A nice classroom case on green economics. An inter-corporate fight has broken out in Jersey (the original small island off Normandy) between Jersey Electricity and Jersey Gas. Gas is all imported in small tankers; since 1984, electricity has been imported from France through a submarine cable, now leaving only vestigial local generation. So Jersey Electricity trumpets its low carbon footprint, as French generation is 78% nuclear and 88% low-carbon (2004). Jersey Gas isn’t happy as there’ s nothing it can do to lower its carbon footprint short of dying.
The States of Jersey are revising building regulations to set a carbon emission standard that gas can’t meet. With its back to the wall, Jersey Gas looked for talking points against electricity. They took out a full-page advert in the local paper:
ELECTRIC HEATING – HOPE YOU LIKE IT, BECAUSE YOU’RE STUCK WITH IT. No matter what the cost to you and the environment
…the States of Jersey Environment Department think that electricity is low carbon and good for the environment. They are wrong. On island electricity generation is high carbon. about 3 times that of gas. The JEC will confirm this. European electricity – what do the French think? The French Secretary of State for Ecology … interviewed in Le Monde, 1 October 2008, “We have a serious problem with electricity heating in France. It was a mistake to develop it …” … THIS IS GOING TO COST: YOU. Electricity is more expensive than other fuels. THE ENVIRONMENT. Electricity is not a low carbon fuel …”.
Jersey Electricity complained to the UK Advertising Standards Authority and won.
What’s worth a second look is the claim Jersey Gas’ lawyers made in the proceedings that since the European electricity market is integrated, the carbon intensity of Jersey’s electricity is that of marginal European supply: which is mostly fossil, and higher carbon than locally consumed gas. If that’s so, then the claimed advantage over gas disappears.
Right or wrong? The answer is relevant to any region within a grid, like California (connected to the rest of the continental USA), the USA (connected to Canada) and China (connected to Russia).
Answers welcome – for once I can guarantee the attention of a live policymaker.
My take: Jersey Gas has one good point but is still wrong.
A market is a social institution, not a technical fact. The European bulk electricity market is being developed under pressure from Brussels, but it hardly existed when the first submarine cable was laid in 1984. The cable was certainly not a standalone investment by entrepreneurs who proposed to buy electricity in a market which didn’t exist. It was part of a long-term contract with EDF: without access to the details, it must be the case that EDF promised to sell and Jersey promised to buy, over a period long enough to amortise the investment. This contract makes Jersey an integral part of the French grid and consequently its demand is part of EDF’s operational management and forward planning.
EDF (or the grid operator RTE, spun off from it at Brussels’ insistence) has different and more arm’s-length relationships with other cross-border counterparties in Belgium, Germany, Spain or the UK over interconnectors. The market here is intraday. EDF will presumably try to sell spot electricity if there’s a surplus from nuclear and wind, otherwise not. So institutionally a French frame of reference is correct, subject to a detailed look at the contracts.
Jersey Gas’ good point is to look at marginal carbon emissions. For gas, it’s a single number: 214 gm/kwh according to the Carbon Trust, a respected UK NGO. For electricity, it depends on the time of day and year. If you are deciding today in Jersey whether to switch on your washing machine at 6 p.m. or midnight, the relevant variable for carbon worry is the short-run – instantaneous – marginal cost. It’s of course desirable for consumers to be given real-time information about these costs. But this is the wrong framing for planning and regulatory purposes. The argument is about public policy, in the form of building and other regulations and taxes designed to affect consumers’ investment decisions and habits of life over a horizon of decades.
A more wonkish approach could go like this. Start with your annual curve of carbon emissions per marginal kwh, supplied by the generators. [Update It should have a mere 8760 data points, one for every hour of the year.] Next, for any given category of appliance, you estimate an annual load curve [for the same time periods]. Finally you sum the marginal carbon emissions over the year [the scalar product], and take the mean. This won’t be the same for a home space heater – daytime and evenings excluding summer – or an electric car – night-time all year. The worries of French greens about space heating are not misplaced. Jersey regulators can’t rely on “electricity is greener”: they should dig properly into typical load curves for different types of use, and insist on smart metering to empower and inform consumers.
You still have to choose between short- and long-run carbon costs. Picking short-run, the advantage is that you can estimate the numbers to some accuracy for France. Installed French nuclear capacity has been steady at 63 GW for decades; it is 85% available. Assume the downtime is flat – an implausible worst case, as I’m sure EDF manage to schedule regular maintenance for the summer. The latest annual consumption curve I could find was for 2003, but the shape won’t change much. Combining the weekly output with the daily variation, I got this only slightly stylised supply and demand curve:
Source: RTE, processed
It looks as if nuclear energy covers all demand for six months of the year, mean demand for nine months of the year, and the nighttime minimum more or less the whole year. The French have so much nuclear capacity that capacity usage is only 77%. For the remaining 7%, equivalent to 25 days a year, some plants are idled because RTE can’t sell the output to anybody, interconnectors or no. Marginal output is therefore non-nuclear for only a fraction of the year – eyeballing the graph, no more than a quarter. Deduct hydro and wind, and you only need made a small correction to carbon intensity for fossil. Taking the mean, French electricity is only 5% carbon-intensive; the summed marginal share may be a few times that, but no more. Jersey regulators are entitled to ask for an exact carbon emission curve.
If it’s investment decisions you are trying to influence, the correct marginal cost is formally the long-term one. This is much more iffy. However, if the correct institutional frame is France, then it’s easy enough to check qualitatively what EDF are planning: more of the same. The plan is to replace the current nuclear reactors with a new standard design over the coming decades. Since the new EPR design is rated at 1650 MW and the oldest of the existing park are 900 MW, EDF can easily increase nuclear capacity if it wants to. EDF will also grow its wind park of 2 GW. Whatever happens to nuclear costs, its carbon intensity will stay negligible.The short-run carbon curve is a good approximation to the long-term one.
Can Jersey Gas win the argument if we concede their use of the European and not the French market as the reference? There is a lot of uncertainty and dithering about new nuclear capacity anywhere outside France, so the long-run European marginal carbon intensity is guesswork. We can say that new capacity will be a mix of nuclear (from France), wind or gas, so long-run marginal cost is a hybrid of these: conservatively ignoring the menu of other jam-tomorrow renewables (PV and thermal solar, tidal, geothermal). The economics dictate that wind output will be taken up first, even if it has to be balanced half the time by gas (or in some places hydro). A third each would be a reasonable guess. Modern gas generation produces 360 gm carbon per kwh (here, p.9). So worst-case long-run marginal carbon intensity is unlikely to be higher than 150 gm/kwh (allowing generously for transmission losses). Jersey’s imported gas comes out at an uncontested 214 gm/kwh, so Jersey Gas can’t win even stacking the deck.