Peter Kelly-Detwiler: 617.875.6575 | Leighton Wolffe: 781.547.1193 pkd@northbridgeep.com, leighton@northbridgeep.com

In March, I moderated a panel at the Energy Thought Summit in Austin entitled Creating the Next Energy Market Structure. There I had a chance to hear the thoughts of Nick Brown, President and CEO of the Southwest Power Pool and Gordon van Welie, President and CEO of ISO New England (ISO-NE). I came away with the distinct impression that running today’s highly complex electricity markets is no easy task.

It became very clear that – while it has always been a significant undertaking – the challenge of keeping the lights on in tomorrow’s wholesale power markets is going to become even more difficult than it has been in the past.

After the ETS conference, I arranged to continue the conversation with Gordon van Welie so that he could further elucidate some of the challenges, and highlight the dynamics of what is going on. van Welie is well-equipped to address these issues. He has headed ISO-NE for well over a decade, and in that time he has witnessed many changes.

The good old days are gone: In the good old days before about 2010, you had the normal hurdles that typically accompanied the coordination of the bulk power system (loosely defined as generation and transmission – the big iron stuff) and wholesale power markets. But big data was a relatively unknown term, distributed generation was fairly new, and renewables were just gearing up. In general, supply was supply and demand was demand. While there were challenges inherent in ensuring that supply was always equal to demand, and that there would be enough generation to meet future needs, the world was fairly (but not entirely) predictable.

Load was growing at a constant and fairly foreseeable rate. Power plants were either base-load oil, coal or nuclear, with perhaps some hydroelectricity thrown into the mix. And natural-gas-fired generation – quick and relatively cheap to build, was increasingly the marginal unit, setting the clearing price in the energy market. These plants were also predictable: they either ran baseload 24×7 or could be dispatched or shut off as necessary to match demand. And politicians generally ignored power markets, as long as electricity rates were affordable.

In general, the kids stayed put in the sandbox, and everybody colored within the lines.

New England’s current challenges: New England is a region that is highly dependent on natural gas and that also literally finds itself at the end of the gas pipeline. That alone puts it in a somewhat precarious position. This year, the mix of the generating capacity is weighted towards natural gas at 44% (a figure projected to reach 57% by 2024), 22% oil (NE is one of very few regions that still rely on oil), 13% nuclear, 11% hydro, 6% coal, and 4% renewables. While oil and coal are still a significant portion of the region’s capacity, those generators are usually more expensive than natural-gas-fired generation and run relatively infrequently: Oil- and coal-fired generation, combined, produced just 6% of the electricity generated within New England in 2015, while natural-gas-fired generation produced 49%.

Image: ISO-NE - keeping the lights on keeps getting harder

Image: ISO-NE – keeping the lights on keeps getting harder

This is about to become more challenging, and there are several reasons why:

  • Natural gas is generally available to generators in the summer, but in the coldest days of winter, much of that gas is allocated to heating – which has first priority because heating customers paid for the gas infrastructure. In some cases, there is simply not enough fuel for both heating and power generation. A new $3 billion pipeline that would have alleviated these stresses was recently planned for the region, but those plans were shelved this year in the face of withering opposition.
  • New England faces pressure to de-carbonize the grid, driven by both federal and state policies. Today, intermittent renewables such as wind and solar make up nine percent of the region’s nameplate generating capacity, but that number could rise significantly in the future.
  • A significant number of existing resources have recently retired or will be retired shortly. These include the coal/oil-fired 1,535 megawatt (MW) Brayton Point facility, and the 677 MW Pilgrim nuclear plant, with the 604 MW Vermont Yankee Nuclear plant recently mothballed and the retired coal/oil 749 MW Salem Harbor station, which will be replaced by a 674 MW natural-gas-fired power plant. .

(Photo by Michael Springer/Getty Images) – Vermont Yankee nuclear plant: one down, more to go

A look at the new proposed resources – 14,000 MW of projects submitted as of April 2016 — suggests that the future challenge does not get easier. 64% of proposed resources are gas-fired (though some are dual fueled, meaning they could burn oil in the winter), while 30% are from wind and 4% from solar (totaling 4,700 MW of intermittent renewables). Further, the Massachusetts Supreme Judicial Court ruled last month that state regulators must set specific limits on greenhouse gases to achieve legally mandated reductions in the state’s Global Warming Solutions Act.

Early market distortions created a problem: van Welie and his team are aware they have their work cut out for them in the years to come, and they have taken a number of proactive steps to address these issues. Critical among them is market design, i.e., how the markets are structured to ensure a competitive outcome. The key element underpinning this market architecture is the capacity market and its ‘pay-for-performance’ rules. In other words, resources will be well-paid for being there when needed; resources that don’t meet their commitment to be available will pay the resources that cover for them.

There is, he notes, a distinct historical reason for having a capacity market in addition to the daily energy markets. Back in May of 2000, northeastern energy markets did not have any wholesale offer price caps as they do today. On May 8th of that year, while numerous power plants were out for maintenance, the region suffered an unusually early heat wave. Demand exceeded supply by nearly 900 MW, and spot market prices soared to $6,000, scorching any participants who were exposed in the real-time market.

The Federal Energy Regulatory Commission (FERC) stepped in and instituted wholesale offer price caps of $1,000 per megawatt-hour (MWh). Caps may be nice to keep consumers from getting economically destroyed but they also essentially distort the market by preventing prices from rising as high as they otherwise would.

van Welie observes that, “The offer caps created the ‘missing money’ problem.” In other words, one must permit energy prices to rise high enough during shortage events to reflect the true value of the energy provided, and for investors to recover the costs of their capital investments with some additional decent rate of return to compensate for risk. There is a challenge in meeting the ‘one-day-in-ten years’ reliability standard (defined as the loss of load for no more than one day over a ten-year period – a probabilistic calculation that helps establish how much generation capacity one must have on hand): you have to provide sufficient incentives to the more expensive resources that are rarely called. If you cap energy prices, nobody will show up for that party. The estimate is that to lure these resources to stay in business, energy prices need to be able to rocket up into the $10,000 to $20,000 per MWh range.

 Two ways to pay – the hired ambulance analogy: There are generally two ways to create incentives and ensure the presence of sufficient generating resources to meet demand: One is to allow energy (kilowatt-hour) costs to rise high enough so that generators can make enough money during certain periods of high demand to be effectively compensated. The other is to cap energy prices at a lower rate, and pay money for the ability to respond (capacity) when necessary.

Perhaps a helpful analogy would be to think of using a similar approach to pay for an ambulance that would be available to you year-round, but which you might never need to use. First, let’s make the (unrealistic) assumption that the driver needs to make at least $12,000 a year, including the cost of owning, maintaining, and operating the vehicle. Second, let’s assume the vehicle only serves you and it is your only way to get to the hospital (no taxis or rides with friends – this is, after all, an emergency).

Under the first model, the driver and ambulance would sit outside your house all year, on call, waiting for you to have a medical emergency. Let’s assume nothing happens until one day in November your appendix bursts and you absolutely need that car. So you climb in and the driver says “that will be $12,000, please.” Of course, you respond incredulously “What, for a single trip to the hospital?” To which the driver responds “Hey, I’ve been waiting here for 11 months with no business. I’ve got a family to feed.” And if you have a relapse a week later, and have to go back to the hospital again, you pay another $12,000.

The second model would be to pay the ambulance and driver a $1,000 monthly retainer for being on call 24×7, and then pay standard taxi-type charges whenever you need to go to the emergency room, to pay for the cost of fuel. If he’s not there when you need him, he pays you back a portion of the retainer you paid to cover the last-minute costs to call another ambulance (otherwise, what are you paying the ambulance for?). This contractual clawback is structured so that if this happens too often, the ambulance driver will lose money, thus creating a powerful incentive to be there when you need him.

The Texas approach to market design – rely solely on energy costs: The Texas market (the Electric Reliability Council of Texas, otherwise known as ERCOT), has essentially adopted the first approach. However, there are several inherent challenges to this structure in the long term. The first problem is that prices may never reach the levels required to lure new generation investment.

van Welie comments that, “You need to have very high energy prices in order to attract the investment.” Part of the challenge, he observes, is that renewables – especially when supported by price distortions such as the $23 per megawatt-hour federal production tax credit for wind and the solar tax credit – have extra financial support outside of energy markets. And with increasingly more renewables added to the system – that don’t have fuel costs – the prices of raw kilowatt-hours tend to get dragged down.

The second challenge is that this is a zero sum “I win-you lose” game with incredibly high stakes. IF those prices are reached for any sustained period, it essentially means that somebody on the demand side may well have suffered a mortal financial blow. $9,000 per MWh (or $9.00 per kilowatt-hour) is over 330 times the 2015 average ERCOT spot market price of $26.77. One could easily imagine companies exposed to such spot market prices going out of business, with jobs lost, and intense political pressure to undo the economic impacts affecting the loser. It is, after all, no accident that FERC required the northeastern grids (New England, New York, and PJM – the mid-Atlantic grid) to impose $1,000/MWh caps in the early 2000s to respond to precisely this type of pricing incident.

van Welie raises concerns that the Texas model may result in problems with respect to future incentives for new generation. “The thing about Texas is that, on average, demand is continuing to grow. But as behind-the-meter investments increase as solar gets more cost-effective, demand will flatten out. If you couple this flattened demand with wind generation, which can drive prices negative at times, then I think it’s going to be difficult to attract new investment.”

The third obvious challenge is that investors cannot know in advance whether they will ever see prices that will allow them to recoup their investments. From that perspective, it is pretty much a game of craps. “With this level of uncertainty,” said van Welie, “the price of new entry should increase considerably because of the risk.”

New England’s approach – a capacity market with performance incentives: By contrast, ISO-NE has taken a decidedly different tack. They essentially have adopted the second model in the above ambulance analogy. ISO-NE breaks compensation up into two discrete elements: capacity (the ability to generate the energy precisely when and where it is needed) and energy (the raw electrons, or kilowatt-hours). And in both the energy and capacity markets, resources that don’t meet their supply obligations must pay those resources that are needed to fill in. The ‘pay-for-performance’ rules in the capacity market, as van Welie describes them, are quite simple: “Resources get paid for their performance in scarcity conditions.”

This capacity concept has become more refined in recent years as challenges to the grid have increased. Originally, ISO-NE paid for physical generating capacity (steel in the ground), as well as demand response (the ability of customers to curtail consumption during critical periods). Over time, the grid operator recognized that this generating capability needed to be year-round and that this was about

“producing energy when we are short of reserves…The capacity market is really an auction to buy an option from suppliers to produce energy or reserves three years out into the future. If you deliver according to the capacity obligation, you get paid what was promised. If you underperform, you get paid less.”

One key difference between the ERCOT and the New England models is that New England’s market design pays for performance on a predictable forward basis that provides investor certainty. Combined, the estimated payment rate for capacity and energy will be approximately $9,000/MWh during scarcity conditions, of which the pay-for-performance element will be $5,455/MWh when the incentives are fully phased in (2024). The assumption is that about 21 hours of scarcity conditions can be expected. van Welie observes,

“This market model gives the investor the ability to look out and ask ‘what am I investing in and will it be capable of meeting those obligations or not? What’s my cost of entering the market, net of energy revenue, that I can expect to earn? Do I have to worry about my performance penalty?”

How does that actually work in practice? van Welie notes that investors need to take a hard look at the performance capabilities of their generating and demand-response assets.

“If a generating unit takes 12 hours to start, that unit may not be around when there is a shortage event. Although most events occur around the peak, that’s not always the case. If you have a 12-hour start, you may be able to capture two-thirds of shortage events, but miss the remaining one-third. You will lose some capacity money, so the system starts rewarding high-performance resources.”

van Welie characterizes pay-for-performance as a financial construct that jus modeled to create predictability and certainty for the investor.

“The capacity market design also includes a seven-year lock on the auction clearing price for new entrants. They have the option of a one-year price lock or up to seven years. That gives the new investor more certainty. The combination of these features allowed us to get over 3,000 MW in new, combined-cycle investments with dual-fuel capability because they know pipelines get constrained from time to time. We wanted people to firm up fuel supply.”

ISO New England implemented this new pay-for-performance approach for several reasons. One was the simple fact that most natural-gas-fired generators had bid into the system as firm power resources, without having a firm gas supply contract. The other issue van Welie cites was the simple under-investment in maintenance and staffing of the legacy generation fleet. The performance challenges became more and more apparent during extended cold snaps, so the ISO launched a strategic planning initiative in 2010 and worked closely with stakeholders to develop the performance incentive rules, which were filed with FERC in late 2013. Van Welie stated that this approach became increasingly necessary.

“Performance was deteriorating quite rapidly, particularly among those generating assets getting toward the end of their useful life. They were collecting capacity payments but leaving us with the performance risk.”

He firmly believes this pay-for-performance approach will become ever more critical as society further de-carbonizes the power grid and more intermittent renewables are added.

“As more renewable energy is put onto the system, infra-marginal revenues in the energy market will go down. In a world where we have to remove 80% of carbon by 2050, there will be little room for fossil fuel in the electric sector. Revenues in energy markets will be low most of the time, but we will still need to call on resources on short notice to maintain grid reliability. So how does that fast-start resource get paid when there is no money in the energy market?… Renewables will displace fossil on average because of price, but you will have to complement variable resources with resources that can supply energy on demand.”

van Welie indicates that no matter how you slice it, resources and capabilities must be paid for. If you don’t pay for them in energy costs, the money has to come from somewhere else (in traditional regulated markets, it will be folded into utility rates). New entrants have to assume that spot market energy prices will fall.

“With low-priced gas and zero marginal cost for renewable energy, you have to take a view of a steadily downward trend for energy market revenues over time.”

At the same time, he states, the ability to perform when required will become more valuable. “There’s going to be a shift from energy markets to resource adequacy mechanisms and reserve markets.”

Where is this headed in the future? Where does the CEO of ISO-NE see this going? Among other things, he believes grid operators will need to be able to better see asset performance in real-time, and not just power plants but more of the distributed and decentralized system as well, such as rooftop solar.

“In the future, as a larger part of the resource base lives outside FERC and NERC jurisdiction, this is going to become even more important. How will these resources participate in the market, and how do we create operational alignment with this set of resources so that it does what we need it to do—meet consumer demand and maintain reliability? This can be accomplished by allowing people to take on a supply obligation, but one of the conditions we need is to be able to see their resources’ performance in real time.”

And, he states, you need to continue to be resource agnostic. It doesn’t matter whether it’s demand response, or batteries or power plants, as long as they perform.

“A construct such as the one we have put in place is going to be key. You need a clearinghouse that establishes an obligation, and resources will perform according to the financial incentive. And the process allows you to reveal the forward cost of reliability. That becomes the reference point for economic decision-making.”

van Welie also believes that the pricing environment will likely get more granular as well.

“Here in New England, supply is already priced on the nodal level. We aggregate that into the energy price, but ideally speaking, it would be nice to evolve towards a place where we also have nodal pricing for demand as well. This locational pricing would work with a dispersed and disaggregated resource base, but there are implicit complications. For example, metering would have to be in place. But as policymakers open the door to more choice for consumers, such as PV and batteries, then the question will become: how do you coordinate the actions of all of these disaggregated resources behind the meter? If we are going to count on them as capacity as opposed to energy displacement over time, they have to meet their obligation.”

van Welie feels that this may be the new role for the demand-response type companies out there that coordinate end-use customer behavior with grid conditions.

“The aggregator model that was pioneered through DR will allow distributed energy resources to participate in wholesale markets.”

The tension between state actions, wholesale markets, and customer choice: van Welie observes that there is another big challenge out there for grid operators, and that comes in the form of individual state policies that may distort markets.

In Massachusetts, lawmakers have drafted legislation calling for 1,200 megawatts of offshore wind over the next ten years, a policy that could be very helpful to getting the offshore wind industry started in the right direction. They are also looking for 1,200 MW of Canadian hydroelectricity.

However, ISO-NE’s leader notes there are clear trade-offs. Policymakers are focused primarily on environmental goals or requirements. But with more intermittent renewables, higher capacity prices to ensure reliability will result. And there is also an inherent contradiction with competitive markets.

“In a scenario where consumers are free to buy energy from wherever they want, and the states are entering into 25 year contracts to meet their renewable energy goals, there will be a trade-off that isn’t well understood because you may find yourself with stranded (uneconomic) contracts over the long run. States have a very legitimate goal of wanting to remove carbon from the economy. But if they want to manage this through a long-term contracting mechanism, this can create problems in capacity market price formation.”

To that end, van Welie and the ISO work hard to explain the potential market implications to politicians and regulators as they work to achieve their policy goals.

“The ‘how’ is very important. It would be more consistent with competition if the desired characteristics, such as clean energy, are valued in a manner that is compatible with wholesale markets. You don’t have to contract for specific resources; resources with the desired characteristics will compete with other resources and will likely clear the market if the characteristics are valued in a systematic and resource-neutral fashion and investors can predict the impact on their profitability and relative competitiveness in the market. Investment risk stays on the investor side of table.”

Unfortunately, it’s not likely to get any easier: In a world where big data becomes increasingly pervasive, the costs of renewables continue to fall, and the pressure to decarbonize the grid increases, the challenges will only grow. Many more players want to have their say concerning what goes on in our power grids. Within the past decade, many things have already changed dramatically. Many actors are now coloring outside the lines, and it seems that just about everybody now wants to play in the sandbox. The ISO’s job is to develop those rules to ensure a steady and reliable flow of electrons. This is not an easy task in the best of times, and it’s likely to get a lot harder…