THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES. August 2008 Edited: January Ambrose H Lee (Hin Fung Ambrose Lee)

Transcription

1 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Augus 2008 Edied: January 2009 Ambrose H Lee (Hin Fung Ambrose Lee) Deparmen of Economics San Francisco Sae Universiy 1600 Holloway Avenue San Francisco, CA ambrosel@sfsu.edu *This disseraion/hesis is approved by he faculy of San Francisco Sae Universiy in he parial fulfillmen of he requiremens for he degree: Maser of Ars in Economics a San Francisco Sae Universiy. Acknowledgemen: I hank Dr. Donald Mar for all he guidance and suppor hroughou his projec. And I offer my special hanks for Dr. Michael Bar, for eaching he models and helping me in consrucing he simulaion, giving me imely advice and asking criical quesions. I also hank Dr. Donald Mar, Dr. Sudip Chaopadhyay, Dr. Michael Bar, Dr. Enchuan Shao, Dr. Yanchun Zhang and Dr. Bey Blecha for heir valuable advices and helps. I would also like o hank Kim Neff, Aurash Alavi, David Pieper, Ashok Chelian, and oher economic graduae sudens in asking criical quesions abou he projec. Lasly, I hank Dr. Donald Mar and Kim Neff in helping me edi his work. Copyrigh by Hin Fung Ambrose Lee

2 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 2 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES (Ediion for he 2009 Pacific Conference for Developmen Economics) Absrac: Ambrose H Lee (Hin Fung Ambrose Lee) San Francisco Sae Universiy 2008 This paper is edied specifically for he 2009 Pacific Conference for Developmen Economics. Full version of he paper is locaed in he San Francisco Sae Universiy Library 1.. The surging demand of oil from China has aroused enormous concerns from he world; however, he lieraure in he field mosly focuses on oil price shocks o he Unied Saes. This paper aemps o answer he quesion ha has no been asked before, i.e. wha is he effec of oil price shocks o China s oupu? Is i he same as a developed counry such as he U.S.? This paper employs a Dynamic Sochasic General Equilibrium (DSGE) model o analyze he effec from he shocks. This DSGE model is an improved model over relaed lieraure, in ha he funcional forms ha he model uses allows calibraion of he parameers using daa, and do no need o borrow parameers from oher calibraed models. Using his DSGE model, his paper finds ha a 27% posiive oil price shock will reduce abou 1.6% of he real GDP of China. This paper also finds ha if he TFP shock is presen a he same ime, he TFP shocks will dominae he shock effecs o he economy. The paper also discovers ha an oil price shock affecs he U.S. and China differenly, his paper hen decomposes he differences beween U.S. and China in order o sudy he role ha economic growh and oil usage play in he differen responses of China and U.S. o an oil price shock. 1 J. Paul Leonard Library of San Francisco Sae Universiy

3 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 3 Inroducion $ per Barrel/CPI Figure 1: Inflaion Adjused Oil Price in U.S. 1969/01/ /01/ /01/ /01/ /01/ /01/ /01/ /01/ /01/ /01/ /01/01 Time Oil is one of he major resources in he world, and, as shown in he figure 1 above, he price of he oil has flucuaed since he 1970s. The effec of oil has always been considered as an imporan opic in economics research. The rich hisory in he economics lieraure abou oil price shocks, which has been developed since he mid-1970s, has shown ha each ime a developed economy experienced an oil shock, significan impacs occur. China, on he oher hand, being a mixed planned and marke economy, a rapidly emerging and developing counry wih a surging hunger of energy, has aroused enormous concerns from he world. While mos sudies focus on researching he effec of oil or an oil shock on he U.S., he quesion sill remains unanswered, i.e. wha is he exogenous effec of oil shock o China? Is his effec he same as U.S. experienced? Answering hese quesions would be useful in complemening he exising research on oil and he economy. To be more specific, his paper is aimed a answering he following quesion: Wha is he effec of an oil price shocks o China s oupu?

4 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 4 In order o answer he above quesion, a number of sub-quesions below need o be invesigaed, namely, Is he oil price shock exogenous o he Chinese economy? Is he effec differen for a developing counry vs. a developed counry? If here is a difference, wha facor(s) conribues he difference? Wha if a TFP shock is presen a he same ime wih oil price shock? This paper is srucured ino seven secions. The firs secion capures he key findings of relaed lieraure abou oil shocks, and provides a lieraure summary o direc our analysis. The second secion inroduces he model and he economic inuiion behind i. The hird secion describes and analyzes he economic srucure of China, discusses he issues of Chinese Macro daa, demonsraes her oil demand and oil share parameer hrough daa and generaes he momens for he calibraion. Secion Four calibraes he parameers ha are used in he esimaion. Secion Five simulaes he effec of boh an oil price shock and a TFP shock by assigning differen correlaions beween hem, his secion also provides a comparaive analysis of China and he Unied Saes and aemps o answer he differen response o an oil price shock beween China s and U.S s model by performing sensiiviy analysis on differen parameers. Secion Six discusses he srenghs and weakness of his paper, and he las secion concludes he findings and develops furher research quesions. This paper employs a Dynamic Sochasic General Equilibrium (DSGE) model o analyze he effec from he shocks. Our model is an improved model from previous lieraure, in ha he funcional forms ha his model uses allows his paper o calibrae he parameers using daa, and does no need o borrow parameers from oher calibraed models. Under his DSGE model, his paper esimaed ha a 27% increase in oil price one sandard error in hisorical prices leads o a decrease of 1.6% in Chinese oupu, i also able o capure he dynamic effec of

5 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 5 posiive oil price shock o oher endogenous variables such as consumpion, capial. Moreover, our model also allows he sudy of a join produciviy and oil price shock, which can be possibly correlaed. The paper also discovers ha an oil price shocks affecs he U.S. and China differenly, his paper hen decomposes he differences beween U.S. and China and sudies he role ha economic growh and oil usage play in he differen responses of China and U.S. o oil price shocks.

6 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 6 1. Relaed Lieraure The relaed lieraure, which mosly focuses on sudying he effecs on he U.S., ackles he quesion by 1) employing economerics/empirical models, or 2) using srucural micro-foundaions models. Table 1: Summary Table for relaed lieraure Mehodology Srucural micro-foundaions models Economerics/empirical models Examples Finn (2000) MMM 2 (2003) Mork(1989), Lee, Ni Rai (1995), Hooker (1999), Hamilon (2003), Kilian (2005) Model Calibraed DSGE Time series Economic Mechanism Main findings Oil is an inpu in he producions funcion. Price of oil is a sochasic variable, wih known disribuion. Oil price shocks affecs he inpu mix and produciviy Long erm effecs; Oil price shocks accoun for some percenage of he observed volailiy of real oupu Prices of oil (and perhaps is lags) ener as regressor in regression o Real GDP. Differen oil shocks variable are used and esed. The effec of oil price shocks o oupu is asymmeric and nonlinear. The majoriy of sudies use economeric models o invesigae he effec of an oil price shock o economy. These economeric models are ime series models such as a Vecor Auoregressive model (VAR) or an Auoregressive Disribued Lags (ADLs) model, where he price of oil, and perhaps is lags, eners as a regressor(s) in a regression using Real GDP as he dependen variable. The main findings from hese prior sudies are ha he effec of an oil price shock o 2 Carlos De Miguel, Balasar Manzano and José M Marín-Moreno (2003)

7 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 7 oupu is asymmeric and nonlinear, in oher words he relaionship beween oil price shocks and GDP is no simple and sraigh forward. As a resul hey conclude ha a simple Ordinary Leas Square (OLS) model is no a suiable analyic approach. They hen developed and esed differen insrumenal oil shock variable o overcome his issue. Using he models suggesed by Hamilon (2003) and Kilian (2005), he economeric models find ha he oil shocks and is lags coefficiens are eiher saisically insignifican or oo small o affec he real GDP of China 3. In order o undersand more explicily on how he oil price shock affecs he economy of China, our paper chooses o follow he srucural micro-foundaion approach. Pas sudies using similar calibraed DSGE model are Finn (2000) and MMM (2003), in which oil is an inpu in he producion funcion and he price of oil is a sochasic variable wih known disribuion. In oher words, an oil price shock affecs he inpu mix and produciviy of he counry. The main finding from Finn (2000) and MMM (2003) are ha, an oil price shock has a long-erm effec on he economy and also accouns for a large percenage of he observed volailiy of real oupu. Boh Finn (2000) and MMM (2003) use parameers from oher calibraed models. The main differences beween hese wo sudies are ha Finn (2000) invesigaes he effec of oil price shock on he U.S. economy, incorporaing a novel relaionship in beween energy and capial. This allows her o capure he effec of an oil price shock on oupu hrough is effec on capial. On he oher hand, MMM (2003) invesigaes he Spanish economy which includes he inernaional rade componen, impors and expors, in heir model. 2. Economic Inuiion behind he model Before a DSGE model is consruced, one needs o undersand how he oil price affecs he economy. This secion describes he economic inuiion behind such a model 3 For he resul of he economeric models, please refer o appendix I of he full version of his paper, which is locaed in he San Francisco Sae Universiy Library.

8 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee Oil Share Parameer Elasiciy of producion funcion The producion funcion relaes he inpus o he oupus in he macro-economy. As will be shown, undersanding he rae of change of oupu wih respec o he rae of change of oil, ha is he elasiciy of producion funcion wih respec o oil, is essenial for measuring he impac of oil price shock on he economy. This framework also gives a general picure on wha effecs should be expeced from oil supply disrupion owards he oupu of he economy. Referring o he framework for he effecs of energy disrupions suggesed in Hamilon (2005), a more generalized producion funcion can be developed as follows: Le he aggregae oupu be modeled in he following producion funcion. Y = F( x1, x2,, xn ) (1) Where Y is he oal oupu of he economy x i is he inpu i Based on his producion funcion, mahemaical proof 4 shows ha he oupu elasiciy of any producion funcion, under he perfec compeiion assumpion, is equal o he relaive spending on ha paricular inpu relaive o he oal revenue. In his paper, suppose ha xi is oil, denoed by( o ). Then he elasiciy of GDP wih respec o oil is equal o he spending on oil as a fracion of nominal GDP: p o o η o = (2) P Y 4 For he deails of his mahemaical proof, please refer o he secion 2.1 of he full version of his paper, which is locaed in he San Francisco Sae Universiy Library.

9 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 9 Where η o is he elasiciy of GDP w.r.. oil as an inpu facor po is he price of he oil o is he quaniy of oil P Y is he nominal GDP In oher words, he elasiciy of oupu wih respec o a given change in oil use can be inferred from he dollar share or he nominal value of he oil expendiures o he nominal oupu of he economy. Cobb-Douglas assumpion beween inpus Nex, mahemaical proof 5 also shows ha if he producion funcion is Cobb-Douglas, hen i implies ha he elasiciy of he oupu per inpu facor is fixed and equal o he inpu share parameerα for he producion funcion. Applying his equaion o he oil inpu equaion: i p o = (4) P Y o η o = α o Whereα is he oil share parameer in producion funcion. o For simpliciy and clariy, le he oil share parameer α 0 = ν hroughou his paper How does oil price affec he producion? Elasiciy of real GDP wih respec o price of an inpu i The elasiciy of real GDP wih respec o price of an inpu i can hen be derived in a similar way 6 Mahemaical proof 7 shows ha if we assume he producion funcion is Cobb-Douglas, hen 5 For he deails of his mahemaical proof, please refer o secion 2.1 of he full version of his paper, wha is locaed in he San Francisco Sae Universiy Library. 6 For he deails of he proof, please refer o appendix III of he full version of his paper, which is locaed in he San Francisco Sae Universiy Library. 7 For he deails of he proof please refer o secion 2.2 of he full version of his paper, which is locaed in he San

10 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 10 he elasiciy of oupu wih respec o he price of oil is a ransformaion from is oil share parameer: η po = α α η0 = η 1 o o 1 0 where η po is he elasiciy of oupu wih respec o he price of he oil These elasiciy equaions, however, only show he parial equilibrium model which neglecs he demand side of he economy. Therefore, in order o consider effec of oil price shock o he real GDP, a general equilibrium model needs o be consruced The Model Economy This secion consrucs a general dynamic sochasic general equilibrium (DSGE) model, under he Cobb-Douglas producion funcion and a perfec compeiion assumpion. These funcional forms allow us o calibrae he parameers using daa, insead of borrowing parameers from oher calibraed models. The consruced DSGE model can also capure boh Toal Produciviy Growh (TFP) shocks and oil price shocks as he shock variables ha will affec he model economy. Again, for simpliciy and clariy, le he oil share parameer α =ν 0 hroughou his paper Srucure Sochasic TFP shock Using he classic definiion, his paper assumes he TFP shock includes a deerminisic rend ( 1+ γ ) A and a sochasic par A = Le ( + γ ) : ln( A e A ) = z z1 1 z e 1, which follows a firs-order auoregressive process: = m + ρ z + ε 1, where : z1, 1 1 1, 1 1, and 1, ~... N 0, ε i i d 2 ( σ ) 1 Francisco Sae Universiy Library

11 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 11 A is he TFP (oal facor produc) a ime γ A is he growh rae of TFP. z 1, is based on an AR(1) auoregressive process Sochasic oil price shock Similar o oher models in he oil lieraure 8, his paper assumes he oil price shock variable p o be he relaive price of oil, and ha also follows a sochasic firs-order auoregressive process z 2, is based on an AR (1) auoregressive process where : p and : m 2, = e and z2, = m2 + ρ 2 z2, 2 z = ln(p ), ln p 0 = z 2, and 1 + ε ρ < 1 and ε 2 2, 2, ~ i. i. d. ( ) N 0, σ ε Social Planner s Problem The Social Planner s Problem becomes 9 [Feasibiliy]: c Where TFP shocks : z Oil price Shocks Where ε1, ε 2, { c, h, o, k } s.. : u = z is he ime variable + k + 1 = ( c,1 h ) θ 1 θ ν ν ( k h o ) + ( 1 δ ) ( c,1 h ) = [ α ln( c ) + ( 1 α ) ln(1 h )] 1, max 2, = m 1 = m 0 ~ N, 0 0 = 0 + p o + ρ z 2 E + ρ z 2 2 σ 1 2 σ 21 β u = A 1 1, 1 2, 1 + ε andε 1, + ε 2 σ 12 2 σ 2 2, 1, andε ~ i. i. d. 2, ~ i. i. d. k 2 ( σ 1 ) 2 N( 0, σ ) α is he percenage parameer ha household allocaes beween consumpion and leisure ime N 0, 2 8 For example MMM (2003) 9 The proof of he deerminisic NGM model o DSGE social planner s problem, please refer o appendix IV of he full version of his paper, which is locaed in he San Francisco Sae Universiy Library.

12 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 12 β is he discoun facor such ha ( 0 < β < 1) c is consumpion a ime k is he capial or physical capial per worker a ime δ is he consan depreciaion rae of capial w is he wage rae in uni of consumpion per uni of ime h is he labor uni per ime k = (1 σ ) k x is he law of moion of capial k > 0 0 represens ha iniial capial is posiive l h = 1 is he ime consrain for labor + o is he quaniy of oil a ime θ is he share parameer of physical capial per worker a ime ν is he share parameer 10 of oil a ime The Sufficien Equaions for he DSGE model Leu c ( 1 h ) = [ ln( c ) + ( 1 α ) ln(1 h )], α, he sufficien equaions 11 become, 1 α c α 1 h θ θ ν ν [ Labor ] : = ( 1 θ ν ) A k h o = ( 1 θ ν ) 1 1 θ 1 1 θ ν ν EE : βe ( θa k h o + 1 δ ) [ ] c = c+ 1 y h [Oil] : p = νa k θ h 1 θ ν ν 1 o y = ν o θ 1 θ ν ν [Feasibiliy] : c + k p o = A ( k h o ) + ( 1 δ ) k = y + ( 1 δ ) k Where : Labor Equaion for Opimal allocaion of ime 10 For clariy purposes, le he oil share parameerα 0 = ν. 11 For he proof of he DSGE model o is sufficien equaion, please refer o appendix V of he full version of his paper, which is locaed in he San Francisco Sae Universiy Library.

13 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 13 EE Oil Feas Euler Equaion Inverse Oil Demand Equaion Feasibiliy Consrain Solving he model numerically, using Time Derended Variables Simulaion echniques rely on he model o converge o a seady sae. In order o calibrae he parameers and o simulae he model wih observed daa, he model herefore, needs o be solved numerically. This can be done using he wo condiions. 12.Firs, if here is an one ime shock and ha is observed, no maer if i is a TFP shock or an oil price shock, he expeced value of is sochasic componen, given enough ime periods, will converge o he value E ( z z ) m = 1 ρ + k. Second, when he model is on he Balanced Growh Pah (BGP), he ime derended variables are all consan; in oher words, he variables are all in seady sae 13 all he inpu facors are growing a he same rae as oupu. Based on hese wo condiions, he seady sae equaions 14 for he DSGE Model are solved o he following: Seady Sae Equaions [ Labor] α cˆ α 1 hˆ 1 1 θ ν Aˆ k ˆ hˆ = θ θ ν ν : ( ) oˆ c EE : ˆ + 1 θ θ ν ν ( + γ ) = β ( θ ˆ ˆ 1 1 A k hˆ ˆ + 1 δ ) 1 o [ ] cˆ [Oil] : ˆ ˆ θ 1 θ ν ν 1 p ˆ 0 = νa k h o ˆ cˆ + kˆ γ + δ + p oˆ = Aˆ k ˆ hˆ oˆ θ 1 θ ν ν [Feasibiliy] : ( ) 0 12 For he mahemaical solving of he model, please refer o appendix of he full version of his paper, which is locaed in he San Francisco Sae Universiy Library. 13 The model would behave as he deerminisic model and approach a saionary soluion, if here is only a one ime shock and i is observed. 14 For he proof of he seady sae equaions, please refer o he appendix IX of he full version of his paper, which is locaed in he San Francisco Sae Universiy Library.

14 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 14 Where : Labor EE Oil Feas Equaion for Opimal allocaion of ime Euler Equaion price of oil a seady sae Feasibiliy Consrain Expeced relaionship The expeced relaionship beween an oil price shock and oher parameers The oil demand equaion is derived from he price of oil a seady sae equaion. o νa k h = p θ 1 θ ν 1 1 ν The oil price is inversely relaed wih he oil demand, which is as expeced by economic heory. When here is a posiive oil price shock, he demand for oil will have o decrease. The economic inuiion behind he oil demand equaion is ha people will consume less oil when here is a price increase in oil. This is a esable equaion. By aking logarihm o boh sides and forming a linear relaion 1 beween quaniy of oil and he price of oil, he coefficien of oil price is, which is 1 ν clearly negaive. Subsiuing he demand for oil ino he seady sae equaions generaes he analyical equaions ha show he relaionships of he oil price shock o he oher parameers. Relaionship wih oupu Subsiuing he oil demand ino he producion funcion gives: y = A ν 1 θ 1 θ ν ν 1 ν 1 ν 1 ν 1 ν k h p

15 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 15 This above equaion illusraes ha in his model, he oil price shock affecs he economy in a similar fashion o a negaive produciviy shocks, assuming ha boh shocks are independen, i.e. he correlaion beween wo shocks is zero. The economic inuiion is ha wo sides of he economy will be affeced when he oil price has a posiive shock. Firs, people are expeced o consume less in he fuure periods, which leads o a decrease in invesmen, and resuls in a decrease in he oal oupu of he economy. Second, as a posiive oil price shock appears, oil demand will decrease; and, since oil is an inpu facor for he producion, oal oupu will decrease immediaely. Then, as he economy regains confidence, consumpion and oupu will sar o converge o he seady sae, along wih he oil demand. Direc relaionship of oupu and oil price shock The relaionship of an oil price shock o oupu can be rewrien as: y = where : ~ A k ~ A h φ 1 φ = A 1 1 ν ν p ν 1 ν θ, φ = 1 ν Taking he logarihm for boh sides gives he direc relaionship: ln y ν 1 = lnν + ln A 1 ν 1 ν θ + ln k 1 ν 1 θ ν + ln h 1 ν ν ln p 1 ν The elasiciy of he oupu 15 ν wih respec o price of oil is. 1 ν Therefore, if he oil inensiy of a counry is higher, hen he prediced loss of GDP will be higher. A regression model may no accoun for hese dynamic effecs as oil prices increases because an economeric model does no specify his relaionship. The DSGE model, aking ino accoun he fac ha oil is one of he subsiues in producion, has he advanage over he 15 For he mahemaical proof, please refer o appendix III of he full version of his paper, which is locaed in he San Francisco Sae Universiy Library.

16 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 16 regression model ha i can predic hese dynamic effecs of oil shock. Relaionship wih Consumpion growh If he expecaion of c +1 is known, ha is when perfec foresigh is assumed, subsiuing he oil demand ino he Euler equaion gives: c 1 ν = β ν 1 1 θ 1+ ν 1 θ ν ν 1 ν 1 ν 1 ν 1 ν + 1 θ A + 1 k+ 1 h δ c p ν 1 ν The above analyical equaion shows ha he oil price shock variable in ime +1 is inversely relaed o he consumpion growh rae. I can be inerpreed as when here is a posiive oil price shock, hen he consumpion growh rae will be reduced, and will sar o converge back o is seady sae when he shock effec is over. The economic inuiion is ha as people experience a posiive oil price shock, hey will consume less and will coninue o consume much less in he fuure period before he price shock effec is over. When he price shock effec is over, he people sar o regain confidence, and he consumpion will sar o converge back in he nex ime period unil i reaches is seady sae. Relaionship wih he capial and he cos of capial Assuming ha a ime he shock is observed and here are no addiional shocks onward, he DSGE model would hen behave as a deerminisic model and approach a saionary soluion. The neoclassical growh model shows ha 16 he ren or cos of capial equaion under he Cobb Douglas assumpion is: r k = θy Where: r is he ren/cos of capial a ime 16 For he deails of he NGM model, please refer o appendix IV of he full version of his paper, which is locaed in he San Francisco Sae Universiy Library.

17 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 17 Since he oil price shock variable is inversely relaed o he oupu, he nominal value of he capial in he producion funcion will also be inversely relaed o oil price shocks. The economic inuiion is ha when he economy is experiencing a posiive oil price shock, oupu will fall and reduce he demand for capial. However, since oil and capial are assumed o be subsiues in he producion funcion, some capial will be used o subsiue for oil in producing goods. Also, by he assumpion ha he capial for he curren ime period is decided in he previous ime period, a posiive oil price shock will no have an immediae effec on he capial, bu capial in he nex ime period will decrease, since he oupu has decreased. When he shock is over and he people again regain heir confidence, and he oupu will converge back o he seady sae. Therefore, he demand for capial is driven up and he cos of capial increases afer he iniial shock. A some poin of his adjusmen ime period, he demand of capial and he cos of capial mus be higher han he seady sae before converging back, indicaing ha he economy is increasing is oupu, and moving back o he seady sae. Relaionship wih wage and working hours Turning o he labor marke effecs and referring o he NGM model 17, he wage equaion for labor is w h ( θ ν ) y = 1 Similar o he analysis in capial and is cos, when all oher inpu variables are kep consan, a posiive oil shock is inversely relaed o oupu; in oher words, i is inversely relaed o he nominal value of labor produciviy, as shown in he analyical funcion above. The economic inuiion is similar o he earlier discussion of he capial and is cos, as labor is also reaed as a perfec subsiue for capial in he producion funcion. As a resul, similar effecs o he capial and ren mus occur in response o a posiive oil shock. However since working hours (h) 17 For deails of he NGM model, please refer o appendix IV of he full version of his paper, wha is locaed in he San Francisco Sae Universiy Library.

18 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 18 is relaively invarian, he proporion and is ime period of convergence mus be smaller and shorer han he capial, and a large proporion of he oil price shock effec will fall o he relaively flexible variable in he labor marke, he wage. The wage is expeced o drop afer he shock effec and move back o he seady sae when he oupu converges. A some poin of his adjusmen ime period, he demand of labor and wage mus be higher han he seady sae before converging back, indicaing ha he economy is increasing is oupu, and moving back o he seady sae Relaionship wih invesmen x = y c The oil price shock affecs invesmen hrough oupu and consumpion. Since invesmen is defined o be he difference beween oupu and consumpion, when boh oupu and consumpion fall afer a posiive oil price shock, invesmen should also drop, converging back when oupu and consumpion regain he seady sae. The economic inuiion is ha when here is a posiive oil price shock, oupu and consumpion will fall, decreasing invesmen. This effec will remain unil he oil price shock effec on oupu and consumpion is over. Relaionship wih he TFP A Y = θ 1 θ ν ν K L o Since he oil price is no direcly relaed o he TFP, bu hrough he oil share in he oupu funcion, i is reasonable o assume ha a posiive oil price shock will no have or will have only a sligh effec on he TFP. A shor and one ime oil price shock should no affec he echnological growh of he economy.

19 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 19 Relaionship wih TFP shocks y = A θ 1 θ ν ν ( k h o ) Since a TFP shock is similar o a negaive oil price shock, one may expec ha if TFP shock and oil price shock are correlaed, posiively or negaively, and if boh shocks are presen in he same ime period, hey will eiher creae a bigger impac o he economy or i may eliminae he effecs of each oher. However, as shown in he equaion above, since TFP shocks affec oupu direcly, when a TFP shock is presen in he same ime period as he oil price shock, he TFP shock is expeced o have a larger effec and dominaes he oil price shock effec on he economy, no maer if hey are correlaed or no. 3. China This secion describes he economy of China wih special aenion o oil. Differences beween China and oher developed counries, such as Unied Saes, wih respec o oil will be highlighed. Lasly, his secion provides analysis and calculaes he daa needed o calibrae he parameers for he model suggesed in he previous secions Economic Srucure of China The People s Republic of China (China), which has one fifh of he populaion in he world, was esablished in 1949 as a cenrally planned economy. Reform o a more marke-oriened economy sared in 1978, shifing China from a planned economy o a mixed economy. Since joining he World Trade Organizaion (WTO) in 2001, China has rapidly developed her economic power, doubling is share of global manufacuring oupu. A he end of 2005, China s economy became he fourh larges in he world; however, wih is large populaion, is esimaed average GDP per person in 2006 is only abou US$2, In erms of inernaional 18 From IMF Ranking

20 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 20 rade, China is he world's larges consumer of seel, concree and copper and also he world's second larges imporer of peroleum. Finally, couning all producs, China is he hird larges imporer and he second larges exporer in he world Alhough here has been considerable economic growh, economic problems include her developing a modern financial sysem and problems accouned wih a relaively fixed exchange rae, an aging populaion, rapid environmenal degradaion, and large income dispariies because of lack of growh in rural areas. Also, he growing Chinese economy is showing sympoms of overheaing, and is expor-dependen srucure and valuaion of he Chinese Yua n are puing limis on he economy. FACTS ON CHINA THAT IS RELATED TO OIL Labor supply in China Before he economic reform in 1978, China s labor supply was rigid and inelasic under he social planning sysem. From 1978 unil he 1990s, he Chinese governmen conrolled he exchange of he residency from non-urban areas o urban areas. Even wihin urban areas, he governmen lierally employed all he workers for is sae-owned enerprises. For he 1990s onward, he labor supply in he urban area of China is considered o be elasic. Li and Zax (2003), sudying he marke for urban Chinese workers, and reached his conclusion in However, hey also find ha he elasiciy of labor o be relaively low compared o developed counries. Operaionally, his characerisic of China means ha he DSGM model should employ an elasic labor supply assumpion raher han an inelasic labor supply assumpion.

21 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 21 Rapid growh in China s oupu The rapid pace of economic growh in China may resul in a differen response o oil price shock compared wih oher developed counries wih a much lower growh rae such as he Unied Saes. Average real GDP growh per capia from he China s official daa 19, for he period , is abou 7.75%, compared o he U.S., which averaged 2% in he same period. 12 Figure 2: China Real GDP growh rae Chinese Saisical Year Book 2007

22 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 22 Surging demand of Oil Since he reform in 1978, China s oil demand has surged along wih her economic growh. This increase in demand of oil has increased he significance of he effec of oil shocks on China. Figure 3 based on BP Plc 20 Plc Saisical Repor 2007 daa illusraes he Chinese rapid increase in oil demand Figure 3: China Oil Impor in housand Barrels The U.S. was a large imporer of oil in China, on he oher hand, was a ne exporer of oil unil he lae 1980s. However, he rapid pace of economic growh led o a shif ino China becoming a ne imporer of oil. Should an oil price shock be reaed exogenously given China s surge of demand in oil? How does his surge of demand in oil affec is componen in China s Cobb Douglas producion funcion? The firs quesion is answered in he following secion, whereas he second quesion is answered in laer secion when esimaing he oil share parameer. 20 BP plc (LSE:BP, NYSE: BP, TYO: 5051, TSX: BP.U), formerly known as Briish Peroleum, is he world's hird larges global energy companies and is a mulinaional oil company ("oil major") wih headquarers in London, England, UK.

23 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee Exogeneiy of Oil Price o China China Oil demand vs. he world demand China has a relaively small demand for oil compared wih developed counries such as he Unied Saes. As shown in he figures 21 below, China in 2006 impored only 5% of he world oil producion, whereas he U.S. impored abou 17% of he world oil producion, riple he size of China s impor percenage. While exising lieraure using DSGE model assumes oil price shock as exogenous o U.S. This relaively small demand for oil versus he world producion makes he exogeneiy of he oil price shocks for China also valid. This assumpion of exogeneiy is imporan in ha i makes modeling easier. Price shocks considered o be endogenous o he Chinese economy would require a significanly more complex srucural model. 5.00% Figure 4: China Oil Impor as a percenage of World Oil Producion 4.00% 3.00% 2.00% 1.00% 0.00% -1.00% -2.00% Year From BP plc Saisical Repor 2007

24 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee % Figure 5: US oil impor as a percenage of World Producion 15.00% 10.00% 5.00% 0.00% Moreover, he world demand for oil does no show an enormous increase in he recen 10 years, even hough China may have conribued a cerain amoun o he increase in world consumpion on oil. In fac, as shown in he graph below, he world oil demand has a seady growh rae a abou 1.6% every year. Also, compared wih he flucuaion of oil price presened a he beginning of his paper, one can conclude ha he oil demand does no drive he flucuaion of oil prices, bu some oher facors, for example poliical evens, wars ec, are responsible. Since he demand of oil is no he mos significan facor affecing he oil prices, oil price effec can be considered exogenous o China s economy. Figure 6: World Oil Conumpion Growh Rae y = x R² = 0.99 World Oil Consumpion Growh Rae

25 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee Issue of China Macro Daa Sysemaic saisical reporing by he Naional Bureau Saisic of China (NBS) sared in1952. Because of he reform in 1978, he daa differs considerably for he years compared o daa afer Daa qualiy is also likely o be worse in a leas some of he pre-reform years. Besides daa issue semming from he reform, he Chinese macro daa sill has oher issues ha need o be addressed. Based on Holz (2003), (2006), who has done a deailed sudy of Chinese macro daa reliabiliy, here are four recurren daa issues for he macro daa in China, namely: he changes in secoral classificaion; he benchmark revision following he 2004 economic census; he muliude of daa ambiguiies if no inconsisencies; and he differen definiions for he fixed asses of China Because of hese daa issues, he daa used for simulaion as well as for he economeric analysis in he Appendix I 22 in his paper are handled wih special care. The assumpion in his paper is ha he laes daa ha comes from he NBS of China is he mos reliable daa. Oher Chinese macro daa are colleced from well recognized inernaional organizaion or research eam, such as he Inernaional Financial Saisic (IFS) of he Inernaional Moneary Fund (IMF), BP plc or he esimaed daa from Holz (2003) and Holz (2006) The oil share parameer of China Elasiciy of producion funcion w.r.. oil vs. w.r. he price of oil Based on he oil share equaion suggesed by he previous secions, ν = p o o, I calculaes P Y China s elasiciy of producion wih respec o oil Using daa from BP Plc, China s oil share 22 For he resul of he economeric models, please refer o appendix I of he full version of his paper, which is locaed in he San Francisco Sae Universiy Library.

26 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 26 over ime is shown below in figure 7: Figure 7: Oil Usage of China υ Elasiciy of Oil Usage in China in housand barrels 9.00% 8.00% 7.00% 6.00% 5.00% 4.00% 3.00% 2.00% 1.00% 0.00% Figure 7 shows ha China s dependence on oil or heir oil usage wih respec o oupu does no change much hrough , alhough here is a surge in he demand of oil in The surge of oil demand appears o be based more on heir high GDP growh, as he share of he oil in oupu does no increase as much. Moreover, since he graph appears o indicae no rend on oil usage in China, i is valid o assume ha he oil usage is saionary and he average of he oil usage can be used for furher analysis. Again, using daa from BP Plc, he U.S. elasiciy of producion funcion is ploed in he nex page. Similar o he Chinese oil usage, he oil usage of he U.S. does no change much hrough Elasiciy of Oil Usage in U.S.in housand barrels 9.00% 8.00% 7.00% 6.00% 5.00% 4.00% 3.00% 2.00% 1.00% 0.00% Figure 8: Oil Usage of US

27 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 27 Thus, China s oil share parameer is, on average 4.6% while he U.S. s oil share parameer is, on average, 3% Physical Capial in China Momens of he Chinese Macro Daa Before moving on o he calibraion of he parameers, he momens of he models mus be compued. The daa is colleced in he period from from he Chinese Saisical Book 2007, he mos updaed official daa from he NBS of China. To verify he accuracy, China s daa from IFS is also colleced and compued. Compued momens from he daa are as follows: Table 2: Momens from China Daa Parameer Esimaed Value Esimaed (Chinese Saisical Value Year book 2007) (IFS Daa) Y/C C/K Y/K The momens generaed from boh daa sources have a sligh difference afer 2 decimal places, bu overall, hese pairs of saisics can be considered as maching well wih each oher. However, while he oupu-consumpion raio (Y/C) is reasonable, he consumpion-capial raio (C/K) and he oupu-capial raio (Y/K) do no appear o be valid. Economic heory saes ha capial could only produce as much as is producion value, which means ha he oupu-capial raio and he consumpion-capial raio should be less han one. Since he daa from boh daa sources are similar, he unreasonable oupu-capial and consumpion-capial raios may herefore come from an incorrec esimae of he common facor, capial (k). According o Holz (2003), he fixed asse definiion of China is differen from he usual

28 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 28 definiion. In China s accouning sysem, he erm fixed asses is (guding zichan) denoed as he sum of he following: (i) (ii) ne fixed asses (guding zichan jingzhi); correcion o fixed asses (guding zichan qingli) due o, for example, he sale, damage, or decommissioning of he fixed asse; (iii) (iv) fixed asses under consrucion (zaijian gongcheng); and unresolved ne losses on fixed asses (dai chuli guding zichan jing sunshi). The firs erm, ne fixed asses, which is he larges componen in Chinese fixed asses, is officially defined as he difference beween he original value of fixed asses (guding zichan yuanzhi) and cumulaive depreciaion (leiji zhejiu). Holz (2003) claims ha hese non-producive or accouning arifac componens in China s definiion of fixed asses means ha he official arifac of fixed asses canno be consiued as a measure of he conribuion of physical capial o producion, as i will lead o an overesimaion of capial value. Therefore, he momens, he oupu-capial raio and he consumpion-oupu raio appear o be invalid in he daa, if he official daa is used. According o Holz (2006), he rue value of physical oupu canno be direcly compued because of wo complicaions from he original value of he fixed asses measures. Firs, only a few daa poins on fixed asses are available for specific enerprise groups of secors, using differen producion processes. Second, he series of revaluaions in 1990s 23, led an unknown mix of fixed asses valued eiher a nominal prices or real prices. Since he value of physical oupu canno be direcly obained from he official or he IFS daa, 23 According o Holz (2006), because of rapid inflaion in he lae 1980s and mid 1990s, in 1993 he governmen asked saed-owned enerprises (SOEs) o revalue all fixed asses purchased before 1991 o marke prices. Enerprises were allowed o spread he revaluaions over several years if hey could no afford o implemen hem immediaely.

29 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 29 his paper uses he esimaed economy-wide fixed asse values 24 in Holz (2003). Using hese values as measures of capial, he momens are compued as follows: Parameer Table 3: Correced Momens for China Esimaed Value (Chinese Saisical Year book 2007) Esimaed Value (IFS Daa) Y/C C/K Y/K By using Holz (2003) physical capial esimaes, he momens generaed from boh daa sources become reasonable and will be used for he calibraion of parameers. As a comparison, he momens of U.S. 25 are also compued and lised below: Table 4: Momens from U.S. Daa Parameer Esimaed Value Y/C C/K Y/K Calibraion of parameers 4.1. For China The daa used in calibraing he parameers for he ime period beween includes: he oil daa from he BP Plc Saisical Repor 2007, he laes official daa published by he NBS China in China Saisics Year Book 2007, he populaion daa from he China 24 This paper chose o use he preferred esimaes series of physical capial, The cumulaive approach wih non-sou effecive invesmen values in Holz (2003), Table 3, under B-C.3 approach. 25 U.S. daa is colleced from BEA Table 6.3 from he U.S. Deparmen of Commerce, Bureau of Economic Analysis (BEA) websie.

30 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 30 Populaion Saisical Yearbook for boh 1995 and 2005, he IFS daa series for China, and he labor share and working hours calculaed in Holz (2006). As menioned in he previous secion, his paper assumes ha he mos updaed version of daa published by he NBS China provides he mos accurae records for Chinese Macro daa. However, o check he validiy of he Chinese Daa, IFS daa is also colleced and compared. Calibraion For Chinese Parameer Based on he model, he parameerθ, γ, δ, β, andν, he mus be calibraed.

31 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 31 Table 5: Summary Table for calibraed parameers for China Technology Parameer Descripion Calibraed Value Official Daa IFS Daa γ Growh rae θ Capial Share θ ν Labor share ν Oil share δ Depreciaion rae (Calibraed) δ * Suggesed Depreciaion rae h Working hours per week 0.44 Oil Shock Variable z 2, = m2 + ρ 2z2, 1 + ε 2, m consan parameer ρ AR1 parameer σ 2 Sandard deviaion forε TFP Shock Variable z 1, = m1 + ρ 1z1, 1 + ε1, m Consan erm ρ ARI parameer σ 1 Sandard deviaion forε Preference parameer α Weigh on uiliy from consumpion β Discoun facor β * Discoun facor as δ = 4.6%

32 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 32 The oil share parameer (ν ) p o The oil share parameer (ν ) is calibraed by he equaion: ν = o. P Y Using eiher he IFS daa or he Chinese Official Daa, he calibraed average oil usageν of China for is The common growh rae ( γ ) The growh rae (γ ) is he average of he real GDP growh rae, deflaed by he Chinese GDP deflaor (Year 2000 = 100). The range of he average real GDP per capia growh rae, based on he IFS daa for , is The range of he average real GDP per capia growh rae, based on he official daa for is The Labor Share ( 1 θ ν ) The labor share in he producion funcion referred in he model is ( 1 θ ν ), based on able 31 and appendix 32-33, of Holz (2006). The average labor share for is calibraed as The Capial Share (θ ) The Capial share in he producion funcion, referred in he models is(θ ). Assuming a Cobb Douglas producion funcion, i is calculaed as 1 minus he labor share and he oil share. Tha is, ( 1 θ α ) α = θ 1. The calibraed capial share is The Depreciaion Rae ( δ ) o o The depreciaion rae δ is he average of he economy-wide depreciaion rae calculaed by using he equaion c k y y + γ + ν = + δ. The depreciaion rae δ calibraed from he IFS daa k k is , and he calibraed value from he official daa is However, according o Holz (2006), in he ime period from 1978 o 2003, a more realisic

33 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 33 value is The working hours ( h ) According o Holz (2006), he average working hour from he official average in he 1995, 1% Populaion Sample Survey is 40.7 hours per week. However, from he Populaion Saisical Year Book 2005, he working hours in 2001 o 2004 average is 45.4 hours per week. As a convenience, he average, 44 hours per week will be used. 44 Thus, h = = The Discoun facor ( β ) The discoun facor is calibraed by he feasibiliy equaion, β = 1+ γ ( θη + 1 δ ) yk Using IFS Daa and is calibraed depreciaion rae, he calibraed β is However, if deprecaion rae are se a 4.6%, hen he discoun facor calibraed is Using he Chinese Official Daa and is calibraed depreciaion rae, he discoun facor is If he depreciaion rae 4.6% as suggesed by Holz is used, hen he calibraed discoun facor is The Weigh on consumpion in Uiliy funcion (α ) 1 α y 1 h α c h The α is calculaed from he opimal labor equaion = ( 1 θ ν ) The calibraed α from he IFS Daa is The calibraed α from he Official Daa is Parameers in he sochasic equaions Oil price shock Following he assumpion in he previous secion, he oil price shock in AR(1) process where : z2, and : m 2 = m2 + ρ 2 z2, 1 = ln(p ) and 0 ρ + ε 2 2, < 1 and ε 2, ~ i. i. d. ( ) N 0, σ ε

34 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 34 The annual oil price daa is colleced from BP plc Saisical Repors 2007, and convered o Chinese Yuan, deflaed by Chinese GDP deflaor (2000 = 100). The calibraed parameers are m 2 = , ρ 2 = and TFP shock TFP shock in AR(1) Process σ ε = A = where : z ( + γ ) z1 1 A e 1, Le = m + ρ z 1 : ln( 1 1, 1 A ) = z1, + ε andε 1, 1, ~ i. i. d. 2 ( σ ) N 0, 1 The TFP daa is consruced from he IFS daa. The annual daa is convered o Chinese Yuan and hen deflaed by a GDP deflaor (2000 = 100). Also, because of is deerminisic componen, a ime-derended TFP variable has been consruced and used. The calibraed parameers are m 2= , ρ = and σ ε = As shown by he able 5, he official daa and he IFS daa esimae very close values, which provides a measure of robusness for he daa. As hey are quie close, he official daa is chosen for use in he simulaion of he model. One of he major puzzles in he daa is he depreciaion rae. The depreciaion rae calibraed from he China daa using he feasibiliy equaion, is abou 0.07 or 7%. However, as suggesed by Holz (2006), which has done a horough analysis and esimaed he depreciaion rae from he province and sae-owned enerprise daa of China, he average annual depreciaion rae should be 4.6%. Differen depreciaion raes will calibrae differen discoun facors β s, since he discoun facor from he calibraed depreciaion rae is very close o 1. Therefore, he more reasonable depreciaion rae 4.6% and is associaed bea are being used in he simulaion of he models in nex secion.

35 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee For he U.S. For he purposes of comparison and sensiiviy analysis, similar calibraions using he same mehods are done for he Unied Saes. Daa for he U.S. parameers come from he BEA 6.3 daa. Table 6: Summary Table for calibraed parameers for he Unied Saes Technology Descripion Calibraed Value Parameer γ Growh rae 0.02 θ Capial Share θ α o Labor share 0.62 ν Oil share 0.03 δ Depreciaion rae calibraed δ * Depreciaion rae suggesed 0.05 h Working hours per week 0.4 Preference parameer α Weigh on uiliy from consumpion β Discoun facor calibraed β * Discoun facor using δ = A similar problem found in he calibraion of China s parameers has he depreciaion calibraed by he model o be only 1.7% wih a discoun facor of The widely acceped depreciaion rae for he U.S. is 5%, wih an associaed discoun facor of For consisency, he 5% depreciaion rae and is accompanying discoun facor will herefore be used in he laer secions o generae simulaion and esimaion resuls. Comparing he wo counries parameers, he differences beween he parameers are he

36 THE EXOGENOUS OIL SHOCKS TO CHINA AND OTHER COUNTRIES Ambrose H Lee 36 growh rae of he counry, which leads o a differen discoun facor, he differen oil usage and he differenα, he weigh on uiliy for consumpion. I consider he following imporan quesions: Do hese differences resul in differen effec of oil shocks o he counries? How abou TFP shocks? 5. Resuls 5.1. Oil Price Shocks This secion esimaes he effec of an one ime oil price shock on he economy. The shock is equal in size o he sandard deviaion of pas price shocks, ε 2, = 0.27 calculaed previously. The suggesed depreciaion raes discussed earlier and heir respecive discoun facors are used for boh counries. Impulse Response Funcion for oil price shock for China The following secions describes he response of he differen endogenous variables of he model in he presence of a posiive oil price shock, which is an iniial increase of 27% in oil price wih respec o is seady sae value, using he parameers of China. When simulaion 26 is done for 100 periods and a one-ime posiive oil shock is inpued a he iniial ime period, he figure shows ha he TFP is no affeced, while he consumpion drops around 0.7% immediaely afer he shock is imposed. Consumpion coninues dropping in he firs few ime periods, hen i sars o converge back o he seady sae. I reaches is seady sae afer 60 ime periods. The working hours drops for 0.5% immediaely and sars o increase in he nex ime period. A ime period 10, he working hours is higher han he seady sae, reflecing he increase in labor as he economy converges back o he seady sae. 26 For deails of he impulse response funcion of he posiive oil price shock o China, please refer o he figure of he full version of his paper, which is locaed in he San Francisco Sae Universiy Library.