Predictability of price movements in deregulated electricity markets

TL;DR

This study demonstrates that electricity prices in certain deregulated markets exhibit anti-persistent, predictable patterns across various time scales, with implications for risk management and forecasting in energy markets.

Contribution

The paper introduces a comprehensive analysis of electricity price predictability using advanced statistical methods across multiple markets and time scales, revealing consistent predictive scenarios.

Findings

Markets show anti-persistent, predictable behavior across hours to a month.

Three types of correlated price movement scenarios identified in Canadian markets.

Predicted extreme price movements follow similar patterns in less volatile markets.

Abstract

In this paper we investigate predictability of electricity prices in the Canadian provinces of Alberta and Ontario, as well as in the US Mid-C market. Using scale-dependent detrended fluctuation analysis, spectral analysis, and the probability distribution analysis we show that the studied markets exhibit strongly anti-persistent properties suggesting that their dynamics can be predicted based on historic price records across the range of time scales from one hour to one month. For both Canadian markets, the price movements reveal three types of correlated behavior which can be used for forecasting. The discovered scenarios remain the same on different time scales up to one month as well as for on- and off- peak electricity data. These scenarios represent sharp increases of prices and are not present in the Mid-C market due to its lower volatility. We argue that extreme price movements in this market should follow the same tendency as the more volatile Canadian markets. The estimated values of the Pareto indices suggest that the prediction of these events can be statistically stable. The results obtained provide new relevant information for managing financial risks associated with the dynamics of electricity derivatives over time frame exceeding one day.