A Free-Market Energy Blog

Electricity Planning: Physical vs. Economic (an exchange with Eric Schubert)

By Robert Bradley Jr. -- March 25, 2021

“The physical operation of the grid can be separated from the economic decisions of what power is generated, transmitted, and sold at wholesale or retail in terms of quantity, cost, and price.”

“Mistakes by ERCOT or the PUCT or Texas legislature make my point–this is a planning failure, not a market failure. Not physics but economics. A historian is weighing evidence about this either being a market or government failure. It is a government failure writ large.” (Bradley, below)

Electricity is different. Power flows must the centrally managed. Ergo, regulators and politicians must manage the grid.

WRONG. The physical operation of the grid can be separated from the economic decisions of what power is generated, transmitted, and sold at wholesale or retail in terms of quantity, cost, and price. Companies themselves, vertically and horizontally integrated, what might be called electricity majors, is the opportunity cost of the present regulatory regime. (The same applies to natural gas–gas majors would have been better coordinated to perform under peak stress conditions.)


The following exchange with power/regulatory expert Eric Schubert, a designer of the current PUCT/ERCOT market, shows how we were/are “talking past each other” on this point. I was arguing against planning in favor of a true free-market order; he was defending Texas’s existing regulatory framework.

Is does not equal ought,” as the philosophers say. Ideas have consequences, the lonely classical liberals say. The Great Texas Blackout of 2021 has created the grand opportunity to “check your premises” regarding one of America’s largest, most important, and most regulated industries.

Classical liberalism and market-process political economy are in play with electricity in a way that was scarcely imaginable several months ago. But many if not almost all classical liberals/libertarians are following an intellectual leader who is so involved in the planning agenda that the “fatal conceit” problem, the knowledge problem, is underappreciated.

More on that in a later post, This post focuses on a learning process between two people who are trying to get the other to “see” what is really going on.

I report, you decide. The Facebook exchange follows:

Bradley: Even free market types have been fooled into electricity planning. The Texas blackout calls for abolition of ERCOT and repeal of state and federal laws governing electricity. Bill Peacock explains: https://www.masterresource.org/…/ercots-planning-snafu…/

Schubert: ERCOT is an RTO [Regional Transmission Organization], which is the interface between commercial energy markets and real-time grid operations that keep the lights on. A successful RTO allows market coordination to address Hayek’s Knowledge Problem while SIMULTANEOUSLY allowing grid operators to address the Common Pool Resource known as reliability in ways that are consistent with the insights of Elinor Ostrom. On the vast majority of days, ERCOT manages both as well as anywhere. The Texas Freeze was a failure of weatherization standards of natural gas from wellhead to burner tip intended to be insurance against extreme and rare cold weather events. That failure simultaneously undermined both reliability and the market.

Bradley: How can a government planning entity like ERCOT under PUC direction qualify as Hayekian? There are physical ways of controlling the flow of power and alternatives of dispatching, but that is an entrepreneurial function that only private entities not subject to PUC regulation or the state legislature can determine. Balancing lowest price and reliability (energy vs. capacity markets) is a planning question in the absence of private property actors in a free market, absent mandatory open-access and PURPA’s must-take rule. And why is electricity a ‘commons problem’ that cannot be solved by non-franchised firms practicing horizontal integration, vertical integration, and interfirm contracting?

Bradley: You state: “The Texas Freeze was a failure of weatherization standards of natural gas from wellhead to burner tip intended to be insurance against extreme and rare cold weather events. That failure simultaneously undermined both reliability and the market.” That is like saying “Enron failed because of negative cash flow and a loss of investor confidence” and behind that, “Enron made many bad investments.” But why–as in the why behind the why? Spectacular failures beg the question of why … in this case why the simultaneous failure of numerous private companies in natural gas and in electricity? That’s where I am focused.

Schubert: Real-time dispatch is an interface between reliability modeling of the centralized grid and decentralized commercial energy markets. The energy markets use bilateral contracting to bring resources on the grid and allow non-dispached resources to be in B a position to respond to real-time prices that ERCOT produces.

Bradley: All that is the ‘planning model’ of ERCOT, yes. But why 80% of the market? Why so heavy on energy versus capacity payments? Why $9,000 per MWH–and for so long? …. Why the greatest failure in the history of the US power market where the planners have reigned? There is surely a free market alternative of electricity and natural gas ‘majors’ to ensure reliability at the peak, winter or summer.

Schubert: Chairman D’Andrea has stated that if the PUCT had been able to anticipate the duration and persistence of the shortage during the freeze, he might have ordered ERCOT to RUC the units and impose a lower clearing price.

Schubert: Because real-time reliability is a common pool resource, you have to have centralized dispatch. A disturbance hundreds of miles away can cause local problems. Power is also non-linear and moves close to the speed of light. Market participants can’t clear the market fast enough (in seconds) and have all the engineering information to clear the market and keep the lights on.

Schubert: What marries the engineering model and commercial energy markets is the duality theory, which states that a non-linear optimization model can mimic the outcomes of a market, all other things being equal. Real-time interface works because the assumption holds in five-minute intervals.

Bradley: This might be the recipe for external planning–which failed in this case spectacularly for reasons of economics, not physics, but why cannot this be done within the firm? The properties of electricity points toward large, even very large, scale economies for the firm. The same physics existed prior to ERCOT and way back in the Edison/Insull era. Why did the mother of all power failures occur in a highly regulated environment? It’s economic planning, not physics.

Schubert: I think we are talking past each other. Also, customer choice is feasible with 21st Century technologies such as wind, solar, demand response, small turbines, and batteries, because they are smaller scale than nuke and coal phantasy.

Schubert: Enron failed because it cooked the books.

Schubert: Private companies and regulated pipelines failed, in part, because they and the Texas Legislature didn’t value the negative interactions between the gas and power segments that led to a cascade of outages during a 1 in 30 cold snap.

Bradley: Mistakes by ERCOT or the PUCT or Texas legislature make my point–this is a planning failure, not a market failure. Not physics but economics. A historian is weighing evidence about this either being a market or government failure. It is a government failure writ large. Can I at least get you to imagine an alternative of electricity majors dealing with reliability in the absence of PUHCA, PURPA, and the rest of it?


A Second Exchange: LinkedIn

At LinkedIn, Eric Shubert shares his latest thinking on correcting market design problems, followed by my query:

In the aftermath of the recent Texas Deep Freeze, the PUCT is currently undergoing a review of the ERCOT Scarcity Pricing Mechanism. This topic is of keen interest to me, as I was co-lead on the initial PUCT rulemaking in 2006.

In response a request for stakeholder comments, the Texas Industrial Energy Consumers (TIEC) filed a proposal at the PUCT that is both theoretically sound (differentiating between shortages resulting from installed capacity vs. operational issues) and works with the current ERCOT market design.

TIEC proposes to keep the HCAP at $9,000 per MWh. The PUCT can order the LCAP at $2,000 per MWh under specific reliability circumstances that result from market disruptions that prevent market participants from effectively responding to price signals. Generation owners can ask for make-whole payments for those hours where their variable costs exceeded $2,000 per MWh.

The rule also would eliminate the Peaker Net Margin concept, which was appropriate fifteen years ago when the rule was first adopted, but no longer has a role in a market increasingly dominated by 21st Century technologies.

I hope the PUCT adopts TIEC’s proposed rule language on the last page. You can find the filing here:

Bradley: Just to be sure, this is more (‘corrective’) central planning in the place of earlier planning that ran into problems. Outsourcing reliability from Grade A corporations to a plethora of companies under the thumb of regulators/politicians is problematic as far as getting the ‘right’ answer between price and reliability.

Is there a ‘market failure’ if we remove franchise protection, PURPA, PUHCA, mandatory open access, and rate regulation?

Shubert: Grid reliability has to be centralized because real-time reliability is a common pool resource (or something that shares most of those characteristics). The pricing TIEC is proposing sends signals to those who are both injecting into and withdrawing power from the grid (the CPR) to take appropriate actions that will reinforce real-time reliability during times of grid stress. At times that reliability is not a serious risk, the real-time pricing often is the marginal cost of the marginal generator plus grid congestion.

Bradley: Separate physical from economics. The ‘common pool’ problem is a reason for vertical/horizontal integration for ‘electricity majors’. It is not a reason to centrally plan and politicize the grid. Planners always say ‘we just got the rules wrong’ and will do better next time–but …

Are you familiar with Austrian-school economics that criticizes the neoclassical notions of marginal-cost pricing, perfect competition, etc. that seem to be behind the current regulatory scheme? Entrepreneurship does not work well under planning and static notions of efficiency.

Shubert: The two are linked. A successful power market needs to simultaneously honor two dynamic, complex phenomena – centralized grid reliability that keeps the lights on and decentralized commercial energy market that drives innovation and experimentation over time. The two meet in real-time, where the RTO real-time market serves as an interface between the engineering model [approximating reliability needs] and the commercial market. Nothing beyond real-time (or possibly real-time and day-ahead) should impede commercial markets from communicating and coordinating information (solving Hayek’s “Knowledge Problem”) and innovating (Schumpeter’s “Creative Destruction.”)

Bradley: Thanks, this is helpful. The problem remains: this is a contrived market: a non-free-market market. It is planned, It is inherently political. And in the case of PUC/ERCOT, it just failed. The greatest failure in the history of the 130-year-old central-station electricity industry, in fact. Now, new rules will be proposed … in the face of politics. The “Knowledge Problem” helps explain the recent failure. Don’t associate Hayek with what happened last month. And “Creative Destruction” in this case is regulatory/political, not capitalistic in the sense that Schumpeter formulated it. (Correct me Michael Giberson or Lynne Kiesling as needed.)

Shubert: There is no way to avoid planning when it comes to grid reliability, because you would be constantly at risk for blackouts as long as you share power across power lines. The failure in February was largely the lack of weatherization of the natural gas supply. The power market itself did not fail. What we have in ERCOT is the closest approximation to a free market that can be had given the unavoidable reliability constraints associated with a meshed AC power grid.

Bradley: This is where we disagree. We need planning as in planning by the firm(s), not planning by the State (regulators, politicians). We need a true free market that would bring economic decisions inside the ‘electricity majors.’ Reliability, etc. is too important to outsource to experts/regulators/politicians. We do not need to ‘approximate’ the free market, we need the real thing. On causes, one must get to the why-behind-the-why. Why the massive failure? Hint #1: electricity is one of the most regulated industries in the U.S. A witches brew of state and federal intervention set up the ‘market’ for failure. I have some posts coming up on this at MasterResource and at IER with more details.

Final Note: If anyone thinks that the Texas electricity market is unregulated and not subject to very difficult questions of pricing and access rules, please consult Public Utility Commission of Texas, A Primer on Wholesale Market Design (51 pages).

One Comment for “Electricity Planning: Physical vs. Economic (an exchange with Eric Schubert)”

  1. kakatoa  

    Thanks for all the references!

    It seems the Challenger disaster failure analysis didn’t make it to the folks using 1/30 events for engineering/market(ing) designs.



Leave a Reply