The impact of China's economic growth on crude oil price: Evidence from structural VAR

Hamendi, Ahmed; Law, Siong Hook Article The impact of China's economic growth on crude oil price: Evidence from structural VAR Review of Economic Analysis (REA) Provided in Cooperation with: International Centre for Economic Analysis (ICEA), Waterloo, Ontario Suggested Citation: Hamendi, Ahmed; Law, Siong Hook (2023) : The impact of China's economic growth on crude oil price: Evidence from structural VAR, Review of Economic Analysis (REA), ISSN 1973-3909, International Centre for Economic Analysis (ICEA), Waterloo (Ontario), Vol. 15, Iss. 3/4, pp. 237-252, https://doi.org/10.15353/rea.v15i3-4.4069 This Version is available at: https://hdl.handle.net/10419/328153 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc/4.0/ Review of Economic Analysis 15 (2023) 237-252 1973-3909/2023237 237 www.RofEA.org The Impact of China’s Economic Growth on Crude Oil Price: Evidence from Structural VAR Empty 15 AHMED TARIK HAMENDI Empty 15 SIONG HOOK LAW Universiti Putra Malaysia Empty 15 This paper examines the impact of Chinese economic growth on the real price of crude oil based on monthly time series data from 1992:01 to 2017:06 using structural vector autoregression (SVAR). The variables of the SVAR model are global crude oil production, index of global economic activity, China’s real GDP and real price of crude oil. Due to a break in the real price of oil series during the 2008 global financial crisis, the data is divided into two intervals. The results for the period prior to the 2008 crisis show that global demand shocks had a significant impact, while shocks from Chinese economic activity and global oil supply were insignificant. However, the results for the post 2008 period demonstrate that demand shocks of the Chinese economy have a significant but delayed impact, while global supply shocks have an immediate impact. The findings indicate a new regime after the 2008 crisis with a resurgence of a supply driven crude oil market structure that is influenced by Chinese economic performance. Keywords: crude oil market, China, SVAR JEL Classifications: Q31, Q32, Q41, Q43 I Introduction Historically, the crude oil market was controlled by an oligopoly of oil majors, even after nationalization of the oil industries in developing oil exporting countries and the subsequent formation of OPEC, the crude oil market continued to be driven from the supply side, allowing for stable oil prices throughout most of the 20th century (see figure 1), except for intermittent periods of supply side disruptions due to political and armed conflict in the middle east - the 1973 OPEC oil embargo and the first gulf war in 1980. However, the oil market dynamics changed in the mid 1980’s with new non-OPEC crude oil supplies entering the market, OPEC share of global crude oil production dropped from 51%  Corresponding author. Department of Economics, Faculty of Economics and Management. [email protected] © 2023 Ahmed Tarik Hamendi and Siong Hook Law. Licensed under the Creative Commons Attribution - Noncommercial 4.0 Licence (http://creativecommons.org/licenses/by-nc/4.0/. Available at http://rofea.org. Review of Economic Analysis 15 (2023) 213-228 238 www.RofEA.org in 1973 to 28% in 1985, which considerably weakened OPEC as an oil pricing cartel (Fattouh, 2011). Coupled with the decline in world demand for crude oil due to global economic recession in the 1980’s, led to a global crude oil market driven by supply and demand market fundamentals that continues to the present time (Kilian, 2009; Fattouh, 2011; Chevillon & Rifflart, 2009). The oil market is experiencing uncertainty with respect to crude oil price movement, this is especially worrisome since the oil industry is capital intensive and future returns on investment need to be established as accurately as possible. The price volatility prevalent in the crude oil market adversely affects market participant’s decisions, therefore, there is a strong need to understand the primary drivers of crude oil price to be better able to predict future oil price trends. This research will help shed new insight into the role of the Chinese economy as a determinant of oil price movement during this past twenty-five-year period. This study analyses the role of Chinese economic activity on crude oil prices over the last twenty five years (1992 - 2017) using a structural VAR model to determine whether there is causal relationship that can help understand future oil price movements. The studies so far have only covered the upsurge in oil prices, i.e., the periods modelled extend to 2014 only, and not the subsequent decline in price after 2014 as is the case in this study. In addition, the model uses monthly GDP series developed by Chang, Chen, Waggoner and Zha (2015) as it is a more accurate measure of the prevalent Chinese economic climate than industrial production index that has been used previously as a proxy of Chinese economic activity (Klotz, Lin & Hsu, 2014; Tian, 2016; Ratti & Vespignani, 2016). Figure 1. Historical crude oil prices from 1950 to 2016 Source: BP statistical review US$2016 HAMENDI, LAW China’s Growth and Oil Price 239 www.RofEA.org The main findings of this study are the prevalence of two different market structures governing the oil price movement, the pre-2008 period is characterized by a demand-driven market, while the post-2008 interval is identified as a supply-driven market that is influenced by Chinese economic performance. While the bulk of crude oil demand originates from developed countries, with OECD crude oil consumption accounting for 58% of world total in 2006, the increase in growth of oil consumption comes from emerging economies with a surge in consumption in China starting from the mid 1990’s leading to China’s imports of crude oil outstripping that of the US in 2015 (BP, 2017) as measured by consistent yearly average of total imports (see table 1). The emerging economies of Asia account for two thirds of the global growth in energy consumption with China accounting for a third of the global increase in demand to become the second largest crude oil consumer after the United States (Chevillon & Rifflart, 2009; Beirne et al., 2013). The compounding effect of continuous double-digit growth in China has had an impact on all commodities including crude oil, see figure 2, as can be seen in the price surge that started in 1998 (Beirne et al., 2013). The impact of a right shift in China’s import demand for a nonrenewable resource was found to raise the price of that commodity (Allen & Day, 2014) as increased oil demand is magnified due to the energy intensive nature of the exports sector that dominates the Chinese economic model (Kahrl & Roland-Holst, 2008). Table 1. China and USA share of global crude oil imports Country 1970 1980 1990 2000 2010 2016 China -0.01% -0.07% -0.07% 2.1% 7.2% 9.1% USA 7.5% 11.2% 12.1% 15.5% 13.1% 7.5% Source: BP statistical review. Note: imports calculated by subtracting consumption from production figures. Negative figures indicate production exceeds consumption. The introduction of vector auto-regression (VAR) provides a distinct advantage to traditional macroeconomic models, by incorporating the dynamics of multiple time series, it allows all the endogenous variables to be jointly examined (Sims, 1980). While the exact specification of VAR models is still open to debate (Beckers & Strom, 2015), it continues to be a popular method of studying the global oil market (Kilian 2009; Chevillon & Rifflart, 2009; Dua, He & Wei, 2010; Ratti & Vespignani, 2013; Klotz et al., 2014; Chen, Yu & Kelly, 2016; Tian, 2016; and Liu, Wang, Wu & Wu, 2016). Review of Economic Analysis 15 (2023) 213-228 240 www.RofEA.org Figure 2. China GDP annual growth (%) and Brent oil spot price (real 2010 USD) Source: World Bank Few empirical studies on crude oil price movement were conducted until the recent oil price volatility, this is probably due to the relatively stable and low oil prices prevalent during most of the 20th century, which led initial work to focus on modelling the supply side of the oil market (Hotelling, 1931 and Pindyck, 1978), the Hotelling model was limited by simplifying assumptions such as known stock of the resource and no cost reduction of extraction due to technology change (Krautkraemer, 1998; Gaudet, 2007). The surge in the oil prices after 2003 drew a lot interest from researchers to analyse the market and better understand oil price determinants but with mixed results. The role of the emerging economies was closely examined, and Chinese economic performance was found to be a primary factor in determining oil price movement (Li & Lin, 2011; Kilian & Hicks, 2013; Beirne, Beulen, Liu & Mirzai, 2013; Ratti & Vespignani, 2013; Klotz, Lin & Hsu, 2014; Ratti & Vespignani, 2016; Liu et al., 2016). On the other hand, many studies have reached the opposite conclusion, i.e., China has had either a secondary or no effect on oil prices (Du, He & Wei, 2010; Tian, 2016; Chen, Yu & Kelly, 2016; Cross & Nguyen, 2017). There is a clear gap in modelling the oil market over both the surge and downturn in crude oil prices, this period is of particular interest as it also coincides with a slowdown in the Chinese economy over the last few years. The rest of the paper is organized as follows. Section 2 describes the methodology and details the identification scheme of the SVAR model. Section 3 presents the results of the 0 20 40 60 80 100 120 0 2 4 6 8 10 12 14 16 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 GDP growth (%) Brent (real 2010 USD) Annual growth (%) Spot price (real 2010 USD) HAMENDI, LAW China’s Growth and Oil Price 241 www.RofEA.org SVAR models together with a detailed discussion of the findings. Section 4 concludes the study and provides policy implications of the findings. II Methodology Under the theoretical framework of supply and demand, this paper extends the structural VAR model proposed by Kilian (2009) by adding Chinese economic activity to model the crude oil market. Thus, the model identifies the determinants of oil price as four structural shocks: oil supply shocks, aggregate demand shocks, Chinese demand shocks and oil specific demand shocks, the latter representing precautionary demand. An increase in uncertainty regarding future oil supply was linked to increase in precautionary demand causing an increase in the price of oil for oil; the concept of precautionary demand was identified with the marginality of convenience yield of physical oil inventory (Kilian, 2009; Chevillon & Rifflart, 2009; Alquist & Kilian, 2010). The first step is to specify the model, next is to ensure that the model satisfies the requirements of stationarity and the absence of autocorrelation and heteroskedasticity as well as model stability before conducting analysis of the impulse response functions and forward error variance decomposition. The Granger causality of the variables on real price oil will be examined using the chi-square statistic of the Wald test with the null hypothesis of no Granger causality. The final step is to carry out robustness checks by changing the lag order (Rydland, 2011; Liu et al., 2016) to ensure that the results from the regression hold under alternative measures. The model assumes a vertical oil supply curve based on the unresponsiveness of supply in the short run to changes in oil price due to the capital-intensive nature of the oil industry (Kilian, 2009; Kilian, 2013). A supply shock by an exogenous shift to the right will have a negative relationship with crude oil price by increasing supply and thereby reducing the price. The three demand shocks on the other hand have a positive relationship with crude oil price, where increase in demand due to shift to the right of the demand curve will increase crude oil price. The proposed structural VAR model is expressed as follows: (1) where Xt is the vector of the four endogenous variables, α is the vector of the intercept terms, j is the optimal lag length and and εt is the vector of structural errors that are serially and mutually uncorrelated structural innovations with zero mean and variance-covariance matrix ∑ε. The vector Xt = (GOSt , GADt , CECt , RPOt), where GOSt is the change in global oil supply, GADt is an index of global real economic activity developed by Kilian (2009), CEGt is the change in Chinese economic activity (GDP) and RPOt is the change in real price of oil. tt j iit XCXC  ++= − = 1 1 0 Review of Economic Analysis 15 (2023) 213-228 242 www.RofEA.org The εt vector includes four structural shocks that affect the real price of oil, one supply shock from the global oil supply and three demand shocks from global aggregate demand, Chinese demand and oil specific demand. The structural model is transformed to reduced form by multiplying the above equation by C0-1 which has a recursive structure. The resultant reduced form vector of errors et is then as follows: et = Co-1 εt. In order for the VAR model to be estimated, a recursive identification scheme with zero short-run restrictions (Cholesky identification) are imposed on the model. The identification restrictions are based on economic theory and are imposed on the model as shown below, (2) The identifying restrictions are made on the following assumptions, global oil supply does not immediately respond to other structural shocks since the industry requires a lag time to adjust production due to the high capital costs needed to increase production, hence it only depends on lags of the other variables, therefore, in the matrix c12=c13=c14=0. Global aggregate demand responds to oil supply shocks as supply shortages can affect the world demand for commodities including oil (Kilian, 2009; Ratti & Vespignani, 2013), however, it will only respond to oil price and Chinese economic activity with a lag period due to (1) size and sluggish nature of the global economy; (2) the largely regional trade relations, hence the impact of a Chinese economic shock would be larger and more significant on its supply partner economies in Asia rather than on global and developed economies (Dinda, 2017); (3) the one-sided nature of the trade relationship, such that Chinese demand cannot have an immediate impact on global demand (Tian, 2016); therefore, c23=c24=0. On the other hand, Chinese economic growth responds to shortages in supply and global aggregate demand due to the export-oriented nature of the Chinese economy, and will only respond with a lag period to oil prices, so c34=0 (Kilian, 2009; Tian, 2016). The sample period of the study covers monthly data from January 1992 to June 2017, the period chosen coincides with both the rise of China’s economic activity and its ascendancy as a major participant of the global economy, as well as the deviation of oil price from their historical trends (see figure 1), this period will cover both the rise of oil prices after 2003 and includes the subsequent decline after 2014. Table 2 summarizes the variables description and the source of data. The global oil supply is represented by the change in cumulative world oil production obtained by differencing the logarithm. Global aggregate demand is given by the global real economic activity index developed by Kilian (2009), it is calculated based on dry cargo single voyage freight rate and                         = t t t t t RPO CEG GAD GOS ccc cc c XC 1 01 001 0001 434241 3231 21 0 HAMENDI, LAW China’s Growth and Oil Price 243 www.RofEA.org has become widely used in the literature as a proxy for cumulative world oil demand (Kilian, 2009; Ratti & Vespignani, 2013; Cross & Nguyen, 2017). Chinese economic growth is represented by the monthly growth of real Chinese GDP per capita obtained by differencing the logarithm. There has been a strong historical correlation between fractional change in oil consumption and fractional change in GDP (Brecha, 2013) and GDP data has frequently been used as a proxy for oil demand (Cross & Nguyen, 2017; Bierne et al., 2013; Kilian & Hicks, 2009). Since data indicates that lower income countries, i.e., emerging economies, have much higher energy intensities (Chevillon & Rifflart, 2009), GDP is taken as a proxy for oil demand in this paper. The GDP series used was developed by Chang, Chen, Waggoner and Zha (2015) and is updated by the Federal Reserve Bank of Atlanta. This GDP series was chosen for the following reasons: available as monthly data series; to avoid discrepancies between the production approach to calculate GDP adopted by China versus the aggregate expenditures approach used elsewhere, even though the World Bank have accepted official Chinese GDP figures since 1998 (Holz, 2014). Table 2. Summary of Variables Variable Description & Data source Global crude oil supply (GOS) World crude oil production in thousand bpd on monthly basis. EIA Energy Review (2017) Global aggregate demand (GAD) Compiled global business index on monthly basis. Kilian’s index at www.umich.edu Chinese economic growth (CEG) Real GDP per capita on monthly basis. Chang et al., 2015 Real price of oil (RPO) Real price of Brent crude oil on monthly basis. EIA (2017) retrieved from Federal Reserve Bank of St. Louis The measure of real price of oil is the monthly change in the spot real price of Brent obtained by differencing the logarithm. The Brent benchmark is used as a reference for almost two thirds of crude oil trading in the world and is therefore more relevant to this study than West Texas Intermediate (WTI). III Empirical Results and Discussion Table 3 presents the results of the Augmented Dickey-Fuller (ADF), Phillips-Perron (PP) and Kwiatkowski-Phillips-Schmidt-Shin (KPSS) tests. Global aggregate demand (Kilian’s index) is a compiled business cycle index that is built to be stationary, it is the only variable that is Review of Economic Analysis 15 (2023) 213-228 244 www.RofEA.org integrated of order zero I(0). The other three variables, global oil supply, Chinese economic growth and real price of oil are non-stationary at the level and integrated of order one I(1). Taking the first difference of the logarithm of these three variables will satisfy the stationarity assumption required to run SVAR analysis. The real price of oil has a break point at October 2008 that corresponds with the financial crisis of 2008. No such break point was found for the Chinese economic growth variable, as GDP per capita continued to grow throughout the financial crisis. Breakpoint testing of the other variables does not yield results that coincide with significant economic events and are thus discarded. The break point selection is done by minimizing the Dickey-Fuller t-statistic. As such, the SVAR model will be split into two time-frames, the first (SVAR I) will cover the period starting from January 1992 till September 2008 and the second (SVAR II) will cover the next period starting from October 2008 till June 2017. Table 3. Unit root test results 3.1 Structural VAR before the 2008 global financial crisis (SVAR I) SVAR I extends from 1992m1 till 2008m09. The lag length is initially selected to be three lags based on the Akaike Information Criterion (AIC), however, this lag length does not satisfy the requirement of no residual serial correlation. As such, the lag length is increased to lag 12 and tested, this lag length is recommended for monthly data and it also satisfies LM autocorrelation test. The roots of the AR characteristic polynomial equation are less than unity indicating that the model is stable. The impulse response functions of the real price of oil are shown in figure 3. An oil supply shock, in the first panel, does not have a significant impact on real price of crude oil and is of a ADF (SIC) PP KPSS GOS -0.614 (0) -0.435 2.044*** ∆GOS -14.758*** (1) -18.868*** 0.032 GAD -3.065** (2) -2.823* 0.316 ∆GAD -13.187*** (1) -13.733*** 0.064 CEG 4.729 (6) 12.007 2.043*** ∆CEG -14.152*** (2) -28.694*** 0.764*** RPO -1.993 (1) -1.821 1.352*** ∆RPO -14.308*** (0) -14.308*** 0.11 a a b *, ** and *** indicate 10%, 5% and 1% level of significance respectively a H : the series has a unit root. b H : the series is stationary. Barlett-Kernel used as spectral estimation method 0 o HAMENDI, LAW China’s Growth and Oil Price 251 www.RofEA.org help provide guidance on near-term oil price movement; (4) China's future energy policy shift towards renewable energy resources will have a large impact on future crude oil prices. References Allen, C. & Day, G. (2014). Depletion of non-renewable resources imported by China. China Economic Review, 30, 235-243. Alquist, R., & L. Kilian (2010). What do we learn from the price of crude oil futures?. Journal of Applied Econometrics, 25, 539 – 573. doi: 10.1002/jae.1159. Beckers, B. & Strom, S. (2015). Forecasting the nominal brent oil price with VARs - One model fits all? IMF WP/15/251. Beirne, J., Beulen, C., Liu, G., & Mirzaei, A. (2013). Global oil prices and the impact of China. China Economic Review, 27, 37-51. BP (2017). Statistical review of world energy. London: BP plc. Brecha, R.J. (2013). Ten Reasons to Take Peak Oil Seriously. Sustainability, 5, 664-694. doi: 10.3390/su5020664. Chang, C., Chen, K., Waggoner, D.F., Zha, T. (2015). Trends and cycles in China’s macroeconomy. NBER Macroeconomics Annual, 30, 1-84. doi: 10.1086/685949 Chen, Y., Yu, J., & Kelly, P. (2016). Does the China factor matter: What drives the surge of world crude oil prices?. The Social Science Journal, 53, 122–133. Chevillon, G., & Rifflart, C. (2009). Physical market determinants of the price of crude oil and the market premium. Energy Economics, 31, 537–549. Cross, J., & Nguyen, B.H. (2017). The relationship between global oil price shocks and China’s output: A time-varying analysis. Energy Economics, 62, 79–91. Dinda, S. (2017). Impact of the People’s Republic of China’s Slowdown on the Global Economy. ADBI WP/784. Dua, L., Hu, Y., & Wei, C. (2010). The relationship between oil price shocks and China’s macro-economy: An empirical analysis. Energy Policy, 38, 4142–4151. EIA (2017). Monthly Energy Review August 2017. US Energy Information Administration. http://www.eia.gov. EIA (2017). Crude Oil Prices: Brent. Retrieved from FRED, Federal Reserve Bank of St. Louis. https://fred.stlouisfed.org/series/MCOILBRENTEU. Fattouh, B. (2011). An anatomy of the crude oil pricing system. Oxford: The Oxford Institute for Energy Studies, ISBN 978-1-907555-20-6. Gaudet, G. (2007). Natural resource economics under the rule of Hotelling. Canadian Journal of Economics, 40(4), 1033–1059. Hamilton, J.D. (2009). Understanding crude oil prices. The Energy Journal, International Association for Energy Economics, 30(2), 179-206. Review of Economic Analysis 15 (2023) 213-228 252 www.RofEA.org Holz, C.A. (2014). The quality of China’s GDP statistics. China Economic Review, 30, 309338. Hotelling, H. (1931). The economics of exhaustible resources. The Journal of Political Economy, 39, 137-175. Kahrl, F., & Roland-Holst, D. (2008). Energy and exports in China. China Economic Review, 19, 649-658. Kilian, L., 2009. Not all oil price shocks are alike: Disentangling demand and supply shocks in the crude oil market. American Economic Review 99, 1053–1069. Kilian, L., 2013. Structural vector autoregressions. In: Handbook of Research Methods and Applications in Empirical Macroeconomics, Hashimzade N., Thornton M. (eds.). Edward Elgar: Cheltenham, UK, pp.515 – 554. Kilian, L., & Hicks, B. (2013). Did unexpectedly strong economic growth cause the oil price shock of 2003–2008? Journal of Forecasting, 32, 385–394. Kim, M. S. (2018). Impacts of supply and demand factors on declining oil prices. Energy, 155, 1059-1065. Klotz, P., Lin, T.C., & Hsu, S-H. (2014). Global commodity prices,economic activity and monetary policy: The relevance of China. Resources Policy, 42, 1–9. Krautkraemer, J. A. (1998). Nonrenewable resource scarcity. Journal of Economic Literature, 36(4), 2065–2107. Li, H., & Lin, S.X. (2011). Do emerging markets matter in the world oil pricing system? Evidence of imported crude by China and India. Energy Policy, 339, 4624–4630. Liu, L., Wang, Y., Wu, C. & Wu, W. (2016). Disentangling the determinants of real oil prices. Energy Economics, 56, 363–373. Pindyck, R. S. (1978). The optimal exploration and production of non-renewable resources. Journal of Political Economy, 86, 641-861. Ratti, R. A., & Vespignani, J. L. (2013). Liquidity and crude oil prices: China's influence over 1996–2011. Economic Modelling, 33, 517–525. Ratti, R. A., & Vespignani, J. L. (2016). Oil prices and global factor macroeconomic variables. Energy Economics, 59, 198–212. Rydland, S.E. (2011). An empirical analysis of how oil price changes influence the Norwegian economy (master’s thesis). Norwegian School of Economics, Bergen, Norway. Sims, C. A. (1980). Macroeconomics and reality. Econometrica, 48(1), 1-48. Tian, H. (2016). The role of China’s real economic activity in oil price fluctuations. Procedia - Social and Behavioral Sciences, 220, 522 – 530.