US Disaggregate Fossil Fuel Consumption and Real GDP: An Empirical Note

Abstract

This empirical note provides a disaggregated analysis of the causal relationship between fossil fuel consumption and real gross domestic product (GDP) in the US using annual data from 1949 to 2006. The Toda-Yamamoto long-run causality tests reveal the absence of Granger-causality between coal consumption and real GDP; positive unidirectional causality from real GDP to natural gas consumption; and positive unidirectional causality from petroleum consumption to real GDP.

Keywords:

• fossil fuel consumption
• Granger-causality
• real GDP

Notes

¹In addition to the studies cited on the US, there have been a number of studies examining the relationship between fossil fuel consumption and economic activity in other countries: (Hoa [1993] for Thailand, Yang [2000] for Taiwan, Aqeel and Butt [2001] for Pakistan, Fatai et al. [2004] for Australia and New Zealand, Wolde-Rufael [2004] for Shanghai, Lee and Chang [2005] for Taiwan, Yoo [2006a,b] for Korea, Zou and Chau [2006] for China, Zamani [2007] for Iran, Hu and Lin [2008] for Taiwan, Jinke et al. [2008] for China, India, Japan, South Africa, and South Korea, Reynolds and Kolodziej [2008] for the Former Soviet Union, and Yuan et al. [2008] for China). For an exhaustive survey of the international empirical evidence on the causal relationship between energy consumption and economic growth see Payne (2010).

² Fatai et al. (2004), Wolde-Rufael (2004, 2005, 2006), Lee (2006), Soytas and Sari (2006, 2010), Mehrara (2007), Soytas et al. (2007), Squalli (2007), Zachariadis (2007), Bowden and Payne (2010), Payne (2009), and Payne and Taylor (2010).

³The test statistics have the standard asymptotic distribution in order to draw valid inferences for the Granger-causality tests.

⁴Unit root tests are available upon request.

⁵Results for the real gross fixed capital formation and employment equations are available upon request.

⁶The Ljung-Box Q-statistics at 24 lags reveal the residuals are free of serial correlation.