My last post examined the key variable costs for combined cycle gas power stations – namely gas and carbon.

This follow-up piece intended to look at ‘fixed costs’ – or costs that don’t vary with output. However, I think it’s more accurate to describe this piece of work as ‘other costs’ – i.e. anything that isn’t gas or carbon.

I am trying to analyse energy in the UK to help improve policy. None of the content should be construed as investment advice. I have done my best to ensure that the content below is accurate – but I am human and will make mistakes – if you spot any, please let me know and I shall update as appropriate.

Methodology:

My original sample of 13 combined cycle gas turbines (CCGT’s) from my capital cost post served as my starting point. Sadly I couldn’t get any detailed disclosure on gas and carbon costs for most of the 13 – which meant working out “other costs” would be near impossible.

I found a different solution. Several of the companies within my sample operated under a tolling agreement – where they are paid an annual operating fee rather than acting as a merchant generator. By isolating these power stations, I was implicitly assuming that gas and carbon costs would be removed from their cost structures, allowing me to examine the rest of their cost base in more detail. Sadly, this meant I was reduced to just three power stations.

I will provide more details in the appendix, but I wanted to capture all costs excluding gas and carbon on a cash basis. I will call my metric “Cash Costs ex G/C” – or cash costs excluding gas and carbon.

I calculated values for the three CCGT’s back to either 2010, or from the first year after the commissioning year.

Into the data:

My first graph expresses my “cash costs ex G/C” metric on a £ per MW per year basis (£/MW/Year) adjusted for inflation to 2023 values. Before explaining the data, a brief background to each of the power stations:

• CCGT A: Commissioned c. 10 years ago, c. 900 MW

• CCGT B: Commissioned c. 30 years ago, c. 400MW – appears to have run at very low capacity factors for the past decade, acting more like an OCGT

• CCGT C: Commissioned around 15 years ago, c. 900MW

The average across all the power station years is c. £31,400/MW/Year in 2023 real terms. You can observe big peaks for each power station:

• CCGT A: The 2020/2021 peak was driven by a big rise in overhaul costs. The cause or nature of this isn’t explained.

• CCGT B: The 2011 peak is associated with a big spike in plant maintenance costs. This was associated with a “periodic overhaul”

• CCGT C: The 2015 peak was associated with gas turbine upgrades and a life extension project. The upgrades were stated to increase the capacity, flexibility and efficiency of the power station.

I think it’s important to include these peaks in costs in my average for the data set - there is clearly an ongoing overhaul process at play for turbine maintenance, as attested by a couple of accounting notes for power stations that were outside my group of three.

Example 1: One CCGT had a 30-year depreciation schedule for “Power station assets” and a 3-6 year depreciation schedule for “Turbine components, other plant and machinery” The depreciation rate for turbine components was based on an operating hours basis.

Example 2: Another CCGT explained that contract work to replace gas turbine components was capitalised and then depreciated over the period to next overhaul.

I also found some third-party evidence that large overhauls can get expensive.

A February 2024 report issued by the Department for Energy Security and Net Zero (DESNZ) entitled “Assessing the deployment potential of flexible capacity in Great Britain” had a section covering life extension decisions for CCGT’s and mentions a “typical one-off major maintenance capex event of 50 £/kW” which equates to £50,000 per MW.1 This is a similar ballpark to the cost peaks in my charts which is reassuring.

Breaking Down the Cost Base:

Whilst I have my headline average of c. £31,400/MW/Year, I don’t have much sense of what that entails, other than periodic spikes in overhaul costs. Luckily, accounting footnotes offered me some insight.

Whilst CCGT A didn’t have much granular cost disclosure, I did get some useful data for CCGT’s B and C. In the case of B, the cost breakdown only ran from 2010 to 2016, but the data splits down into lots of detail, as per the graph below. N.B. these values haven’t been inflation adjusted.

Throughout this period, staff costs were c. 33% of the cost base, plant maintenance c. 25%, system charges c. 15% and other/unspecified are another c. 25%.

The data for CCGT C isn’t as granular but runs for a longer duration.

In this instance, staff costs are proportionately much smaller, at c. 15% of the total cost base over the period. This might be explained by CCGT C running at higher utilisation, and thus having a great proportion of variable ‘other’ costs, whereas staff costs are likely to be relatively fixed in nature.

Staff Costs:

Staff costs are one of the areas that tends to get reasonably widespread accounting disclosure, so I am going back to the original sample of 13 CCGT’s to see what trends I can dig out. Not all the accounts have direct employee costs and instead have a service charge with a related party – I looked at which companies gave staff cost disclosure in 2023 and worked back from there – it gave me a group of 4 CCGT’s.

The chart below shows annual employee costs per MW of capacity, adjusted to 2023 values. Whilst there is a bit of noise in the data, it has been fairly stable though time at around £5,000/MW/Year.

The underlying drivers of staff costs haven’t changed much either – around 5 employees per 100 MW of capacity and employee costs per capita of just under £100,000, expressed in 2023 values (N.B. this will include salary, pension and employer NI costs)

My guess is that direct staff costs are around 15-20% of the non-gas and non-carbon cash cost base for a UK CCGT.

Comparisons

My first chart revealed an overall “Cash Cost ex G/C” metric of c. £31,400/MW/Year, expressed in 2023 values. This was based on data from 3 CCGT’s and 35 individual financial years. As I lamented at the start, I don’t have a breakdown as to which of these costs are fixed in nature, vs variable or semi-variable.

DESNZ also published some cost assumptions for CCGT’s in their “Electricity Generation Costs 2023” report. They break their costs down into four categories:

• Insurance

• Connection and use of system charges

• Fixed operations and maintenance

• Variable operations and maintenance

I am assuming the first three bullet points are fixed costs, and the latter variable (which might be a flawed assumption for use of system charges) I have expressed the three fixed cost line items in 2023 £’s below.

Their fixed costs come in at c. £23,600 per MW per year expressed in 2023 values.

The DESNZ variable operations and maintenance (O&M) cost assumption is £2 per MWh in 2021 values or c. £2.34 per MWh adjusted by CPI to 2023.2 I have converted this to a per MW per year basis for two different load factors – one at 93% (the number referenced in the 2023 generation cost report, and a more “real world” value of 40% - the approximate average load factor for the existing UK CCGT fleet since 2010.3

The bottom right value of c. £8,200 per MW per year is expressing the variable O&M cost per MWh across 1 year of generation at a 40% capacity factor. If I add the c. £8200 figure to the 2023 fixed cost total of c. £23,600 per MW per year, I get to a figure of c. £31,800 per MW per year, which is pretty close to my “cash costs ex Gas/Carbon” figure of £31,400 per MW per year.

Please don’t mistake the closeness of the two figures as any evidence of competence on my part! My data set was noisy and consisted of only three CCGT’s, one of which operated at very low capacity factor for a number of years. There is a degree of chance that the figures matched so closely. However, I haven’t massaged my data– I didn’t do any sums on the DESNZ figures until I had finished my own work.

Conclusions:

Once I have finished my work on gas power station costs, I intend to write an article that knits together their economics in a cohesive post. For now, I just want to illustrate the issue of utilisation for CCGT’s, using 80% and 20% capacity factor assumptions for the DESNZ 2023 cost estimates. My chart assumes that the first three cost buckets are entirely fixed in nature.

The impact of utilisation falling from 80% to 20% drives up non gas/carbon costs by c. £10 per MWh on the DESNZ numbers.

Perhaps an extra £10 per MWh isn’t a massive deal in the context of c. £79 of gas/carbon costs at current conditions.4 But it’s worth noting that the relationship isn’t linear – my final chart extends the table above across a range of different capacity factor assumptions.

It’s interesting that this economic sensitivity to fixed cost utilisation starts to bite quite heavily below 30% capacity factor – and probably explains the logic for the introduction of a capacity market to the UK electricity system in 2014 – but that is a post for another day!

Appendix: Explanation of Cash Costs ex Gas/Carbon metric:

Step 1 was calculating base cost, or subtracting operating profit from revenue. For example, if a power station had £10m of revenue and £2m of operating profit, my base cost figure would be £8m.

Steps 2 and 3 were removing non-cash charges from this £8m figure - namely depreciation and charges associated with changes in decommissioning provision estimates.

Finally I added back any net additions to property plant and equipment. These were all after the commissioning date - I was trying to capture upgrade/overhaul work that might have been capitalised rather than expensed.