Interconnectedness of electricity spot prices: A dynamic network analysis
thesisposted on 28.03.2022, 09:54 by Guan Yan
This present study constructs dynamic Granger causality networks based on the spot prices of five regional markets in the Australian National Electricity Market (NEM). Although transmission lines between adjacent states result in physical interconnection in the geographical sense, the degree of integration of wholesale electricity prices is still equivocal. Based on a data set comprising electricity spot prices from 1 July 2010 to 30 June 2017, this study employs principal component analysis and generates Granger causality networks to examine the degree of interconnectedness of the NEM in a time-varying setting. We find that the derived measures of interdependence can be related to actual market events such as price spikes; unexpected high demand for electricity; sudden increase in price volatility; rebidding of dominant generators; the temporary or permanent outage of major power stations; and upgrades and limitations in transmission capacity. The first measure is the cumulative risk fraction of the first few principal components, which conveys information on risk concentration. Another measure is the dynamic causality index generated from the estimated Granger causality network. In the analysed network, we find that stronger dependence is exhibited by regional markets that are linked by interconnectors, while the direction of the relationship can be related to inter-regional trade. Furthermore, this study examines the usefulness of the derived measures as early-warning indicators for upcoming periods of extreme prices and volatility. Our results suggest limited predictive power of the interconnectedness measures for spot price behaviour in the NEM.