If the weather this coming winter turns out to be as cold as the weather three winters ago, this could result in a total withdrawal from storage of as much as 3,300 – 3,375 BCf before making any provision for a working reserve.
To allow for a working reserve would require total end-of-Refill Season storage of 3,800 – 4,175 BCf – i.e., 350 – 725 BCf greater than total current storage capacity in the U.S.
< b>Importance of Working Reserve
Given the magnitude of the current storage deficit, it might be tempting to dismiss the importance of maintaining a working reserve and to assume that this year’s Refill will be adequate as long as it is sufficient to cover the maximum likely withdrawal that might occur this coming winter (i.e., using our estimate, roughly 3,375 BCf).
Any such assumption, however, would be sorely mistaken, for at least four major reasons:
1. Need to maintain certain minimum quantities of natural gas in storage in order to maintain physical pressures and ensure deliverability.
While it is possible, from a technical standpoint, to withdraw natural gas from storage even when the amount of natural gas in a particular storage facility has been drawn down to very low levels, when storage is drawn down to low levels, pressures begin to drop significantly, and withdrawing additional quantities of natural gas becomes increasingly difficult.
The operating principle that ordinarily is applied, therefore, is that, if at all possible, the amount of natural gas in storage in the aggregate nationally should not be drawn down below 500 BCf, since any reduction below this level might seriously compromise the industry’s ability to ensure physical delivery of needed supplies of natural gas.
2. Need to maintain adequate amounts of natural gas in storage in each geographic market.
Further, this general standard does not take into account the fact that working gas in storage does not sit in a single location, where it can be instantly accessed wherever it is needed anywhere in the U.S.
Instead, it is located in facilities distributed across the U.S and cannot be accessed quickly accept where it is located..
The need to draw upon natural gas in underground storage, however, varies significantly from region to region in any year.
During this past winter heating season, for example, the draw down from storage in the East Consuming Region and the Producing Region was much larger, relative to historical norms, than in the West Consuming Region. Storage in the West Consuming Region never dropped below 40% of total capacity (i.e., specifically, to a low of 167 BCf – just 10% below the 5-year average for the relevant week).
By contrast, storage in the East Consuming Region dropped to as low as 14% of capacity – an all-time record low (and 60% below the 5-year average for the East Consuming Region in the week in which the low point was reached).
As a practical matter, however, the higher relative amounts in storage in the West were not available to help ameliorate the crisis in the East (at least to any appreciable degree). Instead, the tight market conditions that existed nationally caused LDC’s and state regulators in the West to be more reluctant – not less – to release any significant amounts of natural gas from storage for use elsewhere in the country.
During the ‘00/’01 winter heating season, just the opposite occurred – i.e., storage draw downs in the West Consuming Region were much steeper than in the East.
To ensure that the requirements of each geographic market can be met, therefore (including the requirement to be able to ensure physical deliverability), several hundred BCf of additional natural gas in storage above the 500 BCf level, distributed across different locations throughout the U.S., are necessary to ensure that the amount in each market never drops below the level required to ensure physical deliverability.
3. Need to provide reserves to protect against risk contingencies.
Further, even this scenario – as bleak as it is – assumes “business as usual,” with no unusual risk contingencies or other special circumstances that might create a need to draw even larger amounts of natural gas from storage during the course of a particular winter.
From a prudent planning standpoint, no prudent utility planner ever would plan to meet its service obligations without making a specific allowance adequate to take into account the potential impact of these risk factors.
Instead, there are a whole host of risk contingency factors that, while unlikely, could easily result in the need to withdraw even larger amounts of natural gas from storage next winter and generally should be taken into account in establishing reasonable storage targets.
Without attempting to develop a comprehensive list, these factors include, for example:
- The potential for pipeline explosions or extensive equipment damage, that could limit the amounts of natural gas that can be delivered to specific markets for extended periods (such as the explosion on the El Paso system that limited deliveries into the California market at the end of 2000) or even limit flows from a major producing region (as might occur, for example, if there were a failure on the Trans-Canada Pipeline System bringing natural gas in from western Canada);
- Potential heating oil shortages (as almost occurred two winters ago);
- Greater-than-anticipated load growth in the electric utility sector, due either to stronger-than-expected weather-related demand or stronger-than-expected non-weather-related growth or both -- especially in regions in which natural gas is the marginal source of supply (which could result in a disproportionately large increase in the power-sector related need for natural gas);
- A higher than anticipated outage rate at nuclear plants (either because of generic NRC-related safety issues, potentially related to the unexpected corrosion problems discovered at the Davis-Besse and South Texas plants) or due simply to a higher-than-expected forced outage rate or both);
- An unanticipated coal strike, a rail strike and/or freezing river conditions that force one or more utilities to limit utilization of their coal-fired plants (and therefore resulted in higher-than-expected utilization of gas-fired capacity);
- A higher-than-expected forced outage rate at major coal-fired plants (many which have been deferring deep maintenance for unusually long time-periods, due to a combination of high demand for electricity and a desire to minimize maintenance expenditures);
- A threat of a terrorist attack, that might result in a temporary suspension of deliveries of LNG (as occurred with respect to a tanker scheduled to be off-loaded at the Everett Terminal near Boston after the September 11th attacks);
- A potential terrorist attack in the Middle East (such as an attack in the Gulf of Hormuz) that might disrupt oil supplies (and thus prevent dual-fuel facilities from using oil); and/or
- An oil embargo or restrictions on oil production by OPEC.
None of these events, considered in isolation, is necessarily likely to occur this summer or during the next winter heating season.
From the standpoint of prudent risk management, however, it would be nothing short of reckless not to make any provision for the potential for one or more of these events to occur. Instead, if adequate storage capacity were available, a prudent planner almost certainly would want to add several hundred additional BCf to the size of the working reserve to take into account the potential for one or more of these events to occur. The potential risks, from both a public safety standpoint from the standpoint of potential exposure to unfathomable price spikes, are simply too high for any planner to tolerate.
There is ample room for debate as to how large an increase in the size of the U.S. reserve is needed to take into account these factors.
We are living in a world in which the general tendency is to operate “closer to the margins.”
In most circumstances, that may be desirable.
Even a casual review of the list set forth above, however, suggests that the appropriate size of the working reserve almost certainly is much larger than the 500 – 800 BCf target used historically (which appears to be the minimum needed simply to maintain adequate pressures in the lines), not smaller.
This is especially true given: (i) the extreme sensitivity of natural gas consumption to the weather (with the potential for a few days of colder-than-normal winter temperatures to quickly eliminate 100 BCf or more of otherwise “excess” amounts of natural gas in storage; (ii) natural gas’ increasingly pivotal status as the marginal source of supply for meeting our electricity needs (which means that even a small increase electricity demand can lead to a fairly large increase in consumption of natural gas); and (iii) the extraordinary price spikes that occur when supplies begin to tighten, which can add tens of billions of dollars in end-user costs in a matter of weeks in years in which reserves prove to be too small.
4. Need to plan in advance for the next refill cycle.
Finally, as if the factors just listed were not enough, in establishing minimum refill targets for any year, it also is necessary to begin anticipating the next year’s refill cycle.
The further storage is drawn down in any one year, the more difficult it will be to replenish storage during the next Refill Season.
Natural gas supplies cannot be expanded rapidly; even with a record ramp-up in drilling in late 2000 and the first 8 months of 2001, U.S. production increased by less than 3% over the course of more than a year.
As a practical matter, it already is too late to achieve any material increase in supplies available to the U.S. market this year. Instead, even to bring about a relative modest increase will take at least until 2004 or perhaps even 2005 or 2006.
To avoid unacceptable stress on the market in future years, therefore, an end-of-withdrawal season storage level much higher than the 500 – 800 BCf level is required – i.e., arguably at least in the range of 1,000 – 1,200 BCf.
To ensure an end-of-season storage level at least this high, even in a colder-than-normal winter, however, would require an even larger injection during this year’s Refill Season – i.e., potentially in the range of 4,000 TCf (i.e., potential withdrawal of 3,375 BCf + 1,200 BCf end-of-season reserve – 623 BCf starting point = 3,952 BCf required injection) – obviously an unattainable goal.
Any injection below this level, however, creates the potential for an even more formidable challenge in refilling storage to acceptable levels during the 2004 Refill Season.
By any relevant measure, therefore, this year’s refill is likely to fall massively short of an acceptable level of storage.
Severity of the Current Crisis
It bears repeating that, at the start of last winter’s withdrawal season, the amount of natural gas in storage stood at 3,172 BCf – 189 BCf above the 5-year average.
As noted previously, temperatures during the winter heating season were 3% milder than historical norms. While withdrawals during the winter heating season set an all-time record, and end-of-season set an all-time record low of 623 BCf (598 BCf below the 5-year average for the same date), storage never dropped below 623 BCf.
Nonetheless, prices in the day-ahead cash market at Henry Hub still set an all-time record, reportedly reached more than $ 27.00/MMBTU on an intra-day basis on February 25th and closed at $ 18.85/MMBTU. During February and early March, prices at numerous City-gate locations in the East frequently were in the high teens or twenties and occasionally rose above $ 30.00/MMBTU.
Further, as a result of the need during this period to repeatedly withdraw natural gas from storage in order to maintain adequate pressure, at least one major pipeline in the East announced in early March that it was within no more than 2 to 3 days of declaring an event of force majeure under contracts with the LDC’s on its system.
If this had occurred, the pipeline no longer would have been required to fully honor its storage commitments to these LDC’s and instead would have been entitled to reduce its commitments pro rata to each LDC.
It is difficult to imagine how high prices ultimately might have spiked or the severity of the operating problems that might have occurred if the demand for natural gas last winter had been any higher or storage had dropped any further than it did.
Yet, this is precisely what would have occurred if the weather had been closer to normal during the warm spell that occurred between December 10th and January 10th and/or the weather had not suddenly turned Spring-like during the last 17 days of March.
Based upon last winter’s experience, therefore, it should be crystal clear that merely rebuilding storage to the same level reached last fall (i.e., 3,172 BCf) is not adequate to protect against the possibility of colder weather this coming winter or to prevent price spikes that potential could be even more severe than those that occurred last February.
Instead, given the likelihood of a further decline in U.S. production next winter and a further decline in net imports into the U.S., if next winter is just slightly colder than last winter (e.g., a winter that exactly matches historical norms), the net drain on storage easily could increase by an additional 400 – 450 BCf compared to last winter.
If this were to occur, even if storage is filled to capacity by the end of this year’s Refill Season (i.e., total end-of-Refill Season storage of 3,450 BCf), storage could be drawn down to levels below the minimum level needed to ensure physical deliverability (i.e., a minimum of 500 BCf).
This in turn could expose the U.S. market to unprecedented price spikes and to operating problems even more serious than those that occurred this past winter.
If winter temperatures match or exceed the temperatures experienced three winters ago and/or any of the any risk contingencies identified earlier were to occur (i.e., higher-than-expected electricity demand, higher-than-expected forced outage rate at coal or nuclear plants, etc.), the total amount of natural gas required to fully satisfy the needs of U.S. users could increase be another 500 – 600 BCf (i.e., a total increase in demand of 900 – 1,050 BCf above last winter’s level).
This is not by any means an implausible scenario. (Note, for example, that temperatures in November and December of 2000 were more than 350 Heating Degree Days colder than this past November and December. Just as reoccurrence of November and December 2000 temperatures would have added approximately 425 – 450 BCf to last winter’s withdrawal. Further, temperatures in January and March of this year also were well below long-term historical norms, with a net deviation of more than 100 HDD’s for the two month’s combined. If the weather in all four months had been colder, therefore, a 500 – 600 BCf increase in the size of last winter’s withdrawal would not by any means have been out of the question.)
If next winter’s withdrawal were to be this large, however, the amount of natural gas in underground storage might literally be drawn down to zero by the end of the withdrawal season. To balance supply and demand, sustained price increases might be necessary for much of the winter to levels as high as the $ 15.00 -- 20.00/MMBTU prices experienced episodically this past winter at numerous City-gate locations in the Eastern U.S. Even with price increases of this magnitude, as storage begins to be drawn down to record-low levels, operating problems might be frequent and widespread.
To avoid the potential for people freezing to death if the weather turns unusually cold in late March or the first few weeks in April, before the winter is over, it might literally become necessary for the President to declare a national emergency and invoke special rationing provisions for both electricity and natural gas.
The U.S. has never previously experienced a natural gas supply threat this extreme. The time remaining to address it without severe dislocations is dwindling every week.
Current Refill Trajectory
The question that remains, then, is the extent to which storage is likely to be rebuilt over the 175 days (i.e., 4,200 hours) that remain in this year’s Refill Season.
The prognosis is not encouraging – at least if prices during the next few months remain anywhere near current levels (i.e., in recent weeks, cash prices in the day ahead market at Henry Hub in the range of $ 5.25 – 5.75/MMBTU).
As of the last storage report issued prior to the publication of this article (i.e., the Weekly Storage Report issued May 8, 2003 for the week ended May 2, 2003), the amount of working gas in underground storage stood at 821 BCf. This is 825 BCf below – or less than half of -- the figure for the same week last year (i.e., specifically, 1,645 BCf) and lower than the end-of-winter low for all but one winter during the past two decades (i.e., in March of 1996).
Further, since the last storage report in March (on March 28th), the net injection into storage (i.e., specifically, a cumulative injection of 141 BCf) has been nearly identical to the injections during the same 5-week period last year (i.e., 141 BCf this year vs. 145 BCf last year), even though the injection rate during last year’s Refill Season was one of the lowest on record and on an aggregate basis over the 5-week period the total number of heating and cooling degree days was been nearly the same each year.
This paltry level of injections should not be surprising; it is consistent with an unmistakable trend that has persisted for more than a year, in which (as noted previously) there has been a severe supply demand imbalance in the U.S. market that, after adjusting for seasonal variations in demand, generally has become more severe over time.
Critical Next Ten Weeks
How likely is it, then, that storage will be rebuilt to minimally acceptable levels (i.e., at least in our judgment, 3,450 BCf) before the end of the Refill Season?
To a large degree, the next 10 Weekly Storage Reports issued by the Energy Information Agency (EIA) will largely answer this question.
The period from early May through the end of the first or second week in July generally is the heart of the injection season. On average, over the past two years, 50% of the total injection between early May and the end of the Refill Season has occurred during the ten week period between May 2nd and July 18th.
By mid-July, power sector demand for natural gas begins to increase, reducing the size of most injections. Injections then may briefly return to levels typical in May and early June in early September. As the fall progresses, however, heating load begins to return and weekly injections begin to diminish rapidly. Over the past five years, the last four injections of the Refill Season have averaged only 40 BCf per week (i.e., 160 BCf total for the last four weeks).
To have any reasonable prospect of achieving an adequate rebuild this year, therefore, the injection rate has to begin to step up immediately – and accelerate quite dramatically.
Specifically, to reach 3,450 BCf by late October, injections need to average just over 130 BCf per week each week for the next 10 weeks (i.e., a little over 18.5 BCf per day).
To say the least, the prospects of achieving injections this high are not particularly promising. To the best of our knowledge, the highest single-week injection ever previously achieved in the U.S. was in the range of 110 BCf (in mid-May of 2001) – i.e., 20 BCf per week below the average injection that must be achieved over the next 10 weeks for it to be realistic to achieve the 3,450 BCf target. Further, since the period in which that injection was achieved, supplies available to the U.S. market have declined by at least 3.5 – 4.0 BCf/day.
The injections achieved in 2001 and 2002 provide a helpful reference point to use in tracking injections over the next 10 weeks – since 2001 was one of the largest injections on record and 2002 one of the smallest.
2001 Refill Season
In 2001, the total amount of natural gas injected into the storage during the period between May 2nd and mid-July was 1021 BCf (i.e., an average of 102.1 BCf per week), with injections above 100 BCf in eight of the 10 weeks:
During the period between July 13th and the end of October, an additional 1,077 BCf was injected into storage (i.e., an average of 67 BCf per week).
If injections during the remainder of this year’s Refill Season were to exactly match 2001 injections, therefore, end-of-season storage would total 2,919 BCf (i.e., 821 BCf in storage currently + 1,021 BCf + 1,077 BCf = 2,919 BCf).
This is more than 250 BCf below last year’s end-of –season storage level (i.e., 3,172 BCf) and more than 500 BCf below what we believe is the minimum acceptable target for this coming winter.
2002 Refill Season
To date, however, as noted previously, on a cumulative basis, the 2003 Refill Season has been tracking almost precisely last year’s experience, with total net injections since the end of March of 141 BCf (vs. 145 BCf during the same period a year ago, with a similar number of Heating Degree Days and Cooling Degree Days in both years).
This slow pace of injections – while not surprising – ought to be cause for grave concern. During the 10-week period following the first storage report in early May, total injections last year were only 777 BCf – i.e., an average of 77.7 BCf/week (i.e., 52.3 BCf per week below the level needed to achieve what in our judgment is the minimum acceptable storage target going into this coming winter heating season):
This average rate of injections is very consistent with the level of injection reported by EIA in its last Weekly Storage Report (issued May 8th for the week ended May 2, 2003), which reported a net injection of 80 BCf (i.e., just slightly above the 77.7 BCf/week average last year). The weather for the week ended May 2nd was unusually mild -- i.e., more like the weather experienced last year during the 3-week period between May 24th and June 7th (the three mildest weeks of last year’s injection season) when injections averaged 89 BCf/week than the weather that ordinarily would be expected in late April or early May).
If the cumulative injection over the next 10 weeks were to exactly match last year’s injections during the same 10-week period, the total amount in storage as of July 11, 2003 would be only 1,598 BCf as of mid-July (i.e., current storage of 821 BCf + an additional 777 BCf = total storage of 1,598 BCf).
Further, if injections continued on the same pace as last year throughout the remainder of the Refill Season, the total amount of natural gas in storage as of the end of October would be an astonishingly low 2,321 BCf (i.e., 1,598 BCf + an additional 723 BCf between mid-July and the end of October = end-of-October storage of 2,321 BCf).
This is a disastrously low level – i.e., 230 BCf below the actual peak to trough withdrawal last winter (before making any allowance for a working reserve, 851 TCf below last year’s end-of-Refill Season storage level of 3,172 BCf (which proved to be inadequate even in a milder-than-normal winter) and more than 1.1 TCf below what we believe is the minimum acceptable end-of-Refill Season storage target for this coming winter.
Even if this coming winter turns out to be extraordinarily mild, if end-of season storage turns out to be this low, we would expect conditions in the natural gas market this coming winter to be significantly tighter than they were this last winter – when natural gas prices reached record levels and significant operating difficulties were experienced in the Eastern U.S.
If the weather matches historical norms or is colder than average, the pressures on the market could be almost unfathomable – since it could become necessary to back out of the market an additional 850 BCf to 1.1 TCf of demand (beyond the demand that already has been driven out of the market over the past several months) to keep residential and commercial customers from freezing to death before the winter is over.
Key Issues to Monitor
Given these underlying fundamentals, what metrics are most important to monitor over the coming months? And – most critically – how steep an increase in the spot market price of natural gas is likely to be required in order to rebuild supplies to adequate levels and bring supply and demand back into balance?
In terms of indicators to monitor, we recommend focusing particular attention of two sets of data – especially over the critical next 10 weeks of the Refill Season:
1. The cash price in the day-ahead market spot market at Henry Hub and other major delivery points in the Producing Region; and
2. The average weekly rate of injections, as it compares to:
- The 130 BCf/week target needed to meet minimally acceptable end-of-Refill Season storage targets;
- The 102.1 BCf/week injection rate achieved in 2001;
- The 77.7 BCf/week injection rate achieved last year.
Recently, the cash price in the day-ahead market has shown quite a bit of strength, moving up almost 50 cents in the past week, to close at $5.73/MMBTU as of the end of the day on Friday, May 9th at Henry Hub.
This is an all-time record price for early May – made all the more remarkable by the fact that the weather in early May has been fairly mild (i.e., minimal heating load and only slightly greater cooling load).
While the strengthening of the cash price at Henry Hub may have been affected to a degree by warm temperatures in Texas and by the recent outage at the South Texas nuclear plant, we believe it also may be an early indication of LDC’s beginning to enter the market to refill storage and finding it necessary to bid up the price of natural gas in order to free-up needed supplies.
It also is worth noting that the NYMEX futures market for the winter months currently is trading at only the tiniest of premiums relative to the current cash price at Henry Hub – i.e., in the range of 10 – 25 cents/MMBTU for the late fall and winter contracts, depending upon the month despite the far higher level of demand in the winter months and the growing indications that the amounts of natural gas in storage this winter will be far below acceptable levels.
If the cash price in the day ahead market continues to gain strength, or even remains at current levels, we believe it will be a very clear indication that price are likely to rise significantly in subsequent months. (We note, for example, that during the period immediately before prices exploded on February 24th, both in the day ahead market at Henry Hub and in the futures market on NYMEX, the cash price at Henry Hub consistently was running up to $ 1.00/MMBTU above the price for the near-month futures contract on NYMEX, even though the earliest potential delivery date for the near-month contract was only a few days away.)
The average weekly injection rate, of course, will be an even clearer indication of what is likely to occur in the natural gas market – both during the remainder of the injection season and the remainder of the year.
We believe it always is a mistake to draw too strong an inference from the Weekly Storage Report for any one week. There are too many extraneous factors that can affect the reported number for any one week.
It also is necessary to normalize the reported number for each week for deviations from expected weather.
After the weekly number has been normalized for unusual variances in weather, however, the rolling average tends to give a good indication of the overall supply/demand balance in the market.
Further, given the critical importance of the next 10 weeks to achieving total storage targets for the year, a reasonably clear picture of what is likely to occur in the U.S. market during the remainder of this year is likely to emerge fairly soon.
1. If, after the next 4 – 6 weeks, it is clear that injections every week are falling well below (i.e., at least 15 – 20 BCf below) the 130 BCf/week needed to meet minimally acceptable storage targets, there is every reason to expect that conditions in the natural gas market this coming winter will be at least as tight as they were during the second half of last winter (i.e., prices at least in the range of $ 6.00 – 8.00/MMBTU).
In effect, with each passing week, the nature of the problem will be clearer (i.e., available supplies are continuing to fall far short of the levels necessary to achieve acceptable levels of storage, at least without dramatic further increases in the spot market price of natural gas to drive considerable additional demand out of the market).
Further, every week, the deficit will be greater and the time remaining to rebuild storage to acceptable levels will be even shorter. As a result, with each passing week it will become increasingly unrealistic to expect prices to increase rapidly enough to drive out sufficient demand to meet minimally acceptable storage targets in the short time remaining before the Refill Season is over.
2. If, after the next 4 to 6 weeks, injections are falling significantly below the 102.1 BCf/week injection level achieved in 2001, there will be cause for grave concern. The storage deficit almost certainly will be severe enough so that, even under a “best case” scenario for the remainder of the Refill Season, we will be entering into the winter heating season with storage at dangerously low levels (i.e., end-of-Refill Season storage well below the amount of natural gas that LDC’s will actually need to be able to withdraw from storage this coming winter if temperatures are any colder than they were last winter – as easily could occur).
3. If, after the next 4 to 6 weeks, injections continue to be at or near the 77.7 BCf/week injection level achieved last year then, at least in our judgment, it is time to hit the panic button and develop an emergency program at the national level to augment the rebuilding of storage during the remainder of the injection season.
Absent such a program, the risks are simply too great that we will enter the next winter heating season with a deficit of 1.0 TCf or more relative to the minimum amount we may need to make it through the winter without endangering public safety standpoint and/or exposing end use customers to price spikes that could make last winter’s record price levels look tame.
Severity of Likely Price Increase
The question that remains, then, is how severe an increase in natural gas is likely to be required in order to free up the additional supplies of natural gas needed to rebuild storage over the next several months and bring supply and demand back into balance this coming winter.
There can be no certain answer to these questions; we are entering unexplored territory.
This much, however, can be said with a high degree of confidence: the general belief, for many years, has been that the price of natural gas could not rise above the equivalent price of oil on any sustained basis (i.e., typically, in the range of $ 4.00 – 4.50/MMBTU).
In effect, most observers have believed – and, to a remarkable degree, continue to believe -- that, except for brief periods, oil prices effectively put a cap on natural gas prices. The point of parity has shifted in recent years – from the delivered price of residual fuel oil to the delivered price of distillate. The general expectation, however, is that If natural gas prices are significantly above parity with oil for any extended time period, massive fuel switching or other “demand destruction” will occur, quickly bringing natural gas prices back in line with the delivered price of oil.
As recently as two years ago, there may have been more than a small measure of validity in this belief. After all, after the extraordinary price run-up in natural gas prices that occurred in late 2000 and early 2001, industrial demand for natural gas did contract severely – at least by 1.43 TCf (or 22%) over the past two years, if not more.
Most of this reduction did not occur immediately; instead, 90% or more in the reduction in industrial demand occurred after the price of natural gas began to drop in the spring of 2001. Some of it reflected fuel switching. The remainder reflected loss of load – due to manufacturers and industrial users exiting the U.S. market, the delayed impact of energy users instituting energy efficiency measures in response to the price shocks of late 2000/early 2001 and the effects of the severe manufacturing recession that hit the U.S. economy starting in the spring of 2001. But industrial use did contract.
There also is no question that when the cash price for natural gas hit the $ 10.00 – 15.00/MMBTU range this past February and early March, prices well above $ 10.00/MMBTU or even $ 15.00/MMBTU were sufficient to catch everyone’s attention. More than one generator who had previously resisted fuel switching finally decided that the time had come to switch to fuel oil or distillate at certain of their facilities which were duel fuel-permitted and had the flexibility both from a contractual standpoint and a logistical/fuel delivery standpoint to switch to oil. And some – but not all – of those facilities are still burning oil today (more than two months later). In addition, a number of major manufacturers decided either to shut down U.S. factories or to shift a portion of their production overseas (including one of the two best known chemical companies in the U.S., which reportedly shifted certain of its operations in Texas and Louisiana to Germany).
At this point, however, the notion that massive fuel switching or demand destruction will occur if natural gas prices rise above parity with oil (currently, using distillate as a point of comparison, around $ 5.00/MMBTU) is little more than a creative myth, for at least four reasons:
1. Whatever may have been true in earlier years, most major opportunities for fuel switching already have been captured and a high percentage of the most price-sensitive load already has been driven out of the market.
At this point, every industrial facility and generator that is continuing to burn natural gas by definition is doing so even though natural gas is selling at a premium of almost $ 1.50/MMBTU relative to residual fuel oil in most markets and more than $ 0.75/MMBTU relative to distillate.
Further, while some current natural gas users have curtailed operations during periods of high prices and then returned to the market, the great majority of current users have continued to burn natural gas throughout two successive periods when the price of natural gas has exceeded $ 8.00/MMBTU for sustained time periods (i.e., in December and January of 2000 and 2001 and, in many markets, again this past winter).
In addition, natural gas prices have not been below $ 5.00/MMBTU since the beginning of the year and, for most of 2003, have been at least in the $ 5.50 – 6.00/MMBTU range.
The notion that $ 5.00/MMBTU or even $ 5.50/MMBTU is a magic break point at which demand suddenly begins to fall off sharply, therefore, is plainly false.
Further – and perhaps importantly – despite the severe price run-ups that have occurred during two out of the past three winters, there is no clear evidence of a sharp drop-off in consumption at prices anywhere in the range between $ 6.00 – 10.00/MMBTU range.
The belief that massive “demand destruction” will occur once prices breach the $ 4.00 – 5.00/MMBTU threshold (which is the lynchpin of the price forecasts issued by most analysts) appears to be based far more on wishful thinking than on any observed behavior in the market or other empirical evidence of how end use customers respond to increases in price.
2. The amount of fuel switching potential that remains is far less than many observers recognize.
The reasons why fuel switching continues to fall far short of expectations are hardly mysterious.
A number of analysts claim to have identified at least 2.5 BCf/day of remaining potential for industrial fuel switching and at least 4.0 BCf/day of fuel switching potential in the generation sector.
Most such claims, however, are based primarily on a head-count of facilities which are dual-fuel permitted for both natural gas and oil, not on a facility-specific assessment that any particular facility will in fact burn oil when the economics favor switching.
There are numerous factors, however, that prevent or discourage facilities from switching to fuel oil or distillate even when their permits theoretically permit them to burn multiple fuels.
- Some dual-fuel permitted facilities, such as the Manatte plant in Florida (a large powerplant located near the Everglades) have burned large amounts of fuel oil in the past, but recently have been converted to natural gas and are no longer physically capable of burning oil;
- Many other industrial facilities and many generators which have been dual fuel permitted in the past have now had their entitlement to burn oil eliminated during the ozone season (i.e., from May 1st through September 30th, and therefore will be unable to burn oil until the last few weeks of the injection season);
- Many combined cycle units are dual-fuel permitted, but don’t currently have the burners required to burn fuel oil and would not be likely to incur the expense required to install these burners given the likely local opposition to burning oil (which would result in higher emissions of NOx);
- Numerous other industrial facilities and generators are technically dual-fuel permitted, but are reluctant to consider switching, since local opposition would be too great and/or are limited to burning oil in emergencies and/or for brief periods (i.e., often as short a period as 7 or 14 days);
- Finally, many generators may be dual-fuel permitted, and may even have burned substantial quantities of fuel oil within the past year or two, but may be very unlikely to burn oil in the future, since they are not in compliance with the more stringent requirements that apply to oil-burning facilities when burning NOx (especially in the summer months) or lack the emissions credits they would need to offset the higher NOx emission rate associated with burning oil. Still other facilities might have to accept a de-rating of 50% or more in order to comply with the applicable emissions limitations while burning oil (which, as a practical matter, may make it economically prohibitive to burn distillate – even if natural gas prices reach $ 10.00/MMBTU).
These are major limitations, which in all likelihood cut by more than half the remaining potential for fuel switching, both in the industrial sector and in the generation sector.
3. Further, there are a large number of practical impediments that limit both the extent to which dual-fuel permitted facilities will switch fuels and the speed with which fuel switching and/or other “demand destruction” will occur.
Even these factors, however, are just the tip of the iceberg, in terms of practical considerations that deter both industrial users and generators from curtailing use of natural gas.
- Many users are fully or partially hedged, at least for the next 3 – 6 months, and sometimes longer. Even if these users could theoretically make more money by shutting down operations and selling gas supplies into the market, they typically are far more inclined to fill outstanding orders than to disappoint long-standing customers;
- Further, for many generators, economic considerations may strongly deter fuel switching. Burning fuel oil results in significantly greater wear and tear on the generator unit, and therefore significantly higher maintenance costs. These costs generally are not recoverable from ratepayers, and therefore must be absorbed entirely by shareholders in any jurisdiction in which conventional ratemaking still applies. By contrast, reductions in fuel costs typically are flowed through automatically to end use customers. From the standpoint of the utility, therefore, there is much to be lost, and little to be gained by switching fuels – which is precisely why many dual-fuel permitted generators did not seriously consider fuel switching until natural gas prices hit double digit figures;
- Even with fuel price differentials are compelling, fuel switching may not be feasible, either from a timing standpoint or in terms of logistics. Natural gas supply contracts may lock in natural gas purchases for many months (or at least impose significant penalties for failure to take deliveries), or the end user may be required to absorb transportation costs whether or not it takes the gas. Even if lead time is not a significant issue on the natural gas side, it may be on the oil procurement side, especially if large quantities are involved. Extensive lead-time may be required to line-up the required quantities, particularly in a tight market;
- Further, logistics can be a major issue – to the point of making fuel switching utterly unrealistic for many dual-fuel permitted facilities, except perhaps for very brief, emergency conditions (e.g., 24 – 48 hours). When natural gas prices spiked to record high levels in late February and early March, the generators and large industrial facilities that were most likely to fuel switch – and to remain on oil now – were those located on or close to major oil pipelines or refineries. Absent contractual impediments, these facilities often can switch fuels on a few days’ notice, and delivering the required quantities of fuel is not a major issue. Many other generators and large industrial facilities, however, are not as favorably located. Instead, even if a power plant or large industrial facility has river access, it may be necessary to bring in oil by barge, which may require a 1 – 2 month lead time in order to switch fuels, even assuming the necessary haulage capability is available. If the plant does not have river access, the only means to bring in oil in many instances is by truck. This may be perfectly adequate for many smaller industrial facilities (in which case the quantities of natural gas displaced generally will be relatively small). For a large power plant or other major industrial facility, however, as a practical matter it may be a complete non-started to attempt to bring in the huge quantities of oil that may be required by truck, except for brief emergency situations.
4. During the remainder of this year’s injection season, limitations on the availability of distillate are likely to further limit the extent to which fuel switching occurs, even at facilities whose owners might otherwise might be inclined to switch fuels.
Finally, as if these limiting factors were not sufficient, during the remainder of this year a fourth factor is likely to severely limit the amount of fuel switching that occurs in the U.S., especially at facilities that burn distillate: even if generators and/or industrial users were otherwise inclined to switch, the supplies are not likely to be physically available to permit any significant amount of switching to occur.
Inventories of most oil products are currently are relatively low levels in the U.S., in part because U.S. refiners chose not to aggressively bid for additional supplies in late 2002 and early 2003 when oil prices were at exceptionally high levels.
Inventories of distillate, however, are especially low – in part because in recent months substantial amounts of distillate have been processed into gasoline.
Further, while world oil production has increased, so have refinery runs, and inventories are not expected to be replenished significantly any time soon.
As a practical matter, therefore, inventories of distillate are likely to remain at or near current, near record-low levels for at least the next several months.
This already has been reflected in increasing premiums for distillate relative to other oil products (and therefore a lower price differential than otherwise might be expected relative to natural gas).
Even a modest pick-up in fuel switching is likely to put further pressure on distillate prices – and rapidly deplete supplies.
In the end, therefore, all other factors aside, the physical supply simply isn’t available to permit large amounts of switching to distillate, at least over the next 3 – 4 months.
These are not minor issues. Instead, in the aggregate, they suggest that, at least at prices below $ 10.00/MMBTU, the amount of remaining fuel switching that reasonably might be expected to occur may be as little as 1.0 – 1.5 BCf/day, in the industrial sector and the power sector combined.
With less than 175 days left in the Refill Season, even if this fuel switching were to occur overnight, it only would free up another 175 – 250 BCf of natural gas for injection into storage between now and late October – which still would leave end-of-Refill Season storage disastrously short of the levels required to protect public safety and avoid exposure to unprecedented price spikes this coming winter.
Realistically, however, this fuel switching will not occur all at once – to the extent it occurs at all. Instead, even if LDC’s step into the market very aggressively, to attempt to rapidly rebuild storage, it will take a number of weeks for natural gas prices to move substantially higher than current levels and perhaps another two months after these further price increases have occurred before most available opportunities for additional fuel switching have been captured.
At that point, there will be only 10 – 15 weeks left in the Refill Season, and the amount of natural gas in storage may still be well below 2,000 BCf.
This in turn may set the stage for unprecedented price increases, in order to drive further demand out of the market, in an increasingly difficult struggle to rebuild storage to minimally acceptable levels and bring supply and demand back into balance for the remainder of the year.
Two of the most capable firms who follow the natural gas market have taken the position that end-of-Refill Storage will be sufficient if it reaches 2,700 – 2,800 BCf. (Different analysts use different figures.)
Even using these relatively modest targets, these analysts forecast natural gas prices by the end of this summer in the range of $ 6.75/MMBTU – with the potential for significantly higher prices next fall and next winter if end-of-season storage falls short of this range.
Many other analysts duck the storage issue entirely, and confidently predict that natural gas prices will revert to more normal levels before the end of the year.
This analysts in turn generally see the run-up in natural gas prices over the past half year as primarily a product of events in the world oil market, and believe that if natural gas prices are substantially out of synch with oil prices for any extended time period, massive fuel switching and other “demand destruction” will occur – effectively putting a ceiling on natural gas prices at a level no higher than $ 4.50 – 5.00/MMBTU (depending upon the current price of oil).
Time will tell how the market will react.
Fundamentally, however, we believe that both points of view just described are far off the mark.
While we respect a great deal the two major firms that currently are forecasting prices in the $ 6.75/MMBTU range, as discussed earlier in the article, in light of last winter’s experience, we believe it is utterly indefensible to continue to use 2,700 – 2,800 BCf as a reasonable end-of-season planning target.
Comparing next winter to last winter, there are certainly considerations that, all other factors being equal might cause withdrawals next winter to decline modestly compared to this past winter. Some fuel significant fuel switching and demand destruction have already occurred and are likely to persist. The opportunity for additional fuel switching also is likely to be present. Higher prices will have at least some impact on demand. LNG imports are likely to be higher and it is at least plausible (although not necessarily likely) that demand for natural gas in the power sector will decrease.
On the other hand, this past winter, starting with 3,172 BCf of natural gas in storage (i.e., 400 –500 BCf above the range suggested) clearly was not enough to avoid severe price spikes and significant operating difficulties, that easily could have been much worse with just one or two more weeks of cold weather.
Further, the list of countervailing factors that could increase total storage requirements is much longer than the list of potential favorable developments provided.
At this point, we have no way of knowing what next winter’s weather will bring. Under this circumstance, it would be nothing short of reckless not to inject sufficient natural gas in storage to be prepared for a winter at least as cold as the winter we actually experienced just three winters ago. This is turn requires a 600 BCf increase in the amount of natural gas injected into storage, just to account for this factor alone.
Further, this is not the only significant factor. Production of natural gas from U.S. wells almost certainly will be significantly lower next winter than last. Imports from Canada are likely to decline significantly – in part because of the urgent need to retain more natural gas in Canada than was retained last winter. The number of gas-heated homes – and there for residential demand – is continuing to increase rapidly. Total natural gas use in the generation sector is more likely to increase than to decline. Exports to Mexico next winter are likely to increase, at least modestly. And we also clearly need to make a greater allowance for potential risk contingency factors than we have made in the past. Our risk exposure clearly is greater now than it was just a few years ago; further, risk exposure aside, we have learned over the past three winters that, when supplies available to the U.S. market become exceedingly tight (as they did both during the ‘00/’01 winter heating season and again last winter), the cost to customers can be astronomical (i.e., at least $ 40 billion in ‘00/’01 and at least $ 30 –35 billion last winter).
Even the two firms that have recently published price forecasts in the $ 6.75/MMBTU range for this summer, however, have been quick to acknowledge that, if additional injections into storage are required, it may be necessary to increase spot market prices to well above $ 10.00/MMBTU in order to drive more than another 1.0 – 2.0 BCf/day of demand out of the market.
The views of those who continue to believe that natural gas prices will quickly revert to the $ 4.50 to 5.00/MMBTU level require less discussion.
As noted earlier, it may have been true, as recently as two years ago, that considerations of timing aside, it was reasonable to expect oil prices to put an effective cap on the sustainable price for natural gas. But it is no longer true today. The remaining potential for fuel switching is too small and the underlying demand for natural gas that is relatively insensitive to price is too high.
We live in a challenging time.
In two of the past three winters, the natural gas price forecasts issued by most analysts proved to be worthless.
Prior to the ‘00/’01 winter heating season, no one in the industry predicted any significant increase in natural gas prices – and prices quadrupled.
This past winter, a few of us predicted the potential for $ 6 – 8/MMBTU prices as early as September, but the consensus view in January still was in the $ 4 – 5 range.
None of us has a perfect crystal ball.
By now, however, it should be clear that the time has come to toss out most price forecasts, which now have been thoroughly demonstrated to have little or no predictive value.
Further, since most power price forecasts ultimately are derivative of natural gas price forecasts (at least in the summer months), and many forecasters start by simply “taking off the shelf” someone else’s natural gas price forecast (rather than preparing their own ground-up analysis), many electricity price forecasts need to be thoroughly revisited as well.
The potential for severe price dislocations, in both the natural market and regional power markets has huge implications, both for end use customers and for almost every segment of the energy industry.
The time to start revisiting these issues, however, is right now. Any planning or investment decision made today based upon price forecasts that look very much like they may have looked nine months ago has a high risk of being ill-advised.
The even more urgent problem we face, however, is the urgent need to replenish storage this year and begin aggressive efforts to rectify the massive supply demand imbalance we face during the remainder of this decade.
The potential solutions to these challenges will be the subject of future articles.