Michael Giberson
Only after posting my earlier examination of the interaction of wind power, the production tax credit (PTC), and negative power prices in ERCOT did I discover a related analysis, “Curtailment, Negative Prices Symptomatic of Inadequate Transmission,” by Michael Goggin, an American Wind Energy Association analyst, that appeared in September at Renewable Energy World. If you are reading along at home, you may want to take a look.
Also, in October the New York ISO issued a white paper, Integration of Wind into System Dispatch. The NYISO reports that sometimes it has experienced the sudden shut down of wind power during times of negative prices, with in some cases more wind power dropping off the system than would be necessary to relieve the congestion constraint (and so to allow prices to return to positive levels). When “too much” wind power suddenly drops off the system, that drop off puts additional strain on the system operator and the balancing resources available to the market.
The solution that NYISO is pursuing is to better integrate wind power into the system operator’s “security constrained economic dispatch” market model, with a goal of better coordinating wind and non-wind generation along with available transmission capability.
I agree that the PTC is probably not the best policy tool, however, the assumption that value destruction is occurring may not be as warranted as your analysis suggests.
The transmission/market design comments by Lynne and the AWEA analyst capture most of my thoughts on the topic. This is just a milepost on the road to significant market savings as the wider grid gains access to this power. A few additional points…There are non-environmental financial externalities for fuel consumption by individual merchant generators whose plant economics drive them to bid power into the market at negative levels during off-peak periods. Part of the fuel cost of the decision not to shut down is borne by non-producer fuel consumers(Less fuel is available for other uses, hence fuel prices are higher for all fuel users, these costs are borne by the producer only to the extent of his market share, and to the extent the fuel market is homogeneously comprised of similarly situated power producers). To the extent that merchant plant fuel consumption is offset by subsidized wind production, the cost of those externalities is avoided, and some fraction of the subsidy is recouped in lower fuel prices. Effectively, this is energy storage of wind power in the form of natural gas. Further, to the extent that real-time pricing is being utilized, one assumes that these prices are causing investment to allow industrial load-shift, which load shift has the effect of raising these prices, lowering on-peak prices, and reducing the demand for inefficient peaker resources (hence reducing both fuel consumption and wear and tear).
Since much of the subsidy is also being recouped in the differential of the lower wholesale price it does not seem clear to me that the sign of the net value of this power production is negative (i.e. wasteful) despite the negative clearing price.
Interesting comment benamery21, so thanks, even though I’m going to argue with you.
I don’t think we can count the non-environmental financial externalities associated with less fuel use in favor of PTC-subsidized wind generation. Aren’t these simple pecuniary externalities not associated with resource mis-allocation?
Unless, of course, there is some further link to a non-pecuniary externality down the resource chain (say, environmental affects associated with development of fuel resources). So, still, the policy guidance from economics is to try to price the fuel-development externality directly rather than subsidize (some of the) competitors of the fuel users.
It would be interesting to look for evidence of load-shift associated with the negative prices. My intuition is that negative prices are likely fleeting – here for a while, but not worth a long term investment to capture – and relatively difficult to forecast (must forecast both wind output, system load, and zonal transmission congestion which depends on the dispatch of other generation resources on the system). So I wouldn’t think it easy for an industrial user to chase negative prices very effectively. Perhaps I’m underestimating the value available to flexible power consumers.
If I had an electrical load that tended to increase in the Winter and Spring, and higher in early mornings than later in the day, I might explore the possibilities. Heating ski lodges maybe? Unlikely to be big business in ERCOT’s west region.
Mike: ‘Aren’t these simply pecuniary externalities?’ Maybe. I’m not an economist, I’m a utility engineer. But if I owned ALL of the generation and served ALL of the retail load (gas and electric), and had a limited amount of gas available in real-time, I know which generator I’d use when the wind was blowing, and it isn’t the gas plant (even absent subsidies).
Since neither of the substitute goods is simple, my gut says the knee-jerk (don’t use pecuniary externalities in cost-benefit) is missing something here. If you use the gas, the wind and gas are gone. If you use the wind, the gas is not (and gas is not cheap/plentiful right now). The bounds of the gas market and the electricity market in the case are different (less so after the transmission is built). The gas market probably includes the adjacent part of the state where gas is selling to plants producing $100/MWh power. The gas plant bidding in at a negative price probably has a contract fuel price well below the spot gas market (for a fixed amount of fuel).
I would not expect typical load shift to be driven by the presence of negative prices but rather by the summation of the difference between prices in two different hours each day or between the lowest price in any day and a specific time, or the nth lowest (given a sufficiently flexible new process). It is possible that negative prices aren’t raising this differential by much in the aggregate. I was taking negative prices as indicative of the presence of a large number of very low price point hours which sometimes go negative, and the increase in the number of negative points as reflecting an aggregate price shift lower for hours when wind is present, while the preponderance of gas generation in Texas and the price of gas leads me to believe that prices on-peak are still unlikely to be low. My conclusions would be better supported if backed by use of actual data.
I guess the basic question for dispatch if we remove all the market induced complexities by specifying a single utility owns all generation and serves all loads is: Is the avoided fuel cost of running as spinning reserve at no load (some gas used to keep plant hot) rather than producing power in my gas plant more than the increased maintenance cost of producing power in my wind plant. I guess I can’t imagine a world with today’s gas prices where the answer is no.
I know next to nothing about the Texas grid (I’m a WSCC type and not a grid control type anyway).
Now, if there are truly no non-wind resources on in West Texas during these negative price periods, which I doubt, that changes my gut level analysis.
If that is true, then possibly a strong difference in resource mix in two other zones, plus transmission inadequacy due to wind construction having outpaced transmission construction, means that when West Texas is exporting power then power production in one of the other zones is being favored over power production in the third zone. If this is driving use of sufficiently less efficient units in the favored zone it is theoretically possible that the wind generation being on is actually increasing gas consumption in real time. Now THAT’s wasteful. The fact that this might be possible, does point up the distortion by PTC of the market.
I would also note that I can imagine a scenario where wind generators ran, even when prices were more negative than government incentives compensated for, due to the terms of private sector loans.