Submitted on December 2, 2019

The HVDC link will be operating with only one pole (one half) in service for most of the first quarter of 2020, to allow essential maintenance to be carried out.  When this was first announced back in December 2017, futures prices barley moved, but they now show an expectation of a price difference between upper north and mid-south islands in the vicinity of $55/MWh.  What’s changed?

By Greg Sise, 2nd December 2019

This post is longer and more technical than usual, the result of the complex interactions between energy, the need to provide reserve capacity, and the risk characteristics of the HVDC link.

It’s almost two years since Transpower announced that the HVDC interisland link would be operating at lower capacity while maintenance is carried out on the Churton Park section of the link in Q1 of 2020, with the work taking most of Q1.  During this period, only one half of the link will be operating, meaning the total capacity will be reduced below the normal level, but also that additional reserve capacity will be required in the island receiving power over the link:  given the amount of water in the hydro lakes, this is most likely to be the North Island.

The main consequence of having to provide more reserve capacity, which is required to cover the risk of a failure of the HVDC link, is that the cost of the reserves will be reflected in the spot price difference between the two islands.  Ironically, now that the southern lakes are topped up with recent rain, there is likely to be more northward power transfer on the link, meaning more reserve will be needed.

This chart shows the difference between the electricity futures prices at Otahuhu (OTA) in Auckland and Benmore (BEN) at the southern end of the HVDC link for Q1 of 2020, and it shows that the futures prices have only reflected an expectation of a particularly large price difference since the second half of this year.

When the HVDC work was announced, the price difference was around $4/MWh and it didn’t hit $10/MWh until the end of February 2018.  So why is it now $55/MWh?

Large price differences can occur for two reasons:  firstly, if generation in the North Island is very expensive and the HVDC link is loaded to its maximum capacity because there is cheaper generation in the South Island, then there will be a large price difference between BEN and OTA, but the price at BEN will also tend to be elevated.

In the second case, if the cost of reserves to cover the risk of losing the HVDC link is high, this will be reflected in a large price difference between BEN and OTA, but the price at BEN will tend to remain static or fall.

Or it could be a combination of both cases.

The first chart below shows the Q1 futures prices at BEN and OTA, and it shows that the price at BEN has risen along with the price at OTA through to September, but since then it has fallen, and a larger price difference has opened up between these nodes. 

Assuming the futures prices reflect rational expectations (that’s debatable!), then there is clearly a mix of the two cases here:  BEN prices are generally elevated, suggesting North Island generation will be more expensive, but the fall in the BEN price since September suggests a growing expectation of the impact of reserves.

Given the increase in spot prices since 2018, many will be surprised to learn that in our spot market, energy and reserves are “co-optimised” which means that the market works to minimise the cost of providing electrical energy and the reserves, not just energy on its own.  A consequence of this is that if either energy or reserves become scarce, and hence costly, then in certain situations it can impact the cost and spot price of both. 

For example, if reserves are in short supply, then a generator may have its output reduced so that it can provide more reserves, thus requiring more expensive generation to be turned on:  the result of this is a spike in the spot price of energy AND reserves.

So what this suggests, when we look at the futures prices at BEN and OTA for Q1 2020, is an expectation that there will be a shortage of reserves in the North Island, which will tend to push energy prices up as well.

The chart below shows the North Island “sustained” reserves (also known as 60 second reserves) since June 2013, in price bands of up to $10, up to $100, up to $1,000 and then everything over $1,000.


Back in April 2013, there was 962 MW of this reserve offered on average over the month, and 72% of this was offered at less than $10, and 84% at less than $100.  Contrast that with now, and we have around 700 WM offered in total, a drop of 26% (partly explained by the closure of the Otahuhu B power station at the end of 2015), and only 55% of this is offered at less than $10 and 70% at less than $100.

The same chart for North Island “fast” reserves (also known as 6 second reserves) is next, and it shows that in April 2013 70% was offered at $10 or less, but that this has now fallen to 52%.  The two charts also show that since mid-2018 there is also a lot more offered at over $1,000 than at any time since 2007.

Ironically, upgrades to the HVDC link completed at the end of 2013 have reduced the demand for reserves in the North Island, because the risk of losing generation in the North Island can be covered with reserves “shared” across the link from the South Island.  So we can speculate that a large proportion of North Island reserve is no longer needed, and hence not offered into the market.  There is probably a substantial amount of reserves that is still offered but seldom required, so it is offered at a higher price.

But the full benefit of reserve sharing on the link is only captured when the link is fully operational, which it won’t be for most of Q1 2020.

This leads me to think of two possibilities at the extreme ends of the spectrum:  the first is that reserve providers will respond to the large increase in demand by reducing prices and offering all the reserve they have available.  The second is that reserve offers won’t change that much from what they are now, or even increase because of the shortage of reserves.

Only time will tell, but the futures prices for Q1 2020 seem to be saying that there is a significant risk of the latter occurring.

One final chart, this one showing the average spot price difference between OTA and BEN for all quarters from Q3 of 2009 (ending September 2009) through to the current quarter (red-grey columns).  Back in 2009 – 2013 before the HVDC link was upgraded, we can see some big price differences driven by storage (green curve) being above average (high levels of power flowing north on the HVDC link) and below average ((high levels of power flowing south on the HVDC link), in the Mar-11 quarter and Jun-12 quarters, for example.

The correlation between storage deviations and OTA-BEN price differences is nowhere near as strong once the HVDC link was upgraded.  But Q1 2020 will be like going back to the pre-upgrade days, so the 2009-2013 data is more relevant as a comparison:  the largest average price difference for a quarter was in the Mar-11 quarter where it reached $41.

But storage averaged 870 GWh above the long-run average for the entire quarter back then.  Fast forward to now, and storage is currently about 600 GWh above average, having only just reached this point after a relatively dry period earlier this year. 

Are the futures prices telling us that storage is going to stay this high, and in fact go higher, through to the end of March 2020?

I’ve been in this business for over 20 years and no one that I know of has a crystal ball that can tell us that.  Which reinforces the conclusion that the futures prices for Q1 2020 are saying that there is a significant risk of reserve offers staying elevated, and potentially with no expansion in supply above current levels.

There are two mitigating factors in all of this, and one is that market participants have had plenty of time to prepare themselves for Q1 2020, so hopefully there will be plenty of plant available to generate or to provide reserves.  The second is that Lake Taupo is close to full, which increases the likelihood that it will remain at high levels of storage, and if this is the case in Q1 2020 then the stations on the Waikato River will have plenty of water to generate with and to provide reserves when they are not running at full output.

My closing thought on Q1 2020, is that it has the potential to produce some extreme price levels, which will no-doubt prompt calls for the Electricity Authority to investigate what caused them.  Generators and reserve providers that find themselves in a position where they are key in setting extreme prices will come under scrutiny, so all providers would be well served to heed the so-called “good conduct” rules contained in Part 13.5 of the Electricity Industry Participation Code.

Looking at it on a more positive note, it must surely be a good time to brush the dust off old reserve offers and ensure there is plenty of fast and sustained reserve offered into the market in the North Island at reasonable prices!