ORDC reforms bring more smoke than fire
The Public Utilities Commission of Texas (PUCT) has re-drawn the parameters of the Operating Reserve Demand Curve (ORDC) in an effort to improve reliability and decrease volatility, but the impact may be smaller than the PUCT is hoping for.
A major goal of the ORDC is to incentivize new, reliable capacity when needed by the market by creating price spikes. The two major changes made to the curve—a decrease in the price cap from $9,000/MWh to $5,000/MWh, and a shifting outward of the minimum contingency level (the point at which the price is automatically set to the cap)—partially offset one another. As a result, for most of its breadth, the new curve is not terribly different than the old curve:
Price levels will be slightly higher than prior if reserves drop below 6,000 MW, and rather lower if reserves drop below around 2,500 MW. Two potentially beneficial outcomes of the new ORDC may be lower credit requirements and associated exposure for unhedged positions, given the lower price cap, and slightly higher average levels of online reserves (but this is also achieved more directly through recent increases in non-spin procurement). These may be helpful in managing day-to-day operations. However, as one of the main purposes of the ORDC is to give price signals and incentives to new entrants, the new curve is not necessarily much of an improvement.
Simulated grid conditions illustrate lack of price level change
We simulated grid conditions going forward, considering variability in demand, wind and solar output, outages, etc. using a Monte Carlo approach. Statistically, total average price levels throughout the year are not expected to change meaningfully under the new curve:
This indicates that generator revenues under the new ORDC are not likely to change much, and new builds will not be any more incentivized to come online than they were previously. Market equilibrium will still occur somewhere between 10%-12% reserve margin at best. The market will have to send additional price signals to bring new, reliable capacity online. Other pending reforms at the PUCT may achieve this goal, but it is important to note that the ORDC reforms do not help much.
What about volatility?
The question of volatility remains. Will the new curve decrease extreme price outcomes such as occurred in February 2021? The answer is yes, but here again, the impact is more muted than one might expect. In our simulations, extreme outcomes are reduced only somewhat:
While the financial impacts of Winter Storm Uri within the energy market would have been less (approximately 5/9th of what it was), the new ORDC still relies on extreme weather to drive the bulk of needed revenues for new generators, and unhedged short positions remain very risky. Besides, the old ORDC already had a financial safety valve that automatically re-set the price cap to $2,000/MWh after a certain amount of exposure, but this was suspended for various reasons during the storm.
As long as reserve margins in ERCOT continue at ~10%-12% (see commentary below), the market will be vulnerable to crisis no matter how the grid operations and price signals change. At these levels, reliability is subject to the whims of physics (e.g. weather, wind, etc.) even if ERCOT performs perfectly and all capacity is online and responsive. Fundamentally, the PUCT needs to do more to incentivize new capacity. The changes to the ORDC alone may move the needle in the direction of the PUCT’s overall goals, but not by very much.
Why ERCOT’s reserve margin projections are misleading
Recent ERCOT Capacity, Demand, and Reserve Reports (CDRs) have consistently forecasted near-term reserve margins climbing to the 30%-40% range. This is misleading, for several reasons:
Excess reserve crediting for wind and solar: The CDR methodology uses average renewable output levels during top-20 historical load hours. This ignores two critical realities. First, the influx of solar will quickly create a net load shift, especially in summer. While gross load often peaks around 4pm in August, net load will soon peak at 7-8pm. Second, wind is now a large enough collective capacity source that poor wind conditions can cause shortages even if demand is not extreme. For example, August 15, 2019, had no top-20 historical load hours, but very poor wind conditions caused a grid emergency nonetheless. ICF uses an effective load carrying capacity (ELCC) approach that considers the holistic contribution of renewables to grid reliability:
We believe the CDR overstates reserve capacity by at least 7 GW for summer 2022, and the impact only grows over time. The CDR’s credit in winter season are closer to appropriate levels, but still uses a poor methodology and may also overstate the case.
Excess projections for new capacity: The CDR counts all capacity with a signed interconnection agreement (IA) as coming online per the listed COD. Historical trends suggest a signed IA alone is not a reliable metric for on-time success, and the CDR shows absurd results on this basis. Most notably, 15.8 GW of solar is projected to come online in 2023 alone by the most recent Dec 2021 CDR. While solar installations are accelerating, about 4.3 GW were installed in 2021; it is extremely unlikely that build rates will triple in the next two years. This overlaps with the first problem mentioned above since solar builds are so heavily over-counted for reliability purposes.
ERCOT’s true reserve margin remains in the range of ~12% today, leaving the grid still vulnerable. The problems mentioned above are only notional, since the CDR is just a forecast. But signaling matters, and if the PUCT wants to pursue a resource adequacy market (RA, or what is dubbed the “LSERO” option in the latest proposals), the issue will quickly become critical.
