How Do Oil Shocks Impact Energy Consumption? A Disaggregated Analysis for the U.S.

• Author: Hoang, Thi Hong Van

1. INTRODUCTION

   Due to the alarming economic and social issues related to energy security and environmental degradation, concerns have been advocated across the globe for energy sustainability and climate change mitigation. Debates in terms of conventional energy consumption reduction and shift to alternative energy sources have been at the forefront among the policy planners and research scholars. The option of renewable energy has been discussed at length in both developing and developed countries. Scholars like Moomaw et al. (2011) and Dogan and Sekar (2016) have advocated the adoption of renewable energy sources as a substitute to decarbonize the energy system and ensure the environmental protection.

   In the U.S., the recent development of the renewable energy sector has put the country at the 3 (rd) place in the list of the most attractive countries in terms of renewable Energies worldwide (Ernest & Young report in 2017). In the meanwhile, the U.S. is also the biggest oil producer in the world (U.S. Energy Information Administration 2017 report). In this context, an important issue is how oil shocks impact the consumption of renewable energy in the U.S. In the present study, we aim to answer this question. On the other hand, to have a global view of the energy market in the U.S., we also include non-renewable (1) energy consumption for comparison. To consider the specificity of each energy source, we disaggregate them into different sources with hydropower, geothermal, wood and waste for renewable energy consumption; and coal, natural gas and petroleum for non-renewable energy consumption. We also disaggregate oil shocks into different components which are from oil supply, oil demand and the aggregate demand of the economy. For that, we follow the SVAR approach established by Kilian (2009) and Kilian and Park (2009). This distinction of oil shocks is important because each of them can impact the economic system differently (Kilian 2009, Fattouh et al. 2013) and it has been investigated in several previous studies focusing on oil shocks, such as Mutuc et al. (2011), Antonakakis et al. (2017), and Basher et al. (2018).

   On the other hand, the disaggregation of energy consumption in function of the production source is also very important because numerous studies already showed that each energy source behaves differently to economic and political events. Therefore, each energy source potentially responds differently to oil shocks. For example, Wise et al. (2014) found that a widespread of biomass in the U.S. would lead to its imports and limiting biomass imports would modify the balance of trade in agricultural products. On the other hand, Hirth (2015) found that the optimal long-term wind share should be 20% of the energy mix in Northwestern Europe. Regarding the importance to distinguish between different oil shocks and different energy sources, this double disaggregation is an important contribution of our study. To the best of our knowledge, the interaction between oil shocks and energy consumption has not been investigated at a disaggregated level for the U.S. in the previous literature.

   As for the methodological framework, we rely on a complementary analysis based on impulse response functions and the dynamic connectedness measure proposed by Diebold and Yilmaz (2014) to examine the impact of different oil shocks on different components of renewable and non-renewable energy consumption. These two methods are complementary because the first one allows analyzing the response of energy consumption to an unanticipated variation of oil shocks while the second focuses on how oil shocks' variation is transmitted to energy consumption. An additional methodological contribution lies in the computation of time-varying impulse response functions and dynamic connectedness measures. This allows us to investigate the impact of time in the relationship between oil shocks and energy consumption.

   Besides the contribution to the academic literature, our study provides important information to policy makers regarding the energy mix strategy in the U.S. In the context of the Energy Modeling Forum EMF 24 (2) related to "U.S. Technology Transitions under Alternative Climate Policies", the results of this study are helpful to policy makers while considering the impact of oil shocks on renewable non-renewable energy consumption at a disaggregated level. Our findings show that wood and waste energy consumption respond the most to oil shocks while petroleum energy consumption responds to oil supply and demand shocks only. All energy consumption responds the most to oil specific demand shocks. Furthermore, oil shocks spillover the most hydropower consumption while aggregate demand shocks spillover the less to energy consumption. On the other hand, hydropower, waste, coal and petroleum consumption have the highest dynamic connectedness with oil shocks. Finally, in a sensitivity analysis based on a time-rolling window approach, we show that the above results are time-varying. It is thus necessary to consider the changing economic and energy context to adapt the proposed policies in each period.

   The rest of the paper is organized as follows. Section 2 presents a review of the academic literature regarding the oil-related determinants of renewable and non-renewable energy consumption. Section 3 presents the methodology framework and the data used. Section 4 analyzes the empirical findings for the whole period while Section 5 focuses on a time-varying analysis. Section 6 concludes with an extended analysis of policy implications and some perspectives for future research.

2. LITERATURE REVIEW: HOW DOES OIL INFLUENCE ENERGY CONSUMPTION?

   Many studies have been conducted to examine the factors influencing energy consumption (e.g., Kiraly and Lovei 1985, Bhatia 1988, Moroney 1989, Leth-Petersen 2007, Onuonga et al. 2008, Sovacool 2009, Apergis and Payne 2010, Joyeux and Ripple 2011, Aroonruengswat et al. 2012, Salim and Rafiq 2012, Liddle 2013, Zhang et al. 2016, Acheampong 2018, Borozan 2018, Lawley and Thivierge 2018, Mahalingam and Orman 2018, Topca and Payne 2018, and Zhang and Bai 2018). Since we focus on how oil shocks influence energy consumption, the literature review will concentrate on previous studies including oil as a determinant factor of renewable and non-renewable energy consumption.

   Regarding oil-related determinants of renewable energy consumption, Henriques and Sadorsky (2008) reported the existence of a unidirectional Granger causality from oil prices to the stock prices of alternative energy companies. Similarly, Sadorsky (2009) reported that oil price hikes affect renewable energy consumption only marginally but inversely in G7 countries. Popp et al. (2011) reported that fossil fuel production has no impact on renewable energy in 26 OECD countries. Managi and Okimoto (2013) found a direct association between oil and clean energy prices after they identified structural breaks, and an identical market response is observed for both clean energy and technology stock prices. In a panel of 64 countries, Omri and Nguyen (2014) documented that oil price hikes affect the renewable energy consumption aversely in middle-income countries and the whole panel. Omri et al. (2015) found a weak influence of oil prices on renewable energy consumption. Khan et al. (2017) found no impact of oil price declines on the renewable energy sector, while the latter is found to be increasingly cost competitive with the traditional fossil fuel energy. Using the ARDL methodology, Brini et al. (2017) reported a direct impact of oil price hikes on the renewable energy consumption in Tunisia. Lin et al. (2017) found a positive impact of oil price hikes and financial development on the size and share of non-hydro renewable electricity generation in a panel of 46 countries. Shah et al. (2018) investigated the impact of oil prices and macroeconomic factors on the renewable energy market in Norway, the UK and the U.S. from 1960 to 2015. Their results showed that there is a strong relationship between oil and renewable energy in Norway and the U.S. while there is no relationship between them in the UK. The main reason is related to the oil import-export profile of the country. Finally, Troster et al. (2018) studied the causal relationship among renewable energy, oil prices and economic activity in the U.S. from 1989 to 2016 using a quantile approach. Their results showed that there is a lower-tail dependence from changes in oil prices to changes in renewable energy consumption.

   Regarding oil-related determinants of non-renewable energy consumption, Lee and Chiu (2011a) found that nuclear energy and oil are substitutes in the U.S. and Canada while they are complementary in France, Japan and the UK over the 1965-2008 period. Furthermore, the authors indicated that there is a unidirectional causality from real oil prices to nuclear energy consumption, except for the U.S., and a causality from oil consumption to nuclear energy consumption in Canada, Japan and the UK. In the long run, the impact of real oil prices is larger than that of real income on nuclear energy consumption in Canada, Germany, Japan and the US. Lee and Chiu (2011b) examined the short-run and long-run relationship among nuclear energy consumption, oil prices, oil consumption and economic growth for a panel of developed countries from 1971 to 2006. They found that in the long run, oil prices have a positive impact on nuclear energy consumption. However, real income has a higher impact on nuclear energy than oil prices in the long run. Furthermore, there is a unidirectional causality from oil prices to nuclear energy consumption. Adom (2015) used the fully modified OLS and canonical cointegration regressions (CCR) to examine the drivers of energy intensity in Nigeria. The study reported a negative impact of crude oil prices, FDI and trade openness on the energy intensity while...