Volatility Spillovers in Energy Markets.

• Author: Chulia, Helena

1. INTRODUCTION

   Electricity is considered to be a strategic asset because of its extensive use by virtually all sectors in modern economies. Apart from renewable generation sources, which are progressively more and more present in the supply mix of electricity markets across the world, the most important fuels used to generate power are natural gas, oil and coal. These three are generally competitors in the production of electricity, while all four commodities are substitutes for each other in consumption, which may lead to their prices being somewhat linked. Additionally, since 2005, power generators and energy intensive industries from signatory countries to the Kyoto Protocol receive European Union Emission Allowances (EUA) that can be traded. They must report annually on their greenhouse gas emissions and surrender the corresponding number of EUA. Installations cannot exceed their maximum number of emission allowances. In December 2015, 145 countries adopted the Paris Agreement, which entered into force shortly thereafter, on November 2016. The Paris Agreement has reconfirmed the role of emissions trading schemes as an instrument for achieving global climate change goals. According to the United States Environmental Protection Agency (EPA), allowances are fully marketable commodities, since once allocated they may be bought, sold, traded or banked for use in future years. (1) Therefore, it makes sense to extend the study to the interactions between energy and carbon markets, by analyzing potential volatility transmission between them.

   The main purpose of our paper is to assess the extent and evolution of the links between energy markets. Specifically, we are interested in answering the following questions: (1) What is the total volatility spillover effect in energy markets? (2) What is the evolving nature of volatility spillovers? (3) Which markets are exporters (importers) of volatility to (from) other markets? (4) Are volatility spillovers higher within or across energy sectors? (5) Is there evidence of increasing European energy markets' integration over time?

   King and Wadhwani (1990) present volatility spillovers as a consequence of rational agents trying to infer relevant information from price changes across different markets. In the same line, Strohsal and Weber (2015) state that volatility spillovers across assets may indicate the spread of valuable information among fundamentally linked markets. The information content of price movements is not observable, but it might be deducted from observed price changes in one of the assets if they are interpreted as informative enough by traders of the other related assets. In fact, volatility would be zero in absence of relevant news, but price adjustments in an asset provoked by the arrival of new information will increase its volatility. Thus, the volatility spillover effect refers to the impact that events in one market may have on the volatility of other markets, being the information flow connected to volatility whenever observed price changes are used to infer valuable information from price changes in the related market. (2) This research issue is closely related to price discovery, since knowing the direction in which information flows between markets, one can anticipate price movements in relatively less liquid assets that incorporate information less rapidly than others to which they are shown to be linked. It can also be considered the existence of volatility spillovers as an evidence of whether markets within and across regions are integrated, as stated in Bekaert et al. (2005). An integrated electricity market for the whole European Union is a long-term goal of the European authorities. Some voices claim there is significant progress being made in the integration of European energy markets, which is actually hard to assess. The more integrated markets are, the higher the volatility transmission between them. This work is not limited solely to electricity markets but extends the analysis to other energy prices such as natural gas, C02 emission allowances, crude oil and coal to evaluate the current state of integration between European energy markets, as well as to investigate their relationships with other non-European markets

   The linkages across and within energy markets have been widely studied in the literature. Most of the studies focus on analyzing market integration and price relationships and some other papers look at volatility spillovers. (5) Within the first group, Granger causality, cointegration analysis and the Vector Error Correction (VEC) model proposed by Engle and Granger (1987) have been extensively used. Within the second group, multivariate Generalized Autoregressive Conditional Heteroskedasticity (GARCH) models are the most commonly employed econometric techniques. One drawback of these models is that the number of parameters often increases rapidly with the dimensions of the model, which limits their scope of application. Recently, some papers have applied the methodology proposed by Diebold and Yilmaz (2009, 2012 and 2014) to explore spillovers in commodity markets. This approach is based on forecast error variance decompositions in a vector autoregressive framework and does not suffer from the curse of dimensionality of multivariate GARCH models. For example, Chevallier and Ielpo (2013), explore volatility spillovers within commodities, between standard assets and commodities and between commodities and commodity currencies, in the U.S.. Zhang and Wang (2014) analyze spillovers between China and world oil markets. Barunik et al. (2015) analyze volatility spillovers across petroleum-based commodities, differentiating between spillovers due to negative and positive returns. Batten et al. (2015) analyze spillovers among the four main precious metals - gold, silver, platinum and palladium. Kang et al. (2017) examine spillover effects among six commodity futures markets - gold, silver, West Texas Intermediate crude oil, corn, wheat, and rice -. Finally, Diebold et al. (2018) measure commodities volatility connectedness using the framework of Demirer et al. (2018), which build on Diebold and Yilmaz (2014). They include in the analysis four energy commodities, two precious metals, four industrial metals, two livestock commodities, four grains, and three so-called "softs" (coffee, cotton, sugar).

   In this paper, we also adopt the methodology proposed by Diebold and Yilmaz (2009 and 2012) to uncover the links between energy markets. This approach allows us to dynamically capture the extent of linkages as well as their direction. We use 17 series in our analysis, belonging to the electricity, natural gas, C02, oil and coal sectors. Our main findings can be summarized as follows. Firstly, we find that own-sector volatility spillovers account for the highest share of forecast error variance. Furthermore, pairwise directional spillovers are higher within, than across, sectors and the highest pairwise spillovers are observed between crude oil series. Secondly, within sectors, the German electricity market is, in overall terms, the main transmitter of volatility spillovers. Over time, the German Netconnect Germany and the Dutch Title Transfer Facility arise as the two reference price series affecting the rest of the natural gas series. There is a change shown in the role of the crude oil series in the later years of the sample, with Brent having become a net receiver of volatility spillovers from West Texas Intermediate, since 2013. Regarding the coal series, the U.S. Central Appalachian and the European API2 index mutually impact upon one another without the former prevailing over the latter. Interestingly, the linkages between natural gas volatility and the rest of the commodity volatilities are shown to be the greatest. In particular, Title Transfer Facility may be on the way to becoming the benchmark price for natural gas in Europe, overtaking National Balance Point. Last but not least, according to our results, natural gas may be replacing crude oil as a global benchmark for energy commodities. Thirdly, regarding the level of integration between European electricity markets, the most integrated markets appear to be those of Germany, France and the Netherlands, distantly followed by Italy, Spain and the Nordic block. Interestingly, spillovers are shown to be time-varying and seem to increase with economic growth as well as during periods of turmoil.

   The remainder of the paper is organized as follows. The next section summarizes the literature. Section 3 describes our data. Section 4 lays out the methodology we use to analyze volatility spillovers in European energy markets. Section 5 discusses the empirical results and finally. Section 6 concludes.

2. LITERATURE REVIEW

   Interest in energy markets' dynamic relationships has been growing over recent years. The early research on the linkages across energy sectors used cointegration methods. For example, Emery and Liu (2002) analyze the relationship between electricity and natural gas futures prices in the New York Mercantile Exchange, California-Oregon Border and Palo Verde and find that electricity and natural gas futures prices are cointegrated. Interestingly, Asche et al. (2006) report that, in the UK, integration between the wholesale prices of crude oil, natural gas and electricity took place only during that period when the natural gas market had been deregulated but was not yet physically linked to the continental European natural gas market through the interconnector. In a related study, Mohammadi (2009) examines the long-run relations and short-run dynamics among electricity retail prices and fossil fuels (coal, natural gas and crude oil) in the U.S. market. He finds evidence of significant long-run relations only between electricity and coal prices and some evidence of unidirectional short-run causality from coal and natural gas prices to electricity prices.

   ...