Special Report: A Perspective of the Real-Time (LMP) Price Spike of February 2021

As all of us are aware, the extreme cold incursion from February 15th to February 19th of this year was damaging to virtually every part of Texas life. The lack of preparedness was a key element. Making sure that such a calamity isn’t repeated is currently being debated. How price was affected is a significant part of the debate and directly concerns end-users who employed real-time pricing in some degree for procurement. We, along with our industry peers, have long incorporated real-time pricing for suitable clients knowledgeable and understanding of the accompanying risk. That’s a mouthful that doesn’t do justice to the discussion. The following is our attempt to explain.

Real-time pricing (LMP – Locational Marginal Price) is immediate power priced by supply and demand by generators. The market design for pricing is a set of complex algorithms that run independently of human involvement. ERCOT is an energy-only market that follows pricing signals to insure adequate assets for grid needs. In the “normal” course of business, shortfalls of generation are supplied by generators who are compensated through the pricing model. Should additional generation be required, prices increase to incentivize this generation.

We make procurement decisions and manage the accompanying risk guided by data and history. All pricing is an arbitrage that can be and is analyzed. Real-time pricing, prior to February 2021, provided a comfort level in regards to the risk component. There were price spikes that occurred that were painful and surprising (you never knew when a hiccup might occur), but, when averaged with “normal” price expectations, proved to be lower than fixed price alternatives. Below you’ll see a chart of the price patterns from nodal inception (December 2010) through January 2021:

Surprise and inability to forecast these spikes require vetting of which load should consider its implementation. Does the net dollar difference expect to justify the risk? Does the targeted load understand that there will be spikes? Some loads didn’t qualify but pursued this strategy anyway, sometimes that will require correction.

Our experience with LMP was that when the grid was stressed, higher prices would kick in for sufficient time to balance supply and demand. It was temporary. The market design worked and ERCOT provided dependable delivery. The spikes heretofore experienced followed this pattern for the life of nodal up to February. No one we are aware of foresaw a “perfect storm” of circumstances that might result in the unimaginable magnitude of prices experienced in February:

As mentioned above, prices are generated algorithmically and determined under set rules and protocols that change when the grid is in emergency status. Such was the situation in February. There is no doubt of the legitimacy and accuracy of prices that were generated. Arguably, these results weren’t foreseen by designers. Nothing in our experience suggested that prices could go to price caps and stay there for most of the week. Certainly many facilities were disconnected and forced outages were a way ERCOT could maintain grid integrity. However, the consequences were surprising (and shocking) to the industry. It is noteworthy that generators (wholesalers, traders) were generally profited by the spike while REPs and loads were damaged by the spike.

There are two camps in the discussion regarding this price spike. One sees this as an orderly response to circumstances that dictated higher prices. The other sees this as a price response of such magnitude and reach as to be unconscionable whose extreme deserves mitigation. We adhere to the latter. We believe that algorithms should be tempered with common sense. Decisions have yet to be made and the discussion has become political. Time will tell.