Surprised not more focus on marginal pricing during these ERCOT/SPP events. Thread:

Right now if the energy load is satisfied with 50% of generators under $100 and 50% of gas peakers at over $3000, *all* the generators are given the highest price (congestion can complicate)

Implicitly this rewards your cheap generators (coal, nuclear, wind, solar) at the cost of load. And the design also doesn't really incentivize the peakers - they set the price, so their profit margin is significantly less than anybody else.

Typically these price dichotomies are fleeting, but on multi-day 24-hr events, ends up being a blowout transfer of wealth. Some of this is by design, as generators need to recoup the cost of building their plants. Plus we want to reward them for being there when it matters.

However, imagine you didn't have this artificial market design and were a single utility (still exists in parts of the country) that owned generators to cover your own load. Your actual cost increase (passed to retail) would only be your spendy $1-3k gas peakers.

ISOs like to tout significant "cost savings" in various reports (the dark things Excel has seen over the years), but I doubt they will include periods like this where load (if you have an electricity bill that's you!) massively overpays the base-load generator fleet.

Ballpark numbers - SPP had $3k DA prices for most of the week. On Mon, at least 22gw of load was served by coal/wind/nuclear. This is cheap load costing $20-40/MWh. Paying them $3k/MWh is $66M/hr extra, or $1.5B for a single day (and that's assuming *all* gas gen costs $3k)

The ISOs paying the cheap generation high prices is a result of the "model", not a real "cost" that's incurred in these periods. In normal periods it's a reasonable incentive, but in emergency periods where the whole grid goes to a price cap, it's obviously enormous.

Electricity LMP pricing is heavily distorted thanks to 100+ years of work in marginal pricing (MP) models - fundamental inputs (bid/offer) are input to a MP model combining economic and physical limits to give you "prices". A quick overview:…

Each ISO also then has overrides and tweaks to their model that can get so complex that they periodically break. SPP had prices go to $50k (!!) last week that apparently were a result of a model glitch. Other periods the model didn't even "solve" and will be repriced.

The original reason for these pricing models is to increase efficiency as the single utility model also has tons of inefficiencies (i.e. they don't give a f---). But multi-day price cap situations result in such a grotesque distortion, might start to get more attention.

One option would be to have a whole different set of pricing for multi-day emergencies like this where the goal is to minimize pricing for load, i.e. daily or yearly caps where the fallback is a primal price-recovery system once incentives have exceeded some level.

ERCOT was supposed to have such a mechanism where once a year exceeds 30 hours of $9k, the limit drops to $2k for the remainder of the year. However the commission arbitrarily overrode this on Sunday 2/15 (repricing risk is *huge* from political pressure to undo this).

Another alternative will be political pressure to cancel the ISO model and revert to the private utility monopoly model.

This single week in ERCOT changed the average price since 2010 from $30.5 to $43.6, a 42% jump. That's a lot of market "savings" evaporated in a moment.

ISOs do have significant power to reprice, and with potentially their own survival on the line... that's a likely outcome over the next few weeks.

ERCOT just emailed that all settlements/invoices are on hold (*big deal*), so likely large adjustments.

To end, I ponder how much the academics who came up with the marginal pricing models incorporated all these bastardizations into their design.. repricings months later, price caps and constraints changed from day to day, glitchy models making RT signals useless...


• • •

Missing some Tweet in this thread? You can try to force a refresh

Keep Current with Thomas Toppleton

Thomas Toppleton Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!


Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!