Implications for Australia’s NEM of the US Court Decision setting aside FERC Order 745 (about Demand Response)

By a confluence of events, Demand Response featured on both sides of the Pacific in recent weeks – though it has been far from front-page news, even in the business section:

1)  On May 23rd (a little more than 2 weeks ago) the US Court Of Appeal set aside a ruling from FERC (the Federal Energy Regulatory Commission) relating to certain specific mechanisms incentivising Demand Response in wholesale markets across the US.

The mechanism set aside is a sub-set of a diverse range of different mechanisms currently available to energy users in different locations across the US (and in Canada). For those who’d like to go to the source, the US Court ruling is here, and the FERC Order 745 is here.

2)  On 4th June (just last week) I attended a one-day conference in Sydney facilitated by the Energy Users Association of Australia (EUAA) that contained a session on different elements of Demand Response in the NEM (some already operational and some proposed to be implemented).

We heard comments from:
(a)   Paul Troughton of EnerNOC, describing a range of approaches available for demand response, and the benefits these methods can bring for energy users,
(b)  We heard comments from Adam Fricker of Broadcast Australia, who explained how they already are providing Demand Response in the NEM, and in WA (i.e. within the context of current market provisions), and
(c)   We heard from Bruce Mountain of Carbon & Energy Markets, outlining some Energy User perspectives on the proposed Demand Response Mechanism.

Though the developments in the US are literally half a world away from the state of the market in Australia’s NEM, there still do seem to be some similarities that we might learn from.  The following are some of the points that I picked up.

With respect these comments, please note that I’m not an economist, or a lawyer – so I may well have missed, or mistaken, some important points.  If you think this is the case, please do let me know – either offline (tel +61 7 3368 4064), or by adding a comment below?


1)  There’s divided opinion – with logic on both “sides”

The ruling in the US Court of Appeals was made by three Judges (Brown, Edwards and Silberman) – with two judges (Brown and Silberman) supporting the ruling, and one (Edwards) disagreeing.

My first observation is that both sides of the ruling (the 16 pages for, and the longer 28 pages against) are argued with considered logic.

Facilitating active participation of the demand side in electricity markets has been one of the enduring challenges for more than 15 years – with other aspects of this participation including:
(a)  Full Retail Contestability (now largely achieved),
(b)  Facilitating energy efficiency; and
(c)  Ensuring the demand curve can be more elastic – through demand response (for instantaneous balance) and ending retail price controls and other implicit cross-sectoral subsidies.

In my own experience, most people are generally supportive of the principle that the demand curve should be more elastic.   It’s important not to lose sight, in the discussion about how to achieve this, that various methods seeking to do this have all grown out of this core principle.

On p105 of the original FERC Order 745 we see that the dissenting Commissioner (Moeller) noted in March 2011:

“While the merits of various methods for compensating demand response were discussed at length in the course of this rulemaking, nowhere did I review any comment
or hear any testimony that questioned the benefit of having demand response resources participate in the organized wholesale energy markets. On this point, there is no debate. The fact is that demand response plays a very important role in these markets by providing significant economic, reliability, and other market-related benefits.”
(my emphasis added)
This quote was also singled out on p10 of the dissenting Judge’s findings.

Each particular method has its own positives, and its own negatives – it’s important to note that these might apply to a particular method (but perhaps not to the principle, as a whole).  In the same way, it’s important to see that criticism of any particular method may not be criticism of the principle of a more elastic demand curve.

The complexity of these arguments is especially understandable in the case of FERC Order 745, given the trade-offs inherent in establishing a mechanism involving the artificial construct of a “negawatt” whose operation in the wholesale market will naturally have flow-through effects in the retail market.

2)   “Wholesale DR” and “Retail  DR”

To implement the rule in the first place, FERC created (see paged numbered 7 of Order 745) a somewhat artificial distinction between what they labelled “retail DR” and a category they labelled “wholesale DR”.

Part of the reasoning for the need for this distinction seems to have been to account for Jurisdictional limitations on its authority.  This dichotomy does also serve a (perhaps unintended) purpose in identifying how two different groups of people tend to view Demand Response.

2a)  Retail DR

In the US, regulation of electricity retail markets is the jurisdiction of the states, and hence outside of the remit of FERC.  On p6 of the Court Decision:

“FERC acknowledges the first case, ‘price-responsive demand’ is a ‘retail-level’ demand response”
(with the implication being that regulation of this being outside of its jurisdiction)

This basket of initiatives would seem to contain any initiative where the price of the energy consumed changes, and so directly drives a reduction in energy consumed when this commodity price is high.  Hence this would include:
i.  Critical peak pricing; and
ii.  Real-time pricing (or spot price exposure).

In the NEM, this type of “Retail Demand Response” has existed for more than a decade and is, in our experience, growing in popularity.

*  In the interests of Full Disclosure, I note that we have been active for more than a decade in facilitating some forms of this “Retail Demand Response” in the NEM.

Specifically we currently (June 2014) supply software to energy users regularly consuming more than 2,000MW of load – some of which does, on a reasonably regular basis, respond to high wholesale spot prices by curtailing consumption.   Specific curtailment decisions are made by the individual companies based on a diverse range of factors – including electricity prices but also supply & demand variables in their particular industry.   The specifics of these decisions are confidential.

The ruling of the US Court of Appeals did not seem to have anything to say about this form of demand response – which is understandable, as it was not in the scope of the appeal.

To some extent, I suspect, this aspect of demand response is easier to understand and has clearer logic – and hence is less contentious.

2b)  Wholesale DR

In the US, regulation of electricity wholesale markets does fall under the jurisdiction of FERC, under Federal law.  On p6 of the Court Decision we also see:

“In contrast, FERC dubs a reduction in the consumption of energy in response to incentive payments a ‘wholesale demand response’.”
(with the implication being that regulation of this being outside of its jurisdiction)

Using this generic definition, what’s envisaged in the proposed Demand Response Mechanism in the NEM could also be defined as a “wholesale demand response”.  In the same way, some of the complexities that seem to have resulted in its implementation in the US will also apply in the NEM.

To make such an incentive system work, it seems that a new product (not just MWh consumed) must be defined – hence the creation of the “negawatt” concept, which can be dispatched into a wholesale market (effectively a $/-MWh figure).

Applying some form of additional incentive to achieve results that policy makers deem worthwhile is not something unique to such “wholesale demand response” – the wider and more vocal debate about the future of the MRET Scheme in Australia (discussed here) is another topical example, for instance.

In recent times we have seen a number of industry participants caution about the imposition of a number of incentives on the top of the underlying energy market without considering how the energy market will, in aggregate, be affected:
(a)  Paul Simshauser of AGL commented about this in this forum (another aspect of which I commented on here); and
(b)  Keith Orchison notes that Frank Calabria of Origin Energy made similar comments at a forum in Sydney.

One interpretation of comments such as these is that these incumbent gentailers are threatened by developments occurring in the market and, as such, are predisposed to criticize.  However let’s not rush to “shoot the messenger” before we consider what the message is that they are trying to convey?

3)  There were two reasons FERC 745 was set aside in the US

Back in the US, the two judges supporting the ruling are clear that there are two reasons why the FERC order has been set aside.

3a)  Jurisdiction

The more obvious part of the ruling is that FERC was declared to have overstepped its jurisdiction, by regulating operations that were seen to be in the retail market, and not the wholesale market – i.e. the distinction that FERC had made was seen as somewhat arbitrary.  On p11 of the Court Decision we also see:

“Thus, while it is true demand response can occur in two ways – through a response to either price change or incentive payments – nothing about the latter makes it ‘wholesale’.  A buyer is a buyer, but a reduction in consumption cannot be a ‘wholesale sale’.  FERC’s metaphysical distinction between price-responsive demand an incentive-based demand cannot solve its jurisdictional quandary.”

Thankfully Australia does not suffer from that particular type of federal/state demarcation – though we have enough of our own structural inefficiencies to worry about….

3b)  Equity

Perhaps of more universal interest, however, is the second reason given for the rule being vacated – that being that the compensation arrangements specified have been unfair.

On p14 of the Court Decision we also see:

“Alternatively, even if we assume …” [that the above was not the case] “… Order 745 would still fail because it was arbitrary and capricious”.

Specifically with respect to the FERC order, the reasoning given here is that the compensation received by the energy user equates to an explicit payment equivalent to the wholesale spot price, plus an implicit benefit equal to the reduced cost of energy in the retail market priced at the energy user’s retail rate.  On this basis, it is seen as not equitable to the remuneration paid to energy suppliers which contribute to meet the same supply/demand balance.

In the context of the proposed NEM’s Demand Response Mechanism, this specific finding has been avoided because it has been ruled that the energy user will pay the full retail invoice for consumption, assuming that the demand response had not happened.  This seems to be a notable difference between the FERC and NEM approaches.

There were also questions of equity, given the fact that generation supplies have short-run costs of operation, whereas (it is argued) that demand response incurs no similar costs.  However, the dissenting opinion does raise the challenge of barriers to entry (in terms of the establishment of systems for the delivery of demand response) as a means of pointing out that the provision of demand response is not without cost.

It was interesting to read (though certainly not a gripping thriller) as the dissenting judge (Edwards) walked through his own logic on this particular issue.  Many of the 28 pages of dissenting opinion are devoted to this challenge.  Hence it’s not really possible to summarise in a paragraph.

Interestingly, on page 26 of the dissenting judge’s opinion we also see:

“This court has not business second-guessing the Commission’s judgement on the level of compensation” and goes on to quote several examples of where deference has previously been given to FERC’s expert view.

In shorthand, I take this as agreement that this issue is complicated.

Sitting outside this US debate and looking in, it seems probable that the ruling will be appealed.  If this is the case, it is naturally uncertain what the final landing will be.

In more general terms, this particular point of contention seems, to me, to be symptomatic of the difficulties that will inevitably come when seeking to artificially create two markets (one for megawatts, and one for negawatts) when physics dictates that there is really only one commodity delivered.    Whilst it might be decided, by policy makers, that the potential benefits make such an introduction worthwhile, it should not be expected that the implementation will be without hiccups and compromise.

4)   Sample Reactions to the Court’s Ruling

Following the Court’s ruling, and particularly because of the uncertainty surrounding the proposed Demand Response Mechanism in the NEM, I’ve been interested to review the reactions (at least those made public) from those associated with Demand Response in the US.

The following is a sample:

4a)  EnerNOC

Given EnerNOC also operates in Australia, I was particularly interested to read EnerNOC’s comments.   There must be something I am not understanding here, as EnerNOC reveals that the revenue they earn is a small percentage of overall revenue:

“Energy payments that are the subject of Order 745 have not been a material component of EnerNOC’s revenues. Of EnerNOC’s approximately $1 billion of revenue over the last three years, these payments have represented approximately 2% of those revenues”

I have not had time to dig into the details (perhaps someone can help me here?) but it makes me wonder if the EnerNOC revenue is mostly earned in three other areas:

(i)  In Demand Response, more broadly:

With respect to demand response, we believe that the majority of its revenues are supported in other areas – with the big two we can think of being:
1.  network-focused DR (with payments made by the transmission providers and not the market operator), and
2.  in energy-focused DR, but through capacity payments (i.e. payments to be available, without needing to be actually called on), rather than compensations arrangements based actual triggering, and on spot price.

(ii)  Outside of Demand Response

Additionally, we’ve observed EnerNOC’s positioning expand beyond just the provision of Demand Response, and so would believe that some proportion of its revenue is not directly related to demand response.

Perhaps the surprisingly low revenue share might be because EnerNOC has a low MW market share in providing this particular product, or perhaps it might be because US power markets mostly have significantly lower market price caps than those that exist in the NEM, hence lower revenue share for EnerNOC stemming from payments at the LMP (spot price).

Whatever the case, it did surprise me that this type of DR seemed less significant than I expected – prompting a question about how significant the proposed Demand Response Mechanism for the NEM might actually turn out to be?

4b)  Comverge

Another significant demand response aggregator in North America is Comverge (who, as much as I am aware, are not active in Australia currently).  Their approach is somewhat different from that of EnerNOC, but not to the extent that we’d think that they would not be affected by the ruling.

We did find the following public comments relating to the FERC order:
(i)  this tweet here from back in June 2012 noting FERC 745 would play “a meaningful role”;
(ii)  this independent commentary here about Comverge’s business development coincident with Order 745, though not explicitly linked;
(iii)  and a comment from January 2014 that calendar 2013 had been a good year for demand response more generally (and Comverge in particular) – though again no specific attribution to Order 745.

However we could not find any public statement about the recent court ruling, or anything specifically indicating how significant Order 745 was to the Comverge’s provision of DR.

4c)  Johnson Controls (now including the business of EnergyConnect)

It’s our understanding that EnergyConnect (now owned by Johnson Controls) is a Demand Response provider in PJM, and in other locations.  However, we have not been able to locate any comment from them about the court ruling – or with respect to FERC 745 either.

4d)  Landis + Gyr

Landis & Gyr is a company that has been historically focused on provision of electricity meters.  However we also note an involvement with demand response.  In the brief time invested in searching for specific comment about FERC order 745, we could not find anything significant.

4e)  AEMA (Advanced Energy Management Alliance)

The four demand response providers above, along with energy users Walmart and Alcoa, appear to be the main members of the Advanced Energy Management Alliance.

We noticed this comment, published soon after the court ruling, that makes a staggering claim:

“In 2013 alone, demand response saved electricity users in the mid-Atlantic $11.8 billion, according to a report on the market effectiveness of the PJM Interconnection by that region’s independent market monitor.”

Being mainly focused on the Australian National Electricity Market, I recognise that the number ($11.8billion) is greater than the entire volume of electricity traded in the NEM over any given year. Hence this stated claim of benefit is truly staggering.

I presumed that this claim was made with reference to this this report by Monitoring Analytics covering the 2013 calendar year, however I cannot find such a large number inside that report (or, indeed, any statement on the benefit of DR).

Further digging sees that a similar claim was also made 2 months earlier by the AEMA here:

“The Market Monitor has reported that demand response saved consumers in PJM $11.8 billion in the 2013 capacity auction, yet the demand response recommendations in the “PJM State of the Market 2013” report issued on March 13 do not reflect the realities of the state of today’s demand response capabilities.”

In this case, the link included references section 5 of this earlier report by Monitoring Analytics (covering calendar 2012).  Hence I am a little confused…

With reference to the 2012 report, p2 of Volume 1 notes that “In 2012, PJM had total billings of $29.18 billion, down from $35.89 billion in 2011.” Hence an $11.8 billion saving would amount to a massive 40% of the total value of the energy billed by PJM during the year.

In the same report, p18 of the notes that “average real-time generation in 2012 increased by 3.4 percent from 2011, from 85,755 MW to 88,708 MW”.   Across 2012, then, this would equate to 777,082,080MWh.  Spreading the claimed $11.8b benefit across this volume equates to an average reduction in wholesale prices charged of $15.18/MWh – a magnitude that just does not seem credible for what the market price would have to have reduced, across all 8,760 hours (and just attributable demand response).

Perhaps it is a simple case of a typo, or a calculation error, based on data in one of the two reports (2012 or 2013) linked above?

Or perhaps I have grossly misunderstood something very big? If someone can explain what I’m not understanding, here, I would be very keen to know?

4e)  Viridity Inc

I did note that Viridity (another demand response provider) did publish this white paper early in 2012 following the initial FERC ruling – but I can’t find any response to the Court ruling.

4f)  Response from others

Amongst the commentary I’ve come across, following the court ruling, I found the following of particular interest:

(i)  Lawyer Scott Hempling suggests these, as next steps that could be taken to ensure that Demand Response is implemented successfully;

(ii)  I found Jesse Jenkins’ summary useful

(iii)  Michael Kanellos warns that the setback to demand response is likely to be temporary, with the ruling seen as another spur for storage technology and optimisation software.

4g)  Anyone else?

We have looked at a number of other companies facilitating demand response in some way in North America.

However the few comments we can find with respect to FERC Order 745, and the Court’s decision, don’t seem to add anything significant of value, so we have not taken the time to list here.

If you know of other Demand Response providers in the US that have been affected by the recent court ruling, and who have spoken out about it, please let us know about this – either offline, or through the addition of a comment below?

Summing up, there does not seem to be anywhere near the level of concern expressed (at least in public) as we have seen in NEM with the current RET Review process, and the palpable concerns that the review will lead to a massive reduction (or even the complete abolition) of the MRET scheme.

In general terms, it seems that the majority of Demand Response currently provided in US markets is primarily supported by other mechanisms – with the main support seeming to come from Capacity Markets.

The general mood of the commentary I have read seems to be that it’s a hiccup (even if any appeal is unsuccessful) that’s not going to break the bank of demand response providers, and should not derail longer-term momentum – as another commentator here notes that  “Let’s make it clear – Demand Response existed before 745 and it will continue to exist after 745.”

5)  The proposed “Demand Response Mechanism”

The following are some comments made with respect to the proposed new Demand Response Mechanism (DRM) for the NEM.

*  Also in the interests of Full Disclosure, I refer to prior comments about how we strive to remain technology agnostic in the service we provide to a broad range of clients inside, and outside, of the NEM.

Should the proposed new DRM proceed to implementation, it will become one more technology source that AEMO will take into account in balancing supply and demand.  It will be a non-scheduled resource, so AEMO won’t be dispatching the response, but rather taking account of the response in carrying out their responsibilities of operating the market, and keeping the lights on.

At that point (i.e. when Energy Users can actually use it for their benefit) we will be keen to assist energy users utilise this additional approach, as an another avenue they might use (through an aggregator, or direct with the AEMO) to unlock the value in their price-sensitive load flexibility.

If you have other thoughts, we’re keen to hear them from you – either as comments below, or directly with us offline (tel + 61 7 3368 4064)?

5a)  An additional mechanism for Demand Response in the NEM

Already in the NEM, there are a number of different ways in which energy users can operate Demand Response in the context of the NEM.  The following is a starting list of some of these (the list not include DR from the perspective of network utilisation).

(i)  Call option for Demand Response through a Retailer

It’s my understanding that this is one of the more widely used arrangements – whereby a retailer may choose to offer a benefit-sharing arrangement to energy users who respond to calls to reduce consumption when the retailer calls.

Under this arrangement, there needs to be a number of events occur, sequentially, for demand response to be triggered:
1.  The price (presumably) needs to rise above some trigger level;
2.  It needs to be in the retailer’s interests to trigger the demand response (for instance, if the retailer is “long” spot-exposed generation capacity then it might want the price to remain high to maximise its revenue from self-generation); and
3.  If the retailer calls, the energy user needs to opt to respond, and be able to do so in a timely manner.  On occasions it may be that it’s not commercially worthwhile responding, for the energy user (in such cases the energy user does so with no penalty, other than forgone benefit).

Because of the uncertainties involved in the above, this type of response is distinctly non-firm – however balancing this non-firmness is that it might be the easiest arrangement to put in place.

We do know of energy users who operate in this manner – some of whom have told us that they appreciate the payments so much that they try to anticipate when they might be called to the point that they have as much equipment running as possible to maximise their benefit from turning off when called!

We’ve also seen some retailers (which notably don’t have generation assets of their own) focus specifically on this type of approach as a hedging strategy.  Triggering demand response within its own portfolio enables an energy user to address both:
1.  the price risk, because of the mitigating effect DR can have on prices (if it’s large enough); and
2.  the volume risk inherent in the non-firm consumption decisions made by energy users in their portfolio.

(ii)  Spot price pass-through via a Retail arrangement

Over the 15 years we have been in business we have come across a growing number of energy users that take a step further, adopting an approach of accepting some form of spot price exposure in their retail arrangement.  In this case the energy user can avoid the risk premium charged by the retailer for managing their price and volume risk and, through physically hedging (by demand response) avoid the inflationary aspects of the (relatively) few (very) volatile periods that occur each year.

Of these energy users, we work with a significant percentage, providing them software and assistance to help them understand when & how to curtail when prices are high, so that they can save significantly on their average cost of energy.  Back in 2009 we posted this explanation of the benefits of curtailability in relation to the NEM.

Over more than a decade we have seen some energy users become very sophisticated in the approach they take to optimising their operations in relation to a volatile electricity spot price.

On occasions we have come across energy users who opt for spot exposure, without implementing any real form of physical (or financial) risk management strategies.  This is not an approach that we’d advocate, given the risk inherent in the volatility in the NEM.

(iii)  Registering with AEMO as a Wholesale Market Customer, for exposure to the spot price

An energy user might choose to take this approach a step further by eliminating the retail relationship and registering with the AEMO as a wholesale market customer.

There are a limited number of energy users who choose to operate in this manner.  Because of the costs involved in participating in this way, it has tended to be only large energy users who have walked down this path.

See the upcoming update of our “Power Trading Schematic” Market Map to see these energy users that are so registered.

(iv)  Registering with AEMO as a Dispatchable Load

A Wholesale Market Customer, registered with the AEMO, still is a non-scheduled demand response (i.e. it consumes whatever it wants, knowing only that it has to pay the spot price for what it consumes).

It is possible for loads to register as dispatchable loads, which requires them to bid into the market to consume.  Currently it is only the pumped storage loads at pumped storage hydros that operate in this manner.

5b)  … hence the “DRM” is a poor choice of name

I believe that the new mechanism has been poorly named, and that this risks confusing people (inside the NEM, and outside) when the program moves into implementation stage.

There are already a number of ways in which Demand Response can participate in the NEM – the new DRM just provides one more avenue, in the belief that this new mechanism will add more demand response to the amount already active in the NEM.

Hence a name that reflected this additive nature (instead of implying singularity), and specifically reflected what’s different in this new approach, would make it easier to understand – and contrast with existing methods.  Perhaps something like “Self-Scheduled Negawatt Mechanism”?

Whilst on the topic of naming, we also note that the provider of these self-scheduled negawatts is labelled a “Demand Response Aggregator” in the draft operating procedures, even though I note that it might be an individual energy user that chooses to participate directly with the AEMO, rather than through a third-party aggregator.

In an effort to make the new method as clear as possible, could I suggest also that the new Market Participant Classification at the AEMO be named something like a “Negawatt Provider”, making it clearer that that provider might be an energy user themselves, or some aggregator acting on behalf of a number of energy users?  There might be similarities, here, to the Small Generator Aggregator Framework?

5c)  Don’t expect an exponential increase in involvement

In the digging I did into the “State of the Market Report for PJM” for 2013, it was of interest that I read the following passage in Section 1, page 2:

“The high demand days in the summer of 2013 highlighted the fact that demand resources are not full substitutes for generation for several additional reasons. This inadequate definition of demand resources created operational difficulties for PJM in responding to high load particularly in specific local areas. The need for the announcement of emergency conditions, two hour lead time, two hour minimum dispatch period, availability of demand resources only from 12:00-20:00, maximum number of events allowed each delivery year, and lack of nodal mapping are inappropriate limitations on demand resources that should be removed in order to ensure that demand resources serve as capacity resources and are available to resolve reliability issues when necessary. To address these issues the market rules should be modified so that demand resource dispatch is nodal to permit more effective dispatch of such resources, that demand resources are considered an economic resource rather than an emergency resource, that demand resources are available year round and that demand resources have a shorter lead time”
(my emphasis added)

It is a credit to those who have progressed the design of  the DRM for the NEM, thus far, that they have avoided some of the issues flagged above.

However there are a number of reasons why the NEM is likely to see an uptake of negawatt supply under this proposed DRM that builds on, but does not radically transform, the amount of demand response participation that we already see in the NEM today.  Particularly because the unit size of the loads that are drawn into demand response might be an order of magnitude smaller than some we already see operating, the number of suppliers might increase significantly (perhaps through an aggregator) but the increase in MW supplied, especially on a regular basis, won’t increase by the same degree.

This will be the case, because of a number of nuances inherent in the way the NEM has been designed to operate:

(i)  Energy Only Market

The NEM operates as an energy-only market – where generators are supposed to recover revenues sufficient to meet both fixed and variable costs through the half-hourly Trading Price for electricity.  This is different from markets in many other parts of the world (including in most of the US) where a Capacity Market operates in conjunction with the Energy Market, and is designed to cover a generators fixed costs of operation.

It’s our understanding, in the US, that much of the participation of Demand Response is facilitated by Capacity Markets.  For instance, the “State of the Energy Market for PJM” report for 2013 highlights (p197) that “The capacity market is the primary source of revenue to participants in PJM demand response programs”.

The same report for PJM notes registration levels peaking at 3,166MW (May 2011) in the Economic Program from 2010 through 2013 – this program might be seen to be most equivalent to the DRM proposed for the NEM.  In contrast, the PJM cleared well over 10,000MW of demand response resources bid into its capacity market (7% of peak demand in 2013).

In the absence of a Capacity Market, in the NEM, it’s probable that demand response won’t deliver the same levels of involvement as are quoted in markets in North America.

In the NEM, given the larger ongoing debate about the level of support that should be provided to intermittent generation supplies (such as with renewables), it’s possible that calls for the introduction of a Capacity Market in the NEM to support the increased variability in output that might eventuate for all fuel types (should storage not arrive at economical cost over the same time horizon).

Should these calls be acted on, then this would serve as a catalyst for the registration of more demand response resources, that might only operate in extreme cases.

(ii)  Real-Time Pricing

One other potential barrier to entry that will dissuade some demand response resources from participating is that dispatch prices in the NEM are determined only at the start of that five-minute dispatch interval.  As such, generators only receive their dispatch targets at the beginning of each dispatch interval – and, similarly, demand response resources is expected to react instantaneously to receive full benefit from curtailment.

Prior to a dispatch interval, the AEMO does publish predispatch prices – but these prices are non-firm, and (because their purpose is a feedback mechanism to enable supply and demand to optimise their positions leading up till real time) they do tend to vary, particularly when supply and demand are tight (and/or impacted by constraints) and so prices are volatile.

It’s been our observation that some potential demand response providers seek two things in parallel:
1.  Price certainty; and
2.  Some advanced warning (sometimes hours, sometimes even 24 hours) of the need to curtail.
This combination won’t be attainable under the current NEM rules.

In our own experience, in assisting a number of energy users in Ontario in the provision of demand response, we observed the introduction of various programs designed to meet the same paired requirement.  From afar it’s appeared the same requirements have also been noted in other North American markets – as the comment quoted above from the PJM’s market monitor indicate.

Again, I note that I’m not an economist, so the following may be incorrect in some way – however my understanding is that the value of the (supply, or demand) response is highest if it can be delivered at short notice, and that the value of that response reduces the more the notice period required.  Hence, there appears to be a paradox that some demand response resources won’t be able to surmount to participate in the NEM.

Perhaps some aggregators might try to address the shortcomings of these slow responses, by aggregating responses from different providers – however not without accepting considerable risk for the firmness of the aggregated response, and the non-firmness of the pricing outcomes.

(iii)  The 5-minute dispatch and 30-minute settlement issue

Finally, there is the old chestnut of the 5-minute dispatch and 30-minute settlement which means that a wholesale market participant is only really aware of the firm price they receive (for production) or pay (for consumption) until 26 minutes into the 30-minute trading period.

Because of the extreme nature of volatility in the NEM (with dispatch prices capped at $13,100/MWh currently) this can mean a spike in the last five minutes can add over $2,000/MWh to the trading price for the period.  This is especially of concern if the price spike is not forecast – this particular issue for summer 2013 in Queensland was addressed in our extensive review of what happened over that period.

I wonder how much longer we’ll have to wait before we see someone raise a rule change request at the AEMC to address this issue?  It does seem to be in the interests of a broad range of participants to see this resolved.

5d)  Questions about implementation of the DRM

Given our keen interest in assisting energy users to achieve the best outcomes possible for them under the market frameworks that exist at a particular time, we have taken a keen interest in how the proposed DRM has evolved.

This interest has had to be moderated by our need to focus on ways we can help “here and now”. Hence, it may be that we’ve missed some of the discussion that’s been held with respect to the following points – if so, please do let us know (in comments below or offline via tel +61 7 3368 4064).

(i)  Non-firm consumption, and non-firm reductions

It’s our understanding that the demand response triggered under the DRM will not be dispatched by the AEMO – and that this will occur at facilities which are not required to bid firm consumption decisions into the AEMO ahead of time, either.

The energy user (or aggregator acting on their behalf) will inform the AEMO of a curtailment period just prior to, or even within, a trading period in which they have elected to curtail.  The notional “without DR” consumption will be established, after the fact, through a baseline methodology – and from this the deemed DR will also become known.

The fact that we’ll be effectively twinning together two non-firm commodities (Inflated Energy Consumed charged at the retail rate, and Notional Demand Response paid the spot price) raises a number of questions about how the obvious risks of arbitrage can be effectively managed – especially given the large price difference.

In some quick reading to update my understanding of the US situation I did come across this article talking about two settlements at FERC involving large civil penalties for companies (one energy user, one aggregator) that artificially inflated baselines to game demand response payments. Is this a sign of what the DRM might bring in the NEM?

(ii)  Measuring baselines

A properly functioning DRM would seem to resolve around the derivation of an accurate – or at least a generally accepted – baseline for each demand response provider.

It does seem somewhat ironic that the more active the demand response, the harder it would seem to be the task of establishing an accurate baseline of “what we would have done” in the absence of the DR payment.

(iii)  Any move to firm consumption, and firm reductions?

Thinking into the future, perhaps these challenges would be resolved if (as the DRM matures) it evolves to an arrangement whereby:
(a)  Energy Users in advance of a dispatch interval commit to a firm level of consumption – perhaps paid at the firm retail rate arranged with their retailer; and
(b)  Energy Users in advance of a dispatch interval commit to a firm level of demand response from that notional commitment to consume – paid the non-firm spot price for that trading period.

These firm amounts could then be taken account of by the AEMO in the dispatch of the market, and hence reduce another uncertainty that would otherwise act to increase aggregate demand forecast error.  The starting values could also be published by the AEMO, in the same way that generation SCADA data is published within each dispatch interval now, to increase the transparency of this particular technology supply option.

The net consumption figure that would result (as a firm commitment) could be easily compared with the actual metered outcome to enable settlement to be made for any under-consumption or over-consumption.

Putting both halves of the whole together, however, we do wonder how this would really differ from a customer operating as a dispatchable load (the framework for which has long existed in the NEM)?

Perhaps the difference might be that the scheduling process might take place at the customer’s initiative, and not for all dispatch intervals across the year – borrowing the periodicity from the semi-scheduled nature of the newer wind farms in the NEM?

We will watch, with keen interest, to see how this proposed implementation plays out.

About the Author

Paul McArdle
One of three founders of Global-Roam back in 2000, Paul has been CEO of the company since that time. As an author on WattClarity, Paul's focus has been to help make the electricity market more understandable.

11 Comments on "Implications for Australia’s NEM of the US Court Decision setting aside FERC Order 745 (about Demand Response)"

  1. You deserve a gong from the Brits’ dear Queen for the effort above, Paul.
    Much to chew over for many people. Thanks.

    • Hi Steve

      I’d be surprised if old Mrs Windsor had an interest in something as arcane as Demand Response – though I had heard that she takes a keen interest in switching off lights in that big old house of hers.


  2. Grrat discussion on DM Paul.
    There are many parties in the NEM who dont want to see DM succeed that’s why we have had such a long process regarding the Power of Choice review.
    The other improtant role that DM plays in electricity netowrks and its good to see the AER incentivising DB’s for DM solutions; howver there should be a lot more of it.

    • Thanks Peter

      Some methods of Demand Response are already succeeding in NEM (for instance, spot exposure for large industrial energy users – a method we have experience with).

      The proposed new “Self-Scheduled Negawatt Mechanism” (bit of a mouthful – maybe the “Self-Scheduled Demand Sellback Mechanism”?) will be another method available, in the energy market, should it be introduced.

      You note about the network side – this has not really been our focus, to date. Hence we’re less able to comment on how much of that is being used.

      Without some form of dynamic network pricing, I suspect this will be problematic to introduce equitably – but with the debate going on about the see-saw of network flows at a household level, perhaps there will be other reasons to move towards network charges that are:
      1) More related to kW than kWh; and
      2) Change reflecting the loading on the network in that particular part of the world.

      Mike Swanston talked about part of this in his “size of pipe” analogy in the forum I posted about here. Click through to the podcast.

      There are technology offerings that could be developed, if this would be the case, to automate the response. Economic storage would help, as well, with a couple different models in terms of how it could be rolled out.


  3. Hi Paul,

    Thanks for a thought-provoking post! I have a couple of clarifications (as requested), then some comments.

    Significance of the capacity market to DR

    In part 4a, you asked where EnerNOC gets its revenues from if only 2% is subject to Order 745. All the other sources you mention play a part, as does our growing Energy Intelligence Software business, but capacity payments are the most significant. However, I don’t agree with your inference that this means that we’ll see correspondingly low uptake of the DRM in the NEM.

    Where there’s a capacity market as well as an energy market, peaking resources (both conventional generation and DR) derive almost all their revenue from the capacity market. If it’s not mandatory to participate in the energy market (in PJM, it’s not for DR), then they many not bother with it.

    If you don’t have a capacity market, then peaking resources have to make do with the energy market and associated derivatives markets. So long as the market is functioning correctly (I realise that’s a big caveat), at equilibrium they should get the same amount of revenue this way. The NEM isn’t quite a perfect market, so I wouldn’t expect participation to reach as high as it ideally would, based on fundamentals, but the potential is still significant.

    Huge consumer savings from DR in PJM

    In part 4e, you asked whether the reference to PJM saving $11.8b, or 40% of total costs, through DR participation was a typo or a calculation error. It is neither. The independent market monitor calculated this by rerunning the 2013/14 auction without the bids from DR and EE suppliers — i.e. procuring the capacity needed from generation alone:

    “Based on actual auction clearing prices and quantities and make‐whole MW, total RPM market revenues for the 2013/2014 delivery year were $6,708,567,045. If no DR or EE had been offered into the auction, total RPM market revenues for the 2013/2014 delivery year would have been $18,535,847,876, a difference of $11,827,280,831 compared to the total based on actual results.”

    This is from p.52 of Monitoring Analytics’ Analysis of the 2013/2014 RPM Base Residual Auction Revised and Updated. PJM’s auctions are run three years ahead, to give time for proposed resources to be built if they clear. Hence we know the corresponding figures for the next three years: $9.6b, $13.7b, $10.1b. Pretty soon, you’re talking real money. These savings show how much cheaper it is maintain reliability if you involve the demand side, rather than relying purely on the supply side.


    As I hope I made clear in my talk last week, the kind of price-responsive demand with which you assist your customers is by far the best option for the small number of very large customers who are sophisticated enough to cope with the associated risks. By buying at wholesale prices, avoiding paying a margin to the retailer for risk management, they can come out a long way head.

    The purpose of the proposed DRM is to allow many more customers to participate by introducing an alternative approach that doesn’t require them to take market price or volume risks 24x7x365 just because they want to be able to provide some demand response during extreme events.

    I don’t think that introducing the confusing “negawatt” term, or suggesting that it involves the creation of a separate market for a separate commodity, really helps. There’s only energy involved. As you noted in part 3b, there’s no incentive payment; it’s not an incentive scheme.

    One thing I would add: ideally, demand response should be a scheduled resource. This makes it much more useful to the system operator than price-responsive load. I haven’t seen the final version of AEMO’s rule change proposal. However, in the working group discussions before AEMO went away and wrote it, it was clear that the DRM was proposed to be non-scheduled only because AEMO considered it too hard to alter the NEM’s dispatch systems to support load participation in the near term. The intention was that this would be revisited once they had some initial practical experience with the DRM.

    AEMO’s phased approach makes sense. It takes a while to get the implementation details right, just as it does for the market to adjust to broader participation, and for customers to take advantage of their new opportunities. This is why it is important that we start soon, rather than waiting until the next time it’s obvious that we’re paying over the odds due to the low level of customer participation.



    • Excellent, thanks PaulT

      Will think through the rest of what you say later, but want to ensure I understand the savings noted in the PJM capacity market, as that’s what really caught my attention!

      Using the numbers you have quoted, and an installed capacity in PJM of about 180,000MW (think that’s about right – should be right enough to see that I understand the orders of magnitude):

      Under the Reliability Pricing Model (RPM) we have two scenarios:
      Scenario 1) Real = Total payment of $6.7b for a year across 180GW is $37,000/MW p.a.
      Scenario 2) A re-calculated “no DR or EE” scenario = Total payment of $18.5b for a year across 180GW is $103,000MW p.a.

      These numbers are reasonable – in the ballpark of what I’ve seen of the annual price for capacity in Western Australia, for instance.

      Am I correct this far?

      Can I assume that there’s about 20,000MW of Demand Response and Energy Efficiency in PJM? I think it’s less than that but the number should be good enough to allow me to get my head around this.

      Assuming 20,000MW of physical generation pushed out at the top of the bid stack by the DR and EE, this means that there was about that much that bid the difference between the two total revenues.

      This means 20,000MW bidding, collectively, $11,827,280,831 – or $591,000/MW p.a.

      Given it’s a capacity market, I assume this can be peaking capacity, but even if we assume capacity is built for $1,000,000/MW (expensive for a peaker, I would think, even with price escalation) this would mean that there are generation developers out there pitching for a 2 year payback on investment.

      What am I missing?



      • Hi Paul.

        The maths works out a bit differently as it is not a pay as bid auction, everyone is paid at the clearing price.

        The additional cost is spread over all 180 GW of generation, not just the 20 GW in question.

        There will have been 20 GW of generation bidding between $37,000 and $103,000 /MW p.a., which changes the price paid to all cleared bidders, results in increased costs of 180,000 MW x (103,000 – 37,000) $/MW p.a. which is $11.8 billion p.a., making the 20 GW of DR pretty close to the mark.

        The payback period would then be close to 20 years, depending on discount rates etc.

        If those same 20 GW of generators were bidding at up to $591,000, then that would be the clearing rate, and total costs would be ~$120 billion

        • Thanks Scott

          That makes more sense – apologies for not figuring it out myself.

          Are bid prices made public, after the fact? I would be interested to look at the bid stack, to understand more …

          160,000MW of physical plant to bid at or below $37,000/MW p.a. – at 100% utilisation this is only $4.22/MWh capital cost (at most), unless I have made another mistake? That seems way too low to cover fixed costs.

          If this is true, then the approach seems to be to under-recover in the capacity market and recover more in energy (unless the entire fleet is depreciated away)?

          {please excuse the questions, if they are naive – we don’t have a capacity market to work with in the NEM}


          • Hello Paul, generators don’t run at 100% utilization, but the capacity revenues are still a small portion of their revenues compared to the energy market revenues. For DR, however, which is only called a few hours a year, capacity is the vast majority of its payment stream.

            Here is a blog I wrote on the FERC 745 ruling and the recent PJM auction:

          • Hi Paul.

            Unfortunately, PJM doesn’t release bid data, even after the fact. I believe this may have something to do with the fact that their auctions take place over several rounds, with participants able to adjust their bids at several fixed times.

          • In relation to your comments about Capacity revenue matching up with fixed costs and energy revenue matching up with variable costs, I don’t think many generators will approach it from that perspective.

            A peaking generator will earn exactly as much capacity revenue as a baseload generator rated at the same wattage, but will actually earn less energy revenue, because it is only on and earning energy revenue a fraction of the time, whereas baseload is earning energy revenue during low price periods as well as high price periods.

            Peaking generators tend to have lower fixed costs than variable generators, but higher running costs.

            So a peaking generator, with its low fixed costs, high variable costs and low energy revenue may end up approximately covering its fixed costs with its capacity revenue, and then the energy revenue is profit on top.

            A baseload generator however, has higher fixed costs but the same capacity revenue – they likely won’t come close to break even there, and they don’t particularly intend or need to. The energy market is (as Brett points out) where they make the big money.

            DR, which earns energy revenue a lower fraction of the year than a peaker is even more reliant on capacity revenue (again, Brett got to this point before I did)

Leave a comment

Your email address will not be published.