Large-scale wind and solar deployment produces a counter-intuitive effect: it makes gas plants simultaneously more necessary and less profitable. Our modelling of the French power fleet to 2030 quantifies that contradiction and its cost to society.
A statistically very low wind capacity credit
The starting point is the variability of wind output. In 2011 the national capacity factor fell to 2 % for several hours, against an average of around 23 % over four years. Our Monte Carlo simulation, over a thousand draws and two separate fleets, onshore and offshore, shows that only 6 to 7 % of installed wind capacity can be considered firm at nine times out of ten, or 2 GW for a 32 GW fleet.
The implication is direct: intermittent capacity does not substitute for investment in flexible assets. Our four 2030 scenarios point to a need for at least 5 GW of additional gas capacity, assuming the nuclear fleet is maintained at current output. Any reduction in nuclear falls on gas.
The composition of that fleet also shifts. Open cycle turbines, at 30 % of gas capacity, should rise to 40 to 45 % by 2030. Their lower fixed costs, around €65/kW/year against €90 for a combined cycle, suit short call durations, and their faster start-up better matches intermittent variation.
The merit order effect crushes utilisation rates
Must-run output, injected outside the market, shifts the merit order and mechanically displaces peaking and mid-merit assets. Our energy modelling gives combined cycle utilisation of 20 to 25 % in 2030, and close to zero for open cycle turbines. By comparison, UK combined cycles ran at 77 % in 2008, the normative rate for that asset class.
The effect is highly sensitive to wind volume: an extra 8 GW costs combined cycles more than three points of utilisation.
A loss of infra-marginal rent compounds this. Each GW of installed wind lowers market prices by around 0.5 % and infra-marginal rent by 0.8 %. That fall hits nuclear above all, which bears 95 % of it by volume.
Missing money of €900 million a year
The phenomenon was not theoretical even in 2012. Average spot prices had fallen from €69/MWh in 2008 to €49/MWh in 2011 with gas prices flat, and combined cycle utilisation from 72 % in 2009 to 44 % in 2011. The Pont-sur-Sambre plant, commissioned in 2009, entering safeguard proceedings illustrated the position concretely. Of 9 GW authorised in the 2009 investment programme, 4 GW never materialised.
By 2030, cumulative missing money would reach around €900 million a year across the French gas fleet, of which €500m for combined cycles and close to €350m for open cycle turbines, whose shortfall approaches their full fixed costs.
Those €900m a year sit on top of the €4 to 8 billion a year of costs tied to wind and solar feed-in tariffs. They will be borne by consumers or taxpayers.
Market design as the adjustment variable
The capacity market provided for in the NOME law targets security of supply at peak. It does not specifically address financing back-up capacity and could therefore drive sub-optimal investment. Other routes deserved examination.
- Targeted public tenders, as launched in Brittany.
- Reform of renewable subsidies, moving from pure feed-in tariffs to a capacity payment combined with energy sold on the market.
- Incentives for efficient renewable integration, by exposing generators to imbalance costs and so improving output forecasting.
- A higher carbon price, which would lift infra-marginal rent for every technology except coal, in a French market where the price is set by German coal.
- An ancillary services market, where generators responsible for imbalances would buy those services.
To go further, download our full analysis below or contact our experts.
The downloadable PDF of this study is available in French only.
