Tehnial Journal of Engineering and Applied Sienes Available online at www.tjeas.om 2013 TJEAS Journal-2013-3-17/2118-2122 ISSN 2051-0853 2013 TJEAS The effet of oil prie shoks on eonomi growth (Case Study; Seleted Oil Exporting Countries) Mohmmadreza Monjazeb *, Ali Souri and Zahra Shahabi Faulty of Engineering Eonomis, University of Eonomi Sienes Corresponding Author: Mohmmadreza Monjazeb ABSTRACT: This paper examines the impat of oil prie shoks on eonomi growth in oil-exporting ountries. So in this regard, we extrated the Petroleum Exporting Countries Data. Variables of the model are GDP growth, employment ratio, gross apital formation, the atual prie and the real prie of oil and gas. In this paper we use a panel data regression model with fixed and random method. An annual data for the period 1990 to 2009, for 26 oil exporting ountries, inluding Canada, Frane, Iran, Italy, Kazakhstan, Norway, Singapore, Britain, Ameria, Venezuela, Malaysia, Netherlands, Australia, Germany, Indonesia, Bangladesh, Belgium, Japan, Brazil, Kenya, Denmark, Mexio, Egypt, Sweden, India and Thailand have been olleted. The results show that the positive shoks of oil prie has a positive effet and the negative shoks of oil prie has a negative effet on the GDP growth of oil exporting ountries. Keywords: Oil shok, eonomi growth, panel data INTRODUCTION Oil prie shoks has been onsidered by eonomists due to signifiant impat on maroeonomi variables. The world eonomy has experiened various positive and negative flutuations in the oil prie. These flutuations and hanges in world oil pries have affeted on the maroeonomi variables and seriously have hallenged the eonomi situations of these ountries and fored them to onsider alternatives in order to feel seure about the negative effets of these shoks so that the Petroleum Exporting Countries whih are highly vulnerable to negative shoks of oil pries, have established institutions for storing exess foreign exhange revenues from sales of rude oil at high pries to make use it at the time of inidene of negative shoks to oil pries for their own purposes. The first part of this paper is devoted to a review of the literature. In the seond part of the researh methodology is explained. The third setion is devoted to the desription of the estimation proedure and the fourth setion provides a summary and onlusions. A Review of Literatures A study onduted by Hamilton (1983), entitled "Oil and the maro eonomy sine World War II" an be taken as a first study about the effet of inreasing of real oil prie on real inome. Hamilton obtained a meaningful relationship (1948-1972) and (1973-1980) between oil prie hanges and real GDP growth for Ameria's eonomy (sine 1948 to 1980). The final result of this study shows a unidiretional ausality relationship from the produed oil pries toward prodution. Olomola and adejuma (2006) an artile entitled "Oil prie shoks and maroeonomi ativity in Nigeria" Using quarterly data, have analyzed the effet of hanges in oil pries on output, inflation, exhange rate and money supply. Nigeria has been onsidered in the study period 1970-2003.The results that have been obtained using Vetor Auto Regression shows that the impat of oil prie hanges on inflation does not produe a signifiant effet on the exhange rate. Mork (1989) examined the impat of oil pries examined the impat of oil pries on GDP in Ameria based on Granger ausality method deal. Results show that oil pries have a negative impat on GDP growth in Ameria. However, the magnitude of the effet of oil pries on prodution, is less than the impat of rising oil pries. Thus, the impat of prie hanges on output is not symmetrial. Mork and et al (1994)have examined orrelation between oil prie movements and flutuations of GDP in the seven ountries; Ameria, Canada, Japan, Germany, Frane, UK and Norway. They believe that oil pries signifiantly have negative impat on the GDP, whereas this does not apply in the ase of Norway. But on the other hand, oil pries are expressly inreased only in Ameria and Canada's GDP. Farzanegan and Mark Ward (2009) examined the dynami relationship between oil prie shoks and major maroeonomi variables in the VAR method began. They found a positive relationship between the hanges in oil pries and the growth of industrial prodution and the marginal effet of oil prie
flutuations on the atual expenditures have been identified.tang et al (2010) responded to the question of what inreases the effet of oil prie shoks on China's eonomy and the results showed that the impat of rising oil pries on output and investment, and inflation and interest rates has a positive effet. And the impat of short-term effet is stronger. Looney (1990) examines the relationship between oil revenues and the Duth Disease in Saudi Arabia, he supports the differene between the growth setors of the eonomy in this ountry. Sahs and Rodriguez (1999) argue that ountries with natural resoures ompared to ountries that are poor in natural resoures tend to have higher GDP per apita. They introdue a fator of prodution (suh as oil) whih in a rypti growth pattern, expand slower than the other fators of prodution, suh as labor and apital. They show that the eonomy demonstrates the error effet or more than symbolizing. In this onditions, eonomi goes beyond of its equilibrium level in short run and then return to the equilibrium onditions due to redution in the growth of the eonomy. Rodriguez and Sahs(1999) use a dynami general equilibrium model whih shows that negative growth in Venezuela in the period 1972-1993 an be explained by their proposed model. METHODOLOGY In addition to apital and labor inputs, energy is onsidered as one of the leading setors of the maroeonomis.so, the prodution will be a funtion of labor, apital and energy, then we have: Q= F (K, L, E) (1) In this equation, Q is gross produt, K is apital input, L is labor input and E is energy input. So the three inputs of labor, apital and energy level hanges are produed. Energy onsuming whih inludes different arriers supply in energy suh as, oil, gas, oal and eletriity, is inverse funtion of the energy arriers prie. In other words, inreasing energy pries leads to a redution of energy onsumption and onsequently leads to redution of prodution level. F= E (Po, Pg, P, Pe) In the above equation, Po is oil prie, Pg is gas prie, P is oal prie, Pe is eletriity prie. To show the diret and indiret effets of hanges in energy pries, the gross domesti produt, energy osts are deduted, net produt is obtained as follows: Y= Q- Pe.E (2) In the above equation, Peis the energy prie. By using equations (1) and (2) and having MPL=P, we will have: DlnY/DlnPE= [PK.K/Y].DlnK/DlnPE+ [PL.L/Y]. DlnL/DlnPE- [PE.E/Y] (3) In this regard, PL and PK are the pries of apital and labor. So the effet of a hange in the prie of energy is explained by eah agent's share of the ost and the effet of energy pries on the replaement value of labor and apital used in prodution. The right side of equation 3 shows three ways that an energy prie shok may affet the eonomy. The third term, suggesting a diret effet of the oil prie and both the first term and seond term refers to an indiret effets of energy pries. Model Seletion Model is as below: Gdpg= f(, gf,, rop) Whih Gdpgis GDP growth, is hange of the ratio of employment to population (higher than 15 years), gf is gross fixed apital formation, is hange of real prie of gas, rop is real prie of oil. Data of gdp growth are extrated from World Bank and data of oil pries and gas pries are extrated from eia.gov website. Data has been olleted by an annual basis in the period 1990 to 2009 for 26 ountries. ESTIMATION RESULTS The results of unit root test of variables are reported in Table1. Test type variable Table 1. Levin, Lin and Chu stationary test Stationary test results First order differenes results Stat- value prob Stat- value DPG -2.98443 0.0014-17.5459 MP -5.41805-13.5490 C -4.62021-15.1282 RGP L -16.4669-15.4081 OP 1.42978 0.9236-2.39204 0.0084 As you an see the real oil prie is non-stationary, but the first differene of oil prie is stationary. Then we estimate the initial estimation. Then we use F Limer test for hoosing the best model (between Pooling or 2119
panel data) whih is reported in the table2. F- limer test Cross-setion f Cross-setion hi-square Table 2. F Limer test result statistis 7.141511 (25,434) 159.878660 25. Aording to table 2, the null hypothesis is rejeted and the panel model is aepted. Hausman test Cross-setion random Table 3. Hausman test result Chi-sq. statistis Chi-sq.d.f 13.185526. 0.0104 Hausman test indiates that the preferred model is a fixed effets model, so the fixed effets model is seleted. Table 4. fixed effets model variables gf -1.865397 76.06531 0.273006 0.052713 2.738232-3.963510 0.001121 28.42564 0.462233 14.93359 0.0001 0.9991 0.6441 R-squared= 0.77 adjusted R-squared= 0.76 F-stattistis=50.71021 prob (F-statistis)= 00 Durbin-watson stat= 1.71 statistis oefiient t-statistis Fixed effet In this step, we enter dummy variables related to oil prie shoks. Aording to The mean and standard deviation, the real prie of oil based on interval (0.17, 0.55) is determined. Aordingly, in the years 2006, 2007, 2008 and 2009 ourred a positive shok and in year 1998 ourred a negative shok whih show them as dp and dn in the model. Dependent variable: gdpg Method: panel least squared variable oeffiient 3.915331 0.020948 gf 0.269637 33835.98-6.079382 Drop* dp 2.765495 Drop* dn -6.727826 r- squared=0.79 durbin- Watson=1.67 Table 5. estimations results Std- error t- statistis -0.399672 9.796353 0.112463 0.186267 0.009379 28.74907 70062.94 0.482937 1.497961-4.058438 1.022415 2.704867 1.521380-4.422186 f- statistis=50.89243 prob (f- statistis)= 00 0.8523 0.6294 0.0001 0.0071 We will ontinue to investigate other onditions of an aurate model. In the model, Durbin-Watson statisti is equal to 1.67, whih shows the model has autoorrelation so we should solve this problem from the model (by using AR(1) in model). LM Statisti was used for heteroskedastiity test and the result is as below: LM=555.131> 38.8852= χ2 26, 0.05 The result shows that the model has the problem of heteroskedastiity. So for solving this problem we should estimate the model by GLS method. The result of final model is as below: 2120
Dependent variable: gdpg Method: panel least squared variable oeffiient gf Drop* dp Drop* dn Ar(1) R- squared= 0.84 Durbin-Watson= 1.93 Table 6. GLS method Std- error t- statistis 3.894738 0.273340 14.24869 0.199968 0.089696 2.229388 0.250023 0.009544 26.19790 23263.35 43652.61 0.532920-5.738132 0.997508-5.752466 2.253389 0.678526 3.321006-4.112883 0.947725-4.339742 0.113803 0.052136 2.182807 F- statistis= 68.80812 prob (F- statistis)= 00 0.0263 0.5944 0.0010 0.0296 The J-Bof normality test results indiates that the residual terms have a normal distribution. 50 40 Figure 1. standardized residuals Series: Standardized Residuals Sample 1993 2009 Observations 438 30 20 10 0-6 -5-4 -3-2 -1 0 1 2 3 4 5 Mean -0.351844 Median -0.350416 Maximum 5.127397 Minimum -5.726023 Std. Dev. 2.029398 Skewness 0.008963 Kurtosis 2.831466 Jarque-Bera 0.524231 ability 0.769422 The Interpretation of Final Model Interpretation of R2 (oeffiient of determination): the oeffiient of determination R2 with the value of0/84 indiated that 84 perent of hanges of GDP growth of oil exporting ountries is explained by hange of the ratio of employment to population (higher than 15 years), gross fixed apital formation, hange of real prie of gas, real oil prie, labor produtivity, and dummy variable. Interpretation of oeffiient of variable (ratio of employment): It means that if inreases about 100 units, then Y (GDP growth rate) will inrease about 19 units. Interpretation of oeffiient of variable gf (gross fixed apital formation): It means that, if gf inreases about 100 units, then Y will inrease about 25 units. Interpretation of oeffiient of variable (first differene of real oil prie): It means that, if inreases about 100 units, then Y will derease about 573 units. Negative shok: This shok redues the growth rate about 4.14 units. Positive shoks: This shok inreases the growth rate about 2.14 units. Aording to the results, both hypothesis are onfirmed. CONCLUSION This study examines two hypotheses. First; oil prie shoks influene on eonomi growth in seleted oil exporting ountries. seond; effets of oil shoks on the eonomy is different in seleted ountries. Aording to the results,a positive oil prie shok has a positive effet and a negative oil prie shok has a negative effet on the GDP growth rate of exporting ountries and also the fixed effets is different for seleted ountries. 2121
REFRENCES Corden.1982.Booming setor and deindustrialization in small open eonomy, eonomi journal, vol.92, pp 825-848. Farzanegan, Markward T.2008. The effets of oil prie shoks on the Iranian eonomy, Energy Eonomis, vol. 31, pp 134-151. Hamilton.2000. what is an oil shok, Nber working paper no. 7755. Hooker MA. 1996.What Happened to the Oil Prie-maroeonomy Relationship?,Journal of Monetary Eonomis, Vol. 38, pp195-213. Looney R.1990. oil revenues and Duth diseases in Saudi Arabia: differential impat on strutural growth, Canadian journal of development studies, vol. 6, no.1. Manera, Alessandro.2005.Oil Pries, Inflation and Interest Rates in a Strutural Cointegrated VAR Model for the G-7 Countries,IEM International Energy Markets Confrenes. Mork K.1989.Oil shok and the Maroeonomy when prie Go upanddown; Anextension of Hamilton s Results, Journal of PolitialEonomy97, pp 740-744. Rodriguez, Sahs.1999. why do resoure abundant eonomies grow more slowly? A new explanation and an appliation to Venezuela, journal of eonomi growth, vol. 4, pp 277-303. Tang, Wu.2009.oil prie shoks and their short and long term effets on the Chinese eonomy, Mpra paper no. 14703, pp 1-29. 2122