A mini-thread on my and @JesseJenkins working paper, Electricity Markets under Deep Decarbonization (papers.ssrn.com/sol3/papers.cf…).
Hardly a week goes by without someone on #energytwitter speculating that proliferation of zero-marginal-cost variable resources (read: wind and solar) will mean the end of power markets as we know them. After all, how can you base everything on marginal costs when they are zero?
The simple answer is that fuel isn't the only type of marginal cost: the demand side of the market has opportunity costs and storage resources have intertemporal arbitrage opportunities. Even if all supply resources have zero cost, we would still see non-zero prices.
But the story doesn't end there, and it's still possible to argue that power markets in their current form might not work as well with high levels of renewables.
Even if you're not interested in the modeling, you might be curious to read Section 2 of the paper for an overview of these arguments.
In particular, this paper looks at the possibility that even if we get high enough prices on average, they will be so volatile that contracting around them will be exceedingly difficult, necessitating a change in approach.
To assess this possibility, we construct a model where the cost of capital is endogenously determined by interannual volatility in operating profits for supply resources and the degree to which they can hedge that volatility.
When Jesse and I were constructing these numerical tests, we both expected volatility to explode. But the story turned out to be a bit more complicated.
First the bad news: scarcity events do indeed concentrate in a smaller number of the modeled years, leading to higher volatility in profits.
As a result of this volatility, it becomes harder to finance peaking and backup resources in particular, contributing to lower reliability.
But on net, at least with standard estimates of the value of lost load, that's outweighed by the good news: by removing fuel price uncertainty from the system, overall investment risk goes down!
We want to emphasize that these numerical studies are exploratory. But, we hope the modeling framework is helpful for thinking about the issues involved and generating hypotheses.
Lastly, thanks to those of you that took a look last week, and sorry it's taken me until now to do this summary thread!
• • •
Missing some Tweet in this thread? You can try to
force a refresh
