Waiting for the Next Polar Vortex


How recent events could prove a harbinger of winters to come.

Fortnightly Magazine - June 2014

With spring turning to summer, most of us just want to forget about the Polar Vortex, an extraordinary event that was profitable for some but painful for many. Electricity and natural gas demand hit local highs, operators and regulators were forced to take emergency actions, and power plants once thought to be out of the money got a chance to mint gold.

But simply turning the page would be an enormous mistake, papering over a disturbing truth: recent events may be a harbinger of future emergency winter events. In fact, the disturbances of the Polar Vortex might become the new normal in coming winters.

What the Polar Vortex brought to light is that we have had a distorted view of system capacity due to market rules and regulatory assumptions from the U.S. Federal Energy Regulatory Commission (FERC) that have failed to properly value (or consider) reliability. In spite of several FERC decisions since the Polar Vortex to correct these problems, five on-going trends belie assumptions that the grid has sufficient capacity to meet winter peak demands without emergency actions. These trends belie the ability of grid operators to respond to severe winter weather events and thereby raise overall market risks and price volatility:

Coal Plants. Continuing retirements, concentrated in specific markets, possibly accelerating

Natural Gas Delivery. Most gas-fired power plants in deregulated markets lack long-term firm gas supplies1

Figure 1 - U.S. Total Natural Gas Consumption

Back-up Fuel Requirements. Not defined, even for gas-fueled power plants receiving capacity payments2

Demand Response. Overly optimistic expectations for winter capacity contributions from interruptible load programs,3 and

Renewables. Overly optimistic expectations for winter capacity contributions from renewable generation.4

The heightened risk from these trends and FERC policies raises serious questions for utilities, midstream natural gas and natural gas shippers, regulators, and investors alike. Using the Polar Vortex as a lens to examine these trends will help all of these stakeholders to give more urgent consideration to FERC actions to alleviate the risks, including reconsidering the definition of "firm" to reflect the structure of the natural gas market, crediting back-up fuel's contribution to reliability, and implementing new policies to encourage natural gas pipelines to get much needed natural gas supplies to market.5

Figure 2 - Real-Time Energy Prices

Overall, the Polar Vortex created a lot of firsts in the natural gas and power industry. Even a cursory review of key points demonstrates just how dramatic an impact we saw.

A Revealing Drama

Natural gas demand was the highest in January in the history of the market (Figure 1). Seven out of the 10 biggest demand days on record for the U.S. occurred in January 2014.6 According to EIA, total U.S. natural gas consumption in January was over 3,200 billion cubic feet (Bcf), 300 Bcf higher than in any other month on record.

Some locations saw daily natural gas spot price indices spike to the highest absolute level ever seen in the U.S. This fact is not only meaningful in itself, but also revealing in the geographic distribution of the price spikes. The spikes were concentrated in specific downstream market areas (e.g., Transco Z5 and Z6), while supply areas prices (e.g., Dominion South Point, representative of Marcellus area prices) were far more stable. The implication is that that there was plenty of natural gas available, just not enough pipeline capacity secured under contract to meet the heightened demand in the downstream market areas.7

Figure 3 - Unhedged Coal Power Plant Margins Surge

Wholesale power prices also reached very high levels. Power prices hit the price cap in ERCOT at $5,000/MWh which is the highest ever established power price cap (Figure 2). In the Northeast, wholesale power prices that were so high that PJM's counsel issued a dramatic statement: that even though generators purchasing natural gas could not recover their costs for fuel under existing price caps on real time electrical energy prices, and even though FERC was shut down at the time due to extreme weather, it was the "opinion of counsel" that they could in fact eventually recover out-of-pocket fuel costs8

In PJM East, January to February 2014 prices were 3.5 times higher than in 2012 and 2013 over the same period.9 In PJM East, coal plant gross margins over the first 60 days of 2014 were comparable to prices paid to purchase coal plants in 2012 to 2013. In PJM West, prices were 3 times higher compared to 2012 and 2013 (Figure 3).

Furthermore, PJM, the largest FERC-regulated Regional Transmission Organization (RTO),10 experienced record high power plant forced outages and was operating under its emergency protocols. It is fair to say that if ERCOT was experiencing the same conditions as PJM, it would set it price cap to $5000/MWh administratively to signal the grid is operating under scarcity conditions.11 In PJM's view, they were operating under their emergency protocols and were not about to shed load.

More specifically, PJM was briefly 500 MW away from running out of 5-minute-responding operating reserves, though there were 1600-1700 MW of ten-minute reserves at the ready. Also available was the potential to decrease load via voltage reductions.12 Any major erosion in resources (e.g., a 6,000 MW net loss) could make this even more serious. After voltage reductions, which PJM estimates would have provided a 2,000 MW reduction in demand, the next step was for PJM to make emergency purchases (which PJM estimates would have provided another roughly 4,000 MW). Thereafter, PJM would have ordered rolling blackouts, which can have highly adverse consequences to public health and safety during record cold, or further decrease its operating reserves.

Figure 4 - January 2014 Generation Outages in PJM

PJM reported a forced outage rate of 22%13 - far greater than previous years or the average forced outage rate (Figure 4). Forced outage rates were above the historical average throughout January.

While other problems caused the fleet to fail to achieve reliable operation consistent with the expectations of industry planners and regulators, in some key hours up to one-quarter of the outages were due to lack of natural gas, and gas plant outages were half of the total (Figure 5). All natural gas delivery interruptions over this period were of customers relying on interruptible natural gas delivery service. In order to achieve reliable operations, price signals need to be unambiguous that reliable operation is profitable. There must be complete coverage of the year, not just for the summer, and pricing systems must avoid overly relying on planning projections, i.e, pricing must automatically be able to decisively respond to unexpected conditions and threats.

The coal power plant outages were also a serious cause of the problem (35% of outages). However, about half of the coal plant problem seems to be related to not maintaining units scheduled to imminently retire. The concern related to gas plants is that as the grid becomes more reliant on gas, the shortcomings of the gas delivery system will increase unless ameliorated.

Thus, in terms of demand, price, and seeing a grid pushed to into emergency conditions, the Polar Vortex was not only almost unprecedented, but also offers a window into the effects of a set of trends that have decreased future reliability.

Figure 5 - Unplanned Generation Unavailable in PJM

Moreover, the events of the Polar Vortex did not occur in a vacuum. Rather, they followed from a series of changing policies, rapid market changes, and regulatory responses - some dramatic and readily apparent, others more subtle. These factors not only drove the market on their own, but interacted with each other to magnify their individual effects. Together, they have led to a system that is becoming less resilient and reliable than is commonly understood.

The following sections discuss seven of these trends have particularly contributed to the over-estimation and erosion of reliability and are likely to make winter emergencies and volatility the new normal: (1) coal and other power plant retirements concentrated in specific markets, (2) reliance on interruptible natural gas supply, (3) insufficient rules about back-up fuel, and qualifications for receiving capacity payments, (4) the effect of interruptible load programs, (5) muted power price signals understating reliability problems, (6) unrealistic views on renewable supply during the winter, and (7) over-reliance on imports.

Yet there are some positive events that partly ameliorate the above trends, including: (1) high reserve margins in some regions such as PJM, which procures in excess of planning reserve margins, especially in winter, (2) the addition of new gas power plants with high plant reliability (i.e lower forced outage rates if fuel is available), albeit with the fuel firmness concerns discussed, and (3) several new FERC decisions rapidly taken after the Polar Vortex which try to correct FERC policies that have been brewing for years.

Plant Retirements Looming

Starting in 2009, three forces combined to cause the largest retirement of generation capacity in history. U.S. EPA air toxic regulations, referred to as MATS, have a hard April 2015 or April 2016 deadline and are the main driver, although other EPA regulations focused on coal power plants have contributed in the same vein.

Figure 6 - Retiring Unit (Scheduled to Retire Before Winter 2014/2015) Performance on January 7, 7 p.m.

Since 2009, actual and announced retirements in the U.S. are 87 GW (87,000 MW), or about 10% of the nation's capacity.14 These retirements are regionally concentrated in the Northeast and Midwest, creating pockets of more severe capacity constraint. For example, in PJM, 27.5 GW of power has been announced to retire with more than half set to retire in the next three years.

During the Polar Vortex, retiring units provided 7,000 MW, and this will be lost as plants retire (Figure 6). And just as retirements accelerate, EPA is, in June 2014, expected to announce more regulations (i.e, Section 111(d) or New Source Performance Standards), which may usher in another set of retirements. The poor performance of these units (only 43 % on line) is due to low maintenance of retiring units, and explains a significant portion of the outage problems (about half).

Furthermore, in June 2014, EPA is scheduled to announce CO2 emission regulations for existing power plants. Depending on the stringency of these regulations, they could accelerate retirements further.

Some of the informal mechanisms on which the grid unknowingly relies could change very quickly and these shifts have the potential to lead to an acceleration of retirements. Much of the deregulated capacity is still owned by entities that have utility corporate affiliates with obligations to serve. Many of these entities have either initiated or recently completed processes to sell power plants to entities without any corporate affiliates that carry this obligation. There is nothing wrong with this per se; and in some sense this is exactly what deregulation promised. However, regulators should not be surprised if these new entities act to quickly and unexpectedly retire uneconomic plants. If there are rolling blackouts these new entities will not be subject to informal penalties by states or others who regulate their distribution and other businesses.

Figure 7 - Cold Weather Voluntary Demand Response

The problem with quick retirements is greater than commonly recognized. Accommodating already announced retirements is challenging in spite of the lead time for retirements15 in part because there are unexpected events in addition to the retirements (e.g. poor maintenance of units, fuel supply issues for replacement capacity, aggressive preference for demand resources, etc.). In addition, concerns exist because of the long lead time for new transmission and generation. Furthermore, a key mechanism to delay retirement, reliability must-run contracts, are voluntary and RTOs cannot compel power plant owners to accept them.16

These retirements on their own present a challenge to resource adequacy, but this effect was reinforced over the past few years by the fact that the analysis of the regulatory impacts does not seem to have adequately anticipated the potential for grid reliability problems. In the initial 2009-2011 period, this potential was not ignored, but the actions were insufficient, based on flawed assumptions, or inadvertently made things worse than they had to be.

The first response at the time was to place reliance on the FERC-regulated resource adequacy policies and structures in place to ensure that regional planning reserve margins would be met by new power plants if needed.17 It was also widely acknowledged that incremental capacity, to the extent fossil generation was needed, would be plants fired by natural gas - and there was not yet much focus on winter reliability problems with natural gas power plants.

In addition, EPA also agreed to give the states the authority to extend the key MATS compliance deadline by one year, from April 2015 to April 201618 (every request for delay has been granted to our knowledge, though states cannot offer additional extensions beyond April 2016).

Figure 8 - PJM RTO Cleared Capacity Prices vs. PJM Net CONE

In short, a dramatic spate of retirements driven by market and regulatory forces have removed generation capacity in an almost unprecedented way and regulators were unable to fully anticipate the effects or the constraints of the replacement capacity that was being brought on-line, leading to a significant questions about grid reliability especially under contingencies.

Gas/Electric Market Integration

Over the last two years, it has become increasingly clear that there is a natural gas power integration problem that was not adequately anticipated. And it is getting worse as the number of announced coal plant retirements increases, and more gas plants are added.

The specific problem is that the retired capacity is being replaced with new natural gas power plants that do not have firm natural gas pipeline capacity contracts, and lack back-up fuel requirements. A contract for firm pipeline capacity from a liquid supply hub to a plant insures that, excluding a force majeure event such as a pipeline rupture, natural gas supplies will be available to the plant regardless of other demands on the system. Instead, most generators rely on interruptible pipeline capacity that is only available when other firm capacity customers (e.g., the natural gas utilities) are not using all of their capacity. Put another way, coal plants with firm fuel supply are systematically being replaced by natural gas plants with interruptible fuel supplies - i.e, are purchasing delivery service under interruptible contracts.

Anecdotal information suggests that only about 20% of natural gas generating capacity in New England has some amount of firm natural gas supply. This analysis is more than cursory, but not as robust as needed. At this time, there is no government agency or authority that knows what power plants have firm delivery service for future seasons, nor are there commonly accepted notions of what constitutes firm natural gas delivery service. For example, one issue is the extent of firm service. Does it need to extend to the wellhead or to a liquid hub? Furthermore, there is no regional analog in the natural gas delivery industry to NERC and the RTOs which set and regularly measure expected and planning reserve margins, and assess multi-company security, etc.19 Thus, there is no regularly available measure of how stressed natural gas systems are regionally in terms of available interruptible supply.

Natural gas power plants with interruptible natural gas delivery are displacing both coal and nuclear plants with much more secure fuel supply. Coal plants typically maintain 30-60 days of fuel on site. The coal plants purchase 90% of their coal under contracts; little coal is purchased in the spot market. Those large stacks of coal are warm comfort during cold spells. Nuclear units, which are also retiring, typically contract for fuel services for up to 4-5 years.

When non-power natural gas demand is very high - e.g., during colder than average winter periods when space heating demand for natural gas is especially high - natural gas power plants are not necessarily available because there is no interruptible delivery service available. If the infrastructure situation does not change, this interruption will grow in frequency and duration. Adding to the urgency is that over 1.5 Bcf/d of incremental natural gas demand (about 8 GW of new combined cycles) is scheduled to be added in eastern PJM over the next two years alone, and all marginal capacity in PJM is natural gas-fired.

It is important to be clear - even during the Vortex, natural gas power plants that had firm natural gas delivery service had natural gas. In that sense, the natural gas system did what it was designed to do. The problem is that during cold winters, interruptible service can be interrupted for all the interruptible plants in the region together and at once.

Further, power plants lack adequate mechanisms to recover the cost of firm fuel supply. The best example is the calculation by the RTOs of CONE, the Cost of New Entry. This cost is the net cost of building new natural gas power plants calculated line-by-line in great detail. However, the RTOs calculate it without the cost of firm fuel delivery. The CONE estimate is the outcome of a FERC-regulated process, and is used to limit capacity prices. This sends a signal that firm natural gas delivery is not needed.

Grid planners at NERC, RTOs and FERC refer to "all at once" outages as correlated outages. Unfortunately, they have been assuming natural gas power plant outages are independent of each other and the weather. Thus, they have understated the target reserve margins required in the winter. This fault has led to over reliance on summer-only resources.

Also, back-up fuel requirements are not adequate for firm resources qualifying as capacity resources. PJM includes the capital cost of oil storage in its CONE estimate, but does not require back-up fuel be held. The amount of required back-up fuel at the start of a season to qualify as a capacity resource is not establishes. The amount of back-up fuel available for even the next season is highly uncertain in many cases. Price signals that back-up fuel is required need reinforcement.

FERC at Fault?

Concerns about reliability were indirectly - and perhaps also mistakenly - lessened by FERC policy. When these forces were developing, FERC, the entity with lead responsibility for grid reliability, indicated that it was possible that the United States would never need to add additional power plants.20 Former FERC Chairman Jon Wellinghoff suggested instead that demand-side changes such as interruptible load could obviate the need for new power plants. Moreover, he insisted that baseload power plants will become an "anachronism" in the very near future, as new technologies allow control area operators to call on a large variety of renewable resources capable of providing power throughout the entire Eastern or Western interconnections. In that atmosphere, it was more difficult to see as clearly the potential problems that plant retirements might cause.

FERC's and PJM's implementation of interruptible load (demand response) is another factor that is both its own problem and one that magnifies the other trends discussed above. This treatment caused PJM's and ISO-NE's reliance on interruptible load to skyrocket in recent years. This reliance on demand response as a resource suppressed capacity prices for generation capacity, contributing to the retirements of power plants. The problem is especially seasonal in nature: nearly every MW of PJM interruptible load was only interruptible during the summer, and then for only 60 hours annually from 12-6 pm. By design it is not available in the winter.

At first glance, one might have thought the reliability of the grid could have been worse. From this perspective, it was fortunate that FERC denied power plants the opportunity for limited summer seasonal, afternoon-only operation. However, in the face of this discrimination against firm generation, generators could be forgiven for concluding they weren't really needed any longer. Furthermore, this was only part of the widespread and systematic signaling that it was time to retire power plants. Only power plants were required to bid into the day ahead and real time energy markets; interruptible load was exempted. Only existing power plants were required to actually verify unambiguously their ability to operate (though they did not have to show they had firm fuel).

PJM's target planning reserve margin was 16% of peak summer demand (requiring approximately 25,000 MW of planning reserves). For this past winter, 2013/2014, PJM had contracted for 9,300 MW of interruptible load in its BRA auctions (UCAP MW).21 For the planning year 2015/2016, when retirements will be reaching their maximum level, PJM has contracted (BRA auctions) for nearly 15,000 MW of interruptible load, nearly all of which cannot be interrupted during the winter. This is a sixty percent increase in summer-only supply from winter 2013/2014 levels and coincides exactly with the peak of power plant retirements. 15,000 MW equals 60% of the targeted reserve levels, and hence, is extremely large.

During the height of last winter's Polar Vortex, only about 2,000 MW of load reductions occurred out of 9,300 MW of contracted interruptible load (Figure 7).

Concerns about reliability also were indirectly mitigated by the large emphasis on adding renewable power plants. Federal subsidies and other factors were resulting in the largest addition of renewable generation capacity in US history. It was well understood that this generation was intermittently available or effectively interruptible due to periodic lack of wind or solar radiance (e.g., snow covering panels). But the answer was assumed to be that natural gas plants would firm up renewable supply.22 As is discussed below, it was hard to be worried about the interruptibility of natural gas plants if they were widely expected to be the source firming renewables.

Suppressed Power Prices

It might seem paradoxical to indicate that power prices have been suppressed, while at the same time reporting record high power prices. In fact, it was the price signals that capacity would be needed to maintain reliability that were suppressed. This effect can be explained by: (1) the fact that capacity reflects in part decisions made years earlier, and (2) infrequent price spike events have to be very strong to achieve the reliability levels that U.S. customers want. Low capacity prices in FERC regulated markets, combined with natural gas prices sustainably below historic levels have encouraged retirements, and low maintenance.

Examples of price suppression include: (1) the large procurement of interruptible load that lowered capacity prices, (2) "unintended adverse market and reliability impacts of certain capacity market rule changes adopted by PJM in 2011, when PJM first introduced its existing menu of demand response product alternatives (a supplement to PJM's then-existing, limited-availability demand response product)" as described in FERC filings. There were a lack of caps on the amount of interruptible load procured23 favoring interruptible load over generation lowering capacity prices, and (3) policies suppressing electrical energy prices during system emergencies (these spikes can be critical when deciding how much to invest in plant reliability). In Figure 8, PJM capacity prices (in blue bars) are shown to be well below the CONE (black line).

The extent of this suppression is large. On April 18, 2014, the PJM market monitor estimated that capacity prices would have been much higher were it not for policy decisions favoring interruptible load and power imports from other regions over local generation. As can be seen, the extent of the price suppression has been large. Other factors have been responsible for the lower prices as well, but they have been separate from FERC policy (e.g. low gas prices, low demand growth).

Meanwhile, PJM capacity imports will increase rapidly between this winter and the 2016/2017 winter. One would expect that this increase would augment reliability. However, PJM filed recently to lower imports and on April 22, 2014, FERC approved new lower transmission import limits. Thus, previously insufficient limits had been in place.

Positive Trends

Grid emergencies notwithstanding, the grid was able to deliver under adverse circumstances. Why is that? There are four answers.

First, high reserve margins. PJM's procurement of capacity reserves has consistently been above the target, i.e, roughly 20% reserve margin versus a target of 15% and 16%. The reserve margin was even higher for the winter - 27%. This has been critical in limiting the problems described in the previous section. Nonetheless, while it is better to have offsets to reliability problems created by treatment of interruptible load, etc, a more direct treatment would be better.

 Second, lagging demand. The grid has had large amounts of excess capacity due in part to low demand growth. This trend in turn has been due, to some extent, to a poor economy relative to expectations. A significant concern is what happens if load growth is unexpectedly strong?

 Third, the worst is to come. Many of the exacerbating trends have yet to reach their full effect. The most challenging times are still ahead of us.

 Fourth, lack of fuel security. Natural gas-fired capacity additions are expected to have low forced outage rates. However, the fuel situation is an area requiring greater attention.

Also, even in the short time since the Polar Vortex, there have been three FERC decisions attempting to fix these long standing problems. Of particular note are the January 30, 2014, March 3, 2014, and April 22, 2014 decisions.

On January 30, 2014, in approving a PJM filing, FERC stated it was undoing "unintended adverse market and reliability impacts of certain capacity market rule changes adopted by PJM in 2011, when PJM first introduced its existing menu of demand response product alternatives (a supplement to PJM's then-existing, limited-availability demand response product)." This problem was in place for the last three auctions during which time thousands of MWs of plant decided to retire. Even so, the correction still leaves much room for interruptible service to crowd out annual firm generation.24 Additional steps may be necessary, especially if the retirements and demand growth accelerate, and especially in some new emerging load pockets.25

On March 3, 2014, FERC approved a PJM filing requiring that company officers attest to the intent to make DR resources physical. And on April 22, 2014, FERC approved a tightening of PJM import limits, as discussed earlier.

However, all three decisions primarily affect procurement for winter 2017/2018, and do not have immediate effects. Thus, the focal point of concern is the near-term, e.g., the next 2 to 3 years. The expectation is more emergencies and price volatility during cold weather with the most likely scenario a series of close calls.

Implications Going Forward

ICF published a white paper in 2011 indicating that the retirements associated with environmental regulations could be accommodated if they were done carefully and flexibly to account for unexpected problems.26 ICF focused on the potential for unexpected transmission security problems, the newness of deregulation in the power sector, and the limitations of resource adequacy policies and procedures if they are not closely coupled with transmission security analysis.27 ICF's concern was that these models would unexpectedly show local or sub-regional problems under certain contingencies, and that resource adequacy modeling and systems might miss the effect of contingencies. Clearly, and especially in light of the Polar Vortex, a revisiting of ICF's concerns is appropriate, as unexpectedly significant problems have been found that require attention and flexibility. Present events are an example of the set of unexpected transmission security problems that we were concerned might crop up when we wrote our 2011 article, namely summer-only interruptible load, plus interruptible fuel supply. That is especially the case in light of the potential that even more retirements might occur suddenly, and very quickly.

The period of maximum concern is the next 2 to 3 years, before most new FERC policies and market responses can take effect. In this period, in particular, special attention should be devoted to not making reliability planning and operations any more challenging than absolutely necessary. The more challenges the grid is undertaking at once, the greater the potential for inadvertent and unexpected problems. As it is, the grid is already grappling with the largest retirement of plants in world history, while also having massive expansion of new demand resources, using complex forward capacity markets instead of hourly price spikes, pursuing intermittent renewables, favoring only one fossil resource with NSPS changes, pursuing distributed generation, pursuing aggressive rate payer subsidized demand programs, and the overall FERC regulatory direction for many years is to preferentially treat a resource with inferior reliability characteristics.

Events this winter demonstrate that reliability is challenged by non-weather trends. By re-assessing reliability in light of the highly dynamic conditions in the industry, regulators (FERC, EPA and other agencies) will be able to get a more realistic perspective on the current assumptions in the market, and update policy prescriptions. A comprehensive review would incorporate not only operational problems, but also implications for planning, policy, operations, and market structure, and address all resources including transmission imports, demand and supply resources.

An important focal point that regulators should increasingly want to review in-depth is the discrimination on behalf of interruptible load and against generation. Another is the proper treatment of seasonal limited interruptible load. That is true even while acknowledging that FERC's recent decisions for PJM, the decisions of January 30, 2014 and March 3, 2014, are steps in the right direction.

Here are some specific questions that regulators ought to be asking:

 What should be done regarding back-up fuel supply? This problem is likely the most important near-term priority.

 What should be done regarding natural gas supply? This area is challenging, but also deserves attention.

 How should seasonal resources be handled? Should seasonal operation be available to all resources or none?

 Do limited operations (60 hrs/yr) need to be available to all resources, or none?

 Should there be a fresh review of Winter Resource Adequacy?

A careful re-examination of reserve requirements during the winter appears urgently needed in light of under-estimation of winter planning reserve requirements.28 This examination should include a review to ensure proper treatment of the correlation of winter generation outages: when one plant is lost due to lack of firm delivery service and or back up fuel, others will also likely be lost including coal plants in some cases. Similarly, if one plant is lost due to lack of weatherization, others will also likely be lost.

Here are some additional questions for regulators:

 Should all capacity resources be required to participate into the energy markets, both day-ahead and real time, (à la ISO-NE's newly adopted approach), or should none be required?

 Are penalties for non-performance sufficient to handle interruptible load "fatigue?" (That's the tendency of interruptible load resources to fall short of what was promised in the face of repeated interruptions - a tendency that has not been adequately anticipated or tested.) Perhaps load should be prevented from returning to firm power supply for an extended period.

 Should all generator offer caps be a function of natural gas prices times a conversion factor? Such a structure has been in place in other FERC-regulated markets in the past. Otherwise, the consequences of providing large economic incentives not to perform - i.e, not to be able to recover fuel costs - appears to create unnecessary risks to grid reliability.

 Should there be a further consideration of a hybrid model for capacity and resource adequacy?

Regarding this last question, a hybrid model for resource adequacy could serve as a useful construct. It could employ a combination of wholesale electrical energy price spikes (already in place in many markets but with widely varying prices and rationales) and forward capacity prices. Such a hybrid model might better address the need for automatic, real-time and geographically focused responses to resource adequacy problems. And in this context, should the price spikes equal Value of Loss Load (VOLL) during use of operating reserves? Has the complexity of resource adequacy has become too great for administrative mechanisms alone?

The regulatory system is too stacked against building new natural gas pipelines that are desperately needed in many parts of the country. ICF recently concluded that a total of $641 billion in midstream infrastructure would be needed to meet future market demand.29 Generators have been reluctant to contract for firm pipeline capacity because they have no way to recover that cost. The lack of firm capacity threatens electric system reliability, and can also pose risks to the natural gas system when generators withdraw more natural gas than their contractual limits. Fuel is a key component of electric system reliability, and generators need incentives to ensure fuel reliability. More natural gas pipeline capacity is not the only answer - seasonal storage, peak shaving, and oil backup also play a role.

What, then, is the outlook for utility industry investors?

Investors, it appears, can still look forward to the potential for high profits. Even a more mild winter offers large opportunities when combined with the trend of decreasing reliability due to retirements, more summer -only resources, and natural gas plants without firm fuel supply.

ICF recommends investors consider a bifurcated investment strategy: short to medium-term (3 - 5 years) and longer-term. Over the 3-5 year horizon, as the regulatory regime accommodates the shifts in the marketplace, and/or the market adjusts infrastructure, certain assets have the potential for excellent profits from severe weather events. Such assets include power plants with firm fuel supply, firm pipeline capacity, and back-up fuel. Careful tracking of trends will help tailor the investments to market conditions.

At this time, the forward markets do not fully provide support for these investments, and hence, the opportunities will likely continue to exist, but be speculative. Thus, a premium will be on creative balancing of risks and returns.

In the longer-term, assets that position themselves for long term contracts will be well positioned if they are in the right location. In this category, midstream oil and natural gas infrastructure remains a generation investment opportunity along with the backup capacity necessary to ensure adequate capacity.

Markets also need to devote special attention to the risks of power supply, especially retail power supply. Special hedging attention is needed to manage the potential for coincidence of high volumes and high prices.

Lastly, a contrarian strategy is to assume a hard landing where current trends continue for a longer period without triggering major increases in forward market prices and capacity markets due to such factors as continued regulatory intervention to suppress prices and support particular policy initiatives that inadvertently has in reliability. This could lead to a super spike in prices similar to the recent winter.

ICF has successfully worked with clients to anticipate the current opportunities and continues to monitor developments.30


1. Therefore, they are not capable of directly inducing infrastructure expansion.

2. There are no requirements for specified amounts of back up fuel to qualify as capacity in capacity auctions, and no studies as to how much should be available.

3. The formal expectations of summer only resources in planning are zero. However, there are several manifestations of this overly optimistic expectation for interruptible load including similar prices being paid over multiple years for winter (generation) and non-winter resources (most interruptible load).

4. As reflected in lack of penalties for poor winter performance, unreasonable expectations, etc.

5. This includes such questions as to duration of firm, the amount of back up fuel, the origin destination combination needed to be considered firm, etc.

6. http://www.marketwatch.com/story/polar-vortex-brought-record-natural-gas-demand-2014-02-03

7. There may have also been gas delivery system elements which did not perform such as compressor stations.

8. http://www.pjm.com/~/media/documents/reports/20140121-cost-based-offers-into-the-day-ahead-energy-market-on-jan-21.ashx. The reader is encouraged to read the full quotation. The key point is that caps were not indexed to fuel costs, and FERC controls price caps. There is still no full commitment to indexing the myriad number of FERC electric price caps. This is especially a problem because there are no price caps in the gas market.

9. ICF estimates. Gross margin does not account for fixed costs. Some plants in the east may have higher variable costs.

10. Grid operators (e.g., PJM) operate under Federal Energy Regulatory Commission (FERC) regulation and ultimately FERC sets policies, procedures, and market rules. It has not been uncommon for FERC to overrule recommendations of grid operators. FERC on May 9 rejected a PJM proposal designed to diminish incentives for sellers to submit speculative supply offers into its capacity market auctions

11. During scarcity events, ERCOT will set wholesale spot power price at $9,000/MWh in 2015. This sends the signal that ERCOT does not want to use operating reserves to meet energy demand because they are less able to respond to unexpected additional system contingencies.

12. Actual voltage reductions occurred briefly on January 6, 2014, but the greatest stress was on January 7, 2014. Voltage reductions are a serious step, but less serious than involuntary firm power load shedding.

13. MISO also reported high outage rates, but power and natural gas prices did not reach as high as in PJM. http://www.platts.com/latest-news/electric-power/houston/miso-forced-outages-peaked-at-22-of-generation-21084199. Published on Wednesday Jan. 15, 2014. While the winter weather was a one year in ten event, the combination of outages and weather was considered a one in 20 event. However, the question is would PJM have been in the situation if it had purchased generation instead of paying virtually the same price for interruptible summer only load, made other changes, etc.

14. Available to meet planning reserve requirements

15. Typically 30 months.

16. The President via DOE has emergency powers. It is unclear that he could declare a multi month or year emergency.

17. See MJ Bradley and the Analysis Group, August 2010, "The Electric System Has Substantial Excess Generating Capacity and Appropriate Processes in Place to Assure Reliable Electricity Supply to Consumers. Resource adequacy refers to meeting the regional planning reserve margins the power industry establishes to ensure it can meet peak demand during extreme weather and supply conditions."

18. The Clean Air Act was passed after the Federal Power Act and did not carve out exemptions for reliability concerns. Hence, the proper judicial construction is to defer to Congress's revealed latest actions. Thus, the lead agency for determining gird reliability in this context is EPA, not DOE.

19. In general, across the country, there is no entity that knows what the actual 2014 2015 winter reserve margin is. That is, in general, there is no available estimate of delivery capacity divided by expected peak demand e.g. within the market areas in PJM east.

20. FERC Chairman Jon Wellinghoff, April 22, 2013 media briefing in Washington, D.C., sponsored by the United States Energy Association.

21. The BRA (Base Residual Auction) is the key PJM forward capacity market auction. UCAP (unforced capacity) denotes capacity adjusted for generator availability, but not for fuel availability.

22. See earlier footnote.

23. The actual mechanisms involved were complex and subtle.

24. They also did not adjust the reserve margin requirements for meeting the winter peak.

25. The DC Baltimore area (sub-regions in PJM) was ground zero for the vortex. January 6, 2014 voltage reductions were concentrated in the DC area. There are only 7 major non- natural gas fired power plants in this area, and most are possible retirement candidates. Two have announced retirement, two are unscrubbed coal power plants and one is the large Calvert Cliffs nuclear unit which had a serious outage during January 2014 that led to a special review of the plant. This is an example of how things can get worse quickly.

26. White Paper, Retiring Coal Plants While Protecting System Reliability, by Judah Rose, Senior Vice President, Kenneth Collison, Vice President, and Himali Parmar, Senior Manager, 2011.

27. Transmission security is the ability of the grid to operate under contingency conditions (e.g., loss of power plants or transmission system elements such as lines, transformers, etc.). Transmission security is addressed by specialized and highly detailed modeling tools known as Alternating Current (AC) load flow models - these models examine each contingency to see if there would be system overloads which could damage critical equipment, especially long lead time equipment. The key is to focus carefully on what could unexpectedly go wrong.

28. Some regions have had large reserve levels in the winter - e.g., PJM had a 27% reserve margin. However, they made decisions without addressing the correlation in outages such as the maximum level of acceptable DR resources.

29. INGAA Foundation, "North American Midstream Infrastructure through 2035: Capitalizing on Our Energy Abundance", Prepared by ICF International, March 18, 2014.

30. *http://www.iso-ne.com/committees/comm_wkgrps/prtcpnts_comm/pac/mtrls/2013/dec182013/a3_icf_phase_2_gas_study_presentation.pdf

Article image (c) Can Stock Photo