As the Australian generation fleet transition to lower carbon emissions we are seeing the regulatory framework catching up. The regulatory framework is not the only aspect which is changing. We are experiencing an increasing amount of direct power purchase agreements being written between large users and renewable developers.
Economic models are also struggling to catch up. Almost every month a new forecast is published showing that if we just implement the latest proposed change to the market, we will see a return to the good old days of lower electricity prices. Part of this fallacy comes from outdated forecasting models which still look at cost of production as the main input to power prices. Short term price spikes and managing the grid in a sub-five minute period is increasingly important. Most current forecasting techniques fail to capture these factors as well as transient market power where a generator is able to set prices for a period by changing the volume offered in bid bands.
Behaviour is difficult to predict over the long term so most models assume rational behaviour in the long run. This is essentially sensible (it is very difficult to measure irrational behaviour) however begs the question how to model rational behaviour. In the electricity market, we will generally see future costs (over a very long term) trend towards the cost of bringing new plant online. If the long term prices are higher than the cost of new generation, someone should fund additional generation providing a natural cap.
The issue in current forecasting is determining the cost of bringing a new plant online. Traditionally, this has been the total cost of the plant divided by the number of MWh it produced. This provided a neat number expressed in $/MWh and is the basis for levelised cost of electricity (LCOE). The issue with using this number is that price profiles are changing and could be negatively impacted by the type of generation being evaluated. If we look at a solar farm in Queensland, we can estimate the production and the cost of the solar farm. This is the LCOE however it may not be the value of the solar farm, or produce the best estimate of the cap on long term electricity prices. In the event a large number of solar farms are built in the same location, there will be a large production at that time of day and not at others (evenings, mornings and overnight). Other generation sources will be required to fill the gaps. These other sources could have a higher cost and will only be built once there is a price signal to do so. The cost of electricity will be low during daylight hours when all the solar farms should be in full operation. This means that the value of the generation from solar reduces as more is built. This is not captured in the LCOE of generators. A new methodology which looks at the dispatch weighted (when generation occurs) achieved price will have to be considered when evaluating future electricity prices.
The impact of changing long term electricity forecasts are important. Many new plant rely on a component of their generation to be funded from the electricity spot market. With large changes in the assumed cost of electricity as models react to short term announcements it is very difficult to bank a project. On the other side, where consumers are being provided with different assumptions on future prices, it is difficult for them to commit to a purchasing strategy leading to more short term optimisation.
If you are considering a large investment in the electricity market, please call us on 1800 334 336 to learn more about the determination of electricity costs and gain a deeper understanding of the traded market.