A non-parametric structural hybrid modeling approach for electricity prices

Abstract

We develop a stochastic model of zonal/regional electricity prices, designed to reflect information in fuel forward curves and aggregated capacity and load, as well as zonal or regional price spreads. We use a non-parametric model of the supply stack that captures heat rates and fuel prices for all generators in the market operator territory, combined with an adjustment term to approximate congestion and other zone-specific behaviour. The approach requires minimal calibration effort, is readily adaptable to changing market conditions and regulations, and retains sufficient tractability for the purpose of forward price calibration. The model is illustrated for the spot and forward electricity prices of the PS zone in the PJM market, and the set of time-dependent risk premiums are inferred and analysed.

Keywords:

• Electricity market
• Electricity price modelling
• Energy trading
• Supply stack

AMS Subject Classifications:

• Q41
• C14
• C44
• Q47

Acknowledgements

The authors would like to thank Belgacem Bouzaiene-Ayari and Rebecca Zhang for their contributions in this project.

Notes

No potential conflict of interest was reported by the authors.

1 The congestion effect is decreasing over years at least in the PJM market (see Monitoring Analytics LLC Report 2013).

2 www.pjm.com/markets-and-operations/energy/real-time/historical-bid-data/

3 Henry hub natural gas spot price in 2013 (1 January 2013 till 31 December 2013) had an average of with the minimum and maximum . TETCO-M3 natural gas spot price in 2013 had an average of , with the minimum and maximum . Henry hub natural gas spot price from 1 January 2008 till 31 December 2012 varied between and with an average of .

4 As Monitoring Analytics LLC Report (2005) reports, in calendar year 2004 several geographic areas in the PJM Mid-Atlantic region and the Western region experienced frequent congestion and showed high local market concentration. Particularly, the PJM–PS zone experienced 1784 congestion-event hours, the most of any control zone. However, congestion has decreased over years. Day-ahead congestion costs decreased by 43.3% and balancing congestion cost decreased by 5.8% (Monitoring Analytics LLC Report 2013). It has decreased by 51.4% from the first nine months of 2011 to the first nine months of 2012. In fact, PJM has incentive to minimize the cost for everybody, so the decrease in congestion is going to be observed as well over the coming years. Therefore, we expect that congestion will not be the dominant element of in the coming years. This also confirms the importance of focusing on an accurate estimation of .

5 see http://www.rtoinsider.com/pjm-lifts-1k-cap-0114/

6 According to the report dated 3 March 2014, PJM has 1375 generating sources, see http://www.pjm.com/~/media/about-pjm/newsroom/fact-sheets/pjm-statistics.ashx

7 http://www.pjm.com/markets-and-operations/energy/day-ahead/lmpda.aspx

8 If starting from the model in (23(23) ) and (24(24) ) under the physical measure, this corresponds to a time-dependent and piecewise constant market price risk for both gas and coal.

Additional information

Funding

This project is supported by the Public Service Electric and Gas Company (PSEG), through the Andlinger Center for Energy and the Environment at Princeton University.