A Global Perspective on the Renewable Energy Consumption-Growth Nexus

Abstract

This study examines the causal relationship between renewable energy consumption and economic growth for 80 countries within a multivariate panel framework over the period 1990–2007. The results of the panel cointegration test indicates there is a long-run equilibrium relationship between real gross domestic product, renewable energy consumption, real gross fixed capital formation, and the labor force with the respective coefficient estimates positive and statistically significant. The results from the panel error correction model reveal bidirectional causality between renewable energy consumption and economic growth in both the short-run and long-run.

Keywords:

• Granger-causality
• growth panel
• renewable energy consumption

Notes

¹See Ozturk (2010) and Payne (2010a,b) for extensive surveys of the empirical literature on the causal relationship between energy consumption and economic growth.

²In studies on the causal relationship between renewable energy, nuclear energy, economic growth, and emissions, Menyah and Wolde-Rufael (2010) report unidirectional causality from real output to renewable energy consumption while Apergis et al. (2010) find bidirectional causality. In multi-country studies examining the relative impact of renewable and non-renewable electricity consumption on economic growth, Apergis and Payne (2011c,e) find bidirectional causality between renewable electricity consumption and economic growth with respect to developed and developing economies whereas Apergis and Payne (2011d) find unidirectional causality from economic growth to renewable electricity consumption for emerging market economies.

³The time period 1990–2007 was selected to include as many countries with the variables specified by the production model.

⁴Capital is proxied by real gross fixed capital formation given that changes in investment closely align with changes in the capital stock under the assumption of a constant depreciation rate using the perpetual inventory method (Apergis and Payne, 2010a,b; 2011a,b and citations therein).

⁵Details of the respective panel unit root tests are not presented to conserve space, but are available upon request.

⁶See Pedroni (1999) for details on the heterogeneous panel and group mean panel cointegration statistics.

⁷The estimates from either the FMOLS or dynamic ordinary least squares (DOLS) are asymptotically equivalent for more than 60 observations (Banerjee, 1999). The panel data set for this study contains 1,440 observations.

Panel A Notes: Critical values at the 1% significance level denoted by ^a: LLC −0.84; Fisher-ADF 56.09; Fisher-PP 61.15. Panel unit root test includes intercept and trend.

Panel B Notes: Both the panel and group mean panel tests are distributed asymptotically as standard normal. Of the seven tests, the panel v-statistic is a one-sided test where large positive values reject the null hypothesis of no cointegration whereas large negative values for the remaining test statistics reject the null hypothesis of no cointegration. Critical value at the 1% significance level denoted by ^a.

⁸The speed of adjustment is computed as the reciprocal of the absolute value of the coefficient on the respective error correction terms.