Demand-side hedging and market participation: looking back, looking forward
How should consumer side contracts be designed to hedge consumers in this new electricity market paradigm? We review some of the seminal works and newer works that offer insights.
Back to the future …
We tend to think of an active demand side as an emergent concept, but its foundations can be traced back to the start of electricity markets and well before.
Prof Fred Schweppe and his co-authors, penned the seminal book “Spot Pricing of Electricity” in 1988, on the theory of LMP and spot power markets. The enablement of consumer side response occupies a central role in Schweppe’s market architecture, indeed essential ingredient #2 of a successful marketplace.
Ten years prior, Schweppe published ‘Power Systems 2000’ (1978), which envisioned the growth of a vision of consumer side generation and storage, heirarchical control of demand and supply, and consumers and utilities as equals in the marketplace all by the year 2000.1
While the prediction was ~20 years or so too early2 many of the raw materials for ingredient #2 are present today - advanced control, communications, and internet technologies, low-cost distributed energy resources/storage, and notably (as part of the datacenter boom) endogenous demand (see Byers and Billimoria, 2026). The latter is particularly interesting because today’s load growth is economically-driven, as opposed to broadly based demographic. This allows for a clearer valuation of electricity demand, based on the economics of the datacenter / AI business model.
Schweppe’s works are most prominent for establishing the locational marginal price (LMP) as the optimal signal for efficient short-term scheduling and dispatch decisions for both supply and demand, extended in distribution grids to the Distribution LMP or D-LMP. Much work has since been focused upon how to extend those real-time dynamic pricing incentives to energy consumers, but Schweppe also recognized the importance of contractual relationships between the customer and utility, particularly for risk hedging.
The importance of hedging, especially in the face of extreme volatility, was borne out in the aftermath of Uri (Private risk and social resilience in liberalized electricity markets). In the ensueing years, a related workstream many works further developed ideas on how contractual relationships between customers and retailers/utilities could work to provide risk management, but also preserve the dynamic incentives of a spot market. Motivated by Lynne Kiesling’s clarion call for a refocus upon the demand side, we briefly review interesting works (many written by Lynne herself) to inform how such a combination of short-term incentives and risk hedging could be achieved for the future.
Tariff design: by regulation or competitive market?
Before we start to discuss contract structures, there is an point on the framing of the discussion as between the regulated retail and competitive retail markets. While many regions successfully deregulated electricity generation, the restructuring of retail operations has proven more problematic. Retail choice stalled due to the Californian electricity crisis (2000-02) and has since progressed slowly. While retail contestability was initially implemented in markets like Australia, NZ, and the UK, some economic regulation of retail rates has been re-imposed.
The framing of the tariff design problem depends upon the state of retail choice. Under retail competition, retailers/REPs have direct exposure to full-strength marginal prices, and thus have natural incentives to design contracts that balance hedging and real-time demand. They own more of the design decision; they can come up with new, innovative designs; test and pilot them, and modify/change/adapt. Tariff design is thus a part of the dynamic market process. For a regulated retailer, regulators play a more prominent role in the design, structuring and approval of tariffs; including more granular details such as default choices etc. This was convincingly put by Travis Kavulla:
Objectives of an optimal contract
Lets forward to Darryl Biggars’ work last year on “Optimal retail contracts with contractible uncertainty” which provides two conditions for an optimal retail contract (Biggar goes on to discuss what an optimal contract looks like, which we return to later):
That the consumer will make efficient operational (production and consumption) decisions by responding to the spot price signal; and
That the consumer is perfectly insulated from risk.
While these represent a theoretical ideal, which may be harder to achieve in practice (especially perfect risk protection), they still give us a focus when considering tariff design.
Insight #1: the importance of marginality
Flat-rate tariffs, still the default in many regions, provide a perfect price-and-volume hedge for consumers but remove any incentives to respond to spot price signals. Here are few interesting ideas from the past, and [very] near past on that seek to do better on both:
Borenstein has written extensively on dynamic real-time pricing, but this piece written in 2007 zeros in on the marginal incentives at play. It finds that with simple fixed-volume fixed-price hedging strategies, the consumer is protected for volality but still remains exposed and incentivised to dispatch effectively. A key issue for residential and smaller-scale C&I consumers here is projecting volumes.
The “affordability option” as proposed by Tim Schittekatte and Carlos Batlle, (techncially an ‘average-rate’ option). This works effectively as a rate cap to the consumer, not for the power prices in individual intervals, but the average of the rate over a longer period, say a month or year. Here the consumer retains good spot market incentives for marginal usage, even when the rate cap is reached.
This is type of contract is currently implemented in practice in the National Electricity Market (NEM) of Australia. Energy retailer Amber offers a quarterly Bill Guarantee that caps the “maximum average usage price” that the consumer pays. The cost of the hedge is built in to the consumers fixed charge. Trials are expanding to V2G / VPPs etc see Greg Williams’ post here.
Created with ChatGPT
In a similar vein Biggar, from the same paper noted above, proposes a set of caps-and-floors as an optimal contract for prosumers with DER / storage under contractible (or known) uncertainty. He also suggests that contracts should expand to be linked not just to electricity prices, but also other sources of uncertainty such as wind speed etc (a concept I am still getting my head around).
Insight #2: Interruptible contracts, applied to datacenters
Rajnish Kamat and Shmuel Oren back in 2001 (leveraging earlier work from Oren, Gedra, Variya and others) presented models for a suite of interruptible contracts that can be created from a combination of forwards and options, which can be modified to include features like early notification etc. This resonates with the notion of flexibile datacenters which may suit both customer and consumer - a LT forward fixed hedge as the base to provide cost certainty but with options for LSEs to call for response under extreme conditions.
Insight #3: Priority service and differential reliability
Many of the financial contracts rely upon the notion of the spot electricity price being the indicator of technical scarcity or surplus. This suggests a continued focus on removing distortions to ‘full-strength’ price formation is warranted.
However there may also be natural limits to the price formation we enable in spot markets. Work by Hans via the Compute Heat Rate suggests that the value of load for certain AI models could exceed tens of thousans (> $50,000/MWh for frontier models). If regulators are unwilling to fundamentally raise or remove price caps (~$5000/MWh in ERCOT, ~$3700/MWh in PJM, >$20,000/MWh in the NEM), this suggests a potentially new paradigm is required.
Hung-po Chao and Robert Wilson in 1987 proposed the concept of Priority Service in electricity markets. The concept here involves a shift away from dynamic prices and towards a suite of special contracts as the tool for consumer response.3 These contracts would develop tiers of reliability which would be actuated and managed by LSEs/REPs - who could develop multiple tiers of reliability. This could be actuated via a traffic light system (see Papavasiliou’s work below)4.
Shmuel Oren and Joseph Doucet expanded upon this to incorporate a compensatory insurance scheme to charge for different levels of reliability. Lynne Kiesling has written extensively on this topic - see here, here and here. My PhD was also on the same topic integrating insurance with the concept of priority service and DER (keystone papers here and here).
Standing on the shoulders of giants…
The notion of flexible demand was a seed from which modern electricity markets sprung. Our new electricity paradigm will undoubtedly require innovation and new ways of thinking. However, we can stand on a firmer footing by making sure we are aware and knowledgeable about the past.
The technical development of load side flexibility was in development well before the economic linkages were drawn by Schweppe and others. One of the earliest references to a scheme for demand response I could find in the literature was Schiller, P. (1941). The control of the domestic load. Journal of the Institution of Electrical Engineers-Part II: Power Engineering, 88(5), 373-389.… setting out “A new method of centralized load control enables the demand of individual domestic installations to be limited to values that can be adjusted progressively in accordance with the system load.”
I recall the saying forecasting is hard, especially about the future, but still some highly portentous views in the paper. I highly encourage a read.
There is a natural equivalence between spot prices and priority service (see here), but priority service may be superior under uncertainty (a recent paper by Oren, Chao and Wilson), or where even extreme limits to prices are exceeded.
Great article Farhad, you might be interested in our ongoing trial. https://www.ausgrid.com.au/transforming-the-grid/innovating-for-the-future/project-edith