Adversarial Electricity Portfolios

Controlled demolition of a tall smokestack.
Controlled Demolition. (CC-BY-SA Heptagon via Wikimedia Commons)

Can we construct adversarial electricity portfolios made of new zero-carbon resources that undermine the profitability of specific existing fossil plants? Some version of this is already happening, but it’s incidental rather than targeted. The economics of existing coal and nuclear plants are being eroded by flat electricity demand in combination with cheap gas, wind, and solar. Economical storage and dispatchable demand aren’t far behind. But how much faster would the energy transition be if we actively optimized new energy resources to undermine the economics of existing fossil generation?

WattTime and Tomorrow are both trying to estimate real-time marginal emissions by location (folks at CMU have created a similar publicly accessible historical dataset). This information would allow you to program your electric vehicle to charge, or your clothes dryer to run when renewable energy would otherwise be getting curtailed — meaning you’re using zero carbon power (which might also have close to zero cost in a competitive market). This is a great idea, and at least provides consumers the option of explicitly integrating emissions into their energy consumption decisions. If enough people do it, then we shift the overall load curve to dynamically follow the renewable energy production curve. We substitute generation-following load for load-following generation.

However, designing new generation or demand patterns to more specifically undermine the profitability of existing fossil resources could have another kind of non-incremental impact on emissions. If you can deprive a fossil generator of its most profitable hours of generation, maybe you can force it to be decommissioned entirely. Then you’ve not only displaced emissions from those hours you were able to compete with directly, you’ve also removed the emissions from all of the other hours the plant would otherwise have been operating (assuming that what it’s replaced with is less carbon intensive). It would be a kind of weaponization of zero marginal cost economics.

Destructive Creation

There’s more to this idea than just sticking it to the Carbon Barons.

In a lot of ways building new wind and solar is easier than walking away from existing infrastructure. New capital can go into anything profitable, but fossil plants and pipelines are capital that’s already been invested. They have powerful natural constituencies to defend them — the investors. And while new zero carbon power is great, those existing power plants matter. We can build all the renewables we want, but if don’t also shut down existing emitters, atmospheric CO2 (currently 406 ppm) will just keep on rising.

Many economic systems evolve to stable states that depend on participants having similar objective functions. Often you can count on others to not systematically keep doing unprofitable — or even just not-maximally-profitable — things. In many situations it’s hard for profitable but not profit-maximizing agents to take over (even if they do persist), since the other strategies accumulate more capital more quickly.

But in the case of shutting down existing generators, we’re not trying to take over by growing faster. In a sense these plants are already uneconomic, they just don’t know it yet. They’re zombies. Many of them wouldn’t have been built if the decision came up today. Certainly many of them wouldn’t be built if the decision happened 10 years from now. Instead what we’re trying to do is speed up the transition to a new stable equilibrium that’s already evident: lots of cheap, zero carbon, zero marginal cost generation, cost-effective electricity storage, and much more dynamic, dispatchable loads.

With that goal, strategic investments that undermine economics of existing generators — and probably also turn some profit — could make a lot of sense!

Mercenary Generators

So who are there climate mercenaries willing to make this kind of investment? How different would the economics be from profit-maximizing renewables, storage, and DSM? At the very least, a model of displaced emissions could be used to differentiate between otherwise economically similar zero carbon investments.

We could also separate the financial returns and the displacement of existing emissions into different financial “products”. If the project with larger emissions displacement is less profitable, maybe it could be brought up to par with a relatively small amount of additional capital from folks who value the emissions reductions explicitly. An estimate of the displaced emissions and the economics of two different projects would allow you to price the emissions. If project A has a net present value (NPV) of $100 million, and project B has a NPV of $90 million, but project B displaces 1 million more tons of CO2, then the implied emissions price is $10/ton of CO2. (Though really you’d get a probability distribution for the displaced emissions, based on future fuel price scenarios, technology cost curves, etc.)

Maybe the emissions displacement potential could be sold as a kind of carbon offset — a contribution to a Carbon War Chest dedicated to economically targeting existing emitters. Or maybe there’s strategic value for wind/solar/storage/DSM developers in actively forcing existing fossil generation offline, because it opens up space on the grid for new resources that otherwise couldn’t be added, given flat load growth.

What would it take to do this?

There’s a huge amount of public information about existing fossil plants, both economic and operational (which I am helping to compile for public use as part of Catalyst Cooperative‘s Public Utility Data Liberation project.).

How hard would it be to analyze the histories of all of the fossil generators in a competitive market like PJM or ERCOT, and estimate which hours of the year and weather conditions are responsible for the bulk of their profits, and then construct a portfolio of zero carbon, zero marginal cost resources that generates during those hours in the same locations?

How much impact would such a portfolio have on a targeted generator? What kind of generator would be best to target? Would we even find any vulnerable targets? Would the impact only be marginal, since that generator would still be selling into the broader market, and the impacts of this particular portfolio would be spread out across the whole planning area? How location sensitive are electricity prices really? What if the strategy were applied across a whole planning area?

It seems like you could identify some target candidates with a historical analysis, but understanding the consequences of adding a lump of adversarial resources to the grid would require more complicated dispatch modelling. You’d need to look at how other generators would respond to the new energy supply/demand curve. But I wonder if there might be some simple cases too, for existing generators that are particularly isolated and don’t have a lot of options, or that are economically linked to a particular, concentrated load.

Is anybody doing this kind of work already?

Leave a Reply