Are global warming and economic growth compatible? Evidence from five OPEC countries? Ramazan Sari * , Ugur Soytas 1 Dept. of Business Administration, Middle East Technical Univ., 06531 Ankara, Turkey article info Article history: Received 11 September 2008 Received in revised form 2 December 2008 Accepted 5 December 2008 Available online 14 January 2009 Keywords: Autoregressive distributed lag OPEC Cointegration Energy CO 2 abstract In this paper, we investigate the relationship between carbon emissions, income, energy and total employment in selected OPEC countries for the period of 1971–2002. We mainly focus on the link between energy use and income. Employing the autoregressive distributed lag (ARDL) approach, we find that there is a cointegrating relationship between the variables in Saudi Arabia only. The long run forcing variables for income are determined to be employment and energy for Saudi Arabia. In Indonesia, Algeria, Nigeria, and Venezuela, there is no cointegration between income and energy. Secondly we question the long run Granger causality between carbon emissions, energy use, and income. Our results suggest that none of the countries need to sacrifice economic growth to decrease their emission levels. Indonesia and Nigeria may contribute to emissions reduction via energy conservation without negative long run effects on economic growth. Ó 2008 Elsevier Ltd. All rights reserved. 1. Introduction The main source of global warming is emissions of greenhouse gasses (GHG), and the main source of GHG emissions is believed to be energy consumption. Therefore, reducing energy consumption will also decrease the emission levels. However, it is not a simple matter of applying energy conservation methods, since energy con- sumption may have important effects on economic growth. Due to these presumed links between GHG, energy consumption and eco- nomic growth, it is widely believed that decreasing carbon dioxide (CO 2 ) emissions to the Kyoto targets would also reduce the growth of GDP. In other words, emission reduction requires energy conser- vation which hinders economic growth assuming that there is a causal relationship from energy consumption to CO 2 emissions and real income. Because of these presumed links, many countries are hesitant to keep with Kyoto targets. However, there is abun- dant number of empirical studies, employing diverse methods, conducted in several countries, which point out that the link be- tween energy consumption, income and CO 2 may not be unique. Therefore, investigating the temporal relationship between energy use, CO 2 and income in countries separately may be necessary. Stern and Cleveland [44] provide an excellent review of the ear- lier and more recent work on the link between energy consump- tion and economic growth. Stern [43], Masih and Masih [22–24] are some examples that apply relatively stronger time series tech- niques than earlier work. Asafu-Adjaye [2], Hondroyiannis, Lolos and Papapetrou [14], Glasure [12], Soytas and Sari [38–40], Sari and Soytas [36,37], Ghali and El-Sakka [10], Lee [18,19], Lee and Chang [20], Huang, Hwang and Yang [15], Narayan and Smyth [27] and Ewing, Sari and Soytas [9] are examples that utilize rela- tively new time series or panel-data techniques. Even this limited list indicates that the link between energy use and income is a well studied topic. However, despite having many studies employing different techniques, different time periods, and different control variables in different countries, there is a lack of unanimity as to the nature of the relationship between energy use and income. The divergence of results may be indicating that the relationship is too complex and/or its nature differs from country to country. The recent studies on the other hand improved our understand- ing in at least two ways. Firstly, the empirical studies may be suf- fering from omitted variables bias that may yield spurious causality test results. Hence, a multivariate approach should be preferred over bi-variate approaches. Secondly, the temporal rela- tionship between energy use and income may be depending on country specific factors. Furthermore, depending on the nature of the link in concern, alternative policy options may be available to policy makers in different countries. Therefore, studying countries individually may be necessary. There is an abundance of studies that test the environmental Kuznets curve (EKC) hypothesis (see [6,45] for a review) which re- late environmental degradation to economic growth. The hypothe- sis states that as economies grow pollution also grows, but after an income level is reached economic growth is associated with a de- cline in pollution. As Rothman and de Bruyn [35] suggest if the 0306-2619/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.apenergy.2008.12.007 * Corresponding author. Tel.: +90 312 210 3068; fax: +90 312 210 7962. E-mail addresses: soytas@metu.edu.tr (R. Sari), rsari@metu.edu.tr (U. Soytas). 1 Tel.: +90 312 210 2010. Applied Energy 86 (2009) 1887–1893 Contents lists available at ScienceDirect Applied Energy journal homepage: www.elsevier.com/locate/apenergy