01whole.pdf (2.33 MB)
Download file

Risk management in electricity markets: premium dynamics, premium forecasting and industry structure

Download (2.33 MB)
posted on 28.03.2022, 17:50 by Sami G. Aoude
The topic of this PhD thesis is the electricity market with emphasis on the liquidity of electricity futures contracts and the dynamics and forecasting of the futures premium. The futures market is an important tool for managing electricity price risk, particularly for stand-alone electricity retailers. The first research paper titled ‘Electricity Futures Markets in Australia – An Analysis of Risk Premiums during the Delivery Period’ provides an empirical analysis of risk premiums of electricity futures contracts during the delivery period for the major eastern states of Australia. While current research on electricity futures markets typically focuses on risk premiums for the pre-delivery period, a specific feature of the Australian market is that as a contract enters delivery, it continues to be traded until expiry. We develop an approach that decomposes the observed futures price during the delivery period into three parts: the crystallised value of the portion already delivered, the expected average spot price for the remaining days of the contract, and the risk premium for the remaining days of the delivery period. We examine the dynamics of realised risk premiums during the delivery period for quarterly and peak load contracts, as well as drivers of the observed premiums such as liquidity-based measures, time to maturity, current and historical spot prices and the historical behaviour of premiums. We find that risk premiums are positive during the delivery period for the majority of the considered contracts. Further, our results suggest that a model using open interest, time to maturity, as well as recent characteristics of spot prices and risk premiums provides relatively high explanatory power for the observed premiums. Our findings are of interest to market participants such as traders, retailers, producers, consumers and hedgers and are relevant, in particular, for risk management and hedging strategies during the delivery period of futures contracts. The second paper, ‘Electricity Futures Markets in Australia: Generating Density Forecasts for Returns of Low Liquidity Instruments’, examines density forecasts of price changes in electricity futures contracts. These instruments, used for risk management, typically exhibit low liquidity during periods of more than one year prior to delivery. We assess the performance of different density forecasting methods, using conventional approaches that are based on historical returns for the considered instruments. We find that such an approach performs poorly and provides inaccurate predictions for day-ahead densities. The poor performance is due to a reliance on return data from a low liquidity period for making predictions relating to more liquid periods. To deal with this shortcoming, we introduce a new approach which enriches historical data for a contract with data from more liquid trading periods of identical contracts traded over the preceding three years. We find that our data enrichment approach significantly improves the correct specification of density forecasts of daily returns based on various evaluation metrics. Our results are of interest to risk managers and parties with exposure to electricity price risk. Our approach is also relevant for market participants who want to appropriately evaluate the risk of price changes for derivatives exhibiting different phases of return behaviour and liquidity, depending on their time to maturity. The third paper, ‘Vertical Integration of Generation and Retail: Foreclosure in the Electricity Futures Market’, presents empirical evidence of foreclosure in the electricity futures market following vertical integration between the electricity retail and generation stages. This foreclosure limits risk mitigation options open to retailers and other participants and has the potential to reduce retail competition and harm consumers. We find a statistically significant fall in base load energy volume transacted on the Australian Securities Exchange (ASX) relating to a delivery period longer than 12 months. At the same time, we do not find a statistically significant change in the volume within the 12-month horizon, and total volume ignoring the horizon. The horizon beyond 12-months is particularly relevant for the commercial and industrial customer market segments as well as for the residential customer segment on contracts longer than 12 months. The reduction in the volume pertaining to horizons longer than 12 months shows that the structure of the futures market became more short-term; focused on the ≤ 12-month horizon. The sample that we use covers the period from 2007 to 2017 for New South Wales, the largest region in the Australian National Electricity Market in terms of energy volumes traded on the spot and futures markets. The impact of industry structure on stand-alone retailers and the potential to reduce competition is of interest to policy makers, regulators, consumers, and retailers with a net exposure to the spot electricity market.


Table of Contents

1. Abstract -- 2. Introduction -- 3. Electricity futures markets in Australia – an analysis of risk premiums during the delivery period -- 4. Electricity futures markets in Australia : generating density forecasts for returns of low liquidity instruments -- 5. Vertical integration of generation and retail : foreclosure in the electricity futures market -- 6. Thesis summary and conclusions -- References.


Includes bibliographical references Thesis by publication.

Awarding Institution

Macquarie University

Degree Type

Thesis PhD


PhD, Macquarie University, Macquarie Business School, Department of Applied Finance

Department, Centre or School

Department of Applied Finance

Year of Award


Principal Supervisor

Stephan Trück

Additional Supervisor 1

Lurian de Mello


Copyright Sami G. Aoude 2019. Copyright disclaimer: http://mq.edu.au/library/copyright






1 online resource (161 pages) graphs, tables

Former Identifiers

mq:71139 http://hdl.handle.net/1959.14/1271257