Frequent negative power prices in the West region of ERCOT result from wasteful renewable power subsidies

Michael Giberson

What is with all of the negative power prices in the West region of ERCOT?

Frequency of negative prices by data, ERCOT West, 2008 In the first half of 2008, prices were below zero nearly 20 percent of the time. During March, when negative prices were most frequent, prices were below zero about 33 percent of the time. After mostly taking the summer off, negative power prices were back to near 10 percent in October.

[Chart at left shows the number of 15-minute intervals each day that had prices below zero from January through October, 2008.

UPDATE: Charts now revised to include all 2008 data. Contact author to receive full size chart.]

This seems a little crazy. During these negative price periods, suppliers are paying ERCOT to take their power. Consumers (at least at the wholesale level) are getting paid for using power, and the more power consumers use the more they get paid. These prices are a big anti-conservation incentive. You could, as a correspondent put it to me, build a giant toaster in West Texas and be paid by generators to operate it.

In fact most of the regional power markets that are integrated into systems operations (so-called RTOs and ISOs in the U.S.) will produce a negative power price now and then. On the margin, a power supplier should offer power into the market at approximately the net marginal cost of supply, at least in a competitive market. These offers are typically at positive prices and the market will produce a positive price.

Infrequently, a power plant might choose to bid below the short term marginal price in order to stay in the market and avoid shutting down. It can be economically rational for operators of less responsive generation units to offer negative prices in order for it to avoid the costs of shutting down for just a few hours and then start up again when load increases – think coal-fueled or natural gas steam turbine. When energy load is very low, near zero or negative prices can result.

This isn’t the cast in West Texas. Instead, the negative prices appear to be the result of the large installed capacity of wind generation. Wind generators face very small costs of shutting down and starting back up, but they do face another cost when shutting down: loss of the Production Tax Credit and state Renewable Energy Credit revenue which depend upon generator output. It is economically rational for wind power producers to operate as long as the subsidy exceeds their operating costs plus the negative price they have to pay the market. Even if the market value of the power is zero or negative, the subsidies encourage wind power producers to keep churning the megawatts out.

Frequency of negative price by price levelEvidence from market data suggests that wind power producers will accept prices down to about negative $35 MWh before they shut down, since marginal operating costs are very low for wind power we can conclude that the subsidies are worth about $35 – $40 for each MWh of wind output. [UPDATE: Chart now includes data through December 2008.]

Subsidies do this sort of thing – distort the market and lead to waste – and of course to some degree distorting the market is just what is intended when policymakers offer a subsidy. Only usually it isn’t so easy to see the evidence of the waste created by the subsidies. Wind turbines that operate more hours require more maintenance, so these hours spent producing negative-value electric power do consume real resources. At the same time, the conventionally-fueled generation that is forced offline temporarily will also face additional “wear-and-tear” and require additional maintenance because of the effects of shutting down and then restarting the machines. This extra wear-and-tear and extra maintenance also represents wasteful use of resources due to PTC- and REC-subsidized power production.

The subsidy for renewable power may be defended as compensation for avoiding the environmental costs associated with power produced by conventional means, but in this case the link between the payments and the possible reduced emissions effect is tenuous. In Texas the PTC is probably offsetting natural gas generation most of the time, perhaps a relatively efficient combined-cycle gas unit, but maybe an inefficient old steam generator. Sometimes the PTC will displace coal-fired generation. The environmental benefits will vary dramatically depending upon just which kind of unit is displaced by the subsidy, but the cost of the policy is the same. Surely there are more targeted and effective ways of achieving environmental goals.

A second possible defense for the renewable power PTC is that it will spur enough growth in the industry to allow progress in research and development and economies of scale to reduce costs in the future. I think these learning and economies of scale arguments are much abused in renewable policy discussions – treated as if they are somehow automatic if we only spend enough resources now. If learning by doing and economies of scale were automatic, the U.S. auto industry would now be a paragon of efficiency. (A paper on “Learning Curves For Energy Technology and Policy Analysis“, by Tooraj Jamasb and Jonathan Kőhler is on my “to read” list, but I haven’t read it yet.) In the wind energy case, the industry is led by huge international corporations like General Electric, Siemens, and Gamesa. These companies and many others have been in the business for years, and in some cases decades. This is hardly a case of an “infant industry” that needs a handout to grow to maturity.

Maybe there is a public good argument buried in this line of thinking, but like the externality argument my sense here is that some alternative approach would more effectively achieve the desired public policy goals.

I don’t see any easy approaches for Texas. The federal PTC is the main subsidy, and localized evidence of waste due to the PTC in part of Texas in unlikely to derail U.S. Congressional support. Even if more detailed examples of widespread waste could be produced, I’m not sure it would overcome the coming Congress’s warm fuzzy feelings for renewable power. Possibly Texas could take-away the Renewable Energy Credit for wind power generated at negative prices, and that would slightly reduce the waste. But the boom in wind power construction in Texas has already greatly reduced the value associated with a REC in Texas, so taking it away altogether wouldn’t do much. And really, the negative prices in ERCOT’s energy markets are only an especially visible indicator of the waste created by PTC-based distortions, any excessive investment in renewable power or production from existing wind power units at below-cost prices is wasteful.

To be clear, I’m not arguing that wind power or other renewable power projects are inherently wasteful. The policy design is at fault, not the technology. It is the policy that needs repair. Also, I don’t have an estimate of how significant this problem is. Maybe the waste is in the hundreds of thousands of dollars, but could be higher or lower. There may be more significant problems to work on. But the PTC is a key element of renewable power policy, and it is troubling that it causes waste.

Economics provides some guides for fixing the policy: if an externality is the problem, then tax the externality and compensate the harmed parties; if the goal is additional learning, don’t tie the payment to per unit output, tie the payment to progress toward the learning goal.

Renewable power industries are pushing for further expansion of the PTC. Before Congress agrees, it ought to try to find less wasteful ways to achieve intended public policy goals.

Toaster image composed of toast

Advertisements

23 thoughts on “Frequent negative power prices in the West region of ERCOT result from wasteful renewable power subsidies

  1. DOE’s “20% wind power by 2030” should help make the issue more obvious.

    This just reinforces the need for storage in regions served by large percentages of intermittent and contra-cyclical generation.

  2. DOE’s “20% wind power by 2030” should help make the issue more obvious.

    This just reinforces the need for storage in regions served by large percentages of intermittent and contra-cyclical generation.

  3. So – If the problem is too much production from the wind generators… (is that right?) then it seems the “negative cost” is exactly right. Does this not create an incentive to build storage for this generated power? “Draw” it during negative price periods (making money by doing so because of the negative pricing), then “supply” it back to the grid (or to yourself) during high-demand periods.

    Thus, the regulation makes sense. Or, am I missing something?

  4. a little farther out, an “electranet” — or whatever you want to call a smarter, nationwide grid — would guarantee demand for generated power.

  5. a little farther out, an “electranet” — or whatever you want to call a smarter, nationwide grid — would guarantee demand for generated power.

  6. Why would a wholesaler change their habits if their energy costs are low? I can see how it would not lead to increased energy conservation but I don’t think this would lead to an increase in energy usage.

  7. With the country’s high demand for electricity, it sounds crazy to just let that power go to waste. They should be tying it to the national grid, making it available to everyone. Not only will the consumers enjoy lower prices, but the suppliers could make more money.

    I am missing the point here? It just seems so obvious!

  8. There are two other factors that matter in this story:

    1. Time required to build transmission
    2. ERCOT’s transition from zonal to nodal wholesale market design

    As these two factors take effect, those negative prices will change/diminish.

    Is this really such a wasteful way to induce more investment in wind generation capacity?

    Yes, as we have more smart grid capabilities, and hopefully as storage evolves, these negative prices will change.

  9. There are two other factors that matter in this story:

    1. Time required to build transmission
    2. ERCOT’s transition from zonal to nodal wholesale market design

    As these two factors take effect, those negative prices will change/diminish.

    Is this really such a wasteful way to induce more investment in wind generation capacity?

    Yes, as we have more smart grid capabilities, and hopefully as storage evolves, these negative prices will change.

  10. Lynne, you are right that building transmission and improving congestion management (which comes as part of the value of moving from zonal to nodal pricing in ERCOT, sometime next year) will both reduce the frequency of negative prices.

    In fact, ERCOT’s change in congestion management policy adopted approximately June 9 was associated with a fairly dramatic decrease in the frequency of negative prices. It would take a little more data than I have available at present on wind power output to sort out the value of the congestion management change vs. the usual tendency of wind power output to fall during the summer (and other seasonal factors) to sort out the relative impact of the congestion management change. The move to nodal should further improve congestion management and reduce the amount of time excess wind power is, in effect, shut-in in the west region.

    But perhaps related is the fact that in September and October negative prices reached as far as the Houston zone in eighty-five 15 minute pricing intervals, whereas in the Spring there were only two 15 minute intervals of negative prices.

    At least the subsidy is being passed along to more Texas consumers with better congestion management.

    Smart grid & storage can both help overcome related problems. I’m not sure that these negative prices directly provide an incentive for storage, because I don’t anticipate that they will last beyond that transmission build-out and change to nodal congestion management. But even if the negative prices don’t persist, to some extent wind power in West Texas will remain out-of-sync with summer power demands and so smart grids, real time pricing, and storage can provide complementary value (at least given the wind patterns in West Texas, off shore winds tend to have different diurnal patterns).

    While none of this – transmission, congestion management, smart grid, or storage – erases the allocative inefficiency due to the subsidy, they can all diminish the ‘collateral damage’ and thereby reduce the costs of the policy while maintaining the expected benefits.

    And, like I say, I don’t estimate the amount of waste, so don’t claim this is the biggest problem. Maybe this is the best “second best” policy available, but I want to raise a point of information suggesting improvements are possible.

  11. This is a great article and great discussion; thanks, Mike! I was just about to question whether this might be second-best, when I reached the very last line of your last comment. Never mind!

    Maybe this is covered by your reference to nodal pricing, but I’d like to know the extent to which negative wholesale pricing passes through to actual consumers. (I’m assuming that wholesale pricing generally reflects transactions between net producers of electricity.) In other words, does someone running the vacuum cleaner at 3 a.m. potentially get paid to do so, with an actual reduction in his/her monthly utility bill corresponding to that particular fraction of a KwH?

    I realize that time-of-use pricing can’t solve all the issues here. I would guess that one such issue would be the fact that the AC isn’t needed as much at 3AM in March as it is at 3PM in July, so price differentials are unlikely to shift the time of its use. Still, an “efficient” market should reflect marginal prices to end consumers, not just to middlemen.

    Sadly, I also agree with your point that this sort of data is unlikely to change Congressional (or even local legislative, imho) policy decisions. Politics is a blunt instrument — an excavator, not a laser.

    Best regards,
    Jim

  12. This is a great article and great discussion; thanks, Mike! I was just about to question whether this might be second-best, when I reached the very last line of your last comment. Never mind!

    Maybe this is covered by your reference to nodal pricing, but I’d like to know the extent to which negative wholesale pricing passes through to actual consumers. (I’m assuming that wholesale pricing generally reflects transactions between net producers of electricity.) In other words, does someone running the vacuum cleaner at 3 a.m. potentially get paid to do so, with an actual reduction in his/her monthly utility bill corresponding to that particular fraction of a KwH?

    I realize that time-of-use pricing can’t solve all the issues here. I would guess that one such issue would be the fact that the AC isn’t needed as much at 3AM in March as it is at 3PM in July, so price differentials are unlikely to shift the time of its use. Still, an “efficient” market should reflect marginal prices to end consumers, not just to middlemen.

    Sadly, I also agree with your point that this sort of data is unlikely to change Congressional (or even local legislative, imho) policy decisions. Politics is a blunt instrument — an excavator, not a laser.

    Best regards,
    Jim

  13. This is a great article and great discussion; thanks, Mike! I was just about to question whether this might be second-best, when I reached the very last line of your last comment. Never mind!

    Maybe this is covered by your reference to nodal pricing, but I’d like to know the extent to which negative wholesale pricing passes through to actual consumers. (I’m assuming that wholesale pricing generally reflects transactions between net producers of electricity.) In other words, does someone running the vacuum cleaner at 3 a.m. potentially get paid to do so, with an actual reduction in his/her monthly utility bill corresponding to that particular fraction of a KwH?

    I realize that time-of-use pricing can’t solve all the issues here. I would guess that one such issue would be the fact that the AC isn’t needed as much at 3AM in March as it is at 3PM in July, so price differentials are unlikely to shift the time of its use. Still, an “efficient” market should reflect marginal prices to end consumers, not just to middlemen.

    Sadly, I also agree with your point that this sort of data is unlikely to change Congressional (or even local legislative, imho) policy decisions. Politics is a blunt instrument — an excavator, not a laser.

    Best regards,
    Jim

  14. The extent to which a negative price reaches the power consumer will depend on the retail agreement between the consumer and the consumer’s retail supplier. I don’t know what is typical for the west region of ERCOT, but my guess is that retailers gain the benefit in the first instance and whether any is passed through to consumers depends upon the strength of retail competition in the region.

    Time-of-use pricing could be a help here, because it would tend to encourage load shifting to take advantage of changes in the cost of wholesale power. I’ve seen a research paper that estimates the benefits of combining real-time pricing with wind power in ERCOT, but I can’t find the reference at the moment.

  15. The extent to which a negative price reaches the power consumer will depend on the retail agreement between the consumer and the consumer’s retail supplier. I don’t know what is typical for the west region of ERCOT, but my guess is that retailers gain the benefit in the first instance and whether any is passed through to consumers depends upon the strength of retail competition in the region.

    Time-of-use pricing could be a help here, because it would tend to encourage load shifting to take advantage of changes in the cost of wholesale power. I’ve seen a research paper that estimates the benefits of combining real-time pricing with wind power in ERCOT, but I can’t find the reference at the moment.

  16. “With the country’s high demand for electricity, it sounds crazy to just let that power go to waste.”

    It’s entirely rational for all parties involved to let it go to waste. The generator can “bribe” the grid to accept their power and still get compensated by the tax payer; you can get payed to waste excess power by the wind farms.

    There’s no incentive to try and use that excess power for anything.

    “They should be tying it to the national grid[…]”

    There is no such thing as a national grid. Worse, even within each grid the transmission system behaves more like a singly connected graph where electricity goes in one direction from a bunch of predictable sources to a bunch of predictable sinks; it cannot be easily shuffled from anywhere to anywhere. If you tried it would put serious strain on the grid operators and could all to easily cause a black-out due to a careless mistake.

    “[…]Not only will the consumers enjoy lower prices, but the suppliers could make more money.

    I am missing the point here? It just seems so obvious!”

    Yes. You apparently believe that generating electricity is the hard part, but it’s not. Transmission of electricity is already a bigger fraction of the total cost than generation and that’s with numerous small grids and dependable gas, coal, hydro and nuclear sources. If you install “smart grid” technology, criss-cross the country with HVDC lines and pumped storage systems to make the most of wind with the least proportion of fossil fuels the cost for transmission will be astronomical. When the price of electricity is high people have an incentive to shift towards oil and gas wherever possible and heavy industry is eventually forced to flee or file for bankruptcy(see California).

  17. “With the country’s high demand for electricity, it sounds crazy to just let that power go to waste.”

    It’s entirely rational for all parties involved to let it go to waste. The generator can “bribe” the grid to accept their power and still get compensated by the tax payer; you can get payed to waste excess power by the wind farms.

    There’s no incentive to try and use that excess power for anything.

    “They should be tying it to the national grid[…]”

    There is no such thing as a national grid. Worse, even within each grid the transmission system behaves more like a singly connected graph where electricity goes in one direction from a bunch of predictable sources to a bunch of predictable sinks; it cannot be easily shuffled from anywhere to anywhere. If you tried it would put serious strain on the grid operators and could all to easily cause a black-out due to a careless mistake.

    “[…]Not only will the consumers enjoy lower prices, but the suppliers could make more money.

    I am missing the point here? It just seems so obvious!”

    Yes. You apparently believe that generating electricity is the hard part, but it’s not. Transmission of electricity is already a bigger fraction of the total cost than generation and that’s with numerous small grids and dependable gas, coal, hydro and nuclear sources. If you install “smart grid” technology, criss-cross the country with HVDC lines and pumped storage systems to make the most of wind with the least proportion of fossil fuels the cost for transmission will be astronomical. When the price of electricity is high people have an incentive to shift towards oil and gas wherever possible and heavy industry is eventually forced to flee or file for bankruptcy(see California).

  18. “With the country’s high demand for electricity, it sounds crazy to just let that power go to waste.”

    It’s entirely rational for all parties involved to let it go to waste. The generator can “bribe” the grid to accept their power and still get compensated by the tax payer; you can get payed to waste excess power by the wind farms.

    There’s no incentive to try and use that excess power for anything.

    “They should be tying it to the national grid[…]”

    There is no such thing as a national grid. Worse, even within each grid the transmission system behaves more like a singly connected graph where electricity goes in one direction from a bunch of predictable sources to a bunch of predictable sinks; it cannot be easily shuffled from anywhere to anywhere. If you tried it would put serious strain on the grid operators and could all to easily cause a black-out due to a careless mistake.

    “[…]Not only will the consumers enjoy lower prices, but the suppliers could make more money.

    I am missing the point here? It just seems so obvious!”

    Yes. You apparently believe that generating electricity is the hard part, but it’s not. Transmission of electricity is already a bigger fraction of the total cost than generation and that’s with numerous small grids and dependable gas, coal, hydro and nuclear sources. If you install “smart grid” technology, criss-cross the country with HVDC lines and pumped storage systems to make the most of wind with the least proportion of fossil fuels the cost for transmission will be astronomical. When the price of electricity is high people have an incentive to shift towards oil and gas wherever possible and heavy industry is eventually forced to flee or file for bankruptcy(see California).

  19. Storage and transmission are very expensive which commenters seem to be ignoring.

    BTW, where exactly does the storage come from? People keep on mentioning pumped storage. But there is a limited amount of cheap pumped storage available. Also large scaled pumped storage can only be feasibly built in certain areas. Plus pumped storage has an opportunity cost…..you could be using the pumped storage to transfer cheap baseload nuclear power to meet peak power requirements.

    Transmission is the same….it has an opportunity cost.

    The way I think of it is this: you have to have supply and demand on any electricity grid exactly balanced. Mismatch cannot be tolerated. This implies that predictability of supply and demand is highly desirable. On the other hand unpredictability is highly undesirable. Wind is bad precisely because it is unpredictable and cannot be controlled. Nuclear is predictable but not totally controllable (it cannot be turned off quickly). Gas is predictable and controllable. The other side of this is demand which is somewhat unpredictable and totally uncontrollable.

    A lot could be accomplished by focusing on making the demand side more predictable and controllable via demand shedding.

  20. Storage and transmission are very expensive which commenters seem to be ignoring.

    BTW, where exactly does the storage come from? People keep on mentioning pumped storage. But there is a limited amount of cheap pumped storage available. Also large scaled pumped storage can only be feasibly built in certain areas. Plus pumped storage has an opportunity cost…..you could be using the pumped storage to transfer cheap baseload nuclear power to meet peak power requirements.

    Transmission is the same….it has an opportunity cost.

    The way I think of it is this: you have to have supply and demand on any electricity grid exactly balanced. Mismatch cannot be tolerated. This implies that predictability of supply and demand is highly desirable. On the other hand unpredictability is highly undesirable. Wind is bad precisely because it is unpredictable and cannot be controlled. Nuclear is predictable but not totally controllable (it cannot be turned off quickly). Gas is predictable and controllable. The other side of this is demand which is somewhat unpredictable and totally uncontrollable.

    A lot could be accomplished by focusing on making the demand side more predictable and controllable via demand shedding.

  21. At the risk of appearing to be posting purely for self promotion, I will mention that my current research is investigating this very problem in Texas. Using the exogenous variation in wind, I identify the generating substitutes for wind power on the Texas grid. As mentioned in the article above, most of the substitution does indeed come from natural gas plants, but some what of a surprise is that almost 20% of substitution comes from coal plants. This is due to wind’s propensity to blow during off peak demand periods and cut into baseload generation.

    I have another paper, that is currently in progress, that looks at the effect of wind power on start up costs and profits of conventional generators. No firm results there yet.

    http://www.u.arizona.edu/~jcullen/

  22. At the risk of appearing to be posting purely for self promotion, I will mention that my current research is investigating this very problem in Texas. Using the exogenous variation in wind, I identify the generating substitutes for wind power on the Texas grid. As mentioned in the article above, most of the substitution does indeed come from natural gas plants, but some what of a surprise is that almost 20% of substitution comes from coal plants. This is due to wind’s propensity to blow during off peak demand periods and cut into baseload generation.

    I have another paper, that is currently in progress, that looks at the effect of wind power on start up costs and profits of conventional generators. No firm results there yet.

    http://www.u.arizona.edu/~jcullen/

Comments are closed.