Econometrics, Energy and Finance Centre for Econometric Analysis, Cass Business School 106 Bunhill Row, London, EC1Y 8TZ (U.K.) 8 April 2013 Room LG003 PROGRAMME 08:30-09:00 Registration 09:00-10:30 SESSION 1 Chair: Sjur Westgaard Forecasting of daily electricity spot prices by incorporating intra-day relationships Katarzyna Maciejowska (Wroclaw University of Technology, Poland) and Rafal Weron (Wroclaw University of Technology, Poland) Modelling residential electricity demand in Europe with Autometrics TM Elisabetta Pellini (SEEC, University of Surrey, UK) Analysis and forecasting of electricity price risks with quantile factor models Derek Bunn (LBS, UK), Arne Andresen (Norwegian University of Science and Technology, Trondheim, Norway), Dipeng Chen (Centrica plc, UK), Sjur Westgaard (Norwegian University of Science and Technology, Trondheim, Norway) 10:30-11:00 Coffee break SESSION 2 Chair: Nikos Nomikos 11:00-13:00 A model for the optimal management of medium-voltage AC networks with distributed generation and storage devices Maria Teresa Vespucci (University of Bergamo, Italy), Alessandro Bosisio (Politecnico di Milano, Italy), Diana Moneta (RSE SpA Milano, Italy), Stefano Zigrino (University of Bergamo, Italy) 1
Time varying fractional cointegration of energy markets Melanie A. Houllier (Cass Business School, City University London, UK) and Lilian M. de Menezes (Cass Business School, City University London, UK) Calibration of a multifactor model for the forward markets of several commodities Enrico Edoli (Department of Mathematics, University of Padova, Italy and Finalyst SAS, Italy) Davide Tasinato (Department of Mathematics, University of Padova, Italy), and Tiziano Vargiolu (Department of Mathematics, University of Padova, Italy) Extreme value theory and mixed c-vine copulas on modelling energy price risks Nikos K Nomikos (Cass Business School, City University London, UK) and Emmanouil N Karimalis (Cass Business School, UK) 13:00-14:15 Lunch break 14:15-16:15 SESSION 3 Chair: Giovanni Urga The role of speculation in oil markets: what have we learned so far? Lutz Kilian (University of Michigan and CEPR, USA) Recent oil price dynamics: is this time speculation? Rita D'Ecclesia (Sapienza University, Rome, Italy), Emiliano Magrini (European Commission, Brussels, Belgium), Pierluigi Montalbano (Sapienza University, Rome, Italy), Umberto Triulzi (Sapienza University, Rome, Italy) Risk premia in energy markets Almut Veraart (Imperial College London & CREATES) and Luitgard Veraart (LSE). Asynchronous data and volatility spillover with an application to natural gas futures and forward contracts Alessandrto Lanza (LUISS University, Rome, Italy, and Centro Euro-Mediterraneo per i Cambiamenti Climatici, Milan, Italy), Marianna Russo (Cass Business School, UK) and Giovanni Urga (Cass Business School, City University, London, UK and Bergamo University, Italy) Giovanni Urga, 09 April 2013 2
TITLES & ABSTRACTS The role of speculation in oil markets: what have we learned so far? Lutz Kilian (University of Michigan and CEPR, USA) A popular view is that the surge in the real price of oil during 2003-08 cannot be explained by economic fundamentals, but was caused by the increased financialization of oil futures markets, which in turn allowed speculation to become a major determinant of the spot price of oil. This interpretation has been driving policy efforts to tighten the regulation of oil derivatives markets. This presentation reviews the evidence supporting this view. It is shown that the existing evidence is not supportive of an important role of speculation in driving the spot price of oil after 2003. Instead, there is strong evidence that the comovement between spot and futures prices reflects common economic fundamentals rather than the financialization of oil futures markets. Analysis and forecasting of electricity price risks with quantile factor models Derek Bunn (LBS, UK), Arne Andresen (Norwegian University of Science and Technology, Trondheim, Norway), Dipeng Chen (Centrica plc, UK), Sjur Westgaard (Norwegian University of Science and Technology, Trondheim, Norway) Forecasting quantile and value-at-risk levels for spot electricity prices is methodologically challenging because of the distinctive stochastic properties of the price density functions, volatility clustering and various exogenous factors. Despite this, accurate risk measures have considerable value in energy trading and risk management with the topic being actively researched for better techniques. We approach the problem by using a novel multifactor, dynamic, quantile regression formulation, extended to include GARCH properties. This captures the specification effects of mean reversion, spikes, time varying volatility and demonstrates how the prices of gas, coal and carbon, forecasts of demand and reserve margin in addition to price volatility influence the peak price quantiles for GB power. We show how the price elasticities for these factors vary substantially across the quantiles. We also show that a structural linear quantile regression model outperforms and easier to implement than state-of-the-art benchmark models for out-of-sample forecasts of value-at-risk. Modelling residential electricity demand in Europe with Autometrics TM Elisabetta Pellini (SEEC, University of Surrey, UK) This paper estimates residential electricity demand for Austria, Belgium, France, Germany, Italy, Spain, Switzerland, the Netherlands and the UK, for the period 1978-2009. These countries are amongst Europe s largest economies and contribute for two thirds to European electricity consumption. The general unrestricted error correction mechanism with structural breaks is employed to specify a stable relationship between electricity demand and its determinants. The novelty of this paper consists in modelling potential instability factors using the Impulse Indicator Saturation framework and its related extensions, and in estimating the demand function with the search algorithm Autometrics TM. The results provide consistent estimates of price and income elasticities 3
and highlight that electricity is a normal good for European households and that its demand is inelastic to price. Forecasting of daily electricity spot prices by incorporating intra-day relationships Katarzyna Maciejowska (Wroclaw University of Technology, Poland) and Rafal Weron (Wroclaw University of Technology, Poland) In the paper, we show that incorporating intra-day relationships of electricity prices improves the forecasts of(average) daily electricity spot prices. We use half hourly data from the UK power market to model the spot prices directly (via VARX and ARX models) and indirectly (via factor models). Three econometric models are estimated and their forecasting performance is compared to that of a univariate model, which uses only aggregated (daily) data. The results indicate that there are forecast improvements from incorporating the disaggregated data, especially, when the forecast horizon exceeds one week. Additional improvements are achieved, when their correlation structure is explored. A model for the optimal management of medium-voltage AC networks with distributed generation and storage devices Maria Teresa Vespucci (University of Bergamo, Italy), Alessandro Bosisio (Politecnico di Milano, Italy), Diana Moneta (RSE SpA Milano, Italy), Stefano Zigrino (University of Bergamo, Italy) The Forward Search is an iterative algorithm concerned with detection of outliers and other unsuspected structures in data. This approach has been suggested, analysed and applied for regression models and multivariate models in the monographs by Atkinson and Riani (2000) and Atkinson, Riani and Cerioli (2004), respectively, as well as a number of research papers. See also Atkinson, Riani and Cerioli (2010) for a recent overview. So far formal asymptotic analysis has not been undertaken and inferential procedures have been relying on simulation evidence. We will provide an asymptotic analysis of the Forward Search applied to a regression model. Extreme value theory and mixed c-vine copulas on modelling energy price risks Nikos K Nomikos (Cass Business School, UK) and Emmanouil N Karimalis (Cass Business School, UK) In this paper we introduce the idea of mixed canonical vine copulas as an alternative and more flexible way of modelling the joint distributions of energy portfolios and propose an extension of extreme value theory in the context of vine copula modelling that takes into account the asymmetries of the marginal and joint distributions. Our approach combines GARCH models to estimate the conditional volatility, extreme value theory (EVT) for estimating each tail of the innovation distribution of the GARCH models and canonical vine copulas for modelling the dependence structure of the portfolio constituents. In addition, we show that for risk management applications, pair copula selection should 4
not be based on likelihood-based criteria but rather on tail dependence measures. Under this context, we compute risk measures, such as VaR and Expected Shortfall and discuss the implication of our modelling approach for passive and active risk management performance. We show, despite the week degree of tail dependence between portfolio constituents, that the proposed procedure provides better 1-day VaR and Expected Shortfall estimates for lower quantiles than methods which ignore the heavy tails of innovations and the asymmetries of the joint return distribution. Risk premia in energy markets Almut Veraart (Imperial College London & CREATES) and Luitgard Veraart (LSE). Risk premia between spot and forward prices play a key role in energy markets. This paper derives analytic expressions for such risk premia when spot prices are modelled by Levy semistationary processes. While the relation between spot and forward prices can be derived using classical no-arbitrage arguments as long as the underlying commodities are storable, the situation changes in the case of electricity. Hence, in an empirical study based on electricity spot prices and futures from the European Energy Exchange market, we investigate the empirical behaviour of electricity risk premia from a statistical perspective. We find that a model-based prediction of the spot price has some explanatory power for the corresponding forward price, but there is a signi cant additional amount of variability, the risk premium, which needs to be accounted for. We demonstrate how a suitable model for electricity forward prices can be formulated and we obtain promising empirical results. Recent oil price dynamics: is this time speculation? Rita D'Ecclesia (Sapienza University, Rome, Italy), Emiliano Magrini (European Commission, Brussels, Belgium), Pierluigi Montalbano(Sapienza University, Rome, Italy), Umberto Triulzi (Sapienza University, Rome, Italy) Oil price dynamics is at the center of a lively debate. While the recent unprecedented surge in the spot price of crude oil fostered the interpretation of an increasing financialization" of the real price of oil, scholars lack consensus around the long-term fundamentals of the oil market. This paper contributes to the above debate by: i) tracking the most recent oil price dynamics using a structural model approach that takes explicity into account the financial component ii) assessing the distinctive role played by the financialization" of oil market in the short-run vs long-run determinants of oil price. To this hand, a cointegration relationship is estimated between the real oil price, the global economic activity and the world oil production while the short-run vs long-run equilibrium is investigated using an Error Correction Model Framework. Differently from previous works on the issue monthly data of real oil prices are used over the period 1996 to 2010. The results show that the long-run equilibrium of the real oil price can be explained using a standard demand and supply model as integrated by a financial component, while speculation contribute only to the short-run fluctuations. 5
Time Varying Fractional Cointegration of Energy Markets Melanie A. Houllier (Cass Business School, City University London, UK) and Lilian M. de Menezes (Cass Business School, City University London, UK) During the 1990s, the European Union decided to gradually open the electricity and natural gas markets to competition. The first liberalisation directives were approved in 1996 (electricity) and 1998 (gas) to be implemented into Member States' legal systems by 1998 and 2000 respectively (European Commission, 2013). Market liberalization aims to achieve cost-reflective and competitive power prices (Gebhartd and Höffler, 2007), which should translate to common long run dynamics between electricity prices and energy source developments. In this context, the present study revisits previous literature that assesses the extent of energy market integration, by adopting a timevarying fractional co-integration analysis. It uses the semiparametric two-step Feasible Exact Local Whittle estimator (FELW), which is robust to heteroscedasticity (Shao and Wu, 2007). Long run dynamics of spot prices from APX-ENDEX (UK and Netherlands),EEXDE (Germany, Swiss) and EEXFR (France) in the period between February 2009 and November 2012 are compared to oil and gas price movements. Average daily, peak and base-load spot prices are examined and initial results reported. Asynchronous data and volatility spillover with an application to natural gas futures and forward contracts Alessandrto Lanza (LUISS University, Rome, Italy, and Centro Euro-Mediterraneo per i Cambiamenti Climatici, Milan, Italy), Marianna Russo (Cass Business School, UK) and Giovanni Urga (Cass Business School, UK) Market microstructure notion implies that in a market with asymmetrically informed agents, trades convey information causing a persistent impact on the security price. The relevance of this issue is well described in energy market and particularly in markets where future and forward prices are both present such as the European natural gas market. The main aim of the paper is to explore the sources of correlation between the two markets and the way in which information disclosure transmits from one market to the other. Moving from the seminal work of Cox et. al (1981) on the difference between forward and futures prices, we examine the joint dynamics of the returns volatility of the two contracts at different delivery and alternative time intervals using a Dynamic Conditional Correlation model. The analysis of the conditional correlations occurring between the two prices provides useful insights and information about the adjustment process the two markets incur over the period before the delivery time. Our analysis confirms that there is information spillover between futures and forward prices in the European natural gas market providing a measure of this difference and an explanation of when and why one could expect an increase in correlations. 6
Calibration of a Multifactor Model for the Forward Markets of Several Commodities Enrico Edoli (Department of Mathematics, University of Padova, Italy and Finalyst SAS, Italy) Davide Tasinato (Department of Mathematics, University of Padova, Italy), and Tiziano Vargiolu (Department of Mathematics, University of Padova, Italy) We propose a model for the evolution of forward prices of several commodities, which is an extension of the factor forward model in Benmenzer et al. (2007) and Kiesel et al. (2009), to a market where multiple commodities are traded. We then show how to calibrate this model in a market where few or no derivative assets on forward contracts are present, and one is forced to calibrate on historical forward prices. First we calibrate separately the four coefficients of every single commodities, using an approach based on quadratic variation. Then we pass to estimate the mutual correlation among the Brownian motions driving the different commodities, the estimates being based now on the quadratic covariation between forward prices of different commodities. This calibration is compared to a calibration method used by practitioners, which uses rolling time series and requires a modification of the model, but turns out to be more accurate in practice, especially with a low frequency of observed transaction. We present efficient methods to perform the calibration with both methods, as well as the calibration of the intercommodity correlation matrix. Then we test the two methods against simulated data to assess the goodness of both, and calibrate our model to WTI, ICE Brent and ICE Gasoil forward prices. Finally we present how to estimate spot volatility from forward parameters, with an application to the WTI spot volatility. Giovanni Urga, 09 April 2013 7