Modeling and forecasting realized volatility in German–Austrian continuous intraday electricity prices

• Date: 01 September 2017
• Published date: 01 September 2017
• DOI: http://doi.org/10.1002/for.2463
• Author: Peru Muniain,Ainhoa Zarraga,Aitor Ciarreta

Received: 12 January 2016 Revised: 3 October 2016 Accepted: 31 January 2017

DOI: 10.1002/for.2463

RESEARCH ARTICLE

Modeling and forecasting realized volatility in German–Austrian

continuous intraday electricity prices

Aitor Ciarreta1Peru Muniain1Ainhoa Zarraga2

1Department of Economic Analysis II,

University of the Basque Country,

UPV/EHU, Bilbao, Spain

2Department of Applied Economics III,

University of the Basque Country,

UPV/EHU, Bilbao, Spain

Correspondence

Ainhoa Zarraga, Department of Applied

Economics III, University of the Basque

Country, UPV/EHU, Avda.Lehendakari

Aguirre 83, 48015 Bilbao, Spain.

Email: ainhoa.zarraga@ehu.eus

Funding information

Ministerio de Economía y Competitividad

and Fondo Europeo de Desarrollo Regional,

Grant/Award Number: ECO2015-64467-R ;

Departamento de Educación, Universidades

e Investigación del Gobierno Vasco,

Grant/Award Number: IT-783-13 ;

Departamento de Educación, Política

Lingüística y Cultura del Gobierno Vasco.

Grant: Beca Predoctoral de Formación de

Personal Investigador no Doctor

Abstract

This paper uses high-frequency continuous intraday electricity price data from the

EPEX market to estimate and forecast realized volatility. Three different jump tests

are used to break down the variation into jump and continuous components using

quadratic variation theory. Several heterogeneous autoregressive models are then

estimated for the logarithmic and standard deviation transformations. Generalized

autoregressive conditional heteroskedasticity (GARCH) structures are included in

the error terms of the models when evidence of conditional heteroskedasticity is

found. Model selection is based on various out-of-sample criteria. Results show that

decomposition of realized volatility is important for forecasting and that the decision

whether to include GARCH-type innovations might depend on the transformation

selected. Finally, resultsare sensitive to the jump test used in the case of the standard

deviation transformation.

KEYWORDS

GARCH, jumps, realized volatility, volatility forecasting

1INTRODUCTION

With the liberalization of European electricity markets, a

need for transparency in electricity price formation has arisen.

Market participants rely on these prices as a signpost for mak-

ing optimal consumption and production decisions. A better

understanding of their dynamics is therefore crucial. Unlike

other traded commodities, electricity is a nonstorable good,

which makes prices very volatile and leads to frequent price

spikes. Electricity prices are also characterized by mean rever-

sion, seasonality, and stationarity. The need to continuously

balance demand and supply complicates market architecture

and has led to the development of market solutions such as

intraday markets, futures, and other derivatives. Drawing on

the experience of financial markets, financial derivatives are

useful products for sharing and controlling unwanted risks

through properly designed hedging strategies.

The volatility of electricity prices plays a key role in

understanding their uncertain nature. With the availability of

high-frequency, highly volatile data, quadratic variation the-

ory has proved to be a useful approach for analyzingt he daily

realized volatility (RV) of prices (Andersen, Bollerslev, &

Diebold, 2007; Barndorff-Nielsen & Shephard, 2004, 2006).

It is employed to identify significant price jumps and decom-

pose total variation into its jump and nonjump components

nonparametrically, as introduced by Barndorff-Nielsen and

Shephard (2004) and modified by Andersen et al. (2007)

so as to prevent the process of jump detection from being

downward biased. Applications to electricity markets can

be found in Chan, Gray, and van Campen (2008), Ullrich

(2012), Haugom, Westgaard, Solibakke, and Lien (2011),

and Haugom and Ullrich (2012), among others. Addition-

ally, other jump-robust test statistics have been developed

using different types of estimators for continuous variation in

680 Copyright © 2017 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/for Journal of Forecasting.2017;36:680–690.