Economic policy uncertainty, energy consumption and carbon emissions in G7 countries: evidence from a panel Granger causality analysis

We investigate the causal relationship between economic policy uncertainty (EPU) and energy consumption and carbon (CO2) emissions in G7 countries. We employ a bootstrap panel Granger causality test developed by Kónya (Econ Model 23:978–992, 2006), using a yearly data set spanning from 1998 to 2018. Our test results provide significant support for a unidirectional causality running from EPU to energy consumption in Japan; from EPU to CO2 emissions in the USA and Germany; and from EPU to both energy consumption and CO2 emissions in Canada. In Italy, causality runs from CO2 emissions to EPU, but a bidirectional causality between EPU and energy consumption exists as well. We also explore a unidirectional causality that runs from energy consumption to CO2 in the USA. Based on the overall findings, we draw important implications for policymakers and we strongly recommend for G7 countries to take into account possible negative effects of EPU on energy conservation policies, which should be embarked upon to reduce energy consumption and CO2 emissions, as committed in their recent climate mandate.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Subscribe and save

Springer+ Basic €32.70 /Month

Buy Now

Price includes VAT (France)

Instant access to the full article PDF.

Rent this article via DeepDyve

Similar content being viewed by others

Asymmetric effect of economic policy uncertainty, political stability, energy consumption, and economic growth on CO2 emissions: evidence from G-7 countries

Article 04 February 2023

Energy-related uncertainty, financial regulations, and environmental sustainability in the United States

Article Open access 09 August 2024

Interface between energy consumption, CO2 emissions, economic growth, and macroeconomic openness in financial action task force countries through the lens of a causality approach

Article 05 November 2022

Explore related subjects

Notes

This is particularly true when “growth hypothesis” holds where energy consumption causes economic growth. However, literature supports for “conservation hypothesis” (unidirectional causality running from economic growth to energy consumption), “feedback hypothesis” (bidirectional causality), and “neutrality hypothesis” (no causality) as well. We refer to “Theoretical background” for a concise review of the broad literature regarding the energy consumption-economic growth nexus.

Please refer to “Theoretical background” for the origination of the EKC hypothesis as well as its transformation within the energy literature. However, we would like to emphasize that testing the EKC hypothesis is out of the scope of this paper. We rather include CO2 emissions variable both in order to augment our model specifications (see Section 3.2) and more importantly to analyze its relationship with EPU on the basis of CO2 emissions-economic growth nexus (see Section 2.2).

We refer to Khandokar and Serletis (2018) and Su et al. (2019) for a brief comparison of EPU with other sources of uncertainty.

Note that the European EPU index is composed of data pertaining to France, Germany, Italy, Spain, and the UK.

Hajko (2017) argues that one of reasons of the inconclusive evidence is significant methodological omissions.

All indices are available on https://www.policyuncertainty.com/. See Kar et al. (2011) and Wolde-Rufael (2014) and references therein. We refer to Pesaran et al. (2008) for further details. We refer to Pesaran and Yamagata (2008) for further details of Swamy test and its estimators.

Frankly, our trivariate analysis can handle only direct, one-period-ahead causality between any pair of variables ignoring the possibility of indirect causality at longer time horizons. That is, the possibility of indirect causality is disregarded, even though two-step causality may arise when, for instance, EPU causes energy consumption and/or energy consumption causes EPU indirectly, via CO2 emissions. In bivariate systems, however, absence of causality for one-period-ahead implies the same for any horizon. Thus, bivariate and trivariate causality test results are not comparable (Kónya 2004).

We refer to Kónya (2006) for further details of the bootstrap sampling procedure.

Though not reported in order to save space, results from the lag selection procedure are available upon request. In the case of Emirmahmutoglu and Kose’s (2011) procedure, however, the individual lag orders are selected based on the Schwarz Bayesian Criterion (BCI).

In line with this requirement, we determine the maximum integration degree (d) of the series. Briefly, we investigate the time series properties of the variables by employing the augmented Dickey-Fuller (ADF) test and find that d is equal to I(1) for each member country in our panel. Data are available upon request.

Also note that unreported entire panel results obtained from Emirmahmutoglu and Kose (2011) indicate no-causality among variables of interest. This finding suggests that individual countries (i.e. cross-sectional units) are not able to drive the results for G7 as a whole (i.e. entire panel) since time series evidence is not largely consistent with the panel analysis results. Data are available on request.

Banday and Aneja (2019) employ Emirmahmutoglu and Kose (2011) and Dumitrescu and Hurlin (2012) in a comparative manner and conclude with divergent results for individual G7 countries.

This is also confirmed by the entire panel statistics of Emirmahmutoglu and Kose (2011). See footnote 14.

We checked the relationships among our variables by means of standard panel data analyses as well. Unreported results show that there are no statistically significant associations between EPU and energy consumption (with positive coefficient sign) or CO2 emissions (with negative coefficient sign), while positive and significant links are found for energy consumption and CO2 emissions in all cases. These findings corroborate the panel correlation results previously given in Table 2. Data are available on request.

Interestingly, Akadiri et al. (2019) argue that Italy is an energy-independent country and, therefore, energy conservation policies would not harm economic growth. This outcome contradicts with that of Stamatiou and Dritsakis (2019), which posits that a reduction in energy consumption has an adverse effect on economic growth. The former study use data spanning from 1970 to 2014 whereas the latter focuses on the sampling period of 1960–2011. This implies that the causal links may be time-variant.

Note that, Japan, among the G7, is also a huge energy consumer and the country’s energy dependency rate is the highest [around 93%, (Statista 2015)], but one reason for this may be the nuclear plant disaster occurred in Fukushima on March 11, 2011, when the steady declining trend in the energy dependency rate started to reverse. Dissimilar findings we obtain for these two high energy-consuming countries imply that, in the meantime, Japan appears to have stuck to policies to realize its potential in reducing energy intensity and optimizing energy structure without harming growth, probably, by virtue of the advancement of alternative sources of energy (Pereira et al. 2014).

References

Author information

Authors and Affiliations

  1. Department of Business Administration, Hacettepe University, Beytepe, 06800, Cankaya, Ankara, Turkey Burak Pirgaip
  2. Department of Business, Atilim University, Kızılcaşar Mahallesi, 06830, Incek Golbasi, Ankara, Turkey Burcu Dinçergök
  1. Burak Pirgaip