Time Series Analysis of Electricity Prices : A comparative study of power markets

• Main Author: Walter, Erik Løkken
• Format: Others
• Language: English
• Published: Norges teknisk-naturvitenskapelige universitet, Institutt for fysikk 2011
• Subjects: ntnudaim:6575; MTFYMA fysikk og matematikk; Teknisk fysikk
• Online Access: http://urn.kb.se/resolve?urn=urn:nbn:no:ntnu:diva-14046

During the past few decades, the power sectors of several countries have been substantially reorganized, and liberalized markets for trading of electricity have been established. In this report, evidence from six electricity markets are studied in order to identify characteristics of electricity prices. The market structures, statistical quantities, as well as long-term dependence, are investigated. Detrended fluctuation analysis and the average wavelet coefficient method are employed in order to estimate the Hurst exponent, which quantifies the presence of long-termed dependence. Since it is concluded that the price series are periodic on several time scales, all characteristics are investigated for both the original and deseasonalised versions of the time series. In particular, it is confirmed that the electricity prices are volatile, but that a considerable amount of the volatility is caused by the daily and weekly periodicities. Furthermore, the characteristic return distributions, volatility clustering and price spikes are analysed.