Email This Author Comment On Article About The Author More Articles By This Author

Myth that Prices Above $ 6.00 Are Not Sustainable

Once the real causes of this summer’s higher-than-expected injections are better understood, it should be readily apparent that prices this summer could easily have soared far above the $ 6.00/MMBTU level reached early in June. If temperatures this summer had been more like last year, or the surge in the economy that seems to have begun in mid-August had started a few weeks earlier, total power sector consumption of natural gas during the summer months easily could have been 200 BCf or more higher than the consumption that actually occurred.

In fact, the electricity production figures for the last two weeks of August this year show just how narrowly we dodged a bullet this summer, in terms of exposure to higher prices.

The current all-time record for electricity production in the U.S. was set last year, during the week ended August 2, 2002 (a week that was blistering hot in almost every region in the U.S) -- with total electricity production of over 90,000 GWhrs.

Notably, however, electricity production in the last weeks in August this year was almost as high, with total production of over 89,000 GWhrs the week ended August 23rd and 88,400 GWhrs the week ended August 30th – by a wide margin, the second and third weekly production figures in U.S. history.

Further, the near record levels of electricity production in both weeks were due primarily to the resurgence of the economy, not weather. (While temperatures in many regions during both weeks were hotter than normal for late August, they were well within the range that is typical during July and early August; electricity production nonetheless significantly exceeded prior highs for any week in U.S. history other than the week ended August 2, 2002.)

If the tax cuts that became effective this summer had gone into effect three months earlier, therefore, it is entirely plausible that electricity production (and therefore natural gas consumption) would have been comparable to last year all summer long even though temperatures this summer in many key cities were unusually mild. If power sector consumption of natural gas this summer had been 200 BCf higher, however, Local Distribution Companies (LDC’s) still would have had to have met the same PUC-mandated storage targets they met this summer – the main factor driving up spot market prices in the summer months. (The LDC’s would have had no way of knowing that the weather this fall also would turn out to be unusually mild, reducing the need to inject natural gas into storage this summer.)

To meet the same storage targets, however, in a market in which generators consumed an additional 200 BCf of natural gas in June, July and August, the LDC’s would have been required to bid up prices in the spot market high enough to drive out of the market an additional 200 BCf of industrial demand over a period of just 92 days (i.e., a reduction of almost 2.25 BCf/day).

This in turn would have required an additional 15 to 20% reduction in already pared-back industrial demand within a very compressed time frame.

No one knows the exact price that would have been necessary to drive out of the market such a large percentage of the remaining industrial demand in such a short time period; a series of fortunate circumstances spared us from finding out just how steep a price increase might have been required. It is important to remember, however, that spot market prices in the Day Ahead market at Henry Hub averaged $ 5.27/MMBTU this summer.

Further, prices remained at record summer-month levels all summer long even though:

• A high percentage of all of the industrial boilers that could switch to fuel oil already had done so;
• The maximum amount of Natural Gas Liquids that can be left in the gas stream without damaging the pipelines already was being left in the gas stream;
• The fertilizer industry already was operating at significantly reduced capacity; and
• Many other price sensitive industrial users already had left the market.

Given these circumstances, therefore, there is every reason to assume that, if prices above $ 5.00/MMBTU all summer long had only the most minimal impact on industrial consumption, driving an additional 200 BCf of industrial demand out of the market (i.e., 15 to 20% of the remaining industrial load) in the space of less than 12 weeks, could easily have required prices at least in the $ 8.00 to 10.00/MMBTU range – and possibly much higher.

Rather than demonstrating that $ 6.00 prices are sustainable, therefore, this summer’s experience, seen in its proper context, demonstrates how vulnerable we are to far higher than expected prices, even in non-winter months. The Myth that We Are No Longer Exposed to Price Spikes this Winter Because The Amount of Natural Gas in Storage Has Crossed the 3,000 BCf Threshold

The last of the three myths that has so confused the market is the notion that, as long as end of Refill Season storage reaches the 3,000 BCf level, the amount of natural gas in storage should be considered to be adequate and we should not be concerned regarding the potential for price spikes during the winter months.

The notion that anyone would seriously take this position, after last winter’s experience (let alone that it would become the conventional wisdom), quite frankly, perplexes me. Let me confine myself, therefore, to a few simple points.

Even with the corrections the Climate Prediction Center made in its calculations of Heating Degree Days this summer, last winter was hardly a freakishly cold winter. Instead, during the heart of the withdrawal season (i.e., the period between November 1st and March 31st), the number of gas-weighted Heating Degree Days nationally was all of 35 HDD’s (i.e., 0.9%) colder than historical norms:

During this five month period, a total of 2,386 BCf was withdrawn from underground storage (i.e., 3,116 as of 10/31/02 – 730 BCf as of 3/31/03 = 2,386).

Further, approximately 56 BCf was withdrawn during the first two weeks in April, bringing the total withdrawal to 2,442 BCf.

It escapes me as to how anyone could review these figures and conclude that end of season storage of 3,000 BCf would be adequate.

As a practical matter, temperatures that are within 35 HDD’s of historical norms are about as close to a statistically normal winter as we’re ever likely to see. It is true, of course, that if this winter is exactly a statistically normal winter, which would mean 35 fewer HDD’s. That in turn might well translate into 35 to 50 BCf less total consumption.

Even if it did, however, starting the season with only 3,000 BCf (3,172 BCf less than last year) would be nothing short of disastrous. Storage would be drawn down to an all-time record low. And even if we somehow made it through the winter without natural gas prices setting all-time record highs, we’d be entering the Refill Season with only about 600 BCf in storage - and therefore potentially setting ourselves up for an almost impossible task in attempting to Refill Storage next year, when base level electricity demand is likely to be much higher (in recent weeks its been up by about 5% on a year-over-year basis) and summer weather may not be as forgiving as it has been this year.

Further, and just as significantly, no rational planner would plan for the winter season on the assumption that winter temperatures necessarily will be exactly equal to historical norms. To the contrary, there’s every reason to assume that, over the next several years, there will be one or more winters that will be substantially colder than last year (including quite possibly this year).

The last substantially colder-than-normal winter, for example, was just three years ago – in the winter of 2000/2001.

That winter, like last winter, was not in any sense freakishly cold; instead, temperatures were within the range that a planner should anticipate might occur every few years.

The deviation from historical norms three years ago, however, was not 35 HDD’s; it was 356 HDD’s (i.e., 10X as great).

If the same basic temperature pattern were to be repeated this winter – and it is quite possible that it could be – this in turn would be likely to result in total natural gas consumption which is roughly 500 BCf greater than last winter.

Since the supply/demand balance this winter, if anything, is likely to be even worse than last winter, this in turn could necessitate a total withdrawal from storage in excess of 3,000 BCf – i.e., last year’s 2,442 BCf + an additional 500 BCf to serve the increased space heating load + as much as another 100 to 250 BCf to account for continued deterioration in U.S. production, continued declines in imports from Mexico and the addition of approximately 1.0 million new gas-heated homes over the course of the past year. Under this scenario, 3,000 BCf or even the close to 3,200 BCf we have in storage today won’t be even remotely sufficient to cover our needs.

The fortunate set of circumstances that has occurred over the past several months has given us a somewhat larger buffer than we had any right to expect going into this winter.

But we are hardly out of the woods at this point.

Instead, unless the winter weather turns out to be far milder than currently expected this winter, the natural gas market this winter could easily be just as tight as last year – with the potential for it to become much worse if winter temperatures are more like winter weather three years ago. The rejoinder that is usually given to these facts, to the extent that there is any, is to point to the fact that last winter was especially cold in the northeast – as if that were dispositive of the issue.

This observation, however, while accurate as far as it goes, is a red herring, for two reasons:

1. The whole purpose of using gas-weighted heating degree days is to properly weight the impact of differences in temperatures in different regions. At least to a significant degree, therefore, the impact of colder weather in the northeast is already properly taken into account by the use of this methodology.
2. The impact of colder temperatures in the northeast on total gas heating demand is much less than often is assumed. The largest heating load in the country is in the Midwest, which accounts for more than 40% of the total gas heating load in the country. Despite its large population, the gas heating load in the northeast is only about ½ this size. This lower-than-expected gas heating load in the northeast is a direct result of: (I) the moderating effect of the Atlantic ocean (which keeps winter temperatures far milder than in the Midwest – as I can personally testify to having grown up in Chicago); and (II) the significantly lower penetration rate for natural gas in the northeast than in the Midwest.

While temperatures in the Midwest were also above average last winter, the variance was only slightly greater than for the nation as a whole.

Thus, the colder-than-normal temperatures in the northeast were not nearly as important a driving factor last winter as the discussion sometimes suggests. What to Expect This Winter

So what should we expect this winter? Clearly, one of the main lessons we should be learning from our experience in recent years is that natural gas prices – which always have been highly volatile, and always have been highly sensitive to fluctuations in weather – have become even more volatile and even more weather sensitive in recent years and are destined to continue doing so in future years.

This is because a much greater percentage of our total natural gas load all year long is now weather-sensitive than was true in prior years. Over the past three years, a major shift has occurred in the distribution of natural gas use among user categories. Industrial use – which generally does not vary based upon weather – has declined dramatically.

This decline, as noted earlier, began in Q4 of 2000 and has been continuing ever since, with the lion’s share occurring well before this year’s Refill Season. During this same period, however, use of natural gas for residential space heating has grown explosively, due to a combination of record new homebuilding, a high penetration rate for natural gas and aggressive conversion of a large number of existing homes to natural gas -- particularly in the northeast.

During this same period, total supplies available to the U.S. market also have been declining rapidly.

The end result is a market the total size of which is somewhat smaller than in 2000 (the all-time peak year), but in which a much higher percentage of total natural gas use during the year consists of temperature sensitive winter-heating load and, to a lesser degree (at least in 2003) power sector consumption of natural gas during the summer.

At the same time, shoulder--month consumption of natural gas has dropped dramatically, both in absolute terms and as a percentage of total natural gas consumption for the year as a whole.

As a result, when the weather is relatively mild during spring, fall, or early summer, as occurred this year, injections into storage will tend to be significantly higher than in the past, since non-weather driven demand for natural gas is at its lowest level in many years. At the same time, however, once the cold weather kicks in the winter, far larger withdrawals are likely to occur than in the past, especially in weeks in which the weather is particularly cold – just as occurred last winter.

This is due to the combined effect of the significant increase that has occurred in total space heating load coupled with a sharp drop in the net supplies of natural gas available to the U.S. market -- the combined effect of which is to create an unprecedented gap between new pipeline receipts and current demand during weeks of peak demand.

It was the size of this gap (+ the end of a streak of abnormally mild winters) that led to last winter’s all-time record withdrawals, not (as so many have claimed) the fact that winter temperatures were 35 HDD’s above historical norms.

This winter, therefore, once the cold weather hits, we are just as vulnerable to 200 BCf/week + withdrawals as were last winter. Further, it won’t take many withdrawals of this magnitude relatively early in the winter season to have a fairly profound impact on the dynamics of the natural gas market for the remainder of the winter.

The severity of the price spikes this winter, therefore, will depend largely on when the cold weather hits and just how far the temperature drops. We could get lucky; temperatures could stay mild enough so that, given the build-up in storage that has occurred, we could get through the winter without prices ever exceeding $ 6.00/MMBTU.

Or the pressure on the natural gas market could be just as severe as last winter, if not worse. It all depends on the weather.

Much will depend on what happens in December in particular. While it may seem like a distant memory at this point, in December of last year, there was still a modest el Nino effect – which is part of the reason that December temperatures were slightly below historical norms, as indicated in Table 6. This year, there will not be any el Nino effect to keep us warm in December.

Temperatures still could prove to be mild.

There are also many forecasters, however, who believe that this December will be especially cold. If they are correct – and I want to make clear that I have no idea where their predictions will prove to be accurate – temperatures might rival December of 2000, when there were 1058 gas-weighted HDD’s (vs. 841 HDD’s in December of last year).

If this were to occur – and I want to underscore that I am not predicting that it will – it could easily lead to total natural gas consumption next month that is 300 BCf greater than in December of last year, and wipe out in three to four weeks any storage “surplus” compared to last winter that may have developed by the end of this month.

Fundamentally Changed Market Dynamics

A logical next question might be: how would the market be likely to react under this (entirely plausible) scenario that (at least conceivably) could play out over the next six weeks.

Here, I think the answer is that the market almost certainly would react explosively. This, at least in my judgment, is the real story of the past year, which has been lost in the hoopla over “demand destruction.”

Specifically, I believe that, basic dynamics of the market have now fundamentally changed and that the likelihood of severe price spikes is now far greater than it was even last year, for three specific reasons:

1. The “slack in the system,” in terms of available industrial load that can be reduced quickly has been drastically reduced. As noted, earlier, there is now far less industrial load available to decrement when supplies begin to tighten and prices begin to increase than there was just three years ago in Q4 of 2000, when the first of the recent severe price spikes occurred.

   Further, since the beginning of last winter, virtually every dual fuel capable industrial boiler that is allowed to burn fuel oil and had not already done so has switched to fuel oil and retention of Natural Gas Liquids (NGL’s) in the gas stream had been increased to near maximum levels.

   In effect, therefore, industrial consumption already has been pared to the bone and there is virtually no remaining use that can be readily cut. Further, the remaining users already have repeatedly demonstrated their willingness to pay much higher-than-expected prices to continue using natural gas and it may take very high prices to drive them out of the market. (See also # 3 below.)

2. After last winter’s experience, Local Distribution Companies (LDC’s) are likely to be far more cautious in withdrawing natural gas from storage, especially during the first several months of the season. At the same time, after last winter’s experience, in which – despite record prices before the end of the season -- the amount of natural gas in storage in the eastern half of the country was drawn down to perilously low levels, LDC’s are likely to be very cautious in withdrawing natural gas from storage, especially during the first 60 to 90 days of the winter heating season, in order to minimize the risk that supplies of natural gas in storage will prove to be inadequate later in the winter.

   This could be an important factor tending to put a floor on natural gas prices during the early months of the winter, even if temperatures are mild.

   It also could significantly increase upward pressures if weather in December and early January turns out to be unusually cold, since natural gas in storage is likely to be far more “sticky” than it has been in the past.

3. Many industrial users already have locked in pricing and may be reluctant to reduce their consumption of natural gas no matter how high prices climb. Finally, after last winter’s experience, a significant percentage of industrial users who are continuing to use natural gas have locked in pricing for this coming winter by purchasing futures contracts or implementing other hedging strategies. These industrial users are likely to be reluctant to disrupt their operations or default on delivery obligations to customers for the output of their facilities by cutting back on their use of natural gas, irrespective of the market price of natural gas.

   Even a very steep increase in natural gas prices, therefore, may only bring about a relatively small near-term reduction in industrial consumption of natural gas.

As a result, if the winter turns very cold, steep price increases are likely to be required in order to free-up even modest supply increments to meet the increased needs of residential and commercial customers.

Potential Price Impacts of Changed Market Dynamics

Many analysts have not yet picked up on this fundamental change in the underlying dynamics of the natural gas market.

It is important to recognize, however, how profoundly the market has changed in the space of just 36 months. Three years ago, in Q4 of 2000, when cold weather hit in early December and supplies of natural gas began to tighten, there still will a major safety valve available, as there always had been in the past, to relieve the upward price pressure on the market: as soon as supplies began to tighten and prices began to increases, many industrial natural gas users still could – and did -- switch fuels; increased quantities of natural gas liquids still could be – and were -- left in the gas stream and a significant number of price sensitive users (e.g., smelters in the Pacific Northwest, fertilizer producers, etc.) still could – and sometimes did -- quickly shut down.

The combined impact of these actions, in prior years, was to quickly reduce demand and relieve upward pressure on the market price for natural gas.

As a result, in Q4 of 2000, when supplies tightened, while the spot market price of natural gas increased sharply, in the end it only quadrupled – peaking near $ 10.00/MMBTU in late December, and averaging well above in both December of 2000 and January of 2001.

In just 36 months, however, much has changed. There is much less industrial demand available to decrement. LDC’s are likely to be far more cautious in pulling natural gas out of storage. And the industrial users who remain in the market are much more likely to have fully hedged their positions and therefore may not be as quick to cut back on their use of natural gas.

It is possible – although still by no means certain – that if the weather is mild enough this winter we will be able to avoid severe price spikes this winter. Even if prices stay at reasonable levels this winter, however, all of the ingredients exist for a perfect storm again in the very near future – if not this winter, than potentially this summer and even more likely in the 2004/2005 winter heating season.

Given the underlying shifts that have occurred in the natural gas market, from this point forward, in any year in which an extended blast of cold weather occurs early in the winter season, or a severe hot spell occurs in the summer, it is likely to become necessary to bid prices to levels far above 2000/2001 peaks to drive even small amounts of industrial use out of the market.

Even if we dodge the bullet again this winter, therefore (as we were fortunate to do this summer), severe price spikes are virtually certain to occur in some (and perhaps most) future winters during the remainder of this decade.

The exposure to price spikes, however, is not the most severe problem we face.

Instead, our greatest challenge is to develop a strategy for continuing to meet the energy needs of the U.S. economy over the next decade despite the near-certainty of an unprecedented shortfall in supplies of natural gas. We will attempt to begin addressing this issue in a report to be issued by our firm next month.