably lower and more diverse,with pairwise correlations for levels averaging only-0.02,ranging from-0.26 in Empirical Result
192、s28World Oil and Critical Mineral Study:A Global VAR AnalysisTable 8a.Average pairwise cross-sectional correlations of all variables and the associated models residuals.YitDpiteqit LevelsFirst differencesVECMX*residualsLevelsFirst differencesVECMX*residualsLevelsFirst differencesVECMX*residualsArgen
193、tina0.87400.11260.00190.23180.04040.02790.19020.2127-0.0224Australia0.90760.19620.06600.32150.10170.02630.78800.54710.0291Brazil0.90070.11100.02370.26770.0113-0.0220Canada0.90460.27130.00750.32100.07330.04930.74670.55880.0434China0.90960.1499-0.07290.16370.0549-0.0001Chile0.89950.16280.00690.36890.0
194、4700.02390.77220.33750.0207Euro0.89250.32250.01600.40700.12380.04350.76110.5725-0.1070India0.90330.18950.03260.16990.03410.01800.77110.3572-0.0072Indonesia0.89500.09840.01580.06510.04180.0454Iran0.8955-0.0661-0.07230.03810.04610.0377Japan0.84040.19610.02530.30830.06690.00950.32980.4431-0.0909South K
195、orea0.89070.15620.02830.33820.05820.02500.73240.3412-0.0491Malaysia0.90140.18270.00970.23930.12150.04880.63830.40100.0105Mexico0.90360.23270.03760.23850.01860.0210Norway0.89600.1283-0.00850.32050.06880.05180.80250.48930.0190New Zealand0.90080.22070.08130.30230.06560.02880.41250.3261-0.0124Peru0.8604
196、0.07400.02510.2444-0.0370-0.0065Philippines0.88490.14370.04690.21450.0212-0.00530.73400.3790-0.0001Russia0.38460.06840.00230.05090.0073-0.0063South Africa0.89590.28300.07550.35500.09040.04330.76960.46750.0562Saudi Arabia0.87350.06410.00950.00190.03920.0543Singapore0.90230.2488-0.00270.22630.06660.03
197、450.72550.5199-0.0068Sweden0.90450.22940.02770.41140.09610.04120.77510.4988-0.0079Switzerland0.90620.24900.00510.38500.09430.03350.78530.5271-0.0047Thailand0.88460.16220.00690.29890.09990.06150.68110.4296-0.0039Turkey0.90520.1601-0.00430.21380.00350.0288U.K.0.89820.28100.03080.40690.09700.03000.7397
198、0.5654-0.0062U.S.0.89990.2679-0.00470.33130.15180.05880.76980.54440.0019Venezuela-0.46590.10360.0140-0.0414-0.0533-0.0162Source:Internal KAPSARC calculations,2022.Empirical Results29World Oil and Critical Mineral Study:A Global VAR AnalysisTable 8b.Average pairwise cross-sectional correlations of al
199、l variables and the associated models residuals.epitritrlit LevelsFirst differencesVECMX*residualsLevelsFirst differencesVECMX*residualsLevelsFirst differencesVECMX*residualsArgentina0.42600.05590.00600.37170.03690.0378Australia0.79380.36030.23420.58660.11980.06820.75780.34000.0413Brazil0.73460.2296
200、0.14250.43400.03210.0276Canada0.77250.30560.17090.61980.17270.13520.76890.34120.0266China0.59620.08760.04660.55470.04440.0005Chile0.71900.27310.17840.59340.0291-0.0251Euro0.72690.31950.25870.63470.15210.07770.76740.4192-0.0442India0.62460.24970.18740.34940.07850.04250.45070.0865-0.0062Indonesia0.425
201、60.19910.10510.32570.07330.0647Iran-0.13600.03770.0537-0.6318-0.0217-0.0147-0.75310.0576-0.0141Japan0.66390.15260.13460.56950.03030.02050.72670.2365-0.0124South Korea0.76810.27920.15380.60760.04840.04680.73530.0604-0.0319Malaysia0.68430.29830.20030.51130.06020.0283Mexico0.66310.15990.03910.49460.020
202、7-0.0079Norway0.74280.37830.27010.59350.03670.01020.75010.2786-0.0158New Zealand0.79130.35140.24670.55960.04960.01190.71140.18400.0247Peru0.72490.05780.06270.42060.0294-0.0062Philippines0.76740.17890.16290.60910.09250.0442Russia0.68470.05650.00910.49430.06220.04940.73330.2616-0.0533South Africa0.644
203、60.31520.23990.54110.10620.05220.63850.16930.0181Saudi Arabia0.65360.05460.0403Singapore0.78650.36830.27230.56610.07650.0481Sweden0.69010.34030.26010.64680.07510.01420.78040.35870.0385Switzerland0.77910.27800.27350.55850.07310.02280.73310.35030.0164Thailand0.77570.29300.20160.59090.10870.0684Turkey0
204、.75850.15170.07710.40890.06580.0386U.K.0.72220.29000.18700.63820.12670.03780.77180.37180.0114U.S.0.57530.11510.04050.74370.3590-0.0443Venezuela0.2928-0.0185-0.0027-0.21000.04290.0516Source:Internal KAPSARC calculations,2022.Empirical Results30World Oil and Critical Mineral Study:A Global VAR Analysi
205、sthe U.S.and Russia to 0.17 in Iran and Norway.The level pairwise correlations for Saudi Arabian crude oil production are 0.07,falling slightly to 0.05 for first differences and 0.06 for the residuals.For the oil sector and,indeed,the GVAR model as a whole,the correlations calculated for the residua
206、ls are very small and do not appear to depend on the country or variable under consideration.This result supports the proposition that the model is successful in accounting for the factors behind the macroeconomies of the country-specific variables.The sole exceptions to this rule appear for the rea
207、l exchange rate,short-term interest rates,and equity price variables,where the cross-sectional pairwise correlations range from-0.00 to 0.27,from-0.03 to 0.13,and from-0.10 to 0.06,respectively.An investigation into the results for these variables presents an opportunity for future research,and shor
208、t-term interest rates are not included as a foreign variable in the base or reference case.In summary,there is clear evidence of cross-country correlations for all of the variables in the GVAR model.The extent of this correlation depends on the variable and is strongest for real GDP,equity prices,ex
209、change rates,and long-term interest rates.In the oil sector,the cross-country correlations are strongest for crude oil inventories.The effects taper off considerably when residuals are examined,suggesting that the model is successful in capturing the dependencies that exist across the macroeconomic
210、variables considered.Table 8c.Average pairwise cross-sectional correlations of all variables and the associated models residuals.Iit0Qsit0 LevelsFirst differencesVECMX*residualsLevelsFirst differencesVECMX*residualsCanada0.19310.22820.00200.02850.07250.0368China0.1075-0.0252-0.0444Euro0.14770.2619-0
211、.0872Indonesia-0.15760.02350.0232Iran0.1665-0.00300.0086Japan0.31730.30520.0152South Korea0.32000.0658-0.1754Mexico0.03620.04490.0451Norway0.16620.02960.0672Russia-0.26280.04440.0158South AfricaSaudi Arabia0.07240.05080.0634U.K.-0.52900.1426-0.0795-0.13780.06300.0648U.S.0.18240.37890.0020-0.26180.05
212、970.0618Venezuela0.05960.09160.0479Source:Internal KAPSARC calculations,2022.Empirical Results31World Oil and Critical Mineral Study:A Global VAR AnalysisCounterfactual analysis of price shocks:The impact of a positive shock to CM and world oil pricesThe GVAR model has been fully specified and can b
213、e utilized to study the time profile of the effects of shocks to the global system.To illustrate the dynamic properties of the model,we investigate the implications of positive shocks to the CM price and the world oil price.The shocks are analyzed by means of generalized impulse response functions(G
214、IRFs),which consider shocks to individual errors and integrate the effects of all“other”shocks“out,”using the observed distribution of all of the shocks(Koop et al.1996).The impulse response function is calculated from the moving average representation of the GVAR model and is the difference between
215、 the conditional and unconditional forecasts where the conditioning information set is the shock to the variable under consideration(see Appendix A).The stability of the GVAR system can be verified by eigenvalues.The model has 165 endogenous variables with 59 cointegrating relationships.Thus,at leas
216、t 165-59=106 must lie on the unit circle for the system to be stable.In fact,the system has 108 eigenvalues on the unit circle,and all of the remaining values are less than one,which suggests that the system as a whole is stable and that some shocks can be expected to have permanent effects on the e
217、ndogenous variables.The price shock scenarios are run without restrictions on the equations or parameters.Finally,the PPs utilized in the system are illustrated in Appendix A,and all of the countries and variables included in the analysis are listed in Appendix B.Table 9.Significance tests:Selected
218、results by country and region;positive shock to the CM index and Brent.CM price shockOil price shockVariableMedian cumulative changes%Significance*Median cumulative changes%Significance*1 year2 years1 year2 yearsCM price4.81%9.64%FOil price-1.31%-2.88%Inflation-US0.01%0.00%0.52%0.11%DInflation-Saudi
219、 Arabia0.04%0.02%0.06%0.09%Inflation-U.K.0.33%0.40%G0.24%-0.06%CInflation-China0.08%-0.18%0.24%0.25%AInflation-South Korea-0.27%-0.32%E0.27%0.28%BInflation-India0.06%0.19%0.24%0.41%GOil production-Saudi Arabia-0.91%-0.82%2.88%4.17%AOil production-U.S.0.29%0.48%-1.29%-0.20%Source:Internal KAPSARC cal
220、culations,2023.Notes:Median cumulative changes after one year in%,*refers to 90%confidence intervals.A-statistically significant in one quarter.B-statistically significant in two quarters.C-statistically significant in three quarters.D-statistically significant in four quarters.E-statistically signi
221、ficant in five quarters.F-statistically significant in six quarters.G-statistically significant for at least two years.Empirical Results32World Oil and Critical Mineral Study:A Global VAR AnalysisFor our counterfactual analysis sample,we consider a scenario with a one standard deviation positive sho
222、ck to CM prices.Given the period under consideration,1979Q1 to 2022Q2,the shock is roughly equivalent to a 5%increase in CM prices per quarter.While the implications for world oil prices and output are negligible and not statistically significant,the increase is clearly inflationary for most countri
223、es(see Figures 4.a,5.a and 6.a).As illustrated in Figure 5.a,the increase in CM prices results in a significant increase in inflation in most of the countries captured in the GVAR analysis.The impacts are stronger in the first and second quarters,dying out after 3 or 4 quarters.This general observat
224、ion holds for all countries except the United Kingdom and India,where the effects are permanent throughout the forecast period.The effects of the shock are strongest in the United Kingdom,where inflation increases by 0.3%-0.4%in the first two years of the forecast period.Prices fall slightly in Sout
225、h Korea,and this effect is statistically significant(see Table 9).An increase in CM prices has negligible implications for oil production,resulting in a slight increase in crude oil production in the United States of less than 0.5%in the first years after the forecast period.There is a larger effect
226、 in Saudi Arabia,where oil production falls by almost 1%per annum in Figure 4.a.The effects of a positive shock to CM prices on Brent.Note:Bootstrap median estimates with 90%bootstrap error bounds.Source:Internal KAPSARC calculations,2023.Empirical Results33World Oil and Critical Mineral Study:A Glo
227、bal VAR AnalysisFigure 5.a.The effects of a positive shock to CM prices on inflation in selected countries.Note:Bootstrap median estimates with 90%bootstrap error bounds.Source:Internal KAPSARC calculations,2023.the first two years following the CM price shock.These results suggest that the cross-pr
228、ice elasticity of oil with respect to CM prices is positive in the United States,where CMs and oil are substitutes,and negative for Saudi Arabia,where CMs and oil are complements.The results are not statistically significant but suggest that rising prices for CMs and the increased joint cost of CMs
229、and oil as inputs to renewable energy in some regions will reduce the quantity supplied relative to the base case.It is interesting to compare these results to a one standard deviation increase in the price of Brent,roughly equivalent to a 14%increase.The two Empirical Results34World Oil and Critica
230、l Mineral Study:A Global VAR Analysisshocks are definitely asymmetric.The shock to the world oil price has an immediate effect on CM prices,which rise steadily to 5%throughout the forecast period.The results are clearly statistically significant and permanent,lasting throughout the 44-quarter foreca
231、st period.As anticipated,the increase is clearly inflationary and has a strong positive effect on world oil production(see Figures 4.a,5.b,and 6.a).As illustrated in Figure 5.b,the shock to Brent results in a significant increase in inflation in most of the countries captured in the GVAR analysis.Th
232、e impacts are stronger in the first and second quarters Figure 5.b.The effects of a positive shock to oil prices on inflation in selected countries.Note:Bootstrap median estimates with 90%bootstrap error bounds.Source:Internal KAPSARC calculations,2023.Empirical Results35World Oil and Critical Miner
233、al Study:A Global VAR AnalysisFigure 6.a.The effects of a positive shock to CM prices on oil production in selected countries.Note:Bootstrap median estimates with 90%bootstrap error bounds.Source:Internal KAPSARC calculations 2023.for all of the countries and die out after the first two years in the
234、 United States,Saudi Arabia,and the United Kingdom.The effects of the shock are greatest in Brazil(not shown here),where inflation rises by 1.5%in the first quarter following the price shock to over 6%in the first year.The second largest effects are seen in the United States and the U.K.,where infla
235、tion rises to 0.52%and 0.24%,respectively,in the first year following the price shock.The effects on inflation are slightly lower in China,South Korea,and India,where the effects are permanent throughout the forecast period.The results are statistically significant for all of the countries under obs
236、ervation except Saudi Arabia.The results suggest that CMs are starting to have an impact on the world economy at the macro level,due in part to their crucial role in energy transitions and other industrial value chains.This impact is at least as diverse across countries as that of the current major
237、commodityoilperhaps even more so.This is likely due to the unequal distribution of CM reserves and the varying stages of sophistication of nations in the transition to low hydrocarbon or green energy.Moreover,the relationship between CM prices and oil prices is asymmetric.Geopolitical shocks to CM p
238、rices have a small effect on world oil prices,while changes in the price of Brent significantly impact the CM price index.Empirical Results36World Oil and Critical Mineral Study:A Global VAR AnalysisThe varying effect of CM price shocks across countries can be partially explained by differences in t
239、he economic structure and trade patterns.This observation also highlights potential differences in the distribution of CM reserves and relevant industry supply chains across countries:For some economies in our model,the specific commodities included in our CM price index may not be deemed“critical”a
240、ccording to prevailing definitions of CMs.From the policy perspective,this suggests that it is likely to be a“new”geopolitics of energy and redistribution of wealth across nations.In the current phase,there is no universal policy solution to alleviate potential CM price shocks that would be applicab
241、le to all countries(unlike,for example,strategic petroleum reserves and OPEC,in the case of oil).However,it is essential to develop a comprehensive CM strategy that covers the whole value chain and is tailored to a countrys existing and target industrial,economic,and trade parameters.Empirical Resul
242、ts37World Oil and Critical Mineral Study:A Global VAR AnalysisThe GVAR framework is specifically designed to account for the interaction between a large number of countries,each with its own specific political and legislative systems.This is a crucial consideration,as the effects of severe shocks an
243、d global imbalances cannot be contained to one country or region.These disruptions tend to have significant implications for many countries,the severity of which can vary based on the sophistication of capital markets and variations in trade flows.The study contributes to the existing body of GVAR l
244、iterature through(i)an extension of the period under estimation to include post-COVID-19 observations and(ii)an expansion of the traditional GVAR framework to investigate CM markets via the inclusion of a new variable representing CM prices.The GVAR model was estimated for 37 countries and 7 regions
245、 and subregions using quarterly data from 1979Q2 to 2022Q1.The model was tested for weak exogeneity of the global and country-specific variables and for the stability of the estimated parameters.At the 5%level,only 17 out of the 173 exogeneity tests are statistically significant.Considering the fact
246、 that if the weak exogeneity assumption was always valid,one could expect up to 5%of the tests to be rejected,i.e.,approximately 8.65,this result is not an unreasonable number.Overall,the results tend to support the treatment of foreign country-specific variables and the world oil price in the GVAR
247、model(see Appendix C).A key feature of the expanded GVAR model is that CM prices are modeled jointly with oil market and macroeconomic variables.This joint modeling makes it possible to investigate the effects of disruptions to CM markets on oil markets,GDP,inflation,interest rates,exchange rates,an
248、d equity prices worldwide.The complex interrelationships and co-movements of fiscal and monetary policy variables,such as short-term and long-term interest rates,exchange rates,country-specific oil production,inventories,and GDP,can be examined in a global context.Overall,the results tend to support
249、 the treatment of the foreign country-specific variables,the CM price index and world oil price in the GVAR model.While there is evidence of structural breaks,it would appear to indicate structural breaks in the error variances.These issues have been resolved by means of a robust standard error proc
250、edure when calculating the impacts of foreign variables and by using bootstrap measures with confidence intervals to estimate the impulse response functions.The new formulation offers world oil analysts and policy makers the ability to identify and analyze the implications of shocks to CM prices for
251、 individual nations.To illustrate the potential of the model,we perform counterfactual simulations of oil and CM price shocks and identify the implications for CM prices,oil prices,and country-specific inflation and oil production.A CM price shock has positive and statistically significant effects o
252、n inflation in the U.K.and South Korea but only a small effect on world oil prices.On the other hand,a shock to the Brent oil price has a positive and statistically significant effect on the CM price index.These scenario outputs indicate the suitability of the proposed GVAR specification for analyzi
253、ng the relationships between the oil and CM sectors,as well as policy analysis and the development of counterfactual CM price shock scenarios.Conclusion:Key Features of the Methodology and Suggestions for Future Research38World Oil and Critical Mineral Study:A Global VAR AnalysisThe model presented
254、in this study does not contain any restrictions on the equations or estimated parameters.Future studies involving the GVAR structure developed above might consider restrictions on cross-CM price elasticities,as well as the inclusion of more countries or variables.The list of potential restrictions t
255、hat have been utilized in simple VAR studies in the past includes restrictions capturing delays in restoring or increasing crude oil production capacity(Kilian 2009).Such a line of enquiry is relevant today,as it can capture the implications of tight oil markets and low inventory levels due to years
256、 of underinvestment in exploration and production capacity:1)A near vertical crude oil supply curve with a price elasticity of 0.025.2)The response of the real price of oil to a negative crude oil supply shock must be positive for at least 12 months.3)The response of real economic activity to a nega
257、tive supply shock must be negative for at least 12 months.4)Dynamic sign restrictions imposed on GDP shocks due to the implementation of economic and financial sanctions.To gain a deeper understanding of the relationships between CMs,world oil markets,and individual country-specific economic variabl
258、es,topics for future research might include the following:1)The expansion of the data set to include a wider range of countries,including leading CM producers and all OPEC+nations,specifically Iraq,New Caledonia,Gabon,Ghana,Ukraine,Kazakhstan,and Morocco.2)Modifications to the key variables utilized
259、 in the model for a higher frequency of data,such as monthly,weekly or even real-time data.Such modifications will help improve estimates of global inventory forecasts and real economic activity.These variables can be augmented significantly with geospatial estimates of aboveground storage,floating
260、storage,vessel tracking,financial flow analysis,nowcasts and news forecasts.3)The introduction of production data for key CMs for each country represented in the model.Introducing such data would expand the range of potential scenarios to modeling the impacts of production shocks,embargoes,and other
261、 supply chain disruptions.Conclusion39World Oil and Critical Mineral Study:A Global VAR AnalysisEndnotes1 We choose the price index derived from the U.S.Bureau of Labor Statistics based on the import price index of imports under Harmonized System(HS)code 28:inorganic chemicals and organic or inorgan
262、ic compounds of precious metals,of rare earth metals,of radioactive elements or of isotopes instead of a more global representation such as the Rare Earth Monthly Metals Index(MMI)due to numerous factors including the large U.S.share in the world CM market,the existence of a long and comprehensive h
263、istorical time series,and the ability to estimate the index quarterly starting from 1979Q2(see Appendix B).2 Intriguingly,the study shows that real GDP is a relatively unimportant variable compared to exchange rates,interest rates and the oil price.3 The ASEAN-4 countries include Indonesia,Malaysia,
264、Philippines,and Thailand.The NIE-4 include Singapore,Hong Kong,South Korea and Taiwan.4 Once fully specified,the model will provide a stylized representation of the global oil market,and it will have the potential to separate different types of innovative shocks such as political uncertainty,global
265、recessions,changes in interest rates and monetary policy,crude oil supply shocks,shocks to above-ground crude oil inventories(reflecting speculation concerning future levels of crude oil supply and demand),and shocks to CM prices(reflecting the rise of the renewable energy industry).5 The country-sp
266、ecific VAR*models include both domestic variables and foreign(*)variables,where the foreign variables are constructed as weighted averages of the domestic variables across the different countries(Smith and Galesi 2014).6 The results of the statistical tests are available upon request.40World Oil and
267、 Critical Mineral Study:A Global VAR AnalysisReferencesAkaike,Hirotugu.1981.“Likelihood of a Model and Information Criteria ScienceDirect.”Journal of Econometrics 16:314.https:/doi.org/10.1016/0304-4076(81)90071-3Aleksashenko,Sergei,1992.“Macroeconomic Stabilization in the Former Soviet Republics:Dr
268、eam or Reality?”Communist Economies and Economic Transformation 4(January 1):43967.https:/doi.org/10.1080/427734.Andrews,Donald W.K.,and Werner Ploberger.1994.“Optimal Tests When a Nuisance Parameter Is Present Only Under the Alternative.”Econometrica 62(6):13831414.https:/doi.org/10.2307
269、/2951753.Australian Government.2022.“Critical Minerals at Geoscience Australia.”Accessed November 7,2022.https:/www.ga.gov.au/scientific-topics/minerals/critical-mineralsBank of England.2022.“Bank of England Database.”https:/www.bankofengland.co.U.K./boeapps/database/.Bazilian,Morgan.2019.“The Miner
270、al and Metal Foundations of the Energy Transition:Congressional Testimony before the Committee on Energy and Natural Resources.”The Payne Institute for Public Policy,September 17.https:/www.energy.senate.gov/services/files/57CA4D91-142C-44A7-9CE4-583456056FDF.Bradhser,Keith.2010.“China Is Blocking M
271、inerals,Executives Say.”New York Times,September 24.Accessed October 20,2022.https:/ Fadejeva,and Martin Feldkircher.2018.“Spillovers from Euro Area Monetary Policy:A Focus on Emerging Europe.”Working Papers 2018/04,Latvijas Banka.Bartekov,Eva and Ren Kemp.2016.“National strategies for securing a st
272、able supply of rare earths in different world regions.”Resources Policy 59:153-164.https:/doi.org/10.1016/j.resourpol.2016.05.003Bettendorf,Timo.2017.“Investigating Global Imbalances:Empirical Evidence from a GVAR Approach.”Economic Modelling 64(August):20110.https:/doi.org/10.1016/j.econmod.2017.03
273、.033.Bloomberg.2022.Bloomberg terminal.Accessed July 4,2022.BP.2021.“Statistical Review of World Energy 2021.”https:/ Review of World Energy 2022.”60.Burton,Jason,2022.“U.S.Geological Survey Releases 2022 List of Critical Minerals.”February 22.https:/www.usgs.gov/news/national-news-release/us-geolog
274、ical-survey-releases-2022-list-critical-minerals.Burton,Jason.2023.“U.S.Geological Survey Releases 2022 List of Critical Minerals.”https:/www.usgs.gov/news/national-news-release/us-geological-survey-releases-2022-list-critical-minerals.Bussiere,Matthieu,Alexander Chudik,and Giulia Sestieri.2009.“Mod
275、elling Global Trade Flows:Results from a GVAR Model.”SSRN Electronic Journal.https:/doi.org/10.2139/ssrn.1456883.2009.“Modelling Global Trade Flows:Results from a GVAR Model.”SSRN Electronic Journal.https:/doi.org/10.2139/ssrn.1456883.CountryE.2022.“Brazil Central Bank key rates.”Accessed July 22,20
276、22.https:/ M.Hashem Pesaran.2016.“Theory and Practice of Gvar Modelling.”Journal of Economic Surveys 30(1):16597.https:/doi.org/10.1111/joes.12095.Chudik,Alexander,Vanessa Smith,and University of Cambridge.2013.“The GVAR Approach and the Dominance of the U.S.Economy.”Federal Reserve Bank of Dallas,G
277、lobalization and Monetary Policy Institute Working Papers 2013(136).https:/doi.org/10.24149/gwp136.CEIC.2022a.“Singapore Short Term Interest Rate,19952022.”Accessed July 3,2022.https:/ Oil and Critical Mineral Study:A Global VAR Analysis.2022b.“Russia Short Term Interest Rate,20002022.”Accessed July
278、 3,2022.https:/ Short Term Interest Rate,2005 2022.”Accessed July 3,2022.https:/ Real GDP Growth,19892022.”Accessed June 15,2022.https:/ Vank,Ewa Wanda Maruszewska,and Filip Bene.2021.“How economic indicators impact the EU internal demand for critical raw materials.”Resources Policy 74.https:/doi.or
279、g/10.1016/j.resourpol.2021.102417Considine,Jennifer,Abdullah Aldayel,and Emre Hatipoglu.2020.“World Oil and Inventory Study:A Global VAR Analysis.”KAPSARC Methodology Paper KS2020-MP04.doi:10.30573/KS2020-MP04.2022.“The Sensitivity of Oil Price Shocks to Preexisting Market Conditions:A GVAR Analysis
280、.”Journal of Commodity Markets 27(September):100225.https:/doi.org/10.1016/j.jcomm.2021.100225.Dallas Fed.2022.“Index of Global Real Economic Activity.”https:/www.dallasfed.org/research/igrea.aspx.Database on Indian Economy,Reserve Bank of India(DBIE-RBI).2022.“DBIE-RBI:Database of Indian Economy.”A
281、ccessed July 4,2022.https:/dbie.rbi.org.in/DBIE/dbie.rbi?site=publications.Dees,Stephane,M.Hashem Pesaran,L.Vanessa Smith,and Ron P.Smith.2014.“Constructing Multi-Country Rational Expectations Models.”Oxford Bulletin of Economics and Statistics 76(6):81240.https:/doi.org/10.1111/obes.12046.Dees,Step
282、hane,Filippo di Mauro,M.Hashem Pesaran,and L.Vanessa Smith.2007.“Exploring the International Linkages of the Euro Area:A Global VAR Analysis.”Journal of Applied Econometrics 22(1):138.https:/doi.org/10.1002/jae.932.Dungey,Mardi,Faisal Khan,and Mala Raghavan.2018.“International Trade and the Transmis
283、sion of Shocks:The Case of ASEAN-4 and NIE-4 Economies.”Economic Modelling 72(June):10921.https:/doi.org/10.1016/j.econmod.2018.01.008.Dou,Shiquan,Deyi Xu,Yongguang Zhu,and Rodney Keenan.2023.“Critical mineral sustainable supply:Challenges and governance.”Futures 146.https:/doi.org/10.1016/j.futures
284、.2023.103101Eggert,Roderick G.2011.“Minerals Go Critical.”Nature Chemistry 3:688-691.Eurostat,2022.“Statistics.”Accessed July 4,2022.https:/ec.europa.eu/eurostat/databrowser/view/IRT_LT_MCBY_Q_custom_3012440/default/table?lang=en.Fan,John Hua,Akihiro Omura,and Eduardo Roca.2022.“Geopolitics and rare
285、 earth metals.”European Journal of Political Economy.https:/doi.org/10.1016/j.ejpoleco.2022.102356Filatochev,Igor and Roy Bradshaw.1992.“The Soviet Hyperinflation:Its Origins and Impact throughout the Former Republics.”Soviet Studies 44(5):739-759.FRED.1995.“Organization for Economic Co-operation an
286、d Development Long-Term Government Bond Yields:10-Year:Main(Including Benchmark)for the United States.”Federal Reserve Bank of St.Louis.Federal Reserve Bank of St.Louis,January 1,1955.https:/fred.stlouisfed.org/series/IRLTLT01USM156N.2019.“Import Price Index(End Use):Agricultural Products Used for I
287、ndustrial Supplies and Materials.”Accessed June 20,2022.https:/fred.stlouisfed.org/series/IR120.2022.“Import Price Index(Harmonized System):Inorganic Chemicals;Organic or Inorganic Compounds of Precious Metals,of Rare-Earth Metals,of Radioactive Elements or of Isotopes(IP28).”Accessed June 20,2022.h
288、ttps:/fred.stlouisfed.org/series/IP28Foss,Michelle.2023.“Understanding the Minerals Wild West.”U.S.Embassy China Briefing,February 7.References42World Oil and Critical Mineral Study:A Global VAR AnalysisGalos,Krzysztof,Ewa Lewicka,Anna Burkowicz,Katarzyna Guzik,Alicja Kot-Niewiadomska,Jarosaw Kamyk,
289、and Jarosaw Szlugaj.2021.“Approach to identification and classification of the key,strategic and critical minerals important for the mineral security of Poland.”Resources Policy 70.https:/doi.org/10.1016/j.resourpol.2020.101900Greenwood-Nimmo,Matthew,Viet Hoang Nguyen,and Yongcheol Shin.2012.“Using
290、Global VAR Models for Scenario-Based Forecasting and Policy Analysis.”SSRN Electronic Journal,April.https:/doi.org/10.2139/ssrn.2026984.Gross,Marco.2013.“Estimating GVAR Weight Matrices,”Spatial Economic Analysis 14(2).https:/doi.org/10.1080/17421772.2019.1556800.Guliyev,Farid.2022.“The new global e
291、nergy order:shifting players,policies,and power dynamics.”Public Responses to Fossil Fuel Export.Exporting Energy and Emissions in a Time of Transition:25044.https:/doi.org/10.1016/B978-0-12-824046-5.00004-7Harbo,Ingrid,Sren Johansen,Bent Nielsen,Anders Rahbek.1998.“Asymptotic inference on cointegra
292、ting rank in partial systems.”Journal of Business&Economic Statistics 16(4)October:388-399 https:/doi.org/10.2307/1392608Hatipoglu,Emre,Jennifer Considine,and Abdullah AlDayel.2022.“Unintended Transnational Effects of Sanctions:A Global Vector Autoregression Simulation.”Defense and Peace Economics(2
293、022):1-17.https:/doi.org/10.1080/10242694.2022.2073429Hau,Liya,Huiming Zhu,Yang Yu,and Dongwei Yu.2022.“Time-frequency coherence and quantile causality between trade policy uncertainty and rare earth prices:Evidence from China and the U.S.”Resources Policy 75.https:/doi.org/10.1016/j.resourpol.2021.
294、102529He,Yujia.2018.“The trade-security nexus and U.S.policy making in critical minerals.”Resources Policy 59.https:/doi.org/10.1016/j.resourpol.2018.07.010Hensel,Nayantara D.2011.“Economic Challenges in the Clean Energy Supply Chain:The Market for Rare Earth Minerals and Other Critical Inputs.”Busi
295、ness Economics 46(3):17184.https:/doi.org/10.1057/be.2011.17.Hoyn,Kim.2021.“International Spillovers of Shocks and Economic Relationships:A Structural GVAR Approach.”University of Colorado Boulder.International Energy Agency(IEA).2022.“The 2020 EU Critical Raw Materials List.”October 26.https:/www.i
296、ea.org/policies/15274-the-2020-eu-critical-raw-materials-listIFPEN.2023.“Critical Metals and Rare Earths.”Accessed February 12,2023.https:/ Monetary Fund(IMF).2022a.“International Financial Statistics”Accessed June 16,2022.2022b.“IMF Primary Commodity Prices”Accessed June 16,2022.https:/www.imf.org/
297、en/Research/commodity-pricesJots,Marc,and Tovonony Razafindrabe.2017.“Oil Price Shocks and Welfare Social Consequences.”Accessed May 25,2017.https:/pdfs.semanticscholar.org/95f0/f924c306d1426278551a8beddb206c1254fa.pdf.Johansen S.1988.“Statistical analysis of cointegration vectors.”Journal of Econom
298、ic Dynamics 12:231254.doi:10.1016/0165-1889(88)90041-3.Keilhacker,Michael L.and Stefan Minner.2017.“Supply Chain Risk Management for Critical Commodities:A System Dynamics Model for the Case of the Rare Earth Elements.”Resources,Conservation and Recycling 125(October):34962.https:/doi.org/10.1016/j.
299、resconrec.2017.05.004.Kempa,Bernd and Nazmus Sadat Khan.2017.“Spillover Effects of Debt and Growth in the Euro Area:Evidence from a GVAR Model.”International Review of Economics&Finance 49(May):10211.https:/doi.org/10.1016/j.iref.2017.01.024.Khan,Nazmus Sadat.2020.“Spillover Effects of Trade Shocks
300、in the Central and Eastern European and Baltic Countries.”Journal of Economic Integration 35(1):3968.References43World Oil and Critical Mineral Study:A Global VAR AnalysisKilian,Lutz.2009.“Index of Global Real Economic Activity in Industrial Commodity Markets,”https:/www.dallasfed.org:443/research/i
301、grea.KITCO.2022.The worlds largest cobalt producing countries in 2021 report.February 2.https:/ N.,Panayotis G.Michaelides,Efthymios G.Tsionas,and Chrysanthi Minou.2015a.“System Estimation of GVAR with Two Dominants and Network Theory:Evidence for BRICs.”Economic Modelling 51(December):60416.https:/
302、doi.org/10.1016/j.econmod.2015.08.033.2015b.“System Estimation of GVAR with Two Dominants and Network Theory:Evidence for BRICs.”Economic Modelling 51(December):60416.https:/doi.org/10.1016/j.econmod.2015.08.033.Koop,Gary,M.Hashem Pesaran,and Simon M.Potter.1996.“Impulse Response Analysis in Non-Lin
303、ear Multivariate Models.”Journal of Econometrics 74(February):11947.https:/doi.org/10.1016/0304-4076(95)01753-4.Kwok,Chunyeung.2022.“Estimating Structural Shocks with the GVAR-DSGE Model:Pre-and Post-Pandemic.”Mathematics 10(10):1773.https:/doi.org/10.3390/math10101773.Liu,Boyu,Qi Zhang,Jiangfeng Li
304、u,Yawei Hao,Yanyan Tang,and Yaoming Li.2022.“The impacts of critical metal shortage on Chinas electric vehicle industry development and countermeasure policies.”Energy 248.https:/doi.org/10.1016/j.energy.2022.123646.Malaysia Department of Statistics.2022.“Department of Statistics Malaysia Official P
305、ortal.”Accessed June 15,2022.https:/www.dosm.gov.my/v1/index.php?r=column/ctimeseries&menu_id=bnk3bk0wTTkxOXVHaVg3SUFDMlBUUT09.Maral,Emerson Fernandes,Diogo de Prince,Beatrice Zimmermann,Giovanni Merlin,and Oscar Simes.2020.“Assessing Global Economic Activity Linkages:The Role Played by United State
306、s,Germany and China.”EconomiA 21(1):3856.https:/doi.org/10.1016/j.econ.2020.01.001.Maral,Emerson Fernandes,Beatrice Zimmermann,Diogo de Prince,and Giovanni Merlin.2018.“Assessing Interdependence Among Countries Fundamentals and Its Implications for Exchange Rate Misalignment Estimates:An Empirical E
307、xercise Based on GVAR.”Revista Brasileira de Economia 72(December):42950.https:/doi.org/10.5935/0034-7140.20180021.Matthew,Greenwood-Nimmo,Nguyen Viet Hoang,and Shin Yongcheol.2013.“Using Global VAR Models for Scenario-Based Forecasting and Policy Analysis.”In The GVAR Handbook:Structure and Applica
308、tions of a Macro Model of the Global Economy for Policy Analysis,edited by Filippo di Mauro and M.Hashem Pesaran.Oxford University Press.https:/doi.org/10.1093/acprof:oso/9780199670086.003.0007.McAdam,Peter,Kostas Mouratidis,Theodore Panagiotidis,and Georgios Papapanagagiotou.2022.“European Trade an
309、d Growth Imbalances:A Sign-Restriction GVAR Analysis.”SSRN Electronic Journal.https:/doi.org/10.2139/ssrn.4196463.Mining Technology.2022.“European Commission announces new critical minerals legislation.”September 15.https:/www.mining- M.Hashem Pesaran.2016a.“Country-Specific Oil Supply Shocks and th
310、e Global Economy:A Counterfactual Analysis.”Energy Economics 59(September):38299.https:/doi.org/10.1016/j.eneco.2016.08.007.2016b.“Oil Prices and the Global Economy:Is It Different This Time Around?”IMF Working Paper WP/16/210(November):29.Mohaddes,Kamiar,and Mehdi Raissi.2020.“Compilation,Revision
311、and Updating of the Global VAR(GVAR)Database,1979Q2-2019Q4.”Apollo-University of Cambridge Repository.doi:10.17863/CAM.56762.https:/doi.org/10.17863/CAM.27487.Mohaddes,Kamiar,Mehdi Raissi,and Niranjan Sarangi.2020.“Macroeconomic Effects of Global Shocks in the GCC:Evidence from Saudi Arabia.”ERF Wor
312、king Papers Series 1388(April).https:/erf.org.eg/app/uploads/2020/08/1598521492_843_582594_1388.pdf.References44World Oil and Critical Mineral Study:A Global VAR AnalysisNyblom,Jukaa.1989.“Testing for the Constancy of Parameters Over Time.”Journal of the American Statistical Association 84(405):2233
313、0.https:/doi.org/10.2307/2289867.OECD.2022a.“Interest Rates-Long-Term Interest Rates.”Accessed July 4,2022.http:/data.oecd.org/interest/long-term-interest-rates.htm.2022b.“Interest Rates-Short-Term Interest Rates Forecast.”http:/data.oecd.org/interest/short-term-interest-rates-forecast.htm.Park,Heon
314、,and Fuller,Wayne.2008.“Alternative estimators and unit root tests for the autoregressive process.”Journal of Time Series Analysis 16:415429.10.1111/j.1467-9892.1995.tb00243.x.Pesaran,M.Hashem,Til Schuermann,Bjorn-Jakob Treutler,and Scott M Weiner.2006.“Macroeconomic Dynamics and Credit Risk:A Globa
315、l Perspective.”Journal of Money,Credit,and Banking 38(5):121161.https:/doi.org/10.1353/mcb.2006.0074.Pesaran,M.Hashem,Til Schuermann,and Scott M Weiner.2004.“Modeling Regional Interdependencies Using a Global Error-Correcting Macroeconometric Model.”Journal of Business&Economic Statistics 22(2):1296
316、2.https:/doi.org/10.1198/0735009.Pesaran,Hashem,Smith,Vanessa and Smith,Ron,2007.“What if the U.K.or Sweden had joined the euro in 1999?An empirical evaluation using a Global VAR.”International Journal of Finance&Economics 12(1):5587.https:/doi.org/10.1002/ijfe.312.Pesaran,M.Hashem.2015.T
317、ime Series and Panel Data Econometrics.Oxford University Press.https:/doi.org/10.1093/acprof:oso/9780198736912.001.0001.Ploberger,Werner,and Walter Krmer.1992.“The Cusum Test with OLS Residuals.”Econometrica 60(2):271285.https:/doi.org/10.2307/2951597.Proelss,Juliane,Denis Schweizer,and Volker Seile
318、r.2020.“The economic importance of rare earth elements volatility forecasts.”International Review of Financial Analysis 71.https:/doi.org/10.1016/j.irfa.2019.01.010Reboredo,Juan C.,and Andrea Ugolini.2020.“Price Spillovers between Rare Earth Stocks and Financial Markets.”Resources Policy 66(June):10
319、1647.https:/doi.org/10.1016/j.resourpol.2020.101647.http:/.au/files/Macroeconomics_WTI_Crude_Oil_Forecasting_%20 Model.pdf.Reserve Bank of India(RBI).2022.Accessed July 4,2022.https:/www.rbi.org.in/.Salisu,Afees A.,Rangan Gupta,and Riza Demirer.2022.“The Financial U.S.Uncertainty Spillover Multiplie
320、r:Evidence from a GVAR Model.”International Finance.Accessed October 19,2022.https:/doi.org/10.1111/infi.12414.Shaikhmahmud,Abizer.2022.“Rare-Earth Metal Prices Will Skyrocket as Ukraine-Russia Tensions Continue.”Engineering and Technology,March 2.https:/eandt.theiet.org/content/articles/2022/03/rar
321、e-earth-metal-prices-will-skyrocket-as-ukraine-russia-tensions-continue/.Slavic Research Center.2022.Soviet Economic Statistical Series.Accessed July 22,2022.https:/src-h.slav.hokudai.ac.jp/database/SESS.htmlShleifer,Andrei,and Robert W.Vishny.1991.“Reversing the Soviet Economic Collapse,”Brookings
322、Papers on Economic Activity(2):341-360.Smith,L.Vanessa,and Galesi.2014.“Global VAR Toolbox 2.0.”August.https:/ Tarui,and Takashi Yamagata.2021.“Assessing the Impact of COVID-19 on Global Fossil Fuel Consumption and CO2 Emissions.”Energy Economics 97(May):105170.https:/doi.org/10.1016/j.eneco.2021.10
323、5170.Statistics Canada.2022.“Gross Domestic Product by Income and Expenditure:Interactive Tool.”June 1.https:/www150.statcan.gc.ca/n1/pub/71-607-x/71-607-x2021015-eng.htm.References45World Oil and Critical Mineral Study:A Global VAR AnalysisStock,James H.,and Mark Watson.2006.“Forecasting with Many
324、Predictors,”in:G.Elliott,C.Granger,and A.Timmermann(ed.),Handbook of Economic Forecasting,edition 1,volume 1,chapter 10,515-554.Boston:Elsevier.Sznajderska,Anna.2019.“The Role of China in the World Economy:Evidence from a Global VAR Model.”Applied Economics 51(15):157487.https:/doi.org/10.1080/00036
325、846.2018.1527464.T 2022.“Forex Trading A Simpler Way to Trade.”https:/ Economics.2022a.“Argentina GDP Growth Rate-2022 Data-2023 Forecast-1993-2021 Historical.”Accessed June 15,2022.https:/ Crude Oil Production.”Accessed June 15,2022.https:/ GDP Annual Growth Rate-2022 Data-2023 Forecast-1963-2021 H
326、istorical-Chart.”Accessed June 15,2022.https:/ Zealand GDP Growth Rate-2022 Data-2023 Forecast-1987-2021 Historical.”Accessed June 16,2022.https:/ of Labor Statistics.2022.“Import Price Index(Harmonized System):Inorganic Chemicals;Organic or Inorganic Compounds of Precious Metals,of Rare-Earth Metal
327、s,of Radioactive Elements or of Isotopes IP28.”Retrieved from FRED,Federal Reserve Bank of St.Louis November 15,2022.https:/fred.stlouisfed.org/series/IP28.U.S.Energy Information Administration(EIA).2022.“International Data.”Accessed June 15,2022.https:/www.eia.gov/international/data/worldUnited Sta
328、tes Geological Survey(USGS).2021.“Mineral Commodity Summary:Rare Earth.”Accessed November 16,2022.https:/pubs.usgs.gov/periodicals/mcs2022/mcs2022-rare-earths.pdfVargas,Mauricio,and Daniela Hess.2019.“The Caribbean and Its Linkages with the World:A GVAR Model Approach.”IMF Working Papers 256.https:/
329、doi.org/10.5089/9781498326766.001.Vekasi,Kristin.2019.“Politics,Markets,and Rare Commodities:Responses to Chinese Rare Earth Policy.”Japanese Journal of Political Science 20(1):220.https:/doi.org/10.1017/S00385.World Bank.2022.“World Development Indicators.”https:/databank.worldbank.org/s
330、ource/world-development-indicators.The White House.2022.“FACT SHEET:Securing a Made in America Supply Chain for Critical Minerals.”Press Briefing,February 22.https:/www.whitehouse.gov/briefing-room/statements-releases/2022/02/22/fact-sheet-securing-a-made-in-america-supply-chain-for-critical-mineral
331、s/Yu,Guihai,Chao Xiong,Jianxiong Xiao,Deyan He,and Gang Peng.2022.“Evolutionary analysis of the global rare earth trade networks.”Applied Mathematics and Computation 430.https:/doi.org/10.1016/j.amc.2022.127249Zahedi,Razieh,Asghar Shahmoradi,and Ali Taiebnia.2022.“The Ever-Evolving Trade Pattern:A G
332、lobal VAR Approach.”Empirical Economics 63(3):11931218.https:/doi.org/10.1007/s00181-021-02182-5.Zepf,Volker.2016.“An Overview of the Usefulness and Strategic Value of Rare Earth Metals.”In Rare Earths Industry,edited by Ismar Borges De Lima and Walter Leal Filho,317.Boston:Elsevier.https:/doi.org/1
333、0.1016/B978-0-12-802328-0.00001-2.References46World Oil and Critical Mineral Study:A Global VAR AnalysisGiven the country-specific models estimated individually and the world oil and CM price models,the combined GVAR model described above can be estimated as follows:Let=$&(A.1)Equation 8 can be written for each region as follows:111=+A zaa tA z A z uioitioiiitipitpitii(A.2)where:(),(,)1,=A I A for