The future cost of energy
Five minute settlement – opening with a whimper not a bang
We need thoughtful analysis, not campaigning
Chart of the week: Maximum and minimum demand
In May this year the International Energy Agency said no new fossil fuel development should proceed if we were going to meet net zero emissions targets by 2050. It was designed to help frame government ambitions in the lead up to the Glasgow climate summit at the end of this month. Instead, shorting gas supply has been one of the contributing factors to the energy crunch in Europe and China that will frame discussions.
The IEA wasn’t wrong. It wasn’t a directive, but a guide to scale. Shaking reliance on fossil fuels is proving challenging, as many energy experts have continued to warn about, and many governments and activists have stoically ignored. We look at the underlying trend here – the threat of sustained increases in the cost of energy, and how they might be mitigated.
In the meantime the energy crunch in Europe and China continues. Inevitably, everyone interprets it through their own lens, with energy minister Angus Taylor describing it as a wakeup call about the importance of firming renewables. Meanwhile, Australian resource companies are cashing in while they can on the high fossil fuel prices. A single LNG shipment could fetch over $200m.
Will the prime minister go to COP26 at Glasgow or won’t he? There seems to be little for Morrision to be gained by turning up to get browbeaten by other countries over his lack of a net zero target so he’ll stay home and see if he can cut a deal with the Nationals. But if they keep coming up with new conditions, all his secret modelling may be in vain.
A new report argues we can save billions by electrifying our homes and cars. As we explain below, the conclusions need be taken with more than a pinch of salt, even if the premise is worth thinking about.
Five minute settlements have arrived without great consternation, supposedly to better reflect the value and dynamics of utility scale batteries. The rule change is here. Where are the batteries?
Twiggy-watch: the great man is in a fight-picking mood this week. On the one hand he’s slamming CCS just as Santos is talking up its Cooper Basin project, on the other he’s arguing with Angus Taylor whether the $30m handout he got for his Port Kembla plant is for a gas or a hydrogen generator.
The future cost of energy
There are many lessons from the current energy shortages in Europe and China. But the ultimate question for consumers is simple: Is this just an international grade stuff up or an early warning of rising long-term energy costs in the future?
Gas prices in the UK and parts of Europe are setting records because of a combination punch of low summer winds, dry weather across Europe reducing hydro output, increased Chinese demand for gas and an abject lack of planning around the increased variability risk of increased reliance on renewables.
China has summer energy shortages driven by high demand fuelled by heat waves, increased reliance on gas imports, the impact of rejecting Australian coal and attempts to improve air quality and reduce emissions. Energy prices are not on fire in other parts of Asia, North America or Australia.
In May the International Energy Agency advised there should be no new investment in fossil fuel projects in order to meet net zero emissions targets by 2050.
We should not overstate current events, and not should we understate them. Energy is the most traded commodity on earth. Demand for energy is highly price inelastic. It is such a fundamental input to industrial and household activities that when we really need it, we will pay pretty much anything to get it.
The shortages in the UK are being managed by some industrial production being curtailed. But high spot prices pass, and most consumers are not directly exposed to them. Demand eases, higher prices bring on new supply and spot gas prices will probably return to more familiar levels in 2022.
As we discussed in last week’s Boardroom Energy Bulletin, there is plenty of gas in the world. Russia has indicated it will increase supply to Europe this winter. There is additional supply available, most of it in non-OECD countries. It is less clear how free oil and gas companies will be to develop new gas fields to meet increased demand, which may be needed as a balancing fuel for increased renewable generation.
Gas prices historically used to be linked to oil prices. And oil prices have been volatile. They are the result of 50 years of a global oil cartel trying to keep prices high by shorting supply. Sometimes they succeed. Sometimes they don’t.
Gas has an additional demand/price risk overlay. It remains the most flexible large scale energy source, needed to balance intermittent renewable generation. When renewable generation is high, gas demand is low. When it’s low, gas demand increases.
That means demand for gas, and therefore the price of gas, is driven not just by oil, but increasingly by renewable generation. When oil prices are high and renewables supply is low, gas prices can be expected to skyrocket. This could persist for the next decade or so unless there’s either a big expansion of supply or until our electricity systems can be configure to rely less on gas.
Gas fields, once established, produce a flow of gas. Most gas contracts reflect this physics. They strike long term offtake contracts to recover the development and operational costs. Gas cannot be turned on and off like a tap. So this makes gas supply decreasingly compatible with renewable generation as it scales up.
There are a number of regional gas markets with different prices. Where you are selling into the world matters. Australia, as a gas exporter, sells LNG almost exclusively into the Asian markets. The different gas markets are linked, but prices are regional.
So what are the options to constrain prices? How can energy supply-demand balances and therefore prices be less volatile and more stable? There are at least a few options.
Allow more gas development: Shorting gas supply will only push up prices, which exacerbates the impact of current and future supply shortages. Gas is a fossil fuel. But until it can be replaced it remains the most flexible energy source to complement large renewable generation.
More long-term contracts: Europe’s problem over the past five years is it has been it has increased demand for gas while producing less. As renewables generation has increased it has opted to buy much of its gas on spot markets. This has exposed it to extreme price volatility in a way that, say Japan, has not experienced (its gas is purchased in long term supply contracts). By contracting guaranteed supply this will reduce European energy companies to price risk. But it will require other changes.
More gas storage: When the UK closed its massive Rough gas storage facility in 2017, experts warned it could be exposed to more volatile winter prices. Those experts were right. The problem with maintaining gas storages for infrequent price spikes is that it is expensive. But so are gas shortages.
Built more renewables: this is, predictably, the answer trotted out by the renewables developers to pretty much any energy crisis. Renewables will increase alternative energy supply and they will reduce emissions. But they will also increase seasonal volatility. So the more renewables are built, the more strategic energy reserves are needed for those unseasonal years. Which are infrequent, but do happen.
Build other supporting technologies: Where possible other supporting energy technologies like utility batteries and pumped hydro could be scaled up. These are expensive at scale, but will help to mitigate the impact of gas shocks.
Keep flexibility with other fuel sources: Keeping old coal generators available is one way of spreading risk, by having alternative electricity supply sources. From an emissions perspective it’s less than ideal, but running for a few weeks every year or so might be a simple and cost effective way of managing risk at a lower cost. Nuclear generation will probably be a necessary additional source of zero emissions energy for some economies that do not have access to abundant renewables.
Demand management: if we are going to use more renewables, governments need to be more proactive in developing more advanced demand response capacity. This solution is linked to increased use of renewables. Load shifting becomes a critical adjunct to high intermittent energy supply. There is a time dimension here – load shifting across a couple of hours is a different proposition to load shifting across weeks or months to manage seasonal variation.
Most of these solutions add additional infrastructure or contracting costs. Building renewables increases supply of lower cost energy sources, but increases supply risks and hedging costs. This may require much larger gas storage capacity, to balance between the peaks and troughs of demand versus the more stable production capacity.
This is also only the costs associated with ensuring reliability. System security in a high renewables grid will have its own costs. It’s a reasonable bet that the increased risk of an energy market in transition will increase costs. It’s less clear how much a complete future energy system might cost.
Five minute settlement – opening with a whimper not a bang
One of the biggest – and most expensive – market reforms since the NEM was established twenty years ago went live last week. Nobody much noticed. Does that mean it was a smoothly executed reform or a big waste of money?
Five minute settlement is a deceptively simple sounding reform. The NEM wholesale market is based around five minute dispatch intervals (DI). Generators bid for each interval, the market operator solves for physical constraints and then tells each generator how much to dispatch based on the lowest priced bid that provides enough energy to meet demand.
But for a range of reasons, including metering and IT capabilities at the time, the market was set up to settle every half hour (i.e., six DI). The price customers paid, and the revenue generators received was based on averaging the six DI prices. This worked well enough most of the time. But it did result in some odd behaviour. If there was a price spike early in the half hour, then fast-responding generators (gas, hydro) would rebid to the price floor (-$1,000/MWh). This meant they got dispatched for the rest of the half hour and got a share of that earlier price spike, even if they weren’t contributing any supply at that time. Similarly, there appeared to be some cases where a price spike occurred in the last DI of a half hour because a generator or two rebid to higher price bands and there was no time for other participants to react.
While there are rules requiring generators to bid in good faith, these have yet to result in successful prosecution. So, it’s unclear to what extent these outcomes were due to deliberate gaming as opposed to just the dynamics of the market. Nevertheless in 2016 a large user, Sun Metals lodged a rule change to align settlement with dispatch. It argued that along with incentivising the strategic rebidding described above it also disadvantaged fast responding generation and demand side. This quickly morphed into the idea that large batteries were being kept out of the market by the half hour settlement arrangements and the rule change became a cause celebre. The Greens tried to get a parliamentary motion passed in federal Parliament, even though it’s unlikely any of them could have fully explained the rule change if pressed.
On the other side of the debate, several incumbent generators argued that existing gas plant would be disadvantaged if the rule was passed, because although such plant was more flexible than coal, it still couldn’t necessarily ramp up quick enough to take advantage of a single five minute price spike. That might inhibit them from selling caps – the usual way for gas peakers to stabilise their revenue -because they couldn’t be confident of defending them by being dispatched when the price spiked.
The AEMC passed the rule change in 2017, but since the change had implications for everyone’s systems and contracting arrangements it allowed plenty of time for implementation. This was extended by three months due to the challenges caused by COVID restrictions and disruptions. AEMO identified around 70 procedures that needed to be updated due to the Rule change. Analysis by Deloitte for the AEMC estimated implementation costs for each participant at: up to $1m for small retailers, $5-25m for medium sized participants (generators, networks, retailers) and $20-40m for large participants. Including AEMO’s own costs, the total cost is around $500-900m.
While that’s a big sum, the new settlement rules only need to make a modest difference to bidding and generator entry that flow through to prices to accrue enough benefits to cover these costs. The NEM wholesale market is worth billions of dollars each year after all. Roughly speaking a dollar a year price reduction for ten years adds up to a billion dollars saved by customers.
Given the lead time, one might have expected a rush of large-scale battery deployments to be commissioned just in time for the change. There were a few installed two-three years ago – all heavily subsidised – but none have come online since then. There are several under construction, although two of the three largest, the Victoria big battery and the Wallgrove battery are driven by network issues rather than wholesale arbitrage opportunities. This leaves 157MW of batteries expected in the next 12 months. (there’s another 20GW “proposed”, but that is not a clear signal of what will actually get built).
Similarly on the demand response side, there‘s not been a rush to immediately capitalise on the change, even though a new wholesale demand response mechanism is starting later this month. A registration portal set up by AEMO only attracted 56MW of new demand response – although as registration is voluntary, this may not be the full amount.
Conversely there’s no sign of exit by gas peakers who can’t risk selling caps anymore, either.
Early bidding data shows that price spikes have not gone away. Queensland hit the price cap twice on the afternoon of 4 October. These spikes only lasted a single DI. In neither case was there a “race to the floor”, so that type of disorderly bidding seems to have been addressed by the new rule. On the other hand, that bidding behaviour did bring down the average price for the half hour. The main beneficiaries of the price spikes (i.e., those who ramped up for the high priced DI and then back down again) seem to have been Braemar 2 gas peaker and Wivenhoe pumped storage. Queensland doesn’t have any large scale batteries yet -the first is due to start operations in the next few months.
Of course, it’s early to draw any conclusions of the impact of the rule change on investment patterns, bidding behaviour, price outcomes and contract market impacts. There’s no formal mechanism in place for ex post assessments of major rule changes in any case. But an early look suggests the move to five minute settlement was not as radical a change as its strongest advocates or detractors claimed it would be.
We need thoughtful analysis, not campaigning
One way to reduce greenhouse emissions is to replace household consumption of natural gas with a zero emissions replacement. This could be renewable electricity, biomethane or hydrogen. It’s a complex, important discussion that warrants serious thought.
A new think tank called Rewiring Australia gives their game away from the outset. They’ve already decided the answer is to fully electrify households, and they’ve written a report called Castles and Cars to explain how it should work.
And therein lies the problem with this report. Victoria is facing a gas shortage and one possible solution is to replace residential gas demand, which could save around 100PJ of gas a year.
The process of switching this energy load onto the electricity grid is possible, and possibly desirable. But there are more than 5 million households in Australia that use gas for heating air and water and for cooking. And the process of unravelling this is challenging.
The simplest way of decarbonising gas supply is to replace it with low or zero emissions fuels like biomethane or hydrogen. This is not considered in Castles and Cars. The difference between considered analysis and campaigning is analysis explores multiple options in detail and draws conclusions by comparing them. Sometimes the results will be surprising or counter intuitive. Campaigning uses data to justify a pre-determined answer.
Castles and Cars promises extraordinary savings ($46 billion a year by the 2030s) by electrifying households and their cars – up to $6000 per household. This is anchored on the logic that households spend more than $5000 a year on power bills, gas, petrol and most of this would be saved in a distributed, electrified system.
There is the upfront cost of the big rooftop solar PV system, big household battery, appliance replacement and a new EV all to consider. These tens of thousands of dollars per household are paid for initially by a $12 billion government subsidy, but then the falling cost of everything pays for itself. According to the report, electrifying households becomes like a runaway train of free money.
The paper assumes the rollout of 10-12kW rooftop solar and a battery supplied to every household. Even then, households will still rely on the grid for around 20 per cent of their power. This requires reducing average household demand from 102kWh a day to 37kWh, and that’s after including transport needs. The big drop is driven by getting rid of the thermal inefficiency of combustion engines.
It’s important to interrogate the limitations of these reports. The list below is far from exclusive.
- The report does not appear to include the cost of upgrading to 3-phase power which would be likely in many cases given the increased load on each household.
- The removal of households from the gas network will increase gas network costs for residential and commercial gas users remaining on the system. How is this addressed?
- The paper assumes both low cost of EVs and rapidly falling costs and limitless supply of these cheap EVs. In practice the Australian car market is heavily undersupplied for EVs because most OEMs are sending vehicles into the heavily subsidised European market. Accordingly the price premium of the EV models that are available is still very high. This could take several years to resolve
- There is no consideration of retail electricity costs, distribution and transmission charges which will continue.
- Electrification has the potential to increase total system costs. Even if households are drawing less power in total from the grid, the volatility of their demand (or collective export in the middle of the day). There will be periods of high peak demand – either heat waves or dark winter days when electrified households become heavily reliant on grid electricity at exactly the same time. These could require substantial investment in large scale storage, transmission and distribution. The cost of these will need to be shared by energy users. There is no apparent analysis of the extent of these costs in the report.
- There appears little consideration given how to actually address the split incentives between rental and owner-occupied dwellings.
- More than 30 per cent of Australian households are medium or high density and would not be able to support the distribute energy resources identified in the report. Around a quarter of Australian households do not have off street parking to support EV charging.
- Retail and network tariffs will need to be substantially reformed to efficiently exploit the low-cost electricity savings identified here.
- Petrol savings are calculated based on retail fuel costs. But about a third of these are taxes that help to fund road expenditure. The taxes will have to be replaced by road user charges or similar. So savings are greatly overstated. Similarly, electricity prices contain a number of government charges that will have to be recovered somehow.
Electricity is an efficient vector and its appliances do enjoy lower running costs. Electric vehicles are highly desirable and appear to be the most likely successor to internal combustion engines. But it’s a stretch to suggest the savings delivered by the efficiency improvements of electrification cover the capital costs of the transition at current costs. If they did then logically this would already be much more obvious in the current energy market.
It’s an important discussion, one that would be better served by a more robust analysis and less blind optimism.
Chart of the week: Maximum and minimum demand
The NEM has been setting several minimum demand records recently. In the last month, NSW, Queensland and South Australia (ignoring the black system event of 2016) have all set minimum demand records, as has the NEM as a whole. Victoria’s minimum demand record was also set earlier in the year. These tend to occur during the shoulder seasons of spring and autumn on a mild sunny day where there is high rooftop PV output (which is counted as a reduction in demand rather than on the supply side).
Conversely maximum demand records tend to occur on either very hot summer days (the mainland) or very cold winter days (Tasmania). NSW has been running summer peaks but is likely to flip to winter peaks in the next few years. Victoria, South Australia and Queensland have all set new maximum demand peaks in the last two summers.
The difference between maximum and minimum demand is a measure of the flexibility required of the energy system. Chart 1 shows the MW difference between maximum and minimum demand every day this year for each state.
Chart 1: Daily max-min demand, MW, by state
Source: Boardroom Energy analysis from NEM Review
The bigger the state, the more the system needs to be able to ramp between maximum and minimum. At various times this year, NSW has had to ramp almost 6,000MW within a day. That’s a lot of flexible generation. Wind and solar may be helping or adding to the requirement based on the time of day. Victoria, peaked at over 4,000MW during January and Queensland’s minimum demand led to a ramp requirement of almost 4,000MW (see spike on right hand side). Another way to look at it, which adjusts for size is the ratio of maximum/minimum demand. Chart 2 shows that on this metric, South Australia has the proportionally largest ramping requirement, with a daily ratio of over six on minimum demand days, with Victoria a distant second occasionally exceeding a ratio of two.
Chart 2: Ratio of daily maximum: minimum demand, by state
Source: Boardroom Energy analysis from NEM Review
South Australia is unusual in its volatility. While international comparisons are difficult, few systems have anything like the amount of rooftop solar that South Australia has and are more likely to look like Tasmania, comfortably oscillating between 1 and 1.5.