“Will spot prices be high enough to encourage enough investment in renewable generation to meet demand in future?”
I recently presented to the Electricity Authority’s Market Development Advisory Group (MDAG) on the topic of the problem of price formation with 100% renewable electricity. The presentation is available on the Authority’s web sitebut in this post I provide a simple explanation of what I presented.
100% renewable electricity sounds like low emission Nirvana. Imagine the graceful rotations of gigantic wind turbines, gently turned by the sea breeze. The glinting of solar panels lying baking in the hot summer sun. Rain flowing into the hydro lakes where it lingers briefly before winding its way through the varied and beautiful landscapes of Aotearoa, through hydro dams, and ultimately returning to the sea, from whence it came, earlier in the global water cycle.
And what’s best? It’s all absolutely free. But (sound of needle scratching across vinyl record) wait a minute! Are we really going to get free electricity?
This, or a variation of it, is the question that the Authority, and the MDAG in particular, is to probe - details can be found in more detail at the Authority's web site. But essentially, it boils down to this: when the market is 100% renewable generation, and assuming the spot market operates the same way as it does today, in which generators submit competitive offers to generate and the marginal offer sets the price, will spot prices be high enough to encourage enough investment in renewable generation to meet demand in future?
Theory says that in a competitive electricity market generators’ offers will be based on short run marginal cost (SRMC), with the highest offer in any given trading period being marginal and setting the price(s) for all participants in the spot market. The spot market in New Zealand underlies the entire market, and spot prices (or expectations of spot prices) tend to drive prices paid by consumers (ignoring line charges). Hence under our current market structure, the spot market is key.
The SRMC of wind generation is a few dollars per MWh at most, solar about zero, hydro in the south currently less than $10/MWh ($5.33/MWh is due to the HVDC transmission charge), and lower still in the North Is. So, if they all offer their SRMCs, then spot prices will average less than $10/MWh. But if demand increases, requiring more generation to be built to meet this demand, then this level of pricing hardly seems sufficient to incentivise that new generation, which even in the far future will still cost multiple tens of dollars per MWh to build, own and operate.
Which could lead to shortages.
But this is where the story gets a little more complicated. In our market, when there is insufficientgeneration offered into the market to meet demand, then regardless of offers, spot prices get set to very high values, between $10,000 and $20,000 per MWh. This is known as “scarcity pricing” and it occurred once, and only once, on the 9thof August this year when some consumers in the North Is went without power for two hours. Lack of generation was not the only factor that caused these outages, but the point is that scarcity pricing situations are possible.
If there are enough scarcity pricing periods in a year, then the $10,000+ prices could pull the average price up, to a point that justifies investment in new generation.
Two problems arise, however: first, the average price is only pulled up because there are shortages. The experience of the outages on 9thAugust demonstrates once again that shortages are not popular with the public, with business and with politicians.
Second, if enough new generation is built to eliminate the shortages, then the average price collapses back to a few dollars per MWh. As demand grows, the shortages return, and so begins a cycle of over investment, followed by under investment, followed by over investment, and so on.
Would this just be the electricity equivalent of the business cycle? Well yes, but also no, because our modern society is increasingly reliant on electricity, and this will only get the more so as we ditch internal combustion engines in favour of EVs, and convert many fossil-fuelled processes in industry to electricity.
At this point in the explanation, we need to ask: why is 100% renewables so different to the current market in terms of price formation? The main reason is that we still have a fleet of fossil-fuelled generators with a total of 2,133 MW that could theoretically provide 47% of total electricity generation if this generation ran 100% of the time. But it doesn’t: it currently averages around 17% of the generation in each year.
The fuel costs of the eight generators in the thermal fleet are also quite similar, which means that the aggregate offer curve (supply curve) presented to the market by the thermal fleet is what I describe in the MDAG presentation as “wide and flat”. As demand grows, more thermal generation is required, but the prices at which the thermal fleet runs don’t change at a rapid rate, which means that when new renewable generation is built to meet the extra demand, it doesn’t collapse the price.
Some will object that the costs of the thermal fleet rise more than I suggest as demand rises, and that is true in the shorter term, e.g. because they may not have enough fuel contracted at lower prices. But these changes in prices are two or three orders of magnitude less than going from $10 to $10,000+ as I have described above. In the MDAG presentation I included the following charts which contrast between the wide and flat thermal offer curve and the narrow and steep offer curve presented by a combination of a few hundred MW of demand-side response (DSR) offered into the market at between $2,000/MWh and $8,000/MWh and “supply of last resort” (SLR) at $10,000/MWh, which could include some very expensive generation, or ultimately shortage and scarcity prices.
In our market currently, we have a modest amount of storage in the hydro lakes, mostly in a handful of the very largest lakes such as Lake Pukaki near Aoraki-Mt Cook. The cost structure of the water in storage is considerably more complicated than just the few dollars of HVDC charges I mentioned above, because the large hydro generators have a choice – shall we use stored water to generate now? Or shall we store it for use later on when it might be more valuable and attract a higher price, e.g. when the lakes are low and the possibility of shortage rises?
A mathematical technique known as ‘water valuation’ is typically applied to calculate the expected future value of water, and this establishes the water value. The water value is then the opportunity cost of generating now, and both theory and economic reality say that the SRMC of generation offered from the large hydro lakes is equal to this water value, at least at the power stations immediately downstream of the large hydro lakes.
Water values are driven by a range of factors, but most importantly, these include the amount and the cost of fossil-fuelled thermal generation. You can think of the major hydro generators “shadowing” the fossil-fuelled generators in terms of the prices they offer for their output. For example, if the hydro is offered above the price of thermal generation, then the lakes will fill up and there will be excessive spill and hence waste. If the hydro is offered below the price of thermal generation, then the lakes will empty out and there will be a shortage of electricity.
Bottom line: getting water values right is absolutely critical to the secure, cost-effective operation of the spot market.
The calculation of water values in the NZ market can be traced back to the mid-1980s under the old Electricity Corporation of NZ (ECNZ), and the principles developed back then still apply today, and will apply with 100% renewables under the current market structure.
Water values respond to changes in the thermal fleet and to changes in storage, but the big hydro generators are often setting the price at the margin, and this provides a degree of stability in spot prices. At 100% renewables, the presence of storage also potentially provides stability, but only if there is an equivalent to the current thermal fleet – what is this? In the MDAG presentation, I showed that it is the narrow and steep offer curve presented by DSR and SLR.
But as we’ve seen, there is a big problem with this offer curve because it is potentially highly unstable.
So, I went on to look at the alternatives to DSR and SLR or, in other words, what else could replace the current thermal fleet, and provide a wide and flat offer curve, that would stabilise water values and provide more consistent investment signals?
I also stated my view that a 100% renewable market which does not provide more storage than the existing hydro lakes, is not tenable, for the reasons described above: there would be the constant risk of scarcity. Such a market would also have windfarms operating at well below their theoretical capacity for long periods, as they ‘spilled air’. I believe we will need a mix of storage including some or all of the storage mechanisms shown below, including pumped storage, hydrogen storage (perhaps not in dirigibles ha ha), but in a depleted gas field or embodied in a liquid, for example), and many, many chemical batteries of various sizes and technologies.
But the largest component of the SRMC of storage is the electricity price, simply because storage has to charge in order to generate (discharge) at some later time, which introduces a circularity in pricing which doesn’t play a big part in our market today. Our modelling also shows that storage devices struggle to recover their costs just based on buying low and selling high. But storage certainly helps to stabilise prices and investment signals, and to prevent shortages.
I suggested in the presentation that biomass-fuelled thermal generation and hydrogen generation could have a key role to play in future. There are efforts being made to determine how much wood waste there is that could be used as fuel for existing or future thermal generators, with Genesis announcing recently that they plan a pilot program running biomass at the Huntly power station.
The following diagram, prompted by work done byFirst Gas, shows how hydrogen could help provide thermal generation, without the circularity issue. A hydrogen plant, labelled ‘electrolyser’, would produce hydrogen using electricity from renewable generation supplied to it at a fixed price, either by direct connection to the plant or via suitable power purchase agreements. The hydrogen would be stored in a depleted natural gas field (assuming this is feasible – yet to be proven), and available to go to the gas transmission pipelines for use in direct gas-firing applications and thermal electricity generation.
I haven’t said it yet, but if there is no solution to the price formation problem under the current market design, then market enhancements may be required. But there is a lot of water to flow through the penstocks before we come to that conclusion.