Two alternate theories of why capacity was withdrawn

This article was written today in conjunction with this pictorial review of bidding behaviour over the past 16 days … but (because we’ll be referring back to it frequently), I thought it would be useful to publish as a separate article in its own right.

Readers should keep in mind that I’m only part-way up a steep learning curve with respect to this aspect of NEM operations there may well be errors or other shortcomings in this quick note.  If you’re a more knowledgeable reader, then please feel free to add a comment below to correct or clarify any shortcoming you see here.

Based on my current understanding, there seems to be two quite different theories that might account for why a significant number of DUIDs withdrew capacity* after Administered Pricing commenced on Monday 13th June for NSW, SA and VIC … leading to the AEMO decision to Suspend the Market on Wednesday 15th June.

* note ‘withdrew capacity’ refers to the NEMDE dispatch/pricing run … this capacity was still bid available to AEMO to enable them to Direct as requires

There may well be other explanations, as well – but (even just for these two theories) it seems quite important that stakeholders are aware that they both exist, and both seem quite plausible in various ways.


Question… Theory #1
(based on compensation arrangements)
Theory #2
(the need not to run out of fuel or energy stored)

Who’s talking about this theory?

Pretty much everyone!

In simple terms it seems that this has quickly become the generally quoted theory for why participants withdrew significant volumes of capacity offered into the market

… but just because it’s most talked about, that does not necessarily mean it’s correct in all cases!

Not many people at all, it seems, had even thought of this one (though a number of people have relayed to me their thanks for writing about it).

…. but that does not necessarily mean it’s incorrect!

Brief description of the theory

This theory is based on the underlying consideration that the Administered Price Cap under Cumulative Price Threshold (which is currently set at $300/MWh) is insufficient to cover short-run marginal costs (SRMC) for ongoing operations.

These costs are largely:

1)  driven by fuel costs, for thermal plant … which have been driven up in the ‘2022 Energy Crisis’ by external and internal pressures; or

2)  driven by opportunity cost of energy stored, for battery storage and hydro … which have been driven up in the ‘2022 Energy Crisis’ by factors including the above.

In both cases, there would be an allowance for variable O&M costs for each MWh of energy production (e.g. chemicals and other input materials consumed onsite).

This theory holds that this places the participant in an invidious position – whereby the participant is effectively confronted with a choice between two compensation regimes:

Compensation Regime #1 is one via the AEMC that is supposed to operate under Administered Pricing … which will have been used rarely in the past:

(a)  given that CPT triggers are rare; and

(b)  this situation even rarer, if you think about occasions when $300/MWh Administered Price Cap would not cover  operating costs!

Compensation Regime #2 is one via AEMO for compensation under Direction

… which has been used much more frequently.

This theory holds that some participants have been:

1)  opting or Regime #2 by withdrawing capacity from the price stack and waiting to be directed;

2)  with the motivation being that they want Compensation Regime #2 (and not Compensation Regime #1).

Assuming this theory is correct (at least in the case of some DUIDs) … and note that this is an assumption in this thought bubble exercise … then my next question would be to wonder what’s better about Regime #2 than Regime #1:

1)  Is it that it’s just more well known, and so viewed as ‘less subject to risk’?

2)  Are the timeframes for payment more attractive?

3)  Are the (likely) magnitudes of payment more attractive?

4)  Is it for some other reason?

I don’t know the answers to these questions, but will be interested to learn more.

On 16th June 2022, partly in response to a question I’d received online and partly as a result of a briefing on GenInsights21 to the EUAA Electricity Committee, I posted ‘Trended bidding from BESS discharge (NEM-wide) … highlighting APC and Market Suspension’.

In that article I wrote:

‘This means that, if a battery were to bid itself available for any period in the market, NEMDE would take all if could give and run it dry without any consideration of whether that energy would be better saved for later.’

My sense was that this had been the first time a number of readers had thought of this alternate interpretation of events … unfortunately.

The following day, Allan looked more broadly with ‘Stepping to suspension’ and in that article explained:

‘Generators with limited supplies of fuel or water generally manage supply levels by re-bidding market offers to higher prices as their remaining fuel or water stocks drop, which normally reduces their market dispatch levels as lower-priced generation is dispatched before them. This still leaves their re-priced capacity available to run in the market if necessary.

But once enough generators start withdrawing significant capacity offered to the market, this can require remaining generators to “pick up the slack”, and re-pricing of offers might become a less effective mechanism for managing fuel or water levels. This can end in a “rush for the exits” or prisoners’ dilemma, as more generators withdraw volumes from the market in response to rundown of fuel or water stocks, essentially forcing remaining “energy-limited” generators to follow.’

At the end of his article, Allan explains:

‘The NEM dispatch process has no inbuilt recognition of cross-temporal energy constraints and doesn’t seek to internally optimise limited energy reserves.  Normally generators manage their fuel or water position through pricing of their market offers, repricing upwards – even to the market cap – if fuel is scarce. But in periods where there is a near-deficit of generation – because of withdrawn capacity – everything offered available to the dispatch process would be dispatched and offer price may no longer be an effective rationing mechanism for generators with limited fuel or water. So if too many generators have already “headed for the exits” by withdrawing capacity, remaining energy-limited generators may almost necessarily have to do the same.’

For more discussion

With particular reference to Theory #1 (about potential trade-offs between Compensation under APC and Compensation under Direction) this prior article from 15th June 2022 is a good place to start.

 … then follow that through to two articles on LinkedIn from Christiaan Zuur and Farhad Billimoria that seem to deliver value in explaining some of the considerations with this respect.

With respect to this theory:

1)  it seems no coincidence that the AEMC published ‘how to lodge a claim for compensation’:

(a)  it’s unfortunate that this occurred on 13th June …  after the CPT had been breached in QLD and just prior to NSW, SA and VIC;

(b)  in the reviews after this event, it may well  that recommendations will be made with respect to:

i.  Better ‘war gaming’ to identify scenarios that might unfold; and

ii.  A more proactive communications regime to support this possibility

…. in writing this I am thinking back to the ‘NOUS Report’ into the blackout of January 2007 that flagged a number of communications improvements to be implemented.

2)  it’s also not much of a stretch to imagine that a number of participants were belatedly reaching for the rule book in the days prior around this time to understand how the process works … as they will likely never have used the process before.

3)  Worth remembering that some smart readers asked the question ‘What would happen if…?’ about this very thing back on 30th May 2022 (and there were useful comments made on the related thread on LinkedIn as noted in that article)  … but even then this was only 2 weeks prior to it occurring.

In addition to the coverage above on WattClarity, worth also highlighting the Ben Kefford wrote ‘Caught between a rock and a reservoir – why withdrawing capacity in the market made unfortunate sense of some generators’ on LinkedIn… specifically with respect to the Snowy Hydro operations.

Ben flags both Theory #1 and Theory #2 in his article as possible reasons why Snowy did what it did.

Once again note the caveat that this area of the market is outside of my prior level of understanding

1)  I’ve tried to come up the learning curve a long way in a short space of time, but there may well be errors or misunderstandings in this article.

2)  Based on my current level of understanding:

(a)  Whilst it seems plausible that some participants might have been behaving as described in Theory #1.

(b)  It seems certain that a number of participants were behaving as described in Theory #2.

Like many other stakeholders, the authors at WattClarity® will wait with keen interest to see what is determined from the official reviews.


PS1 at ~16:50 on Sunday 26th June

Readers here might appreciate reading the comments being added over on this LinkedIn conversation about this same article…. and perhaps also here on Twitter.



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.

3 Comments on "Two alternate theories of why capacity was withdrawn"

  1. We may learn some lessons from another energy sector: oil.

    Peak oilers have long been thinking what will happen when the global oil peak occurs. There is the theory of a Seneca Cliff after the peak in which the whole system declines because everything depends on a smooth supply of oil.

    The first global peak we had in 2005 (as predicted by Deffeyes) which caused the US recession in 2007 and the global oil price shock in 2008. High Chinese oil demand for the Olympic games also played a role. At the same time banks were so imprudent to have invested in car and therefore oil dependent US suburbia just as the peak was happening.
    High oil prices give a strong signal to motorists to economize the use of fuel with fuel prices at the pump following oil prices after 2 weeks or so. If demand destruction happens in all countries then oil prices will fall again. If these prices are too low for oil companies to make money then oil production may suddenly drop, potentially below levels needed for the safe functioning of the world economy. This did not happen yet because of US shale oil but this oil has astronomical decline rates and will also peak.

    Drawing parallels to the power sector, capacity of dispatchable generation may have peaked (including from coal depletion in mines near Eraring and Liddell/Bayswater) whereby this is a function of the price. We then have a cold snap with higher demand (similar to China’s 2008 Olympic Games). All the while Australian governments were driving up evening peak power demand by increasing immigration over 15 years, requiring hundreds of new power hungry apartment towers (similar to expanding US suburbia). The surge in renewables is similar in function to the new frontier of US shale oil and occasional/unpredictable recoveries of oil production from unreliable suppliers like Libya, Nigeria etc. and new fields.

    The big difference between oil and power:

    In oil markets, the feedback loop for lower demand via fuel prices at the pump is comparatively short as mentioned above. On the fuel supply side, oil companies are sitting on inventories and have a certain flexibility in case crude production is short. So this gives at least a short to medium term window to balance supply and demand.

    In power markets, however, generators must balance supply and demand instantaneously while the feed-back loop to lower demand is up to 3 months when the next power bill arrives. On the supply side, generation depends on variable renewables without sufficient pumped storage to provide dispatchable spinning reserve.

    So in the power market, there is a higher chance that the system collapses (=load shedding)

    Interested in responses to this.

  2. Looking at this rather unfortunate state of affairs from the other side of the country, I suggest the main drivers are fairly obvious. There were a few coal fired units out for maintenance then a few more were out of service from forced outages – either plant related or fuel supply related. Oh and lets not forget that there didn’t seem to be a lot of wind around for a few days. Now there is some gas fired gas turbine plant around but nothing like the proportion of the fleet we have here in WA. But that only works if you have contracted fuel supplies and hence a sensible price. Given most OCGT’s will have a heat rate of between 10 & 12 GJ/MWhr then if you cap your price at $300/MWhr then as soon as your gas price exceeds $25-$30/GJ you are out of the money before you have even paid for your starting costs on the unit. Spot gas prices that I saw exceeded this so clearly there is little contracted gas reserved for these peakers. Liquid fuels are even worse.
    This is one of the serious limitations of an energy only market and why Western Australia doesn’t have one. The NEM used to have lots of spare capacity but this has been hollowed out over the last decade or so as older stations have been retired and a market regularly trading at SRMC results in little incentive to invest in sustaining capital. I have heard much nonsense about “aging coal fired power stations” – coal fired power stations do age but are like Grandad’s axe. A new head or handle and it is still Grandad’s axe. Unfortunately when the economics of reinvestment have been so damaged by the hollowing out of daytime energy served the plants become unreliable. Their eventual retirement (whenever that might be) is also a bit of an are as no one retires a pristine plant – the glide slope is always tricky.
    This is where a capacity and energy market works. It guarantees a revenue stream, but it has to come with strings attached seeing at the end of the day the punters (via the retailers) are paying. There has to be firm fuel supply arrangements in place – if the generators want to be paid there are no excuses. Seeing the capacity market pays for some of the fixed costs of the plant then the market price caps get pulled right in. In WA they have always been $300/MWhr based on a GT burning liquids – probably about half the number it needs to be right now but the point is numbers like $15,000/MWhr just break the little guys who have little chance of hedging with a generator.
    Capacity markets work – they simply need to be carefully designed and calibrated (in terms of $$) to get the correct outcome and prevent gaming by people whose business model is only capacity payments. They can also provide good drivers for preferentially locating (say) a wind farm where it delivers energy best aligned to the daily load profile. Ultimate energy produced might be less but the capacity payment keeps the selection whole.

  3. Paul, 2 things I appreciate from you:
    1. Trying your best to stay neutral and wholistic
    2. Summarizing hours’ and days’ worth of research into a short, concise and structured article
    3. And sharing for free. Well with some subtle marketing but who wouldn’t!

    Ok I can’t count but you get the idea.

Leave a comment

Your email address will not be published.