Determinants for European Energy Markets Intra-day Volatility using Dynamic Panel Data Models and GMM-type Estimators
Abstract—The volatility of the Scandinavian international future energy market is examined based on a well known intra-day range-based measure of volatility. The main purpose of the paper is to identify determinants of the energy market’s intra-day volatility. Firstly, the investigation is a contract-by-contract, range-based volatility measure. The findings are (1) long memory in volatility the longest contracts (year), but not for the shortest (month), (2) the first difference contemporaneous and one-day lagged trading volume measures influences daily volatility positively and consistently over all contracts, (3) trading volume seems to dominate the Samuelson (maturity) and the open-interest hypotheses thus manifesting earlier findings of future markets. Secondly, the paper applies a relatively new methodology to cope with dynamic panel data models, where every contract in the market is organized as a panel. The GMM-type estimator for dynamic panel-data of Arellano [1] is therefore implemented. The results show appropriate specification residual tests and demonstrate firstly the importance of volatility serial correlation. This estimator also shows that the change in daily trading volume is the major determinant of intra-day volatility. In contrast to the segmented analyses, time to maturity shows increased intra-day volatility influence while the measures of bid-ask spread and open interest show decreased significance. The main GMM findings are therefore that the change in trading volume (including lags) and time to maturity are major determinants of intra-day volatility.
Index Terms— Electricity Forward/Future Markets, Intra-day Volatility, Trading Volume, Open Interest, Bid-Ask Spread, Market Depth, Time to Maturity
[1] INTRODUCTION
International literature has shown that future prices have at least four statistical characteristics. First, future prices are not normally distributed, but the distribution of percentage changes is leptokurtic. Second,
Abstract—The volatility of the Scandinavian international future energy market is examined based on a well known intra-day range-based measure of volatility. The main purpose of the paper is to identify determinants of the energy market’s intra-day volatility. Firstly, the investigation is a contract-by-contract, range-based volatility measure. The findings are (1) long memory in volatility the longest contracts (year), but not for the shortest (month), (2) the first difference contemporaneous and one-day lagged trading volume measures influences daily volatility positively and consistently over all contracts, (3) trading volume seems to dominate the Samuelson (maturity) and the open-interest hypotheses thus manifesting earlier findings of future markets. Secondly, the paper applies a relatively new methodology to cope with dynamic panel data models, where every contract in the market is organized as a panel. The GMM-type estimator for dynamic panel-data of Arellano [1] is therefore implemented. The results show appropriate specification residual tests and demonstrate firstly the importance of volatility serial correlation. This estimator also shows that the change in daily trading volume is the major determinant of intra-day volatility. In contrast to the segmented analyses, time to maturity shows increased intra-day volatility influence while the measures of bid-ask spread and open interest show decreased significance. The main GMM findings are therefore that the change in trading volume (including lags) and time to maturity are major determinants of intra-day volatility.
Index Terms— Electricity Forward/Future Markets, Intra-day Volatility, Trading Volume, Open Interest, Bid-Ask Spread, Market Depth, Time to Maturity
[1] INTRODUCTION
International literature has shown that future prices have at least four statistical characteristics. First, future prices are not normally distributed, but the distribution of percentage changes is leptokurtic. Second,